CN117933593B

CN117933593B - Heating system thermal dispatching management method and system

Info

Publication number: CN117933593B
Application number: CN202311673685.1A
Authority: CN
Inventors: 赵永芳; 李晓琴; 刘大海; 李昔真
Original assignee: Beijing Zhongneng North Technology Co ltd
Current assignee: Beijing Zhongneng North Technology Co ltd
Priority date: 2023-12-07
Filing date: 2023-12-07
Publication date: 2024-09-03
Anticipated expiration: 2043-12-07
Also published as: CN117933593A

Abstract

According to the heating system thermal power dispatching management method and system provided by the embodiment of the invention, the feasibility and the interpretability of the heating power dispatching suggestion viewpoint of the heating power equipment operation state data of the heating power station of the heating power network are determined according to the system dispatching decision score of the heating power equipment operation state data of the heating power station of the heating power network, so that guidance can be provided for the subsequent coordinated dispatching of the heating power system, the dispatching management intelligent degree of the heating power station of the heating power network is improved, and the heating effect is ensured.

Description

Heating system thermal dispatching management method and system

Technical Field

The invention relates to the technical field of data processing, in particular to a heat supply system thermal power dispatching management method and system.

Background

With the acceleration of the urban process, the urban heat supply network scale is continuously enlarged, and effective management and optimization of a heat supply system become keys for improving energy efficiency and resident life quality. The heat supply system heat distribution is used as an important link for guaranteeing the stability and economy of heating, and the management method and the intelligent level of the heat supply system heat distribution directly influence the operation effect of the heat supply system.

The traditional heat supply system scheduling management is mostly dependent on manual experience to start and stop equipment and parameter adjustment, and the mode is slow in response when a complex and dynamically-changed heat supply network is processed, so that the optimal operation is difficult to realize. Although some automation techniques have been introduced, they are generally limited to the monitoring and control of a single or a few parameters and do not fully reflect the operating state of the entire thermodynamic system.

In order to solve the above problems, a new thermal schedule management method needs to be developed.

Disclosure of Invention

In order to improve the problems, the invention provides a heating system thermal dispatching management method and a heating system thermal dispatching management system.

In a first aspect, a thermal schedule management method of a heating system is provided, and the method is applied to a thermal schedule management system, and includes:

Performing state text semantic mining on the thermodynamic equipment operation state data of the heat supply network thermodynamic station to obtain a state text semantic relation network with a size of a; wherein a represents the number of thermal parameter data in the thermal equipment operation state data, b represents the number of operation state text semantics of each thermal parameter data in the thermal equipment operation state data, and a and b are integers not less than 2;

determining heat supply scheduling decision feature with the size of 1*b of each piece of thermodynamic parameter data in the thermodynamic equipment operation state data based on the state text semantic relation network with the size of a.b, and obtaining a heat supply scheduling decision feature relation network with the size of a.b; the heat supply scheduling decision feature of the c-th heat supply parameter data in the heat equipment operation state data is based on the c-th heat supply parameter data being judged to belong to the e-th heat supply scheduling label in the d-th heat supply scheduling labels, and the heat supply scheduling decision feature corresponding to the e-th heat supply scheduling label in the d-th heat supply scheduling decision feature corresponding to the c-th heat supply parameter data is an integer not smaller than 2, c is an integer not smaller than 1 and not larger than a, and e is an integer not smaller than 1 and not larger than d;

Determining a coordinated scheduling priority list with the size of b of each thermodynamic parameter data in the thermodynamic equipment operation state data based on the state text semantic relation network with the size of a and the heat supply scheduling decision feature relation network with the size of a, and obtaining a coordinated scheduling priority list;

And determining a system scheduling decision score of the thermodynamic equipment operation state data based on the a coordination scheduling priority list.

Optionally, the determining, based on the state text semantic relation network with the size of a×b, the heat supply scheduling decision feature with the size of 1*b of each piece of heat distribution parameter data in the heat distribution equipment operation state data to obtain a heat supply scheduling decision feature relation network with the size of a×b includes:

Determining heat supply dispatching suggestion viewpoints of all thermal parameter data in the thermal equipment operation state data based on the state text semantic relation network with the size of a x b and the heat supply dispatching decision feature relation network with the size of d x b to obtain a heat supply dispatching suggestion viewpoints; wherein the heat supply schedule suggestion viewpoint of the c-th heat supply parameter data in the heat power equipment operation state data is used for representing that the c-th heat supply parameter data belongs to the e-th heat supply schedule tag in the d-th heat supply schedule tags, the heat supply schedule suggestion viewpoint of the c-th heat supply parameter data in the heat power equipment operation state data is determined based on operation state text semantic knowledge of a size 1*b used for representing the c-th heat supply parameter data and heat supply schedule decision feature of a size 1*b corresponding to the e-th heat supply schedule tag, the state text semantic relation network of a size a×b comprises operation state text semantic knowledge of a size 1*b used for representing the c-th heat supply parameter data, and the heat supply schedule decision feature relation network of a size d×b comprises heat supply schedule decision feature of a size 1*b corresponding to the e-th heat supply schedule tag;

determining a heat supply scheduling decision feature with a size of 1*b corresponding to each of the a heat supply scheduling suggestion viewpoints as a heat supply scheduling decision feature with a size of 1*b of each of the thermodynamic parameter data in the thermodynamic equipment operation state data; wherein a heating schedule decision feature of size 1*b corresponding to a c-th heating schedule advice point of the a-th heating schedule advice point of the thermal plant operating state data is a heating schedule decision feature of size 1*b corresponding to the e-th heating schedule label.

Optionally, the determining the heat supply schedule suggestion views of each thermal parameter data in the thermal equipment operation state data based on the state text semantic relation network with the size of a×b and the heat supply schedule decision feature relation network with the size of d×b to obtain a heat supply schedule suggestion views includes:

Inputting the state text semantic relation network with the size of a x b into a target heating power system dispatching analysis network for completing debugging to obtain a target heating power dispatching proposal output branch, and obtaining a x d initial heating power dispatching proposal views; the target heat supply schedule proposal output branch determines a d initial heat supply schedule proposal viewpoints based on a state text semantic relation network with a size of a x b and a set heat supply schedule decision feature relation network with a size of d x b, d initial heat supply schedule proposal viewpoints corresponding to the c-th heat supply parameter data in the a x d initial heat supply schedule proposal viewpoints represent possibility that the c-th heat supply parameter data belongs to each heat supply schedule label in the d heat supply schedule labels, d initial heat supply schedule proposal viewpoints corresponding to the c-th heat supply parameter data are respectively determined based on running state text semantic knowledge with a size of 1*b used for representing the c-th heat supply parameter data and the heat supply schedule decision feature relation network with a size of d x b;

among d initial heat supply schedule advice views corresponding to each thermodynamic parameter data in the thermodynamic equipment operation state data, determining the initial heat supply schedule advice view with the highest characterization possibility as the heat supply schedule advice view of each thermodynamic parameter data in the thermodynamic equipment operation state data; the heat supply schedule advice view of the c-th heat supply parameter data is the initial heat supply schedule advice view with highest possibility of being characterized in d initial heat supply schedule advice views corresponding to the c-th heat supply parameter data, and the c-th heat supply parameter data in the d heat supply schedule tags has highest possibility of belonging to the e-th heat supply schedule tag.

Optionally, the determining a system scheduling decision score of the thermodynamic device operation state data based on the a pieces of coordinated scheduling priority lists includes:

Executing heat supply coordination feature mapping on a coordination scheduling priority list corresponding to each thermal parameter data in the a coordination scheduling priority lists to obtain a heat supply coordination feature values;

and determining a system scheduling decision score of the thermodynamic equipment operation state data according to the a heat supply coordination characteristic values.

Optionally, the determining a coordinated scheduling priority list of b size of each thermal parameter data in the thermal equipment operation state data based on the state text semantic relation network of a size of b and the heating scheduling decision feature relation network of a size of a, to obtain a coordinated scheduling priority list includes:

Inputting the state text semantic relation network with the size of a x b and the heat supply scheduling decision feature relation network with the size of a x b to finish the target scheduling decision scoring output branch of debugging to obtain the a coordination scheduling priority list; the target thermodynamic system scheduling analysis network is a neural network obtained by debugging a scheduling decision scoring output network to be debugged through a historical thermodynamic equipment running state data set.

