CN116326654B

CN116326654B - Cooperative control method and system for units in primary tea making process

Info

Publication number: CN116326654B
Application number: CN202310609865.7A
Authority: CN
Inventors: 曹向阳
Original assignee: Weishan Bihechun Biotechnology Co ltd
Current assignee: Weishan Bihechun Biotechnology Co ltd
Priority date: 2023-05-29
Filing date: 2023-05-29
Publication date: 2023-09-19
Anticipated expiration: 2043-05-29
Also published as: CN116326654A

Abstract

The invention relates to the technical field of equipment control, in particular to a cooperative control method and a cooperative control system for each unit in a tea primary making process, wherein the method collects a fixation temperature vector and a stress variation variance sequence of a batch of tea; dividing a batch of tea leaves into at least two small parts, wherein one small part corresponds to one rolling machine, and obtaining the stress fluctuation range of the batch of tea leaves according to the stress distribution vector of one circle of rotation of each rolling machine; the rolling speed of the rolling machine corresponding to the next batch of tea is adjusted by using the stress fluctuation amplitude, and then a stress variation variance sequence of the next batch of tea is obtained; obtaining normal batches of tea leaves and abnormal batches of tea leaves in the batches of tea leaves according to the difference of the stress variation variance sequences; the normal batch of tea leaves and the abnormal batch of tea leaves are utilized to train the two classifiers, and the rolling speed of the rolling machine is controlled through the trained two classifiers. The invention reduces the delay of finding abnormality, stops the damage to the abnormal condition of the tea primary making process in time, and reduces the resource waste.

Description

Cooperative control method and system for units in primary tea making process

Technical Field

The invention relates to the technical field of equipment control, in particular to a cooperative control method and system for each unit in a tea primary making process.

Background

In the past, the primary processing technology of Pu' er tea is basically the same as that of the green tea with stir-frying, and the primary processing technology is divided into three steps of fixation, rolling and drying. Because of some irrational matters in the processing technology, the quality of the tea product is difficult to master, the color of the finished tea is not emerald, the aroma is mixed, the taste and the taste are poor, the product shape is not beautiful enough, and the higher requirements of people on the pan-fired green tea cannot be met.

At present, the thinking of green tea is abandoned at the primary tea making stage of the tea making process, and the quality control of puer tea is studied instead. The quality of the primary tea has a great relation with the production process parameters, and in the enzyme deactivation unit (enzyme deactivation process), the control of the process parameters is required to be enhanced, and the process operation is standardized. The heating temperature in the fixation unit is important to influence on the product, and further the time length of rolling is controlled based on the fixation unit, so that the units are cooperatively controlled to meet the uniform quality of puer tea.

In the primary making stage of the tea making process, when any process parameter of any unit is abnormal, operators hardly find the process directly, and the process continues to produce after the problem occurs, so that equipment is easy to damage, and more tea leaves are disqualified in the later stage. At present, tea images in the primary making process are collected, and the tea images are matched with tea images with standard quality so as to detect whether technological parameters are abnormal or not, and further, the industrial parameters are regulated and controlled so as to ensure the quality of finished tea. However, the abnormality judgment of the method has larger delay, so that the adjustment of industrial parameters is not timely, and great resource waste is caused.

Disclosure of Invention

In order to solve the problem of resource waste in the primary making stage of the tea making process, the invention aims to provide a cooperative control method and a cooperative control system for each unit in the primary making process of tea, and the adopted technical scheme is as follows:

in a first aspect, an embodiment of the present invention provides a method for cooperatively controlling units in a tea primary making process, where the method includes the following steps:

the method comprises the steps of deactivating enzymes of a batch of tea to obtain a deactivated temperature vector, dividing the deactivated batch of tea into at least two small parts, wherein one small part corresponds to one rolling machine, and obtaining a stress distribution vector corresponding to each circle of each rolling machine according to radial stress change of a material disc of each circle of the rolling machine;

setting the number of turns of the rolling machines, equally dividing the number of turns into at least two analysis units, and calculating the stress variation variance among all the rolling machines under the same analysis unit by combining the radial stress variation of each rolling machine corresponding to each turn in each analysis unit, thereby obtaining a stress variation variance sequence of a batch of tea leaves;

obtaining local outlier factors of each circle of stress distribution vector of each rolling machine according to each circle of stress distribution vector of each rolling machine, and further forming a local outlier factor set corresponding to each rolling machine; obtaining stress fluctuation amplitude of a batch of tea leaves according to the local outlier factor set of each rolling machine; the rolling speed of a rolling machine corresponding to the next adjacent batch of tea is adjusted by using the stress fluctuation amplitude of one batch of tea, so that a stress variation variance sequence of the next batch of tea is obtained;

Continuously obtaining stress variation variance sequences of at least two batches of tea leaves, and obtaining normal batches of tea leaves and abnormal batches of tea leaves according to the difference of the stress variation variance sequences between adjacent batches of tea leaves; training the two classifiers by using the de-enzyming temperature vectors corresponding to the normal tea leaves and the abnormal tea leaves and the local outlier factor set of each rolling machine, and controlling the rolling speed of the rolling machine by the trained two classifiers.

Further, the fixation of a batch of tea leaves to obtain a fixation temperature vector comprises:

collecting the temperature of a batch of tea leaves in the enzyme deactivating process based on the sampling frequency to obtain a temperature sequence;

calculating the range and variance of all temperatures in the temperature sequence, calculating the addition result of the range and a preset value, calculating the addition result of the variance and the preset value, and taking the product of the two addition results as a temperature change index;

acquiring the mode of the temperature according to all the temperatures in the temperature sequence, and acquiring a common temperature index by utilizing the mode of the temperature;

calculating the temperature median value and the temperature mean value of all the temperatures in the temperature sequence, and taking the absolute value of the difference value of the temperature median value and the temperature mean value as a temperature bias index;

and forming a batch of fixation temperature vectors of the tea by the temperature change index, the common temperature index and the temperature bias index.

