CN117052970A - Intelligent control system and method for pneumatic ball valve assembly - Google Patents
Intelligent control system and method for pneumatic ball valve assembly Download PDFInfo
- Publication number
- CN117052970A CN117052970A CN202311150968.8A CN202311150968A CN117052970A CN 117052970 A CN117052970 A CN 117052970A CN 202311150968 A CN202311150968 A CN 202311150968A CN 117052970 A CN117052970 A CN 117052970A
- Authority
- CN
- China
- Prior art keywords
- valve knob
- force
- time sequence
- valve
- force application
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 239000013598 vector Substances 0.000 claims description 198
- 238000005070 sampling Methods 0.000 claims description 51
- 238000009826 distribution Methods 0.000 claims description 14
- 238000000605 extraction Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 10
- 238000005457 optimization Methods 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 5
- 230000008859 change Effects 0.000 description 9
- 238000001914 filtration Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 9
- 230000014509 gene expression Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 230000001419 dependent effect Effects 0.000 description 7
- 238000010606 normalization Methods 0.000 description 7
- 238000013507 mapping Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000012549 training Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 238000007789 sealing Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000009434 installation Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000265 homogenisation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 239000011265 semifinished product Substances 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F16—ENGINEERING ELEMENTS AND UNITS; GENERAL MEASURES FOR PRODUCING AND MAINTAINING EFFECTIVE FUNCTIONING OF MACHINES OR INSTALLATIONS; THERMAL INSULATION IN GENERAL
- F16K—VALVES; TAPS; COCKS; ACTUATING-FLOATS; DEVICES FOR VENTING OR AERATING
- F16K31/00—Actuating devices; Operating means; Releasing devices
- F16K31/12—Actuating devices; Operating means; Releasing devices actuated by fluid
- F16K31/122—Actuating devices; Operating means; Releasing devices actuated by fluid the fluid acting on a piston
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F16—ENGINEERING ELEMENTS AND UNITS; GENERAL MEASURES FOR PRODUCING AND MAINTAINING EFFECTIVE FUNCTIONING OF MACHINES OR INSTALLATIONS; THERMAL INSULATION IN GENERAL
- F16K—VALVES; TAPS; COCKS; ACTUATING-FLOATS; DEVICES FOR VENTING OR AERATING
- F16K37/00—Special means in or on valves or other cut-off apparatus for indicating or recording operation thereof, or for enabling an alarm to be given
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F17—STORING OR DISTRIBUTING GASES OR LIQUIDS
- F17D—PIPE-LINE SYSTEMS; PIPE-LINES
- F17D3/00—Arrangements for supervising or controlling working operations
- F17D3/01—Arrangements for supervising or controlling working operations for controlling, signalling, or supervising the conveyance of a product
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Indication Of The Valve Opening Or Closing Status (AREA)
Abstract
The application discloses an intelligent control system and a method for assembling a pneumatic ball valve, which analyze the degree of tightness of the assembly by acquiring a force value applied to a valve knob acquired by a force sensor so as to intelligently judge whether the degree of tightness of the assembly meets preset requirements.
Description
Technical Field
The application relates to the field of intelligent control, and more particularly relates to an intelligent control system for pneumatic ball valve assembly and a method thereof.
Background
The pneumatic ball valve mainly comprises a valve body, a valve core, a valve cover, a valve knob and a sealing ring. According to the field investigation, on the first workbench, a worker sequentially installs a valve knob, a valve core, a sealing ring and a valve cover into a valve body by hand. The finished semi-finished product is conveyed to the next workbench, the pneumatic ball valve is clamped by manually rotating the clamp, the valve cover is mounted manually or in an auxiliary mode, and meanwhile, a manual wrench is required to reciprocally rotate a valve knob of the pneumatic ball valve to feel the assembly tightness degree. When the valve cover is loosely installed, the sealing element of the pneumatic ball valve does not bear certain installation pretightening force, so that oil leakage phenomenon occurs in the working process; if the valve cover is too tightly installed, the installation pretightening force is too large, the opening and closing of the valve are very difficult, and meanwhile, the damage to the sealing element is also very large.
Because the staff judges whether the installation meets the requirements according to own experience, the error is great, and the requirement on the proficiency of the worker is high. In the prior art, automatic apparatuses such as a mechanical arm can be utilized to replace manual work to rotate the clamp so as to tighten the pneumatic ball valve. However, if an automated approach is desired to replace the traditional manual control, an important problem arises: how the degree of tightness of the assembly should be measured and analyzed. Thus, a solution is desired.