Optionally, the method further comprises:

On the basis that the historical thermal equipment operation state data set comprises Q historical thermal equipment operation state data, carrying out state text semantic mining on each historical thermal equipment operation state data in the Q historical thermal equipment operation state data to obtain Q state text semantic relation network debugging samples; the number of the thermodynamic parameter data in the operation state data of each historical thermodynamic device is a, the f-th state text semantic relation net debugging sample in the Q state text semantic relation net debugging samples is a state text semantic relation net debugging sample which is obtained by carrying out state text semantic mining on the f-th historical thermodynamic device operation state data in the Q historical thermodynamic device operation state data and has a size of a x b, Q is an integer not smaller than 2, and f is an integer not smaller than 1 and not larger than Q;

Determining a heat supply scheduling decision feature relation network debugging sample corresponding to each historical thermodynamic equipment operation state data in the Q historical thermodynamic equipment operation state data based on the Q state text semantic relation network debugging samples, and obtaining Q heat supply scheduling decision feature relation network debugging samples; the f-th heat supply scheduling decision feature relation network debugging sample in the Q heat supply scheduling decision feature relation network debugging samples is a heat supply scheduling decision feature relation network debugging sample with a size of a x b determined based on the f-th state text semantic relation network debugging sample; wherein, in the f-th heat supply scheduling decision feature relation network debugging sample, a heat supply scheduling decision feature debugging sample with the size of 1*b of each piece of heat parameter data in the f-th historical heat equipment operation state data is included, the heat supply scheduling decision feature debugging sample of the g-th heat parameter data in the f-th historical heat equipment operation state data is based on that the g-th heat parameter data is judged to belong to the h-th heat supply scheduling label in the d-th heat supply scheduling label, and in d heat supply scheduling decision features corresponding to the g-th heat supply parameter data, the heat supply scheduling decision feature debugging sample corresponding to the h-th heat supply scheduling label is g which is an integer not less than 1 and not more than a, and h is an integer not less than 1 and not more than d;

and debugging the scheduling decision scoring output network to be debugged based on the Q state text semantic relation network debugging samples and the Q heat supply scheduling decision feature relation network debugging samples to obtain the target scheduling decision scoring output branch.

Optionally, the debugging the scheduling decision score output network to be debugged based on the Q state text semantic relation network debugging samples and the Q heat supply scheduling decision feature relation network debugging samples to obtain the target scheduling decision score output branch includes:

executing ith debugging on the scheduling decision scoring output network to be debugged by using the following operation; wherein i is an integer of not less than 1:

Inputting a j-th state text semantic relation network debugging sample and a j-th heat supply dispatching decision feature relation network debugging sample used by the i-th debugging into a dispatching decision scoring output network of the i-th debugging to obtain a coordinated dispatching priority list dispatching sample with the size of b x b of each piece of thermodynamic parameter data in the j-th historical thermodynamic equipment operation state data, and obtaining an a coordinated dispatching priority list dispatching sample; j is an integer which is not less than 1 and not more than Q, the Q state text semantic relation network debugging samples comprise a j state text semantic relation network debugging sample, the Q heat supply scheduling decision feature relation network debugging samples comprise a j heat supply scheduling decision feature relation network debugging sample, and the Q historical thermal equipment operation state data comprise j historical thermal equipment operation state data;

Determining a debugging cost variable of the ith debugging based on the j-th state text semantic relation network debugging sample, the j-th heat supply dispatching decision feature relation network debugging sample and the a-th coordination dispatching priority list debugging sample;

On the basis that the debugging cost variable of the ith debugging does not meet the set debugging standard reaching requirement, improving the neural network variable in the scheduling decision scoring output network of the ith debugging to obtain the scheduling decision scoring output network of the (i+1) th debugging;

And stopping debugging on the basis that the debugging cost variable of the ith debugging meets the debugging standard reaching requirement, and determining the scheduling decision scoring output network of the ith debugging as the target scheduling decision scoring output branch.

Optionally, the determining the debugging cost variable of the ith debugging based on the jth state text semantic relation network debugging sample, the jth heat supply dispatching decision feature relation network debugging sample and the a coordination dispatching priority list debugging sample includes:

determining, on the basis that the jth state text semantic relation network debugging sample includes an operation state text semantic knowledge debugging sample with a size of 1*b corresponding to each piece of thermal parameter data in the operation state data of the jth historical thermal equipment, the jth heat supply scheduling decision feature relation network debugging sample includes a heat supply scheduling decision feature debugging sample with a size of 1*b corresponding to each piece of thermal parameter data in the operation state data of the jth historical thermal equipment, and the a-th coordinate scheduling priority list debugging sample includes a coordinate scheduling priority list debugging sample with a size of b corresponding to each piece of thermal parameter data in the operation state data of the jth historical thermal equipment, according to an operation state text semantic knowledge debugging sample with a size of 1*b corresponding to each piece of thermal parameter data in the operation state data of the jth historical thermal equipment, a heat supply scheduling decision feature debugging sample with a size of 1*b corresponding to each piece of thermal parameter data in the operation state data of the jth historical thermal equipment, and a coordinate scheduling priority list debugging sample with a corresponding to each piece of b corresponding to each piece of thermal parameter data in the operation state data of the j historical thermal equipment;

And determining the debugging cost variable of the ith debugging based on the a debugging cost variables.

Optionally, the performing state text semantic mining on each of the Q historical thermal device operation state data to obtain Q state text semantic relation network debug samples includes:

Inputting all the historical thermodynamic equipment operation state data in the Q historical thermodynamic equipment operation state data to complete target state text semantic mining branches in a target thermodynamic system scheduling analysis network for debugging, so as to obtain Q state text semantic relation network debugging samples; the target thermodynamic system scheduling analysis network is a neural network obtained by debugging the thermodynamic system scheduling analysis network to be debugged through at least part of historical thermodynamic equipment operation state data in the historical thermodynamic equipment operation state data set.

Optionally, the determining, based on the Q state text semantic relation network debugging samples, a heating schedule decision feature relation network debugging sample corresponding to each historical heating power equipment operation state data in the Q historical heating power equipment operation state data, to obtain Q heating schedule decision feature relation network debugging samples includes:

Inputting each state text semantic relation net debugging sample in the Q state text semantic relation net debugging samples into a target heating power system dispatching analysis network for completing debugging, and outputting a target heating dispatching suggestion to obtain Q heating dispatching decision feature relation net debugging samples; the target thermodynamic system scheduling analysis network is a neural network obtained by debugging the thermodynamic system scheduling analysis network to be debugged through at least part of historical thermodynamic equipment operation state data in the historical thermodynamic equipment operation state data set.

In a second aspect, a thermal schedule management system is provided, comprising a processor and a memory in communication with each other, the processor being configured to retrieve a computer program from the memory and to implement the method of the first aspect by running the computer program.

In a third aspect, a computer readable storage medium is provided, on which a computer program is stored, which computer program, when run, implements the method of the first aspect.

According to the heat supply system heat distribution management method and system provided by the embodiment of the invention, the heat supply system heat distribution analysis network which is debugged in advance is input into the heat supply system heat distribution equipment operation state data of the heat supply network heat station, so that a state text semantic relation network with a size of a x b and a heat supply distribution decision feature relation network with a size of a x b for representing each heat supply parameter data in the heat supply system heat supply equipment operation state data of the heat supply network heat station are sequentially obtained; and determining a coordinated scheduling priority list of each piece of thermodynamic parameter data in thermodynamic equipment operation state data of the thermodynamic station of the heat supply network through a state text semantic relation network with the size of a and a heat supply scheduling decision feature relation network with the size of a, and determining a system scheduling decision score of the thermodynamic equipment operation state data of the thermodynamic station of the heat supply network. Based on the method, the feasibility and the interpretability of the heat supply schedule proposal viewpoint of the heat supply network heat station heat supply equipment operation state data are determined according to the system schedule decision score of the heat supply network heat station heat supply equipment operation state data, so that guidance can be provided for subsequent heat supply system coordinated scheduling, the schedule management intelligent degree of the heat supply network heat station is improved, and the heat supply effect is ensured.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flowchart of a heat supply system thermal schedule management method according to an embodiment of the present invention.

Detailed Description

In order to better understand the above technical solutions, the following detailed description of the technical solutions of the present invention is made by using the accompanying drawings and specific embodiments, and it should be understood that the specific features of the embodiments and the embodiments of the present invention are detailed descriptions of the technical solutions of the present invention, and not limiting the technical solutions of the present invention, and the technical features of the embodiments and the embodiments of the present invention may be combined with each other without conflict.

Fig. 1 shows a thermal schedule management method of a heating system, which is applied to a thermal schedule management system, and includes the following steps 10-40.

Step 10: and carrying out state text semantic mining on the thermodynamic equipment operation state data of the heat supply network thermodynamic station to obtain a state text semantic relation network with a size of a.

In step 10, a represents the number of thermodynamic parameter data in the thermodynamic device operation state data, b represents the number of operation state text semantics of each thermodynamic parameter data in the thermodynamic device operation state data, and a and b are integers not less than 2.

First, the key terms mentioned in step 10 are explained.

A heat grid heating station is an energy supply point responsible for delivering heat energy to a user in the form of steam or hot water. Typically, a heat grid thermal station will contain thermal equipment such as boilers, heat exchangers, pumps, valves, etc.

The thermodynamic device operation state data refers to data recording the working conditions of the thermodynamic device in a specific time, such as parameters of temperature, pressure, flow rate, etc.

State text semantic mining is a technique for understanding the operational state of a thermodynamic device by analyzing textual descriptions. It may involve natural language processing techniques to identify concepts, entities, and relationships related to device states.

The dimensions refer to the dimensions of a matrix or a relational network. a represents the number of different thermodynamic parameter data, and b represents the number of corresponding operational state text semantics.

The state text semantic relation net is a structure representing the relation between each parameter in the operation state data of the thermodynamic equipment and the semantic description thereof, and can be regarded as a two-dimensional array or matrix.

Thermal parameter data are specific values that measure the performance of a thermal device, such as temperature, pressure, flow, etc.