Further, the obtaining the stress distribution vector corresponding to each circle of the rolling machine according to the radial stress change of the material tray of each circle of the rolling machine comprises:

for the process of rotating one circle of any one twisting machine, acquiring the average radial stress of the collecting disc based on the sampling frequency, and obtaining an average radial stress sequence corresponding to the rotating one circle;

calculating the average value of the average radial stress according to all the average radial stresses in the average radial stress sequence, and taking the average value as a first stress distribution index;

obtaining the maximum value and the minimum value of all average radial stresses in the average radial stress sequence, and respectively serving as a second stress distribution index and a third stress distribution index;

calculating the polar difference and variance of the average radial stress according to all the average radial stresses in the average radial stress sequence, calculating the addition result of the polar difference and a preset value, calculating the addition result of the variance and the preset value, and taking the product of the two addition results as a stress variation index;

and forming the stress distribution vector of the blanking disc of the rolling machine rotating for one circle by the first stress distribution index, the second stress distribution index, the third stress distribution index and the stress variation index.

Further, the method for obtaining the stress variation variance sequence of the batch of tea leaves comprises the following steps:

For any analysis unit of any one twisting machine, obtaining the maximum average radial stress of the corresponding ring according to the average radial stress sequence corresponding to each ring in the analysis unit, and obtaining the maximum average radial stress set corresponding to the analysis unit; calculating the average value of all the maximum average radial stresses in the maximum average radial stress set, and recording the average value as a first characteristic value of the analysis unit;

numbering all analysis units of each rolling machine respectively, calculating variances of all first characteristic values under the same number, and marking the variances as stress variation variances among all rolling machines under the corresponding number;

and forming a stress variation variance sequence of a batch of tea leaves by using the stress variation variances under all numbers.

Further, the step of obtaining the stress fluctuation amplitude of a batch of tea leaves according to the local outlier factor set of each rolling machine comprises the following steps:

for any one twisting machine, acquiring a maximum local outlier factor according to a local outlier factor set corresponding to the twisting machine, taking the maximum local outlier factor as the maximum value of the value range of the local outlier factor, and forming the value range of the local outlier factor corresponding to the twisting machine by taking 0 as the minimum value of the value range of the local outlier factor;

Dividing the value range of the local outlier factors into a set number of intervals in an equipartition mode, counting the number of the local outlier factors contained in each interval in a local outlier factor set, and constructing an LOF density map of the twisting machine by taking the interval as an abscissa and the number of the local outlier factors contained in the interval as an ordinate;

performing curve fitting on the local outlier factor quantity corresponding to each interval of the LOF density map to obtain a corresponding fitting curve, and collecting the numerical value corresponding to a first preset number of points on the fitting curve as a characteristic value of the twisting machine;

acquiring all characteristic values of each twisting machine; and obtaining stress error indexes among the rolling machines according to all the characteristic values of each rolling machine, and obtaining the stress fluctuation amplitude of a batch of tea according to the stress error indexes.

Further, the step of obtaining the stress fluctuation amplitude of a batch of tea leaves according to the stress error index comprises the following steps:

arranging all characteristic values of each rolling machine from large to small to form a first characteristic value sequence of each rolling machine; arranging all the characteristic values of each rolling machine according to the position sequence on the fitting curve to form a second characteristic value sequence of each rolling machine;

Taking any one of the rolling machines as a target rolling machine, respectively calculating the L2 distance of a first characteristic value sequence between the target rolling machine and each other rolling machine, calculating an L2 distance average value according to the L2 distances corresponding to all the other rolling machines, and taking the L2 distance average value as a first stress error index of the target rolling machine; respectively calculating cosine distances of a second characteristic value sequence between the target rolling machine and each other rolling machine, calculating cosine distance average values according to the cosine distances corresponding to all the other rolling machines, and taking the cosine distance average values as second stress error indexes of the target rolling machines;

respectively obtaining a first stress error index and a second stress error index of each rolling machine; acquiring the median value and the maximum value of the first stress error index according to the first stress error index of each rolling machine, and taking the ratio of the median value and the maximum value of the first stress error index as an integral first error index; acquiring the median value and the maximum value of the second stress error index according to the second stress error index of each rolling machine, and taking the ratio of the median value and the maximum value of the second stress error index as an integral second error index;

taking the product of the integral first error index and the integral second error index as the stress fluctuation amplitude of a batch of tea leaves.

Further, the obtaining the normal batch of tea leaves and the abnormal batch of tea leaves according to the stress variation variance sequence difference between the adjacent batches of tea leaves comprises the following steps:

calculating the difference value of a stress variation variance sequence between two adjacent batches of tea leaves by using a DTW function, wherein one batch of tea leaves corresponds to two difference values;

taking the first occurrence of continuous second preset number of difference values as a starting point, taking the time when the subsequent difference values become smaller as an ending point, taking the process from the starting point to the ending point as the whole increasing trend, and recording all the difference values under the whole increasing trend as target difference values;

when all the difference values of the two batches of tea leaves corresponding to the target difference value belong to the target difference value, determining that the corresponding batch of tea leaves are abnormal batches of tea leaves; when only one of the two difference values of one batch of tea leaves belongs to the target difference value in the two batches of tea leaves corresponding to the target difference value, determining the batch of tea leaves as normal batch of tea leaves; the tea leaves corresponding to the difference values except the target difference value are taken as the normal tea leaves.

Further, the training of the classifier by using the de-enzyming temperature vector corresponding to the normal batch of tea leaves and the abnormal batch of tea leaves and the local outlier factor set of each rolling machine comprises the following steps:

For normal batches of tea leaves and abnormal batches of tea leaves, the fixation temperature vector of each batch of tea leaves and the LOF density map corresponding to each rolling machine form a high-dimensional characteristic vector of the corresponding rolling machine under the corresponding batch of tea leaves;

taking the high-dimensional characteristic vector of each rolling machine corresponding to the normal batch of tea leaves as a positive sample, and taking the high-dimensional characteristic vector of each rolling machine corresponding to the abnormal batch of tea leaves as a negative sample; the two classifiers are trained using positive and negative samples.