Disclosure of Invention
The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides an intelligent control system and a method for assembling a pneumatic ball valve, which are used for analyzing the degree of tightness of the assembly by acquiring a force value applied to a valve knob acquired by a force sensor so as to intelligently judge whether the degree of tightness of the assembly meets preset requirements.
According to one aspect of the present application, there is provided an intelligent control method for pneumatic ball valve assembly, comprising:
acquiring force values applied to the valve knob at a plurality of predetermined time points within a predetermined time period acquired by the force sensor;
performing data preprocessing on force values applied to the valve knob at the plurality of preset time points to obtain an up-sampling valve knob force application time sequence input vector;
performing feature extraction on the up-sampling valve knob force application time sequence input vector to obtain a valve knob force application time sequence context correlation feature vector; and
based on the valve knob apply force timing context associated feature vector, it is determined whether the assembly tightness meets a predetermined requirement.
According to another aspect of the present application, there is provided an intelligent control system for pneumatic ball valve assembly, comprising:
a data acquisition module for acquiring force values applied to the valve knob at a plurality of predetermined time points within a predetermined time period acquired by the force sensor;
the data preprocessing module is used for preprocessing the data of the force values applied to the valve knob at a plurality of preset time points to obtain an up-sampling valve knob force application time sequence input vector;
the feature extraction module is used for carrying out feature extraction on the up-sampling valve knob force application time sequence input vector so as to obtain a valve knob force application time sequence context correlation feature vector; and
and the result generation module is used for determining whether the assembly tightness degree reaches a preset requirement or not based on the valve knob force application time sequence context correlation characteristic vector.
Compared with the prior art, the intelligent control system and the method for assembling the pneumatic ball valve provided by the application analyze the degree of tightness of the assembly by acquiring the force value applied to the valve knob acquired by the force sensor so as to intelligently judge whether the degree of tightness of the assembly meets the preset requirement.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing embodiments of the present application in more detail with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and together with the embodiments of the application, and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
FIG. 1 is a flow chart of an intelligent control method for pneumatic ball valve assembly in accordance with an embodiment of the present application;
FIG. 2 is a system architecture diagram of an intelligent control method for pneumatic ball valve assembly in accordance with an embodiment of the present application;
FIG. 3 is a flow chart of substep S2 of the intelligent control method for pneumatic ball valve assembly according to an embodiment of the present application;
FIG. 4 is a flow chart of sub-step S3 of the intelligent control method for pneumatic ball valve assembly according to an embodiment of the present application;
FIG. 5 is a flowchart of sub-step S31 of the intelligent control method for pneumatic ball valve assembly according to an embodiment of the present application;
FIG. 6 is a flowchart of substep S4 of the intelligent control method for pneumatic ball valve assembly according to an embodiment of the present application;
FIG. 7 is a block diagram of an intelligent control system for pneumatic ball valve assembly in accordance with an embodiment of the present application.
Detailed Description
Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
As used in the specification and in the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.
A flowchart is used in the present application to describe the operations performed by a system according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
In the prior art, automatic apparatuses such as a mechanical arm can be utilized to replace manual work to rotate the clamp so as to tighten the pneumatic ball valve. However, if an automated approach is desired to replace the traditional manual control, an important problem arises: how the degree of tightness of the assembly should be measured and analyzed. Thus, a solution is desired.
In the technical scheme of the application, an intelligent control method for assembling the pneumatic ball valve is provided. Fig. 1 is a flow chart of an intelligent control method for pneumatic ball valve assembly according to an embodiment of the present application. Fig. 2 is a system architecture diagram of an intelligent control method for pneumatic ball valve assembly according to an embodiment of the present application. As shown in fig. 1 and 2, an intelligent control method for pneumatic ball valve assembly according to an embodiment of the present application includes the steps of: s1, acquiring force values applied to a valve knob at a plurality of preset time points in a preset time period, wherein the force values are acquired by a force sensor; s2, carrying out data preprocessing on the force values applied to the valve knob at a plurality of preset time points to obtain an up-sampling valve knob force application time sequence input vector; s3, carrying out feature extraction on the up-sampling valve knob force application time sequence input vector to obtain a valve knob force application time sequence context correlation feature vector; and S4, determining whether the assembly tightness degree reaches a preset requirement based on the valve knob force application time sequence context correlation characteristic vector.