The operation state text semantic refers to the description of the operation state of the thermal equipment, and is converted into understandable and meaningful semantic information.

Next, step 10 is further explained with an example.

For example, an intelligent integrated energy technology company has a heat grid heat station that includes a plurality of heat power devices, each of which generates operational state data. In step 10, this data now needs to be processed in order to better understand the operating conditions of the device and to optimize the scheduling decisions.

For example, the heat network station has 4 thermodynamic parameter data (a=4), temperature, pressure, flow and humidity, respectively. For each parameter, the company analyzed the corresponding run state description using natural language processing techniques, resulting in 5 different text semantic tags (b=5), such as "normal", "above normal", "below normal", "abnormally elevated" and "abnormally lowered".

In combination with these data, a 4*5 (a x b) state text semantic relationship network is constructed. The grid matrix shows the relationship between each thermodynamic parameter and the text semantic tags for different operating states. Such a relational network may more clearly characterize the operating condition of each device and identify areas that may require further inspection or adjustment.

In more detail, in the thermal schedule management of a heating system, an operating state feature (or referred to as an operating state text semantic) refers to the description and characterization of the current operating conditions of the system equipment. These features can be identified by analyzing the various data collected to convert complex, quantized operational data into more easily understood textual descriptions or classifications. The following are some examples.

Temperature state description: for example, "normal operation" means that the temperature is within a predetermined range, "overheating" may mean that the temperature exceeds a safety threshold, while "low temperature" may mean that the system fails to reach the desired heating temperature.

Description of pressure level: pressure is also a critical parameter and may be described as "pressure stable", "pressure high" or "pressure low", reflecting the pressure state of the fluid (typically water or steam) in the heating system.

Flow condition description: the state of flow may be described as "sufficient flow", "insufficient flow" or "over-flow", indicating whether the flow of the thermal medium in the system meets the thermal load demand.

Description of device efficiency: for thermodynamic devices such as boilers, the efficiency descriptions may include "efficiency optimization," "normal efficiency," or "reduced efficiency," which help operators understand whether the device requires maintenance or adjustment.

Fault and warning description: when a potential problem is detected, the system may provide descriptions such as "sensor failure," "leak warning," or "equipment shutdown," which are literally summarised of the abnormal status of the equipment.

Description of energy consumption conditions: depending on the energy consumption data of the system, the state may be described as "energy saving mode", "standard consumption" or "high energy consumption", indicating whether the system operation meets the energy efficiency goal.

By converting these quantified parameters into textual semantic descriptions, the operational status features allow non-professionals to quickly understand the current operational status of the system and take corresponding action. In intelligent scheduling systems, such descriptions may also be used to automatically generate reports, raise alarms, or as input to advanced decision support systems.

Step 20: and determining the heat supply scheduling decision feature with the size of 1*b of each piece of thermodynamic parameter data in the thermodynamic equipment operation state data based on the state text semantic relation network with the size of a.b, and obtaining the heat supply scheduling decision feature relation network with the size of a.b.

In step 20, the heat supply scheduling decision feature of the c-th heat supply parameter data in the heat power equipment operation state data is based on the c-th heat supply parameter data being determined to belong to the e-th heat supply scheduling tag in the d-th heat supply scheduling tags, and on the basis of the d-th heat supply scheduling decision feature corresponding to the c-th heat supply parameter data, d is an integer not less than 2, c is an integer not less than 1 and not more than a, and e is an integer not less than 1 and not more than d.

The key terms mentioned in step 20 are explained as follows.

The heat supply scheduling decision feature refers to a key feature for assisting in making a heat supply scheduling decision after analyzing the operation state data of the thermal equipment. These features are extracted from the original thermal parameter data and the operational state text semantics and they are directly related to the scheduling strategy of the heating system.

The heat supply scheduling decision feature relation network is a relation network which is further determined based on a state text semantic relation network, wherein each piece of thermal parameter data is associated with a vector consisting of d heat supply scheduling labels. The matrix of a x b is formed as a whole, wherein each row represents the relationship between one thermal parameter data and all its possible heating schedule decision features.

The heat supply schedule tag is a predefined set of tags, each representing a heat supply schedule state or an operational recommendation to which thermal parameter data may be associated. For example, "increase heating", "decrease heating", "keep the current state", and the like.

Continuing with the example of some intelligent integrated energy technology company, state text semantic mining is completed in step 10, and a state text semantic relationship net with a size 4*5 (i.e., a=4, b=5) is obtained. Now, in step 20, a heating schedule decision feature needs to be determined from this relationship network.

First, d=3 heating schedule tags are set, for example: "increase heating", "decrease heating", "maintain current state". Each thermal parameter data (such as temperature) is determined to belong to one of the three heating schedule tags. For example, if the current temperature is too low, the corresponding heating schedule tag may be "add heating".

Taking the temperature parameter as an example, let c=1, this means the first thermal parameter. Assuming that temperature parameters are associated with 5 run state text semantics in a state text semantic relationship network, heating schedule decision features corresponding to these 5 text semantics are now determined.

Thus, for the temperature parameter, a heating schedule decision feature vector of 15 (i.e. size 1 b) will be generated, where each element is a heating schedule tag, such as: the heat supply is increased, the heat supply is reduced, the current state is kept, the heat supply is reduced, and the heat supply is increased.

Thus, a heating schedule decision feature relation network of 45 (i.e. ab in size) can be obtained for all 4 thermal parameters. The relation network can guide more accurate heat supply dispatching and optimize the performance of the whole heat supply network.

In other possible examples, a state text semantic relationship network has been constructed by step 10. The present objective is to determine the heating schedule decision characteristics corresponding to the respective thermal parameter data.

For example, some intelligent integrated energy technologies limited monitor and record 4 thermal parameters (a=4), respectively: a) temperature, B) pressure, C) flow, D) humidity.

For each parameter, the company sets 3 heat supply schedule tags (d=3), for example for temperature, the tags may be: "suitable", "too low", "too high".

Now consider the 2 nd thermodynamic parameter, namely pressure (c=2). It is assumed that the pressure is determined to belong to the heat supply schedule tag "too high" by the system analysis (e=3, if "suitable", "too low", "too high" are regarded as 1 st, 2 nd, 3 rd tags, respectively, in order).

Next, based on the pressure parameter (the c-th parameter) and its corresponding heating schedule tag (the e-th tag, "too high"), it is necessary to determine the heating schedule decision feature associated therewith. That is, of the 3 heating schedule decision features (because d=3) related to pressure, the feature corresponding to the "too high" state will be focused.

This heating schedule decision feature may include taking steps to reduce pressure, such as reducing heat source supply, opening an overflow valve, or other adjustment mechanism.

Finally, a heating schedule decision feature relation network with a size 4*3 (a x d) is obtained. In this relational network, each row represents a thermal parameter, each column represents a heating schedule tag, and each cell in the grid corresponds to a particular parameter and a decision feature of a particular schedule tag.

By such analysis, specific scheduling strategies can be set for different operating states, thereby optimizing the performance of the heating system.

Step 30: and determining a coordinated scheduling priority list with the size of b in the operation state data of the thermodynamic equipment based on the state text semantic relation network with the size of a and the heat supply scheduling decision feature relation network with the size of a, and obtaining a coordinated scheduling priority list.

The coordinated scheduling priority list is a list for determining the relative importance and the execution sequence of each piece of thermal parameter data when the thermal parameter data are scheduled in the heating system. Based on a state text semantic relation network and a heat supply scheduling decision feature relation network, different thermodynamic parameter data are ordered according to the emergency degree and priority of the thermodynamic parameter data on the whole heat supply system.

The list has a size b x b, where b represents the number of run state text semantics defined in the previous step. The list shows the priority relationships between different operating status features during the heating schedule, and which status features should be prioritized for more efficient energy allocation and use.

Each thermal parameter will have its own list of coordinated scheduling priorities, so if a is the number of thermal parameters, a such lists will be obtained.

Continuing with the case of some intelligent integrated energy technologies, a heating schedule decision feature relation network of size 4*5 (i.e. a=4, b=5) has been obtained in step 20. Now, in step 30, a coordinated scheduling priority list for each thermal parameter needs to be determined.

Taking a temperature parameter as an example, a coordinated scheduling priority list 5*5 is generated according to the heating scheduling decision feature of the parameter (such as "increase heating", "decrease heating", "keep current state", etc.). This list may be as follows:

Priority-increase heating-decrease heating-keep current status.

Heating is increased by-1-3-2.

Reducing heat supply by-4-2-5-..

Maintaining the current state-2-5-3.

...

In this list, the numbers represent priorities, and smaller numbers represent higher priorities. For example, "increase heating" has a higher priority than "keep current state", so at the time of actual scheduling, if there is a suggestion of "increase heating" and "keep current state" at the same time, the system should give priority to "increase heating".

Each thermal parameter will have a similar list, eventually yielding 4 (i.e. a) coordinated scheduling priority lists, which will be used to guide the actual heating system scheduling, ensuring that the system is both efficient and stable.

In addition, in intelligent scheduling management of heating systems, a coordinated scheduling priority list (or called a coordinated scheduling priority matrix) is a tool for determining the relative importance and priority of individual thermal parameter data when scheduling in the entire system. Such a matrix helps an operator or an automated system to determine which operations are most urgent in a particular situation and which can suspend processing.