Further, the controlling the rolling speed of the rolling machine by the trained two classifiers includes:

the method comprises the steps of obtaining response values of positive samples input into a second classifier, sorting the response values according to the sequence from large to small, taking the sorted preset proportional response values, and calculating an average value to serve as a response threshold;

the method comprises the steps of obtaining high-dimensional feature vectors of each rolling machine under real-time batch tea, inputting the high-dimensional feature vectors into two trained classifiers to obtain corresponding real-time response values, and reminding people to check the tea quality when the real-time response values are larger than or equal to response thresholds; and when the real-time response value is smaller than the response threshold value, continuously regulating the speed of the rolling machine and rolling.

In a second aspect, the embodiment of the invention also provides a cooperative control system for each unit in the tea primary making process, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to realize the steps of any one of the methods.

The embodiment of the invention has at least the following beneficial effects:

in order to intuitively embody the quality condition of tea leaves in a primary making process, firstly, a fixation temperature vector of a batch of tea leaves in a fixation process is collected, and then, for a batch of fixation tea leaves, a stress variation variance sequence of the batch of tea leaves under the set number of turns is obtained through radial stress variation of a charging tray of the batch of tea leaves in each turn of different rolling machines, so that the integral stress characteristics of the batch of tea leaves are reflected; in order to analyze the stress change condition of the twisting machines in the process of twisting tea leaves, so as to check whether the quality of the corresponding small tea leaves is in mismatch, according to the stress distribution vector obtained by the radial stress change of the material tray in the process of each twisting machine rotating round, a corresponding local outlier factor set of each twisting machine is obtained, and the stress fluctuation condition of each twisting machine under the set number of rotating rounds is shown; because the small tea corresponding to each rolling machine is the tea from the same batch after fixation, the stress fluctuation amplitude of a batch of tea is obtained according to the local outlier factor set of each rolling machine, and the stress matching condition of the tea during rolling after fixation of the batch of tea is accurately reflected; in order to ensure the stress matching of the tea leaves after the subsequent rolling machine is rolled and de-enzymed, the rolling speed of the rolling machine corresponding to the next adjacent batch of tea leaves is adjusted by using the stress fluctuation amplitude of one batch of tea leaves, so that the stress variation variance sequence of the next batch of tea leaves is obtained, and the effect of optimizing the quality of primary tea leaves can be achieved; in order to reduce the delay of judging the batch of tea leaves and improve the quality of primary tea leaves, stress variation variance sequences of at least two batches of tea leaves are continuously obtained, normal batch of tea leaves and abnormal batch of tea leaves are obtained according to the stress variation variance sequence differences between adjacent batches of tea leaves, and the classifier is trained by using fixation temperature vectors corresponding to the normal batch of tea leaves and the abnormal batch of tea leaves and local outlier factor sets of each rolling machine, so that the accuracy of training data of the classifier and the accuracy of training effects are ensured; furthermore, the rolling speed of the rolling machine is controlled through the trained two classifiers, multiple attempts in the primary making process are avoided, abnormal delay is reduced, abnormal conditions of the primary making process of tea are prevented from being damaged in time, and resource waste is reduced.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flow chart of steps of a method for cooperatively controlling each unit in a tea primary making process according to an embodiment of the present invention.

Detailed Description

In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of the specific implementation, structure, characteristics and effects of the method and system for controlling each unit in the tea primary making process according to the invention in conjunction with the accompanying drawings and the preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

The invention provides a method and a system for cooperatively controlling each unit in the primary tea making process, which are specifically described below with reference to the accompanying drawings.

Referring to fig. 1, a flowchart of a method for cooperatively controlling each unit in a tea primary making process according to an embodiment of the invention is shown, and the method includes the following steps:

and S001, carrying out fixation on a batch of tea to obtain fixation temperature vectors, dividing the fixation batch of tea into at least two small parts, wherein one small part corresponds to one rolling machine, and obtaining stress distribution vectors corresponding to each circle of rolling machine according to radial stress changes of each circle of material disc of the rolling machine.

Specifically, the embodiment of the invention firstly carries out fixation treatment on a batch of tea, and records fixation temperature information of the tea in the fixation process, and specifically comprises the following steps:

the heating capacity of the heating element is changed according to the temperature change of the heating element in the tea leaf fixation process, the heating quantity of the heating element is changed according to the temperature measured by the temperature measuring element, if the temperature measuring element is not abnormal, the temperature is accurate, the change condition of the heating temperature of the heating element is normal, and if the temperature measuring element is abnormal, the temperature which is reacted to the heating element is wrong, and the temperature which is regulated by the heating element is wrong.

If the heating temperature is abnormal, the quality of the tea leaves in the subsequent rolling process can be changed, so that the temperature change condition of the heating element is collected in the de-enzyming process, and the de-enzyming temperature condition of the tea leaves in the de-enzyming process can be intuitively known.

In the embodiment of the invention, a batch of tea leaves is taken as an example, and the quantity of the batch of tea leaves is set according to the implementation scene of an implementer, so that the invention does not require; the tea de-enzyming is carried out by using the boiler, namely, a batch of tea leaves are put into the boiler for de-enzyming, the temperature in the boiler is collected in the de-enzyming process, and the temperature in the boiler can not be accurately analyzed in consideration of directly reading the temperature in the boiler. Further, by collecting the temperature of the heating element in the de-enzyming process, a temperature sequence is obtainedWherein, the method comprises the steps of, wherein,the temperature collected for the 1 st turn of the boiler,for the temperature collected for the 2 nd turn of the boiler, For the temperature collected in the 3 rd turn of the boiler,the temperature acquired for the t-th turn of the boiler is t, and the total number of turns of the boiler in the de-enzyming process is represented by t.

Based on the temperature sequence I acquired in the de-enzyming process, the element difference in the temperature sequence I is analyzed to detect the working state of the heating element, and the method specifically comprises the following steps:

calculating the polar differences and variances of all temperatures in a temperature sequence, calculating the addition results of the polar differences and preset values, calculating the addition results of the variances and the preset values, taking the product of the two addition results as a temperature change index, and adopting a specific calculation formula of the temperature change index as follows:

wherein,,is the temperature change index;is a polar difference function;as a variance function; i is a temperature sequence; 1 is a preset value.