In particular, in step S1, force values applied to the valve knob are acquired by the force sensor at a plurality of predetermined points in time within a predetermined period of time. It will be appreciated that the force sensor can accurately measure the force value applied to the valve knob, providing an objective measure of the degree of tightness of the assembly. Compared with the traditional subjective judgment, the force sensor can provide accurate numerical data, and errors of artificial judgment are eliminated.
Notably, a force sensor is a device or apparatus for measuring the force exerted by an object. Which is capable of converting the force experienced by an object into a measurable electrical signal or other form of output. Force sensors are widely used in the fields of industry, scientific research, medical treatment, robotics, and the like.
Accordingly, in one possible implementation, the force values applied to the valve knob at a plurality of predetermined points in time within a predetermined period of time acquired by the force sensor may be obtained by, for example: the period of time during which data needs to be acquired, for example from a start time to an end time, is first determined. Then determining a plurality of preset time points according to the requirements; the force sensor is correctly installed at the position of the valve knob. Ensuring that the sensor is in close contact with the valve knob and is able to accurately measure the force exerted on the knob; the force sensor is connected with the data acquisition system. This may involve the use of a suitable cable, interface or signal converter to transmit the output signal of the sensor into the data acquisition system; and (5) performing corresponding setting according to the specification of the sensor and the requirement of a data acquisition system. This may include selecting an appropriate measurement range, sampling frequency, data storage means, etc.; before the preset time period starts, starting the data acquisition system to ensure that the data acquisition system operates normally and is ready to receive the data of the sensor; at a predetermined point in time, the force value data output by the sensor is recorded. An interface or command provided by the data acquisition system can be used to trigger data acquisition and ensure accurate acquisition at each point in time; the collected force value data is stored in a suitable format, such as a database, spreadsheet, or text file. The corresponding relation between the data and the time point is clear and definite; the collected data is analyzed and interpreted as needed. This may include calculating average force values, maximum/minimum force values, force trend, etc.
Specifically, in step S2, the force values applied to the valve knob at the plurality of predetermined points in time are data pre-processed to obtain an up-sampled valve knob force application timing input vector. That is, the time-series discrete distribution of force values applied to the valve knob is converted into a structured vector representation. In particular, in one specific example of the present application, as shown in fig. 3, the S2 includes: s21, arranging the force values applied to the valve knob at a plurality of preset time points into a valve knob force application time sequence input vector according to a time dimension; and S22, carrying out up-sampling processing on the valve knob force applying time sequence input vector to obtain the up-sampling valve knob force applying time sequence input vector.
Specifically, the step S21 is to arrange the force values applied to the valve knob at the plurality of predetermined time points into a valve knob force applying time sequence input vector according to a time dimension. It should be appreciated that the effect of arranging the force values applied to the valve knob at a plurality of predetermined points in time dimension into the valve knob force application timing input vector is to establish a relationship of force values to time so that dynamic information over time can be better understood and utilized in subsequent data analysis and modeling.
Accordingly, in one possible implementation, the force values applied to the valve knob at the plurality of predetermined points in time may be arranged in a time dimension as a valve knob force application timing input vector, for example: determining the unit and the precision of the time dimension; ordering the preset time points according to a time sequence, and ensuring that the preset time points are arranged in an incremental manner according to a time dimension; for each predetermined point in time, corresponding force values are extracted from the raw force value data and arranged together in a time sequence to form a time-sequential input vector.
Specifically, the step S22 performs an up-sampling process on the valve knob force timing input vector to obtain the up-sampling valve knob force timing input vector. It should be appreciated that by upsampling, the time interval in the original timing input vector may be reduced, thereby improving the time resolution. Thus, the change process of the force value can be captured more finely, and finer dynamic characteristics are revealed; the up-sampling can make the force value change smoother, and reduce the jump with larger time interval. This helps to eliminate noise or outlier interference with the model analysis, making the data more reliable and stable.