For example, an intelligent integrated energy technology company is responsible for managing a complex heating network, which includes a plurality of heating stations and corresponding equipment. And creating a coordinated scheduling priority matrix by using the established heat supply scheduling decision feature relation network.

The following is an example of how this matrix may be created.

Setting thermal parameter data: temperature, pressure, flow and humidity, there are therefore 4 thermodynamic parameters (a=4).

For each parameter, based on the analysis in step 20, a heating schedule decision feature associated with each thermal parameter has been obtained. For example, the temperature may be characterized as follows: "too low", "normal", "too high".

Next, each decision feature needs to be given a priority based on how much they affect overall system stability and efficiency. These priorities are organized into a matrix.

Each thermal parameter has a b x b matrix (assuming b=5) with rows and columns representing different aspects of the heating schedule decision characteristics of the parameter, and the values within the cells represent priorities.

Through the priority matrix, which problems are most urgent in the real-time scheduling process can be rapidly identified, and decision and resource allocation are carried out according to the problems, so that the operation efficiency and reliability of the whole heating system are effectively improved.

Step 40: and determining a system scheduling decision score of the thermodynamic equipment operation state data based on the a coordination scheduling priority list.

The system scheduling decision score is a comprehensive evaluation index and is used for measuring the overall scheduling decision effect of the heating system. It scores the scheduling scheme of the whole heating system based on the a coordination scheduling priority lists obtained in the previous step and possibly considering other factors such as energy efficiency, cost, reliability and the like, so as to evaluate the performance and feasibility of the whole heating system.

The score may be determined by quantitative analysis, such as by calculating a priority score for each scheduling action by a specific algorithm, and giving a final system scheduling decision score based on how much they affect overall system operation. This score may help operators identify optimal or suboptimal scheduling strategies, as well as where improvements are needed.

Step 40 is further specifically explained with respect to a case of a certain intelligent integrated energy technology company.

For example, the construction of the state text semantic relation network, the heating schedule decision feature relation network has been completed according to the previous steps, and 4 (i.e. a=4) coordinated schedule priority lists (one list for each parameter, list size 5*5, i.e. b=5) have been generated from these data.

In step 40, the 4 list of cooperative scheduling priorities will be used to comprehensively evaluate the heating schedule decisions for the entire heating station. The scoring process may include the following steps:

(1) Weight distribution: first, a weight is assigned to the importance of the system operation based on each thermodynamic parameter. For example, temperature and pressure may be more critical to stable operation of the system, so they are given higher weights;

(2) Priority weighting: secondly, the priority score in each coordinated scheduling priority list is multiplied by a corresponding weight to reflect the importance of different parameters;

(3) Comprehensive scoring: then, all weighted scores will be aggregated, and a total system scheduling decision score will be calculated. This score may be a percentage, a full score, or any other form to indicate the effectiveness of the current scheduling scheme for the entire heating system.

(4) Decision evaluation: finally, whether the current heat supply schedule is reasonable or not is evaluated according to the system schedule decision score, and whether adjustment is needed to improve efficiency or reduce cost is judged.

For example, if the system scheduling decision score shows that the current heating scheduling scheme is very close to full score, then it means that the current heating scheduling strategy is efficient; if the score is low, the operator is prompted to check and adjust scheduling policies, such as control policies that optimize certain parameters, to promote overall performance of the system.

By the method, the running state of the heating system can be monitored and optimized regularly, and efficient utilization of energy and the heating requirement of a user are met.

By adopting the technical scheme, the state text semantic relation network with a size of a x b and the heat supply scheduling decision feature relation network with a size of a x b for representing each thermodynamic parameter data in the thermodynamic equipment operation state data of the heat supply network thermodynamic station are sequentially obtained by inputting the thermodynamic equipment operation state data of the heat supply network thermodynamic station into a target thermodynamic system scheduling analysis network which completes debugging in advance; and determining a coordinated scheduling priority list of each piece of thermodynamic parameter data in thermodynamic equipment operation state data of the thermodynamic station of the heat supply network through a state text semantic relation network with the size of a and a heat supply scheduling decision feature relation network with the size of a, and determining a system scheduling decision score of the thermodynamic equipment operation state data of the thermodynamic station of the heat supply network. Based on the method, the feasibility and the interpretability of the heat supply schedule proposal viewpoint of the heat supply network heat station heat supply equipment operation state data are determined according to the system schedule decision score of the heat supply network heat station heat supply equipment operation state data, so that guidance can be provided for subsequent heat supply system coordinated scheduling, the schedule management intelligent degree of the heat supply network heat station is improved, and the heat supply effect is ensured.

In detail, through the above technical scheme, the following beneficial effects can be achieved.

Data driven decision: by converting thermodynamic device operational state data into a state text semantic relationship network and a heating schedule decision feature relationship network, the process utilizes a large amount of data information to guide decisions. The method avoids uncertainty of decision making purely depending on experience or intuition, and provides a more objective and scientific decision making mode based on data analysis.

Enhanced decision interpretability: the complex data information is converted into text semantics and labeled form, so that non-professional staff can understand the running state of the system and interpret the heat supply scheduling decision. This helps to improve operator decision confidence while also facilitating the management layer to monitor and evaluate the rationality of scheduling decisions.

Optimizing resource allocation: by coordinating the scheduling priority list (or matrix), it is possible to ascertain which operations are most urgent and important, thereby making resource allocation more efficient. Such prioritization helps reduce energy waste and improves the energy efficiency ratio of the overall system.

The response speed of the system is improved: in an emergency, the system can quickly respond to the parameter which needs to be adjusted most through a preset coordinated scheduling priority list, so that decision delay time is reduced. Such an automated decision process significantly improves the overall reaction capacity of the system.

Intelligent scheduling management: the automation and intelligence level of the whole process is improved, which means that the dispatching management of the heating power station can process more data and more complex scenes, and meanwhile, the risk of human errors is reduced.

The heat supply effect is ensured: since the scheduling suggestions are generated based on the results of the detailed analysis, this ensures that the heating effect can meet the needs of the user. For example, energy waste due to excessive heat supply or a decrease in comfort due to insufficient heat supply is avoided.

System scheduling decision scoring: by giving a systematic scheduling decision score, the validity of each scheduling decision can be quantitatively assessed. This helps to continuously improve the scheduling algorithm, improving the stability and reliability of the overall thermodynamic system.

Feasibility and interpretability of heating schedule advice: the heating schedule suggestions are generated based on the system scores and the priority list, and the feasibility and the interpretability of the heating schedule suggestions are verified. These suggestions can be directly used as a decision to perform or can be used as a basis for further analysis and optimization of the heating system.

In some possible embodiments, determining the heat supply scheduling decision feature with the size 1*b of each piece of thermal parameter data in the thermal equipment operation state data based on the state text semantic relation network with the size a×b described in step 20, and obtaining the heat supply scheduling decision feature relation network with the size a×b, including steps 21-22.

Step 21: determining heat supply dispatching suggestion viewpoints of all thermal parameter data in the thermal equipment operation state data based on the state text semantic relation network with the size of a x b and the heat supply dispatching decision feature relation network with the size of d x b to obtain a heat supply dispatching suggestion viewpoints; the heat supply schedule suggestion viewpoint of the c-th heat supply parameter data in the heat equipment operation state data is used for representing that the c-th heat supply parameter data belongs to the e-th heat supply schedule label in the d-th heat supply schedule label, the heat supply schedule suggestion viewpoint of the c-th heat supply parameter data in the heat equipment operation state data is determined based on operation state text semantic knowledge of a size 1*b used for representing that the c-th heat supply parameter data and heat supply schedule decision feature of a size 1*b corresponding to the e-th heat supply schedule label, the state text semantic relation network of the size a×b comprises operation state text semantic knowledge of a size 1*b used for representing that the c-th heat supply parameter data, and the heat supply schedule decision feature relation network of the size d×b comprises heat supply schedule decision feature of a size 1*b corresponding to the e-th heat supply schedule label.

Step 22: determining a heat supply scheduling decision feature with a size of 1*b corresponding to each of the a heat supply scheduling suggestion viewpoints as a heat supply scheduling decision feature with a size of 1*b of each of the thermodynamic parameter data in the thermodynamic equipment operation state data; wherein a heating schedule decision feature of size 1*b corresponding to a c-th heating schedule advice point of the a-th heating schedule advice point of the thermal plant operating state data is a heating schedule decision feature of size 1*b corresponding to the e-th heating schedule label.

In step 21, the "heating schedule advice point" is an operation advice for each thermal parameter derived from the textual semantic knowledge of the operating state and the established heating schedule decision feature relation network. Briefly, this step is to give targeted scheduling suggestions in combination with the current specific operating data of the device (e.g., temperature, pressure, etc.), and semantic understanding of such data (e.g., temperature "too high" or "appropriate").

For example, a smart integrated energy technology company manages a heating system that includes a plurality of thermal parameters, a=4 for four thermal parameters (e.g., temperature, pressure, flow, and humidity), and b=5 for five different operational state text semantics (e.g., "too low", "slightly low", "normal", "slightly high", "too high"). A network of heating schedule decision characteristics has been established with d x b, where d may represent the number of heating schedule tags, e.g. d=3 for three scheduling tags (e.g. "decrease heating", "keep current heating", "increase heating").