It should be noted that, the temperature fluctuation condition in the de-enzyming process is represented by the extremely poor and variance of the temperature sequence, the aim is that when the temperature measuring element has a problem, the read temperature value is high and low, the fluctuation is large, and the temperature under normal conditions has large deviation, at the moment, the temperature measuring element feeds back to the heating element, so that the heating temperature of the heating element also changes along with the change, and under normal conditions, the temperature fluctuation in the de-enzyming process is very small, and the temperature fluctuation is almost in a stable range, so that the temperature change index W1 can reflect the temperature characteristic in the de-enzyming process, and the influence of the de-enzyming temperature on the tea quality is indirectly represented.

According to all temperatures in the temperature sequence, the mode of the temperature is used for obtaining the common temperature index, when only one mode exists, the mode of the temperature is used as the common temperature index W2 in the de-enzyming process, namelyWhereinAs a mode function; when a plurality of modes appear, calculating the average value of all temperatures in the temperature sequence, and taking the mode corresponding to the minimum absolute value of the difference value of the average value as a common temperature index W2. The mode of the temperature refers to the temperature with the largest occurrence number in the temperature sequence, and the mode of the temperature can reflect the temperature condition obtained by the temperature measuring elementWhen the temperature of the temperature measuring element is lower than the normal value, the value of the common temperature index W2 is smaller.

Calculating a temperature median value and a temperature mean value of all temperatures in a temperature sequence, taking the absolute value of the difference value of the temperature median value and the temperature mean value as a temperature bias index in the water-removing process, and reflecting the bias of the temperature data distribution in the water-removing process, wherein the calculation formula of the temperature bias index is as follows:

wherein,,is a temperature bias index;is a mean function;is a median function; i is a temperature sequence;to take an absolute function.

In the de-enzyming process of a batch of tea, the temperature change index W1, the common temperature index W2 and the temperature bias index W3 are respectively obtained according to the temperature sequence acquired in the de-enzyming process, and then the temperature change index, the common temperature index and the temperature bias index form a de-enzyming temperature vector of a batch of tea 。

After fixation, the tea leaves after fixation are required to be placed into a material tray of a rolling machine for rolling, and as the quantity of the tea leaves after fixation is large, the tea leaves after fixation are required to be divided into a plurality of small parts, and the small parts are respectively placed into material trays of different rolling machines for rolling, wherein the tea leaves can be equally divided into a plurality of small parts according to the weight, and the tea leaves can be equally divided into a plurality of small parts according to the volume of the material tray of the rolling machine, and an operator can divide the tea leaves according to own scenes.

A small part is rolled by a rolling machine, the stress characteristics of the material trays in each rolling machine are analyzed according to the weight condition of the material trays in the rolling process of each rolling machine, specifically, if the heating temperature in the de-enzyming process is too high, the tea leaves are heated at high temperature for a long time, the situation of excessive softening can occur, the resistance and the pressure of the material trays in the rolling process are smaller, the load of the material trays for rotating the rolling machine is reduced, and the radial stress of the material trays is also reduced; on the contrary, if the heating temperature in the de-enzyming process is too low, the inner stems of the tea leaves are not fully softened, the leaves are in an unbalanced softening stage, at the moment, the leaves are broken or adhered with other leaves under the pushing of the rolling machine, and the serious situation is likely to be blocked, in the process, the load of the rolling machine is increased, and the radial stress of a charging tray is also increased. Therefore, the radial stress of the material disc can embody the quality of tea leaves during rolling, the radial stress refers to stress along the radial direction, and because the material disc is generally circular, a plurality of radial directions exist, the embodiment of the invention calculates the average radial stress according to the radial stress corresponding to each radial direction to serve as the stress characteristic of the material disc.

Taking any small part of tea leaves as an example, the embodiment of the invention takes one circle of rolling machine corresponding to the small part of tea leaves as an analysis unit, sets the sampling frequency to be 20Hz, and acquires the average radial stress of a material tray in the process of one circle of rolling machine rotation based on the sampling frequency, so that the average radial stress sequence corresponding to the jth circle of rolling process can be obtained，The average radial stress sequence corresponding to the j-th circle,for the average radial stress collected at sample time 1 in the j-th turn,for the average radial stress collected at sample time 2 in the j-th turn,the average radial stress collected at the 3 rd sampling instant in the j-th turn,the average radial stress acquired at the nth sampling time in the jth circle is obtained.

According to the average radial stress condition of the material tray in the process of rotating one circle of the twisting machine, analyzing the stress distribution corresponding to one circle, specifically:

the magnitude of the average radial stress can reflect the basic condition of the tea after being treated by a certain temperature environment in the early enzyme deactivation process, so that the average value of the average radial stress is calculated according to all the average radial stresses in the average radial stress sequence, and the average value is used as a first stress distribution index B1. When the heating temperature is too high, the tea leaves become soft, the accumulated tea leaves are mutually stuffy, further the tea leaves continue to soften and collapse, and the average value of the average radial stress is smaller; on the contrary, if the heating temperature is low, the softening condition of the tea leaves is not obvious, the tea leaves are relatively hard, the resistance in the rolling process is larger, after the tea leaves are twisted together to form a knot, the stress can be further increased, and finally the average value of the average radial stress is larger.

Because the softened and collapsed tea leaves cannot have overlarge stress difference in one circle, but the softening condition is not obvious, and relatively hard tea leaves can have larger difference between the maximum value and the minimum value of stress, the maximum value and the minimum value of all average radial stress in an average radial stress sequence are obtained and are respectively used as a second stress distribution index B2 and a third stress distribution index B3. The second stress distribution index B2 and the third stress distribution index B3 are used for describing the limit range of average radial stress in one circle, so that qualitative description is made for the quality of tea leaves in one circle.