Notably, upsampling (Upsampling) is a signal processing technique used to increase the sampling rate of a signal or to increase the time resolution of data. In the up-sampling process, the sampling points of the original signal are inserted into new sampling points, thereby increasing the details and accuracy of the signal. In digital signal processing, upsampling typically involves the steps of: interpolation: the sampling rate or time resolution is increased by inserting new sampling points between the original sampling points. Common interpolation methods include nearest neighbor interpolation, linear interpolation, spline interpolation, and the like. The choice of interpolation method depends on the application scenario and signal characteristics; and (3) filtering: after interpolation, the signal typically needs to be filtered to remove high frequency noise or artifacts introduced during interpolation. Common filtering methods include low pass filtering or band-limited filtering to ensure signal smoothness and accuracy. The purpose of upsampling is to increase the detail and accuracy of the signal so that the signal more closely approximates the original continuous signal. It has found wide application in a number of fields including audio processing, image processing, data compression, and the like. In machine learning and deep learning, upsampling is also commonly used to increase the sample size of training data, improving the performance and generalization ability of the model.
Accordingly, in one possible implementation, the valve knob force timing input vector may be upsampled to obtain the upsampled valve knob force timing input vector by, for example: first it is necessary to determine the multiple of the up-sampling, i.e. how many smaller time intervals each time interval in the original input vector is subdivided into; a new point in time is inserted between each time interval of the original input vector using a suitable interpolation method. Common interpolation methods include nearest neighbor interpolation, linear interpolation, spline interpolation, and the like. The purpose of interpolation is to add more data points in the time dimension; after interpolation, filtering is performed to remove high frequency noise or artifacts introduced during interpolation. Common filtering methods include low pass filtering or band-limited filtering to ensure signal smoothness and accuracy; and according to the up-sampling multiple, adjusting the position of the interpolated data point on the time axis, so that the time resolution is higher. Thus, the change process of the force value can be captured more finely; and combining the data points subjected to interpolation and filtering to obtain an up-sampled valve knob force application time sequence input vector.
It is worth mentioning that in other examples of the present application, the force values applied to the valve knob at the plurality of predetermined time points may be data pre-processed to obtain an up-sampled valve knob force timing input vector, for example: acquiring the acquired force value data from a data acquisition system, and ensuring that the corresponding relation between the data and a time point is accurate; the desired up-sampling frequency, i.e. the number of samples per second, is determined. This determines how many times the force value is interpolated in a second; the raw force value data is interpolated using an interpolation algorithm, which is up-sampled from the original point in time to a higher frequency. Common interpolation algorithms include linear interpolation, spline interpolation, and the like. The purpose of interpolation is to fill in reasonable force value data between two points in time to obtain smoother and more continuous time sequential input vectors; and generating a time sequence input vector according to the up-sampled data. The time series input vector is a time series of force values which are arranged in time sequence and can be used as the input of a model. According to specific requirements, a sliding window with fixed length can be selected, and a force value sequence in the window is used as a time sequence input vector; the generated time series input vector is subjected to data normalization processing, and the force value range is mapped to a proper interval, such as [0, 1]. Common normalization methods include min-max normalization and normalization; checking whether the generated time sequence input vector accords with the expected or not, and carrying out necessary correction. For example, whether an abnormal value or a missing value exists or not may be checked, and processing is performed according to actual conditions; the preprocessed time series input vector is stored in a suitable data file for subsequent use.
In particular, in step S3, feature extraction is performed on the upsampled valve knob apply force timing input vector to obtain a valve knob apply force timing context associated feature vector. That is, the pattern of temporal variations and the profile of force values that are embedded in the up-sampling valve knob force application temporal input vector are captured. In the application scenario of the present application, the time-series variation of the valve knob applied force may be represented as a relatively steady state if the degree of tightness of the assembly meets a predetermined requirement. And the valve knob is assembled too loosely, so that the force applied by the valve knob has a larger variation trend; if the assembly is too tight, the force value may be large in value, small in change and the like. In particular, in one specific example of the present application, as shown in fig. 4, the S3 includes: s31, extracting local time sequence characteristics of the up-sampling valve knob force application time sequence input vector to obtain a sequence of up-sampling valve knob force application time sequence input sub-vectors; and S32, extracting correlation features between the sequences of up-sampling valve knob force application time sequence input sub-vectors to obtain the valve knob force application time sequence context correlation feature vectors.