In practical application, when the running state of the c-th thermal parameter (such as the 1 st parameter, i.e. the temperature) is monitored to be "slightly higher", based on the state text semantic relation network, it can be judged which heating schedule label this state belongs to, and the assumption is "heating increase". Then, based on the heating schedule decision feature corresponding to "add heating" (e.g., turn on additional heat sink facilities), a heating schedule recommendation perspective of the temperature is determined.

Step 22 further translates each heating schedule suggestion viewpoint into a specific heating schedule decision feature. These decision features directly affect the actual operation and control strategy of the heating system.

In the above example, when a heating schedule recommendation perspective for a temperature parameter is determined (e.g., "turn on additional heat sink"), this recommendation perspective is taken as a final heating schedule decision feature for the temperature. This means that in actual operation, when the temperature monitoring result shows "slightly higher", the control system will automatically perform the corresponding operation, such as starting the fan or opening the window to cool down.

For example, an intelligent integrated energy technology company has a heating network managed by an intelligent control system. Companies want to optimize their heating efficiency and ensure user comfort. Temperature, pressure, flow and humidity data of each heating station are collected by using a sensor, and a state text semantic relation network and a heat supply scheduling decision feature relation network are constructed by using an advanced analysis tool.

In a typical winter, the system monitors the following conditions:

Temperature (1 st parameter): the monitored value is slightly higher than the ideal range, so the semantic tag is "slightly higher";

Pressure (parameter 2): in an ideal state, the semantic tag is "normal";

Flow (parameter 3): the monitoring value is lower than the ideal range, and the semantic label is 'too low';

Humidity (4 th parameter): slightly beyond the ideal range, the semantic tag is "slightly higher".

Based on the state text semantic relationship net, the intelligent control system of the company first determines the heat supply scheduling proposal viewpoints corresponding to the semantic tags:

temperature: "open additional heat sink;

Pressure: "keep current heating setting";

Flow rate: "increase Hot Pump speed";

humidity: "start dehumidification procedure".

The intelligent control system then translates these suggested perspectives into specific heating schedule decision features and automatically adjusts the system settings in response to step 22:

the temperature monitoring system activates a fan to reduce the temperature within the thermal station;

the pressure monitoring system confirms that adjustment is not needed, and original setting is kept;

the flow monitoring system increases the output of the hot water pump and improves the hot water flow of the heating network;

The humidity monitoring system activates the dehumidifier to reduce the humidity in the air.

Thus, the high-efficiency operation of the heating system is ensured, the comfort level of a user is improved, the energy consumption is reduced through intelligent optimization and adjustment, and the energy conservation and the cost effectiveness are realized.

In some exemplary embodiments, determining the heat supply schedule advice point of view of each thermal parameter data in the thermal plant operation status data based on the state text semantic relation network with a×b size and the heat supply schedule decision feature relation network with d×b size described in step 21, to obtain a heat supply schedule advice point of view, including step 211-step 212.

Step 211: inputting the state text semantic relation network with the size of a x b into a target heating power system dispatching analysis network for completing debugging to obtain a target heating power dispatching proposal output branch, and obtaining a x d initial heating power dispatching proposal views; the target heat supply schedule proposal output branch determines the a.d initial heat supply schedule proposal viewpoints based on the state text semantic relation network with the size of a.b and a set heat supply schedule decision feature relation network with the size of d.b, d initial heat supply schedule proposal viewpoints corresponding to the c-th heat supply parameter data in the a.d initial heat supply schedule proposal viewpoints represent the possibility that the c-th heat supply parameter data belongs to each heat supply schedule label in the d heat supply schedule labels, and d initial heat supply schedule proposal viewpoints corresponding to the c-th heat supply parameter data are respectively determined based on the running state text semantic knowledge with the size of 1*b for representing the c-th heat supply parameter data and the heat supply schedule decision feature relation network with the size of d.b.

Step 212: among d initial heat supply schedule advice views corresponding to each thermodynamic parameter data in the thermodynamic equipment operation state data, determining the initial heat supply schedule advice view with the highest characterization possibility as the heat supply schedule advice view of each thermodynamic parameter data in the thermodynamic equipment operation state data; the heat supply schedule advice view of the c-th heat supply parameter data is the initial heat supply schedule advice view with highest possibility of being characterized in d initial heat supply schedule advice views corresponding to the c-th heat supply parameter data, and the c-th heat supply parameter data in the d heat supply schedule tags has highest possibility of belonging to the e-th heat supply schedule tag.

To further understand and illustrate the scheme described in step 21, the description of the related terms will be continued with the "certain intelligent integrated energy technologies company" as a case background.

State text semantic relationship net: this is a network structure obtained by analyzing the thermodynamic device operating state data, which expresses the correlation between the various thermodynamic parameters and their operating states. In this network, each node represents a textual description of a thermodynamic parameter or its state, while the edges represent relationships between different nodes.

Heating schedule decision feature relation network: this is a network defining the relationship between the individual heat supply schedule tags and the thermal parameters. It makes possible scheduling decisions for each parameter based on the actual operating state of the thermal parameter and the intended target.

Heat supply schedule advice point of view: this is a specific operational recommendation based on the current system state, which directs how to adjust the operation of the thermal plant to meet the heating demand and ensure system stability.

Now, steps 211 and 212 will be described in detail by the following examples:

For example, an intelligent integrated energy technology company manages a heating grid heat station that contains a plurality of thermal parameters (a=4, such as temperature, pressure, flow, humidity). A state text semantic relation net with a size 4*5 is obtained through state text semantic mining, and a heat supply scheduling decision feature relation net with a size d x b is established based on the data, wherein d is possibly the number of heat supply scheduling labels, and b is the number of text semantics.

When the heat supply scheduling is needed, the state text semantic relation network is input into a target heat supply scheduling proposal output branch in a target thermodynamic system scheduling analysis network after debugging. The output branch analyzes the current thermodynamic parameter state and combines the heat supply dispatching decision feature relation network to generate a.d initial heat supply dispatching proposal views. Each thermal parameter corresponds to d initial suggested views reflecting the likelihood that the parameter belongs to each heating schedule tag.

For example, for this parameter of temperature (c=1), if there are 3 heat supply schedule tags (d=3), such as "too low", "normal", "too high", the system will generate 3 initial heat supply schedule advice views, corresponding to the actions that should be taken in case of too low, moderate and too high temperature, respectively.

Next, one final heating schedule proposal viewpoint will be selected for each thermal parameter among the a x d initial heating schedule proposal viewpoints. This suggestion is the one of the d initial suggestions that has the highest likelihood.

Still taking temperature as an example, if the likelihood of the current temperature being "too low" is found to be highest after analysis, the final heating schedule advice viewpoint will be advice for the "too low" case, such as increasing the heating capacity to raise the temperature to the normal range.

Through the steps, a series of accurate heat supply schedule suggestions can be obtained, and each suggestion is based on the current state and historical data of the thermal parameters. These suggestions are used to adjust the operating state of the thermal plant in real time, optimize the heating efficiency, reduce the energy waste and ensure that the user's heating needs are met.

In some alternative embodiments, the system scheduling decision score of the thermal plant operational status data is determined in step 40 based on the a-coordination scheduling priority lists, including steps 41-42.

Step 41: and executing heat supply coordination feature mapping on the coordination scheduling priority list corresponding to each thermal parameter data in the a coordination scheduling priority lists to obtain a heat supply coordination feature values.

Step 42: and determining a system scheduling decision score of the thermodynamic equipment operation state data according to the a heat supply coordination characteristic values.

To further explain how the system scheduling decision scores for the thermodynamic device operational status data are determined in the intelligent heat supply scheduling system, steps 41 and 42 will be described in detail and key terms therein will be explained. This process is a very important part of the overall intelligent heating management system, which allows optimizing the heating scheduling decisions in a data driven manner.

In step 41, the "heating coordination feature map" refers to a process of mapping each item in the coordination schedule priority list to a specific numerical value, that is, a "heating coordination feature value". This mapping process takes into account the differences between the current operating conditions of the various thermal parameters and the target conditions they should achieve, and the impact of such differences on the overall operating efficiency and safety of the system.

For example, "some intelligent integrated energy technologies limited" has four thermal parameters a=4 (as described above, possibly temperature, pressure, flow and humidity), and each parameter has a corresponding list of coordinated scheduling priorities. When the real-time monitoring data come from the heating station, the system of the company calculates the heat supply coordination characteristic value of each heating power parameter according to preset rules and algorithms. For example, if the temperature is slightly higher, the system may give a relatively low characteristic value based on the coordinated scheduling priority list of temperatures, as this indicates that cooling is required, but not an emergency.

Next, in step 42, a "system scheduling decision score" is a comprehensive evaluation index that measures the overall effectiveness of the current scheduling scheme of the heating system based on all of the heating coordination characteristic values. This score may help the company identify the most efficient scheduling policy and provide a reference for continued improvement.

Specifically, the system will use a specific weighting and summarization method to calculate a total score in combination with each heating coordination characteristic value obtained in step 41. This summary reflects the operating status of the entire heating station or heating network under the current conditions, as well as the corresponding urgency and adjustment requirements of the scheduling scheme.