If the temperature in the de-enzyming process is high or low, the heating degree is uneven, so that the heating of the tea leaves is uneven, the final rolling state is changed, the stress is also changed, and the change phenomenon is that softened tea leaves are twisted with crushed insufficiently heated tea leaves. Therefore, according to all average radial stresses in the average radial stress sequence, calculating the polar difference and variance of the average radial stress, calculating the addition result of the polar difference and a preset value, calculating the addition result of the variance and the preset value, taking the product of the two addition results as a stress variation index B4, wherein the larger the value of the stress variation index B4 is, the larger the temperature deviation amplitude measured in the fixation process is, and the calculation formula of the stress variation index B4 is as follows:

Wherein,,is the stress variation index;is a polar difference function;as a variance function;the average radial stress sequence corresponding to the j-th circle; 1 is a preset value.

The quality of the tea after fixation is reflected by the variance and the extremely poor of the average radial stress sequence, the quality of the tea after fixation is correspondingly improved from the variation of the average radial stress of the charging tray in the rolling process, the larger the extremely poor is, the larger the variance is, the larger the corresponding stress variation index is, the quality of the tea after fixation is poorer, and therefore the unstable heating temperature in the fixation process is reflected.

So far, according to the average radial stress change of the material tray in the j-th turn of the rolling machine corresponding to any small tea, a first stress distribution index B1, a second stress distribution index B2, a third stress distribution index B3 and a stress change index B4 are obtained, and then the first stress distribution index and the second stress are obtainedThe distribution index, the third stress distribution index and the stress variation index form a stress distribution vector corresponding to the j-th ring blanking disc of the rolling machine。

Step S002, setting the number of turns of the twisting machines, equally dividing the number of turns into at least two analysis units, and calculating the stress variation variance among all twisting machines under the same analysis unit by combining the radial stress variation of each twisting machine corresponding to each turn in each analysis unit, thereby obtaining a stress variation variance sequence of a batch of tea.

Specifically, the number of the rotating rings of each rolling machine is 220, and the number of the rotating rings is not required and can be set by oneself. Based on the average radial stress sequence of the material tray obtained in each turn of the rolling machine in the step S001, an average radial stress sequence corresponding to each turn of the rolling machine can be obtained, and then an average radial stress sequence corresponding to each turn of the rolling machine in 220 turns of the rolling machine is collected, and one turn corresponds to one average radial stress sequence.

For a batch of tea leaves after fixation, in order to analyze the integral stress characteristics of the batch of tea leaves, the quality of the batch of tea leaves is indirectly reflected according to an average radial stress sequence collected under 220 circles of rotation of each rolling machine, specifically:

for any analysis unit of any one twisting machine, obtaining the maximum average radial stress of the corresponding ring according to the average radial stress sequence corresponding to each ring in the analysis unit, and obtaining the maximum average radial stress set corresponding to the analysis unit; calculating the average value of all the maximum average radial stresses in the maximum average radial stress set, and recording the average value as a first characteristic value of the analysis unit; numbering all analysis units of each rolling machine respectively, calculating variances of all first characteristic values under the same number, and marking the variances as stress variation variances among all rolling machines under the corresponding number; and forming a stress variation variance sequence of a batch of tea leaves by using the stress variation variances under all numbers.

Taking a rolling machine as an example, every 10 turns are taken as an analysis unit, 220 turns correspond to 22 analysis units, namely the first 10 turns of the rolling machine are taken as a first analysis unit, 11-20 turns are taken as a second analysis unit, and the last 10 turns are taken as 22 analysis units. Obtaining the maximum average radial stress of the corresponding ring according to the average radial stress sequence corresponding to each ring, obtaining a maximum average radial stress set corresponding to one analysis unit, and further marking the average value of all the maximum average radial stresses in the maximum average radial stress set as a first characteristic value of the corresponding analysis unit, wherein the calculation formula of the first characteristic value is as follows:

wherein,,the first characteristic value corresponding to the kth analysis unit;the average radial stress sequence corresponding to the 1 st circle in the kth analysis unit;the average radial stress sequence corresponding to the 10 th circle in the kth analysis unit;is a mean function;as a function of the maximum value.

Based on the first characteristic value acquisition method, the first characteristic value corresponding to each analysis unit of each rolling machine is acquired.

For the first analysis unit, acquiring a first characteristic value corresponding to each rolling machine in the first analysis unit, calculating variances of the first characteristic values according to all the first characteristic values, and recording the variances as variances of stress variation among all the rolling machines under the first analysis unit; for the second analysis unit, obtaining a first characteristic value corresponding to each rolling machine in the second analysis unit, calculating variances of the first characteristic values according to all the first characteristic values, recording as stress variation variances among all the rolling machines in the second analysis unit, and the like, so as to obtain stress variation variances among all the rolling machines in 22 analysis units, namely, obtaining 22 stress variation variances, and further forming a stress variation variance sequence, wherein the length of the stress variation variance sequence in the embodiment of the invention is 22.

Step S003, local outlier factors corresponding to the stress distribution vectors of each circle of the rolling machines are obtained according to the stress distribution vectors of each circle of the rolling machines, and then local outlier factor sets corresponding to the rolling machines are formed; obtaining stress fluctuation amplitude of a batch of tea leaves according to the local outlier factor set of each rolling machine; and adjusting the twisting speed of a twisting machine corresponding to the next adjacent batch of tea by using the stress fluctuation amplitude of one batch of tea, so as to obtain a stress variation variance sequence of the next batch of tea.

Specifically, according to step S001, a stress distribution vector of a batch of tea leaves corresponding to each turn of the rolling machine can be obtained. For any one twisting machine, according to the stress distribution vector corresponding to each circle of the twisting machine, the stress distribution vector is taken as a data point, and 220 data points are corresponding to one twisting machine. The L2 distance between the stress distribution vectors is calculated as the distance d between the corresponding two data points, the local outlier factor (LOF value) of each data point is calculated by using the distance d, the local outlier factor (LOF value) of each data point is recorded as the local outlier factor of the corresponding stress distribution vector, the greater the LOF value is 1, the lower the density of the position of the data point is smaller than the density of the position of the neighborhood data point, and the more likely the data point is the abnormal data point. The calculation of the LOF value of each data point by using the distance between the data points belongs to the known technology in the art, and the embodiments of the present invention will not be described in detail.