Specifically, the step S31 extracts the local timing characteristics of the up-sampling valve knob force timing input vector to obtain a sequence of up-sampling valve knob force timing input sub-vectors. In particular, in one specific example of the present application, as shown in fig. 5, the S31 includes: s311, vector segmentation is carried out on the up-sampling valve knob force application time sequence input vector so as to obtain a sequence of up-sampling valve knob force application time sequence input sub-vectors; and S312, passing the sequence of up-sampling valve knob force application time sequence input sub-vectors through a force application feature extractor having a multi-scale one-dimensional convolution structure to obtain the sequence of multi-scale time sequence valve knob force application feature vectors.
More specifically, the step S311 performs vector slicing on the up-sampling valve knob force timing input vector to obtain a sequence of up-sampling valve knob force timing input sub-vectors. It will be appreciated that vector slicing the up-sampled valve knob apply force timing input vector can improve time resolution, improve model performance, smooth data, and preserve detailed information to better understand and analyze patterns and trends in force value changes.
Notably, vector slicing refers to the process of dividing a vector into multiple smaller sub-vectors. The purpose of the segmentation is to break down the original vector into smaller parts for more convenient processing, analysis or presentation of the data.
Accordingly, in one possible implementation, the up-sampling valve knob force timing input vector may be vector sliced to obtain a sequence of up-sampling valve knob force timing input sub-vectors, for example: firstly, determining the length or window size of a subvector to be segmented; the step length of the splitting window, namely the distance of each sliding window, is determined. The step size may control the degree of overlap between the sub-vectors. Smaller steps may increase overlap between sub-vectors, while larger steps may decrease overlap; using the determined window size and step size, a sliding window is started from the upsampling valve knob apply force timing input vector. Intercepting corresponding sub-vectors at each window position as a segmented result; and (3) repeating the step 3 until the sliding window covers the whole up-sampling valve knob force applying time sequence input vector according to the step length setting. Ensuring that all time points are contained in the split sub-vectors; and forming a sequence of sub-vectors obtained by segmentation according to the segmentation sequence, namely, a sequence of the sub-vectors input by the force applying time sequence of the up-sampling valve knob.
More specifically, the step S312 is to pass the sequence of up-sampled valve knob apply force timing input sub-vectors through an apply force feature extractor having a multi-scale one-dimensional convolution structure to obtain the sequence of multi-scale timing valve knob apply force feature vectors. In one example, the sequence of upsampled valve knob applied force timing input sub-vectors is input into a first convolution layer of the applied force feature extractor having a multi-scale one-dimensional convolution structure to obtain a sequence of first neighborhood scale time sequence valve knob applied force feature vectors, wherein the first convolution layer has a first one-dimensional convolution kernel of a first length; inputting the sequence of upsampled valve knob applied force timing input sub-vectors into a second convolution layer of the applied force feature extractor having a multi-scale one-dimensional convolution structure to obtain a sequence of second neighborhood scale time sequential valve knob applied force feature vectors, wherein the second convolution layer has a second one-dimensional convolution kernel of a second length, the first length being different from the second length; and cascading the sequence of the first neighborhood scale sequential valve knob application force feature vector with the sequence of the second neighborhood scale sequential valve knob application force feature vector to obtain the sequence of the multi-scale sequential valve knob application force feature vector.
Notably, in conventional one-dimensional convolutional neural networks, a fixed-size convolutional kernel is typically used to extract features. However, patterns and features in one-dimensional data may have different time scales. To better capture these different-scale features, a multi-scale one-dimensional convolution structure introduces a number of different-sized convolution kernels. The basic idea of a multi-scale one-dimensional convolution structure is to use multiple convolution kernels in parallel in the same layer, each convolution kernel having a different size. These convolution kernels may be one-dimensional convolution kernels having different sizes or one-dimensional convolution kernels having different receptive fields. By performing convolution operations on different scales, the multi-scale one-dimensional convolution structure can capture features at different time scales simultaneously. Smaller size convolution kernels may better capture local detail and fast-varying features, while larger size convolution kernels may better capture global trends and long-term dependent features.