For example, an intelligent integrated energy technology company is responsible for an urban heating network comprising a plurality of heating stations. Advanced sensor and monitoring technology is adopted to continuously collect data on parameters such as temperature, pressure, flow and humidity. Each thermal parameter has a coordinated scheduling priority list that contains all possible states from most urgent to least urgent.

At a particular moment, the following are real-time data of four thermal parameters and the heating coordination characteristic values calculated by the system:

Temperature: "slightly higher", eigenvalue 30;

pressure: "normal", eigenvalue 50;

Flow rate: "too low", eigenvalue 20;

Humidity: "slightly higher" feature value 40.

The system then calculates a system scheduling decision score based on these eigenvalues and their weight on overall system performance. If the score indicates that there is a problem with the current heating schedule, such as the characteristic value 20 indicating that the flow is "too low" condition requires urgent adjustment, the system or operator can quickly respond by adjusting the pump speed or opening a backup line to increase the flow.

By such a process, stability and efficient operation of the heating system can be ensured. Meanwhile, the system scheduling decision scoring provides a quantization tool, so that the system scheduling decision scoring can evaluate and improve the heat supply scheduling strategy, finally achieve the aim of energy conservation and emission reduction, improve the service quality and ensure the comfort level of users.

In some preferred embodiments, determining a coordinated scheduling priority list of size b of each thermal parameter data in the thermal equipment operation state data in step 30 based on the state text semantic relation network of size a×b and the heating scheduling decision feature relation network of size a×b, to obtain a coordinated scheduling priority list, including step 300.

Step 300: inputting the state text semantic relation network with the size of a x b and the heat supply scheduling decision feature relation network with the size of a x b to finish the target scheduling decision scoring output branch of debugging to obtain the a coordination scheduling priority list; the target thermodynamic system scheduling analysis network is a neural network obtained by debugging a scheduling decision scoring output network to be debugged through a historical thermodynamic equipment running state data set.

For example, a "smart integrated energy technology company" is responsible for operating a complex heat grid system consisting of a plurality of heat stations, each of which has a plurality of critical heat parameters to monitor and adjust. An intelligent scheduling system is adopted to achieve optimal heating efficiency and customer satisfaction.

A neural network model is developed by a certain intelligent comprehensive energy science and technology company and is used for automatically analyzing historical and current thermodynamic equipment operation state data, so that debugging of a scheduling decision scoring output network is completed. In this way, the neural network learns and predicts the most efficient scheduling strategy.

During operation, when the operation state of each thermal parameter is monitored, for example, the temperature may be in one of five states ("too low", "slightly low", "normal", "slightly high", "too high"), and these states are converted into a state text semantic relationship network and input into the debugged target scheduling decision score output branch.

At the same time, each thermal parameter is also associated with a set of heating schedule decision features, such as temperature adjustment may include "increase heating", "maintain current level", "decrease heating". These features form a heat supply scheduling decision feature relation network and are also input into the scoring output branches.

Through analysis of the scoring output branches, the intelligent comprehensive energy technology limited company obtains a coordinated scheduling priority list of each thermal parameter for each state. For example, if the temperature is "slightly higher" and the pressure is "normal" and the system analysis considers the temperature problem to be more urgent, a higher priority score may be given to instruct the dispatcher or the automated control system to first lower the temperature.

Over time, the neural network-based system can further optimize the coordinated scheduling priority list by continuously learning historical data and real-time feedback, so as to ensure that the thermodynamic device is always adjusted according to the most priority order. Thus, the overall performance of the system is improved, the energy waste is reduced, and the reliability of the heating system and the comfort level of users are increased.

In some alternative embodiments, the method further comprises steps 41-43.

Step 41: on the basis that the historical thermal equipment operation state data set comprises Q historical thermal equipment operation state data, carrying out state text semantic mining on each historical thermal equipment operation state data in the Q historical thermal equipment operation state data to obtain Q state text semantic relation network debugging samples; the number of the thermodynamic parameter data in the operation state data of each historical thermodynamic device is a, the f-th state text semantic relation net debugging sample in the Q state text semantic relation net debugging samples is a state text semantic relation net debugging sample which is obtained by carrying out state text semantic mining on the f-th historical thermodynamic device operation state data in the Q historical thermodynamic device operation state data and has a size of a x b, Q is an integer not smaller than 2, and f is an integer not smaller than 1 and not larger than Q.

Step 42: determining a heat supply scheduling decision feature relation network debugging sample corresponding to each historical thermodynamic equipment operation state data in the Q historical thermodynamic equipment operation state data based on the Q state text semantic relation network debugging samples, and obtaining Q heat supply scheduling decision feature relation network debugging samples; the f-th heat supply scheduling decision feature relation network debugging sample in the Q heat supply scheduling decision feature relation network debugging samples is a heat supply scheduling decision feature relation network debugging sample with a size of a x b determined based on the f-th state text semantic relation network debugging sample; the f-th heat supply scheduling decision feature relation network debugging sample comprises heat supply scheduling decision feature debugging sample samples with sizes of 1*b of all heat parameter data in f-th historical heat equipment operation state data, wherein the heat supply scheduling decision feature debugging sample samples of g-th heat parameter data in f-th historical heat equipment operation state data are heat supply scheduling decision feature debugging sample samples corresponding to h-th heat supply scheduling tags in d-th heat supply scheduling decision features corresponding to g-th heat supply parameter data on the basis that the g-th heat parameter data are judged to belong to the h-th heat supply scheduling tags in d-th heat supply scheduling tags, g is an integer which is not less than 1 and not more than a, and h is an integer which is not less than 1 and not more than d.

Step 43: and debugging the scheduling decision scoring output network to be debugged based on the Q state text semantic relation network debugging samples and the Q heat supply scheduling decision feature relation network debugging samples to obtain the target scheduling decision scoring output branch.

In order to provide an exhaustive explanation and example, how a certain intelligent comprehensive energy science and technology company utilizes historical data to perform state text semantic mining and building of a heat supply dispatching decision feature relation network in a heat supply dispatching system of the intelligent comprehensive energy science and technology company is analyzed step by step, and finally debugging of a dispatching decision scoring output network is achieved.

First, state text semantic mining is a data analysis technique that extracts valuable information and patterns from historical operating data. In this process, the intelligent integrated energy technology company analyzes each of the Q historical thermodynamic device operational state data sets. Each history contains data for a thermal parameter, such as temperature, pressure, etc., associated with b different state text descriptions, such as "normal", "high", or "low".

Q state text semantic relation network debugging samples can be obtained through the mining process, wherein each sample reflects the operation condition of the thermodynamic equipment at a specific historical moment.

For example, assume that the historical data set contains 100 records (q=100), each record involving 4 thermodynamic parameters (a=4), each parameter having 5 possible state descriptions (b=5). The company's analysis tool performs text semantic mining on the data to generate 100 sample instances, each representing a state text semantic relationship network at a particular point in time.

Next, in step 42, "heating schedule decision feature network debug sample" refers to a schedule decision feature sample generated for each thermal parameter data based on the state text semantic relationship network debug sample. These examples demonstrate how thermal parameters should be scheduled and the relationship with the heating schedule tag under different historical thermal plant operating conditions.

In the example, the intelligent integrated energy technology limited company will utilize these state text semantic relationship network debugging samples to generate corresponding heating schedule decision feature relationship network debugging samples. If d heat schedule labels are considered, each thermal parameter will have a set of heat schedule decision feature schedule samples corresponding to it, where each sample reflects the features of the parameter that belong to a particular schedule label.

Finally, in step 43, a "scheduling decision score output network" is the core component of the heating scheduling system, which is responsible for outputting the effect scores of the scheduling decisions. By using Q state text semantic relationship network debug samples and Q heating schedule decision feature relationship network debug samples, a company can debug and optimize this network.

The debugging process involves machine learning algorithms and a large amount of historical data. These data training and validation models are used by the limited smart integrated energy technology community to ensure that the scheduling decision score output network can accurately give a valid score when new operational state data is entered.

For example, heating data of past winter is collected, a model is constructed through the above steps, and it is commissioned. When a new winter starts, the real-time monitored operating state data is fed into the commissioned model, which gives a score according to the current situation. This score reflects the merits of the current heating schedule and helps the operator or the automated control system make quick and accurate adjustment decisions.

Through the systematic method, the intelligent comprehensive energy science and technology limited company can effectively manage the heat supply service, optimize the energy consumption and ensure that a user obtains stable and reliable heat supply. In addition, by continuously analyzing the historical data and adjusting the model, the performance of the intelligent heating system can be continuously improved, and the intelligent heating system caters to the changing heating requirements and environmental conditions.

In some preferred embodiments, the debugging of the scheduling decision score output network to be debugged in step 43 based on the Q state text semantic relation network debugging samples and the Q heat supply scheduling decision feature relation network debugging samples to obtain the target scheduling decision score output branch includes performing an ith debugging on the scheduling decision score output network to be debugged by using the following steps 431-434; wherein i is an integer not less than 1.