The larger the difference of stress distribution vectors between two data points, the larger the corresponding distance d, and the calculation formula of the distance d is as follows:

wherein,,distance between data point a and data point b;an L2 distance between the stress distribution vector for data point a and the stress distribution vector for data point b;stress distribution vectors for data point a;the stress distribution vector for data point b.

Obtaining local outlier factors of each data point of the rolling machine, and further forming the local outlier factors corresponding to each data point into a local outlier factor set.

And similarly, according to the acquisition method of the local outlier factor set corresponding to one rolling machine, acquiring the local outlier factor set corresponding to each rolling machine.

Based on the local outlier factor set corresponding to each rolling machine, analyzing the stress change condition of each rolling machine in the tea rolling process so as to check whether the quality of the corresponding small tea is mismatched, specifically:

(1) Constructing LOF density maps of the corresponding rolling machines according to the local outlier factor sets corresponding to each rolling machine, wherein the LOF density maps are local outlier factor density maps; and obtaining a first preset number of characteristic values of the corresponding rolling machines according to the LOF density map of each rolling machine.

Specifically, taking a rolling machine as an example, because the upper and lower bounds of an LOF density chart are empirical values, and the tray of the rolling machine and the bracket thereof have different dead weights and tea types, the upper and lower bounds of the corresponding LOF density chart are also different, so that the maximum local outlier factor is obtained according to a local outlier factor set corresponding to the rolling machine, the maximum local outlier factor is taken as the maximum value of a local outlier factor value range, 0 is taken as the minimum value of the local outlier factor value range, the local outlier factor value range corresponding to the rolling machine is formed, then the local outlier factor value range is divided into a set number of sections in an equipartition mode, the set number of the LOF density chart of the rolling machine is 10 in the embodiment of the invention, the local outlier factor number contained in each section is counted in the local outlier factor set, and the local outlier factor number contained in the section is taken as the ordinate.

And performing curve fitting on the number of local outliers corresponding to each interval of the LOF density map to obtain a corresponding fitting curve, and acquiring the numerical value corresponding to the first preset number of points on the fitting curve as the characteristic value of the rolling machine, wherein the first preset number is 100 in the embodiment of the invention, namely, acquiring the numerical value corresponding to 100 points on the fitting curve as 100 characteristic values of the rolling machine.

(2) Acquiring all characteristic values of each twisting machine; and obtaining stress error indexes among the rolling machines according to all the characteristic values of each rolling machine, and obtaining the stress fluctuation amplitude of a batch of tea according to the stress error indexes.

Specifically, arranging all characteristic values of each rolling machine from large to small to form a first characteristic value sequence of each rolling machine; and arranging all the characteristic values of each rolling machine according to the position sequence on the fitting curve to form a second characteristic value sequence of each rolling machine, wherein each rolling machine has a first characteristic value sequence and a second characteristic value sequence.

Taking any one of the rolling machines as a target rolling machine, respectively calculating the L2 distance of a first characteristic value sequence between the target rolling machine and each other rolling machine, calculating an L2 distance average value according to the L2 distances corresponding to all the other rolling machines, and taking the L2 distance average value as a first stress error index of the target rolling machine.

The calculation formula of the first stress error index is as follows:

wherein,,first stress error finger for the xth twisting machineA number;the total number of other rolling machines except the x-th rolling machine;a first characteristic value sequence for the ith other rolling machine; A first characteristic value sequence for the x-th twisting machine;is the L2 distance of the first eigenvalue sequence between the xth rolling machine and the ith other rolling machine.

It should be noted that, the first characteristic value sequence is the radial stress variation of the tray during the process of twisting the corresponding small parts of tea leaves by each twisting machine, and because the small parts of tea leaves corresponding to each twisting machine are all tea leaves from the same batch after fixation, the quality difference of the tea leaves, namely fixation effect, is indirectly reflected according to the difference between the first characteristic value sequences corresponding to the twisting machinesThe larger the value, the larger the corresponding first stress error index, which indicates that the fixation effect of the small part of tea leaves in the corresponding rolling machine is poorer than that of other small parts of tea leaves.

And similarly, respectively calculating the cosine distances of the second characteristic value sequences between the target rolling machine and each other rolling machine, calculating a cosine distance mean value according to the cosine distances corresponding to all the other rolling machines, and taking the cosine distance mean value as a second stress error index of the target rolling machine.

The calculation formula of the second stress error index is as follows:

wherein,,a second stress error index for the x-th kneader;the total number of other rolling machines except the x-th rolling machine; A second characteristic value sequence for the ith other rolling machine;a second characteristic value sequence for the x-th twisting machine;as a cosine distance function.

It should be noted that, the cosine distance of the second characteristic value sequence between the target rolling machine and all other rolling machines is calculated and is used for indicating the stress influence of the tea leaves on the material tray in the rolling process, the larger the cosine distance is, the larger the second stress error index is, and the different stress distribution among the rolling machines is considered, namely, the inconsistent tea leaf quality after fixation and inconsistent fixation effect are obtained.

According to the method for acquiring the first stress error index and the second stress error index of the target rolling machine, the first stress error index and the second stress error index of each rolling machine are acquired respectively.

The first stress error index of each rolling machine is combined to calculate the integral first error index of all rolling machines, specifically: and obtaining the median value and the maximum value of the first stress error index according to the first stress error index of each rolling machine, and taking the ratio of the median value and the maximum value of the first stress error index as the whole first error index.

The calculation formula of the integral first error index is as follows:

wherein,,first error as a whole An index;is the maximum value of the first stress error index;is the median of the first stress error index;is a first stress error index.

It should be noted that, the ratio of the median value and the maximum value of the first stress error index is used as the overall first error index, so as to analyze the variation situation of the rolling stress distribution of a batch of tea leaves in different rolling machines, and if the stress variation is too early or too late, the quality of the tea leaves is not matched.