It is worth mentioning that in other examples of the present application, the local timing features of the upsampling valve knob force timing input vector may be extracted to obtain a sequence of upsampling valve knob force timing input sub-vectors, for example: it is first necessary to determine the length of each sub-vector, i.e. the time range of the local timing characteristics. This length may be determined according to the specific application requirements and may be a fixed length of time or a fixed number of data points; moving the force time sequence input vector of the up-sampling valve knob by using a sliding window mode, and taking out a window with the length of a subvector each time; for each sliding window, a set of local timing characteristics is calculated to describe the force value changes within that window. Common local timing characteristics include mean, variance, maximum, minimum, slope, etc. These features may provide statistics and trends regarding force value changes; the calculated local timing characteristics are combined to form a sequence of up-sampling valve knob force application timing input sub-vectors. Each sub-vector corresponds to a sliding window, and the sub-vector contains local time sequence characteristic information describing the change of the force value.
Specifically, the step S32 extracts correlation features between the sequence of up-sampling valve knob force application timing input sub-vectors to obtain the valve knob force application timing context correlation feature vector. That is, in the technical solution of the present application, the sequence of the multi-scale time series valve knob force application feature vectors is passed through a transducer-based valve knob force application inter-correlation feature extractor to obtain the valve knob force application time series context correlation feature vector. In one example, the sequence of multi-scale sequential valve knob application force feature vectors is one-dimensionally arranged to obtain a global multi-scale sequential valve knob application force feature vector; calculating the product between the global multi-scale time sequence valve knob application force characteristic vector and the transpose vector of each multi-scale time sequence valve knob application force characteristic vector in the sequence of the multi-scale time sequence valve knob application force characteristic vectors to obtain a plurality of self-attention correlation matrices; respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; weighting each multi-scale time sequence valve knob application force characteristic vector in the sequence of multi-scale time sequence valve knob application force characteristic vectors by taking each probability value in the plurality of probability values as a weight so as to obtain the plurality of context semantic multi-scale time sequence valve knob application force characteristic vectors; and cascading the plurality of context semantic multi-scale sequential valve knob apply force feature vectors to obtain the valve knob apply force time sequence context correlation feature vector.
It is worth mentioning that in other examples of the present application, the correlation feature between the sequence of up-sampled valve knob apply force timing input sub-vectors may be extracted to obtain the valve knob apply force timing context correlation feature vector, for example: first, a method for extracting correlation features between sub-vector sequences is determined. Common methods include statistical feature computation, timing models (e.g., cyclic neural networks, long and short term memory networks), convolutional neural networks, and the like; the sequence of upsampling valve knob applied force timing input sub-vectors is divided into fixed size windows. The size of the window may be determined according to specific requirements, typically selected according to the number of sub-vectors and the time span; for the sequence of sub-vectors within each window, the associated features are computed using the selected associated feature extraction method. These features may include statistics (e.g., mean, standard deviation, maximum, minimum), hidden states of the sequence model, output of convolutional neural network, etc.; the associated features are associated with the sub-vectors corresponding to the window center to establish a contextual association. The context correlation may be achieved using different methods, such as stitching, weighted addition, or other means of correlating features with the center subvector; for each window, an associated feature vector is generated based on the contextual relevance of the resulting features. This feature vector may consist of the features of the window center sub-vector and other features associated therewith; and repeating the steps of window division, feature calculation, context association and feature vector generation according to the step length of the window until the whole sub-vector sequence is covered.
It is worth mentioning that in other examples of the present application, the upsampling valve knob force timing input vector may be feature extracted to obtain a valve knob force timing context associated feature vector, for example, by: the upsampled valve knob apply force timing input vector is divided into fixed size time windows. The size of the time window can be determined according to specific requirements, and the common window size comprises a fixed time length or a fixed data point number; for each time window, a set of features is computed to describe the force value changes within that window. Common features include mean, variance, maximum, minimum, slope, etc. These features may provide statistics and trends regarding force value changes; the context-dependent feature may be introduced in view of the timing nature of the valve knob applied force. These characteristics may be the difference between the current time window and the first few time windows, the rate of change, or other relevance indicators. By introducing the context-dependent features, trends and patterns of force value changes can be captured; the calculated features are combined to form a valve knob applied force timing context-dependent feature vector. Each time window corresponds to a feature vector, and the vector contains feature information describing the change of the force value and the context association; and carrying out standardization processing on the feature vectors so as to eliminate dimension differences among different features. Common normalization methods include mean normalization, standard deviation normalization, and the like.