Step 431: inputting a j-th state text semantic relation network debugging sample and a j-th heat supply dispatching decision feature relation network debugging sample used by the i-th debugging into a dispatching decision scoring output network of the i-th debugging to obtain a coordinated dispatching priority list dispatching sample with the size of b x b of each piece of thermodynamic parameter data in the j-th historical thermodynamic equipment operation state data, and obtaining an a coordinated dispatching priority list dispatching sample; j is an integer which is not less than 1 and not more than Q, the Q state text semantic relation network debugging samples comprise a j state text semantic relation network debugging sample, the Q heat supply scheduling decision feature relation network debugging samples comprise a j heat supply scheduling decision feature relation network debugging sample, and the Q historical thermal equipment operation state data comprise j historical thermal equipment operation state data.

Step 432: and determining the debugging cost variable of the ith debugging based on the j-th state text semantic relation network debugging sample, the j-th heat supply dispatching decision feature relation network debugging sample and the a-th coordination dispatching priority list debugging sample.

Step 433: and on the basis that the debugging cost variable of the ith debugging does not meet the set debugging standard reaching requirement, improving the neural network variable in the scheduling decision scoring output network of the ith debugging to obtain the scheduling decision scoring output network of the (i+1) th debugging.

Step 434: and stopping debugging on the basis that the debugging cost variable of the ith debugging meets the debugging standard reaching requirement, and determining the scheduling decision scoring output network of the ith debugging as the target scheduling decision scoring output branch.

Before explaining how to debug the "certain intelligent integrated energy technologies limited" scheduling decision score output network, several related terms are specified in step 43.

State text semantic relationship net debugging sample: the examples are selected from historical thermodynamic device operation state data, each example comprises state descriptions of different thermodynamic parameters, and semantic information of the examples is expressed in text form. These examples are used for neural network inputs to aid the network in learning to identify and process various states.

Heating schedule decision feature relation network debugging sample: similar to the state text semantic relationship network examples, these examples contain heating schedule decision features related to thermal parameters, which are also input data for network debugging in order to enable the network to understand and predict heating schedule decisions.

Scheduling decision score output network: the method is a model formed by a neural network, and can generate a coordinated scheduling priority list according to an input state text semantic relation network and a heat supply scheduling decision feature relation network and score the coordinated scheduling priority list so as to guide an actual heat supply scheduling decision.

Steps 431 through 434 will now be described in detail, and this process will be illustrated in conjunction with the case of "certain intelligent integrated energy technologies limited".

Suppose a company has Q sets of historical data, each set containing a state text semantic relationship net and a debug sample of a heating schedule decision feature relationship net. At the ith debug, the system will select the jth set of data as input. Based on these inputs, the dispatch decision scoring output network generates a set of dispatch priority list dispatch samples, each representing a thermal parameter tuning priority.

After each debug, the system calculates a debug cost variable that reflects the gap between the current network output's coordinated dispatch priority list and the expected result. For example, if the prioritization of a certain debug sample does not match the actual best order, the debug cost will increase accordingly.

If the debug cost variable does not reach the preset standard, the network is required to be further optimized. In this case, the automatic debugging program adjusts parameters such as weight and bias in the network according to the feedback of the cost function, so as to reduce the debugging cost. Then, the i+1st debug is started.

When the debugging cost variable meets the set standard reaching requirement, the performance of the neural network is good enough, and an effective coordinated scheduling priority list can be accurately generated. At this point, the debugging is stopped and the current network is validated as the target scheduling decision scoring output branch.

A new intelligent heat supply scheduling system is being developed by certain intelligent comprehensive energy science and technology Co. The goal of the company is to automatically optimize the operating state of each thermal station through a neural network model. To achieve this goal, heating data from the past winter was collected, which included the status of different temperatures, pressures, flows and humidities, and corresponding heating schedule decisions.

These historical data are used to debug the neural network. In the first round of commissioning, the neural network may not be able to properly generate a list of suitable coordinated scheduling priorities for each thermal parameter. When the temperature is lower and the traffic is higher, the network does not give the appropriate priority to increase the temperature. Therefore, they adjust the network parameters and re-debug.

After multiple iterations, the performance of the neural network is significantly improved. It is now possible to predict more accurately which thermal parameters are more urgently to adjust under different operating conditions. When the debugging cost reaches an acceptable low level, the debugging is completed, the neural network model is integrated into the intelligent heat supply dispatching system of the company, and the actual dispatching decision is guided. Over time, this model continues to learn and adapt, thereby continually improving heating efficiency and customer satisfaction.

In some possible embodiments, determining the debugging cost variable of the ith debugging in step 432 based on the jth state text semantic relation net debugging sample, the jth heating schedule decision feature relation net debugging sample, and the a-th scheduling priority list debugging sample includes steps 4321-4322.

Step 4321: and determining a heat supply dispatching decision feature dispatching sample with a size of 1*b corresponding to each piece of thermal parameter data in the j-th historical thermal equipment operation state data according to an operation state text semantic knowledge dispatching sample with a size of 1*b corresponding to each piece of thermal parameter data in the j-th historical thermal equipment operation state data, a heat supply dispatching decision feature dispatching sample with a size of 1*b corresponding to each piece of thermal parameter data in the j-th historical thermal equipment operation state data, and a heat supply dispatching priority list dispatching sample with a size of b corresponding to each piece of thermal parameter data in the j-th historical thermal equipment operation state data, wherein the heat supply dispatching decision feature dispatching sample with a size of 1*b corresponding to each piece of thermal parameter data in the j-th historical thermal equipment operation state data, and the heat supply dispatching priority list dispatching sample with a size of 1*b corresponding to each piece of thermal parameter data in the j-th historical thermal equipment operation state data are based on the j-th historical thermal equipment operation state data.

Step 4322: and determining the debugging cost variable of the ith debugging based on the a debugging cost variables.

In order to describe and explain in more detail how "certain intelligent integrated energy technology limited" determines the debugging cost variable through historical data and debugging samples in its heat supply scheduling system, steps 4321 and 4322 will be specifically analyzed, and a practical scheme example will be given in combination with related terms.

Debugging cost variable: in the optimization or debugging process, the debugging cost variable is an index for measuring the difference between the performance of the model and the expected target. It may evaluate the cost paid in the debugging process based on various factors, such as error magnitude, response time, or energy consumption, etc.

Operational state text semantic knowledge sample adjustment: this is the knowledge extracted from the j-th historical thermodynamic device operational state data, including the corresponding state description for each thermodynamic parameter data. The state description for each thermal parameter may be a vector of size 1*b, where b represents the number of possible states.

Heat supply schedule decision feature schedule sample case: similar to the state text semantic knowledge scheduling sample, it represents the scheduling decision feature for each thermal parameter in the j-th historical thermal device operating state data. These features may also be represented as a vector of size 1*b.

Coordination scheduling priority list scheduling sample: this is a matrix of size b x b representing the scheduling priority relationship of different thermal parameters in the j-th history. Each parameter has a priority score for its possible states that is coordinated with the states of the other parameters.

Suppose that a significant update of its heating dispatch system is being performed by a smart integrated energy technology company. In this update, it is desirable to use historical data to optimize the performance of the scheduling decision network.

The first thing to deal with is the debugging cost variable for a single thermal parameter. For the j-th historical thermodynamic equipment operation state data, operation state text semantic knowledge scheduling sample cases, heat supply scheduling decision feature scheduling sample cases and coordination scheduling priority list scheduling sample cases of all thermodynamic parameters in the state need to be checked. These three sample sets would provide enough information for each thermal parameter to calculate its debugging cost variables.

For example, if there are 4 main thermal parameters (a=4) and there are 5 possible state descriptions for each parameter (b=5), then the debug cost variables for each parameter would be calculated separately. This may involve taking into account the deviation between the current state of the parameter and the ideal state, the increased energy consumption that may result from the continued operation of the parameter in the wrong state, and the emergency adjustment measures required according to the coordinated scheduling priority list.

Next, in step 4322, an overall debug cost will be determined based on the debug cost variables for all a thermal parameters obtained in the previous step. This global debug cost variable reflects the global cost required to debug the entire system.

In this process, a weighted average method may be used to comprehensively consider the debugging cost of all parameters, or a more complex mathematical model is used to ensure that high risk or high cost debugging work is properly emphasized.

Finally, through the comprehensive debugging cost variable, a more intelligent decision can be made in the ith debugging, the system performance is optimized, and the resource consumption in the debugging process is controlled.

By the method, the heat supply dispatching system can be continuously improved, so that the heat supply dispatching system is more accurate and economical, and finally, more reliable and efficient heat supply service is provided for customers. In addition, regular system debugging and optimization can also help companies predict and prevent problems which may occur in the future, and ensure long-term stable operation of the heating system.

In some exemplary embodiments, the state text semantic mining is performed on each of the Q pieces of historical thermal device operation state data in step 41 to obtain Q pieces of state text semantic relationship net debug samples, including step 411.

Step 411: inputting all the historical thermodynamic equipment operation state data in the Q historical thermodynamic equipment operation state data to complete target state text semantic mining branches in a target thermodynamic system scheduling analysis network for debugging, so as to obtain Q state text semantic relation network debugging samples; the target thermodynamic system scheduling analysis network is a neural network obtained by debugging the thermodynamic system scheduling analysis network to be debugged through at least part of historical thermodynamic equipment operation state data in the historical thermodynamic equipment operation state data set.