Similarly, the second stress error index of each rolling machine is combined to calculate the integral second error index of all rolling machines, and the integral second error index is specifically: and obtaining the median value and the maximum value of the second stress error index according to the second stress error index of each rolling machine, and taking the ratio of the median value and the maximum value of the second stress error index as the whole second error index.

The calculation formula of the integral second error index is as follows:

wherein,,is the integral second error index;is the maximum value of the second stress error index;is the median of the second stress error index;is a second stress error index.

It should be noted that, taking the ratio of the median value and the maximum value of the second stress error index as the overall second error index, analyzing the variation condition of the rolling stress distribution of a batch of tea leaves in different rolling machines, if the stress variation is too early or too late, the quality of the tea leaves is not matched.

And combining the integral first error index and the integral second error index to obtain the stress fluctuation amplitude of the batch of tea, namely taking the product of the integral first error index and the integral second error index as the stress fluctuation amplitude of the batch of tea.

The calculation formula of the stress fluctuation amplitude is as follows:wherein, the method comprises the steps of, wherein,for the stress fluctuation amplitude, the larger the error index is, the larger the stress fluctuation of a batch of tea leaves is, and the larger the corresponding stress fluctuation amplitude is.

Amplitude of stress fluctuationEmbody the stress matching condition of the tea leaves of the corresponding batch after fixation, and in order to ensure the stress matching of the tea leaves after fixation by the subsequent twisting machine, the stress matching is further carried out according to the stress fluctuation rangeAnd adjusting the twisting speed of the twisting machine. The rolling speed of all rolling machines is uniform and set, i.e. the initial rolling speed is fixed, and thus the amplitude of the stress fluctuationsSimultaneously, the rolling speeds of all the rolling machines are adjusted to optimize the quality of primary tea products, and meanwhile, the gradual aging condition of the heating element in the de-enzyming process can be judged in an auxiliary mode.

The rolling speed of the rolling machine when the next batch of tea leaves are rolled is adjusted to be the product value of the stress fluctuation amplitude and the initial rolling speed by using the stress fluctuation amplitude obtained by one batch of tea leaves, so that effective adjustment is achieved.

Under the scene of the rolling speed of the next batch of tea, namely according to the rolling speed adjusted by the previous batch of tea, the stress variation variance sequence of the next batch of tea is obtained according to the methods of the step S001 and the step S002.

Step S004, stress variation variance sequences of at least two batches of tea leaves are continuously obtained, and normal batches of tea leaves and abnormal batches of tea leaves are obtained according to the stress variation variance sequence differences between adjacent batches of tea leaves; training the two classifiers by using the de-enzyming temperature vectors corresponding to the normal tea leaves and the abnormal tea leaves and the local outlier factor set of each rolling machine, and controlling the rolling speed of the rolling machine by the trained two classifiers.

Specifically, by using the method of steps S001-S003, stress variation variance sequences of at least two batches of tea leaves are continuously obtained, when a heating element in the de-enzyming process is gradually invalid, namely, when the reading of a temperature measuring element is abnormal, stress fluctuation of the tea leaves in the twisting process is relatively large, and the fact that the tea leaves in adjacent batches of tea leaves have relatively large differences is indirectly reflected, whether the differences of the quality of the two batches of tea leaves are relatively large or not can be found by comparing the stress variation variance sequences of the adjacent batches of tea leaves, and then all the batches of tea leaves are divided into normal batches of tea leaves and abnormal batches of tea leaves, and the method specifically comprises the following steps: calculating the difference value of a stress variation variance sequence between two adjacent batches of tea leaves by using a DTW function, wherein one batch of tea leaves corresponds to two difference values; taking the first occurrence of continuous second preset number of difference values as a starting point, taking the time when the subsequent difference values become smaller as an ending point, taking the process from the starting point to the ending point as the whole increasing trend, and recording all the difference values under the whole increasing trend as target difference values; when all the difference values of the two batches of tea leaves corresponding to the target difference value belong to the target difference value, determining that the corresponding batch of tea leaves are abnormal batches of tea leaves; when only one of the two difference values of one batch of tea leaves belongs to the target difference value in the two batches of tea leaves corresponding to the target difference value, determining the batch of tea leaves as normal batch of tea leaves; the tea leaves corresponding to the difference values except the target difference value are taken as the normal tea leaves.

As an example, in the embodiment of the invention, a stress variation variance sequence of 11 batches of tea leaves is obtained, the second preset number is 3, 10 differential values can be calculated by using a DTW function, one differential value corresponds to two batches of tea leaves, one batch of tea leaves corresponds to two differential values, namely, one differential value exists between the ith batch of tea leaves and the i-1 th batch of tea leaves, one differential value also exists between the ith batch of tea leaves and the i+1 th batch of tea leaves, the ith batch of tea leaves corresponds to two differential values, one differential value exists between the ith+2 th batch of tea leaves and the i+1 th batch of tea leaves, the ith+1 th batch of tea leaves corresponds to two differential values, and the 11 th batch of tea leaves corresponds to 10 differential values under the condition that one differential value corresponding to the ith batch of tea leaves and one differential value corresponding to the i+1 th batch of tea leaves exist; if the difference value is always in an increasing trend, the heating element in the de-enzyming process is gradually disabled, so that the quality of the corresponding batch of tea leaves is greatly abnormal, namely the de-enzyming process is abnormal, and then the stress of the rolling process is abnormal, the continuous 3 difference values are used as starting points when the increasing trend appears for the first time, the subsequent difference values become small as end points, the process from the starting point to the end point is used as the whole increasing trend, and all the difference values in the whole increasing trend are recorded as target difference values. When all the difference values of the two batches of tea leaves corresponding to the target difference value belong to the target difference value, determining the corresponding batch of tea leaves as abnormal batch of tea leaves, and when only one of the two difference values of one batch of tea leaves in the two batches of tea leaves corresponding to the target difference value belongs to the target difference value, determining the batch of tea leaves as normal batch of tea leaves; the tea leaves corresponding to the difference values except the target difference value are taken as the normal tea leaves.