In particular, in step S4, it is determined whether the degree of tightness of assembly meets a predetermined requirement based on the valve knob apply force timing context-dependent feature vector. In particular, in one specific example of the present application, as shown in fig. 6, the S4 includes: s41, performing feature distribution optimization on the valve knob force application time sequence context associated feature vector to obtain an optimized valve knob force application time sequence context associated feature vector; and S42, enabling the optimized valve knob force application time sequence context associated feature vector to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the assembly tightness degree meets the preset requirement.
Specifically, the step S41 performs feature distribution optimization on the valve knob force applying time sequence context associated feature vector to obtain an optimized valve knob force applying time sequence context associated feature vector. In the technical scheme of the application, each multi-scale time sequence valve knob force applying characteristic vector expresses multi-scale time sequence local correlation characteristics of valve knob force in local time domain, so that after a sequence of the multi-scale time sequence valve knob force applying characteristic vectors passes through a transducer-based valve knob force inter-force correlation characteristic extractor, context correlation characteristics among local time domains in global time domain are further extracted, and thus, when the valve knob force applying time sequence context correlation characteristic vectors are subjected to classification regression through a classifier, regression probability mapping of scale heuristic is also performed based on unit time sequence expression scales of the local time domains in global time domain, but considering that in each officeOn the expression scale of the partial time domain, the valve knob force application time sequence context correlation feature vector not only comprises multi-scale time sequence correlation distribution representation under the scale of the partial time domain, but also comprises time sequence correlation distribution representation in each local time domain and between local time domains under the global time domain, which can cause the reduction of the training efficiency of the classifier. Based on the above, when classifying the valve knob force application time sequence context associated feature vector by a classifier, the applicant of the present application performs semantic information homogenization activation of feature rank expression on the valve knob force application time sequence context associated feature vector, specifically expressed as:wherein (1)>Is the valve knob force application time sequence context associated feature vector,>is the +.o. of the valve knob force application time sequence context associated feature vector>Personal characteristic value->Representing the two norms of the valve knob apply force timing context associated feature vector,is a logarithm based on 2, and +.>Is a weight superparameter,/->Is the +.o. of the optimized valve knob force application time sequence context associated feature vector>And characteristic values. Here, consider the valveDoor knob force application timing context-dependent feature vectorFeature distribution mapping of the feature distribution in the high-dimensional feature space to the classification regression space can present different mapping modes on different feature distribution levels based on mixed time domain features, so that the optimal efficiency cannot be obtained based on a scale heuristic mapping strategy, and therefore, rank expression semantic information based on feature vector norms is uniform instead of scale feature matching, similar feature rank expressions can be activated in a similar manner, and the correlation between feature rank expressions with large difference is reduced, so that the problem that the valve knob applies force time sequence context correlation feature vectors are solved>The problem that the probability expression mapping efficiency of the feature distribution under different space rank expressions is low is solved, and the training efficiency of the valve knob applied force time sequence context correlation feature vector in classification regression through a classifier is improved.
Specifically, the step S42 is to pass the optimized valve knob force application time sequence context-related feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether the assembly tightness degree meets a predetermined requirement. That is, after the optimized valve knob force applying time sequence context associated feature vector is obtained, the optimized valve knob force applying time sequence context associated feature vector is further used as a classification feature vector to pass through a classifier to obtain a classification result for indicating whether the assembly tightness degree reaches a preset requirement, and specifically, a plurality of full connection layers of the classifier are used for full connection coding of the optimized valve knob force applying time sequence context associated feature vector to obtain a coding classification feature vector; and passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
A classifier refers to a machine learning model or algorithm that is used to classify input data into different categories or labels. The classifier is part of supervised learning, which performs classification tasks by learning mappings from input data to output categories.
Fully connected layers are one type of layer commonly found in neural networks. In the fully connected layer, each neuron is connected to all neurons of the upper layer, and each connection has a weight. This means that each neuron in the fully connected layer receives inputs from all neurons in the upper layer, and weights these inputs together, and then passes the result to the next layer.
The Softmax classification function is a commonly used activation function for multi-classification problems. It converts each element of the input vector into a probability value between 0 and 1, and the sum of these probability values equals 1. The Softmax function is commonly used at the output layer of a neural network, and is particularly suited for multi-classification problems, because it can map the network output into probability distributions for individual classes. During the training process, the output of the Softmax function may be used to calculate the loss function and update the network parameters through a back propagation algorithm. Notably, the output of the Softmax function does not change the relative magnitude relationship between elements, but rather normalizes them. Thus, the Softmax function does not change the characteristics of the input vector, but simply converts it into a probability distribution form.