In some exemplary embodiments, the step 42 of determining a heating schedule decision feature relation network debug sample corresponding to each of the Q historical thermal device operation state data based on the Q state text semantic relation network debug samples, to obtain Q heating schedule decision feature relation network debug samples includes a step 421.

Step 421: inputting each state text semantic relation net debugging sample in the Q state text semantic relation net debugging samples into a target heating power system dispatching analysis network for completing debugging, and outputting a target heating dispatching suggestion to obtain Q heating dispatching decision feature relation net debugging samples; the target thermodynamic system scheduling analysis network is a neural network obtained by debugging the thermodynamic system scheduling analysis network to be debugged through at least part of historical thermodynamic equipment operation state data in the historical thermodynamic equipment operation state data set.

In the above embodiments, "some intelligent integrated energy technology limited" is utilizing advanced data mining and neural network technology to optimize its thermodynamic system scheduling. The method aims at extracting useful text semantic information by analyzing historical operation state data and converting the information into decision features which have guiding significance for future heat supply dispatching. The following is a detailed explanation of the key sub-steps 411 and 421 in steps 41 and 42, and a scenario example.

State text semantic mining: this process involves analyzing historical thermodynamic device operational state data, identifying semantic information (e.g., "normal," "high temperature," "low pressure," etc.) related to device performance, and building a textual semantic relationship network.

State text semantic relationship net debugging sample: examples of textual semantic information extracted from each piece of historical thermodynamic device operational state data are used to train and debug a thermodynamic system scheduling analysis network.

Target thermodynamic system scheduling analysis network: the neural network obtained after training and debugging by using the historical thermodynamic equipment operation state data can process the new operation state data and output related text semantic relation networks and heat supply dispatching suggestions.

In step 411, the intelligent integrated energy technology company collects a large amount of historical thermal plant operating state data. And using the data to debug the target thermodynamic system scheduling analysis network. Once debugging is complete, the network can automatically perform state text semantic mining tasks.

For example, there may be a data set (q=5000) containing 5000 records, where each record contains operational status information for a plurality of parameters. The debugged network can be used for analyzing the records one by one, extracting semantic information of the running state of each thermodynamic equipment, and constructing 5000 state text semantic relation network debugging samples. These samples can then be used for further analysis and decision support.

Heating schedule decision feature relation network debugging sample: and generating a set of heat supply scheduling decision features based on the analysis result of the state text semantic relation network debugging sample. These features help define how the various portions of the heating system are adjusted to respond to different operating conditions.

Target heat supply schedule advice output branch: and a part of the target thermodynamic system scheduling analysis network is responsible for generating a heat supply scheduling suggestion according to the input state text semantic relation network debugging sample.

In step 421, the intelligent integrated energy technology limited inputs the 5000 state text semantic relationship net scheduling samples generated in step 411 to the heat supply schedule advice output branch of the target thermodynamic system schedule analysis network. The network analyzes the samples and generates corresponding heating schedule decision feature relation network debugging samples.

For example, when the network analyzes a sample whose status indicates "high temperature", it may infer that it is desired to lower the temperature setting of a certain heat exchanger, or to reduce the heat supply network flow. The network will output these inferences as characteristics of the heating schedule suggestions to assist the operator or the automated system in making more efficient scheduling decisions.

Through the steps, the intelligent comprehensive energy science and technology limited company can create a powerful tool which can not only understand the semantic level of historical operation data, but also provide data-driven suggestions for immediate heat supply scheduling decisions. The method not only improves the heat supply efficiency, but also saves the cost for the company.

On the basis of the above, a thermal schedule management system is provided, comprising a processor and a memory in communication with each other, the processor being adapted to retrieve a computer program from the memory and to implement the above-mentioned method by running the computer program.

On the basis of the above, a computer readable storage medium is provided, on which a computer program is stored, which computer program, when run, implements the method described above.

The foregoing is merely exemplary of the present invention and is not intended to limit the present invention. Various modifications and variations of the present invention will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the invention are to be included in the scope of the claims of the present invention.

Claims

1. A thermal schedule management method for a heating system, the method being applied to a thermal schedule management system, the method comprising:

Performing state text semantic mining on the thermodynamic equipment operation state data of the heat supply network thermodynamic station to obtain a state text semantic relation network with a size of a; wherein a represents the number of thermal parameter data in the thermal equipment operation state data, b represents the number of operation state text semantics of each thermal parameter data in the thermal equipment operation state data, and a and b are integers not less than 2; wherein the state text semantic mining is used for understanding the operation state of the thermodynamic device by analyzing the text description, and comprises the steps of identifying concepts, entities and relations related to the state of the device based on natural language processing technology; the state text semantic relation net is a structure for representing the relation between each parameter and semantic description of the parameter in the operation state data of the thermodynamic equipment, and the state text semantic relation net is a two-dimensional array or matrix;

determining a system scheduling decision score of the thermodynamic device operation state data based on the a coordination scheduling priority list;

the determining, based on the state text semantic relation network with the size of a×b, the heat supply scheduling decision feature with the size of 1*b of each piece of heat distribution parameter data in the heat distribution equipment operation state data to obtain the heat supply scheduling decision feature relation network with the size of a×b, including:

Determining a heat supply scheduling decision feature with a size of 1*b corresponding to each of the a heat supply scheduling suggestion viewpoints as a heat supply scheduling decision feature with a size of 1*b of each of the thermodynamic parameter data in the thermodynamic equipment operation state data; wherein a heating schedule decision feature of size 1*b corresponding to a c-th heating schedule advice point of the a-th heating schedule advice point of the thermal plant operating state data is a heating schedule decision feature of size 1*b corresponding to the e-th heating schedule label;

The determining a heat supply schedule suggestion viewpoint of each thermal parameter data in the thermal equipment operation state data based on the state text semantic relation network with the size of a×b and the heat supply schedule decision feature relation network with the size of d×b, to obtain a heat supply schedule suggestion viewpoints, including:

Inputting the state text semantic relation network with the size of a x b into a target heating power system dispatching analysis network for completing debugging to obtain a x d initial heating power dispatching proposal views; the target heat supply schedule proposal output branch determines a d initial heat supply schedule proposal viewpoints based on a state text semantic relation network with a size of a x b and a set heat supply schedule decision feature relation network with a size of d x b, d initial heat supply schedule proposal viewpoints corresponding to the c-th heat supply parameter data in the a x d initial heat supply schedule proposal viewpoints represent possibility that the c-th heat supply parameter data belongs to each heat supply schedule label in the d heat supply schedule labels, d initial heat supply schedule proposal viewpoints corresponding to the c-th heat supply parameter data are respectively determined based on running state text semantic knowledge with a size of 1*b used for representing the c-th heat supply parameter data and the heat supply schedule decision feature relation network with a size of d x b;

Among d initial heat supply schedule advice views corresponding to each thermodynamic parameter data in the thermodynamic equipment operation state data, determining the initial heat supply schedule advice view with the highest characterization possibility as the heat supply schedule advice view of each thermodynamic parameter data in the thermodynamic equipment operation state data; wherein the heating schedule advice view of the c-th heating parameter data is an initial heating schedule advice view with highest possibility of being characterized in d initial heating schedule advice views corresponding to the c-th heating parameter data, and the c-th heating parameter data in the d heating schedule tags has highest possibility of belonging to the e-th heating schedule tag;

The determining a coordinated scheduling priority list of each thermal parameter data with a size b in the thermal equipment operation state data based on the state text semantic relation network with a size a×b and the heat supply scheduling decision feature relation network with a size a×b, to obtain a coordinated scheduling priority list, including:

Inputting the state text semantic relation network with the size of a x b and the heat supply scheduling decision feature relation network with the size of a x b to finish the target scheduling decision scoring output branch of debugging to obtain the a coordination scheduling priority list; the target thermodynamic system scheduling analysis network is a neural network obtained by debugging a scheduling decision scoring output network to be debugged through a historical thermodynamic equipment running state data set;

The determining a system scheduling decision score of the thermodynamic device operation state data based on the a coordination scheduling priority lists comprises the following steps:

2. The method of claim 1, wherein the method further comprises:

3. The method of claim 2, wherein the debugging the scheduling decision score output network to be debugged based on the Q state text semantic relationship network debug samples and the Q heating scheduling decision feature relationship network debug samples to obtain the target scheduling decision score output branch comprises:

stopping debugging on the basis that the debugging cost variable of the ith debugging meets the debugging standard reaching requirement, and determining the scheduling decision scoring output network of the ith debugging as the target scheduling decision scoring output branch;

The determining the debugging cost variable of the ith debugging based on the jth state text semantic relation network debugging sample, the jth heat supply dispatching decision feature relation network debugging sample and the a coordination dispatching priority list debugging sample comprises the following steps:

4. The method of claim 2, wherein performing state text semantic mining on each of the Q historical thermal device operational state data to obtain Q state text semantic relationship net debug samples, comprises:

5. The method of claim 2, wherein determining the heating schedule decision feature relation network debug samples corresponding to each of the Q historical thermal device operational state data based on the Q state text semantic relation network debug samples, obtaining Q heating schedule decision feature relation network debug samples, comprises:

6. A thermal schedule management system comprising a processor and a memory in communication with each other, the processor being adapted to retrieve a computer program from the memory and to implement the method of any of claims 1-5 by running the computer program.