In order to reduce the delay of judging the batch of tea leaves and improve the quality of primary tea leaves, the classifier is trained by utilizing the fixation temperature vectors corresponding to the normal batch of tea leaves and the abnormal batch of tea leaves and the local outlier factor set of each twisting machine, and specifically comprises the following steps: for normal batches of tea leaves and abnormal batches of tea leaves, the fixation temperature vector of each batch of tea leaves and the LOF density map corresponding to each rolling machine form a high-dimensional characteristic vector of the corresponding rolling machine under the corresponding batch of tea leaves; taking the high-dimensional characteristic vector of each rolling machine corresponding to the normal batch of tea leaves as a positive sample, and taking the high-dimensional characteristic vector of each rolling machine corresponding to the abnormal batch of tea leaves as a negative sample; the two classifiers are trained using positive and negative samples.

As an example, since there is one de-enzyming process and one twisting process per batch of tea leaves and one twisting process is performed by using a plurality of twisting machines, one de-enzyming temperature vector W corresponding to the batch of tea leaves and a local outlier factor set corresponding to one twisting machine can be obtained. Because the LOF density map of each rolling machine is constructed by utilizing the local outlier factor set in the step S003 and is used for representing the stress change condition of each rolling machine in the process of rolling tea, whether the quality of the corresponding small tea is matched or not is indirectly reflected, the high-dimensional feature vector of the corresponding rolling machine is formed by the water-removing temperature vector obtained under one batch of tea and the LOF density map of each rolling machine corresponding to the batch of tea in the embodiment of the invention, the high-dimensional feature vector of each rolling machine corresponding to the normal batch of tea and the high-dimensional feature vector of each rolling machine corresponding to the abnormal batch of tea can be obtained in the same way, and then the high-dimensional feature vector of each rolling machine corresponding to the normal batch of tea is taken as a positive sample, and the high-dimensional feature vector of each rolling machine corresponding to the abnormal batch of tea is taken as a negative sample.

The two classifiers in the embodiment of the invention select an Adaboost two classifier, and the training method of the Adaboost two classifier is mainly related to the construction of positive and negative samples, so that the Adaboost two classifier is trained by utilizing the positive samples and the negative samples, wherein the training method of the Adaboost two classifier points is unique and well known, and the invention is not repeated.

Because the output of the Adaboost second classifier is a response value, the response value of the positive sample input into the Adaboost second classifier is obtained, the response values are ordered from large to small, the ordered preset proportional response values are taken, and the average value is calculated and is used as a response threshold value. The preset proportion of the embodiment of the invention is ten percent of the total response value.

The method comprises the steps of obtaining a high-dimensional feature vector of each rolling machine under real-time batch tea, inputting the high-dimensional feature vector into a trained two-classifier to obtain a corresponding real-time response value, and determining that the quality of the tea in the rolling machine is abnormal when the real-time response value is greater than or equal to a response threshold value, directly reminding people to check the quality of the tea, and not adjusting the rolling speed and carrying out subsequent rolling; otherwise, when the real-time response value is smaller than the response threshold value, the speed of the rolling machine is continuously regulated and the rolling is carried out, so that the tea quality is aligned.

Based on the same inventive concept as the above method, the embodiment of the invention also provides a cooperative control system for each unit in the tea primary making process, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to realize the steps of any one of the cooperative control methods for each unit in the tea primary making process.

It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.

The foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.

Claims

1. The cooperative control method for each unit in the primary tea making process is characterized by comprising the following steps of:

Continuously obtaining stress variation variance sequences of at least two batches of tea leaves, and obtaining normal batches of tea leaves and abnormal batches of tea leaves according to the difference of the stress variation variance sequences between adjacent batches of tea leaves; training the two classifiers by using the de-enzyming temperature vectors corresponding to the normal tea leaves and the abnormal tea leaves and the local outlier factor set of each rolling machine, and controlling the rolling speed of the rolling machines by the trained two classifiers;

wherein, carry out the fixation to a batch of tealeaves and obtain the fixation temperature vector, include:

the temperature change index, the common temperature index and the temperature bias index form a fixation temperature vector of a batch of tea;

The method for obtaining the stress distribution vector corresponding to each circle of the rolling machine according to the radial stress change of the material disc of each circle of the rolling machine comprises the following steps:

the first stress distribution index, the second stress distribution index, the third stress distribution index and the stress variation index form a stress distribution vector of a feeding disc of the rolling machine rotating for one circle;

the method for acquiring the stress variation variance sequence of the batch of tea leaves comprises the following steps:

the stress variation variances under all numbers form a stress variation variance sequence of a batch of tea;

the method for obtaining the stress fluctuation amplitude of a batch of tea leaves according to the local outlier factor set of each twisting machine comprises the following steps:

acquiring all characteristic values of each twisting machine; according to all characteristic values of each rolling machine, stress error indexes among the rolling machines are obtained, and stress fluctuation amplitude of a batch of tea leaves is obtained according to the stress error indexes;

the step of obtaining the stress fluctuation amplitude of a batch of tea leaves according to the stress error index comprises the following steps:

taking the product of the integral first error index and the integral second error index as the stress fluctuation amplitude of a batch of tea leaves;

The rolling speed control of the rolling machine is carried out through the trained two classifiers, and the rolling speed control method comprises the following steps:

2. A method of controlling co-operation of units in a primary tea making process according to claim 1, wherein said obtaining normal and abnormal batches of tea leaves based on differences in stress variation variance sequences between adjacent batches of tea leaves comprises:

3. The method for collaborative control of units in a primary tea making process according to claim 1, wherein training the classifier by using de-enzyming temperature vectors corresponding to normal and abnormal batches of tea leaves and a local outlier factor set of each rolling machine comprises:

4. A coordinated control system of units in a tea primary making process, comprising a memory, a processor and a computer program stored in the memory and running on the processor, characterized in that the processor, when executing the computer program, realizes the steps of a coordinated control method of units in a tea primary making process according to any one of claims 1-3.