It is worth mentioning that in other examples of the application, it may be determined whether the degree of tightness of the assembly meets predetermined requirements, for example, based on the valve knob applied force timing context-dependent feature vector, by: first, a set of sample data is collected that is known to determine whether the degree of tightness of the assembly has reached a predetermined requirement. These sample data should include valve knob force application timing data and corresponding assembly tightness labels; and performing feature engineering on the valve knob force application time sequence context associated feature vector according to specific requirements. This may include feature selection, feature scaling, feature transformation, etc. operations to extract more useful features; the collected sample data is divided into a training set and a test set. Typically, most of the data is used to train the model, while a portion of the data is reserved for evaluating model performance; selecting a machine learning or deep learning model suitable for a task, such as a support vector machine, a random forest, a neural network and the like; the trained model is evaluated using the test set. Calculating indexes such as prediction accuracy, recall rate, F1 score and the like of the model on the test set so as to evaluate the performance of the model; the trained model is used to predict the new valve knob apply force timing context correlation feature vector. The model outputs a predicted value to indicate whether the assembly tightness degree reaches a preset requirement; and setting a threshold value according to the predicted value to judge whether the assembly tightness degree meets the preset requirement. If the predicted value exceeds the threshold value, the assembly tightness degree does not meet the requirement; otherwise, the assembly tightness degree is indicated to meet the requirement.
In summary, an intelligent control method for assembling a pneumatic ball valve according to an embodiment of the present application is explained, which analyzes the degree of tightness of the assembly by acquiring a force value applied to a valve knob acquired by a force sensor to intelligently determine whether the degree of tightness of the assembly meets a predetermined requirement.
Further, an intelligent control system for pneumatic ball valve assembly is also provided.
FIG. 7 is a block diagram of an intelligent control system for pneumatic ball valve assembly in accordance with an embodiment of the present application. As shown in fig. 7, an intelligent control system 300 for pneumatic ball valve assembly according to an embodiment of the present application includes: a data acquisition module 310 for acquiring force values applied to the valve knob at a plurality of predetermined time points within a predetermined time period acquired by the force sensor; a data preprocessing module 320, configured to perform data preprocessing on force values applied to the valve knob at the plurality of predetermined time points to obtain an upsampled valve knob force application time sequence input vector; a feature extraction module 330, configured to perform feature extraction on the upsampled valve knob force timing input vector to obtain a valve knob force timing context associated feature vector; and a result generation module 340 for determining whether the degree of tightness of the assembly meets a predetermined requirement based on the valve knob apply force timing context associated feature vector.
As described above, the intelligent control system 300 for pneumatic ball valve assembly according to the embodiment of the present application may be implemented in various wireless terminals, such as a server or the like having an intelligent control algorithm for pneumatic ball valve assembly. In one possible implementation, the intelligent control system 300 for pneumatic ball valve assembly according to embodiments of the present application may be integrated into a wireless terminal as a software module and/or hardware module. For example, the intelligent control system 300 for pneumatic ball valve assembly may be a software module in the operating system of the wireless terminal, or may be an application developed for the wireless terminal; of course, the intelligent control system 300 for pneumatic ball valve assembly could equally be one of the many hardware modules of the wireless terminal.
Alternatively, in another example, the intelligent control system for pneumatic ball valve assembly 300 and the wireless terminal may be separate devices, and the intelligent control system for pneumatic ball valve assembly 300 may be connected to the wireless terminal through a wired and/or wireless network and transmit interactive information in a agreed data format.
The foregoing description of the embodiments of the present disclosure has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (8)
1. An intelligent control method for assembling a pneumatic ball valve is characterized by comprising the following steps:
acquiring force values applied to the valve knob at a plurality of predetermined time points within a predetermined time period acquired by the force sensor;
performing data preprocessing on force values applied to the valve knob at the plurality of preset time points to obtain an up-sampling valve knob force application time sequence input vector;
performing feature extraction on the up-sampling valve knob force application time sequence input vector to obtain a valve knob force application time sequence context correlation feature vector; and
based on the valve knob apply force timing context associated feature vector, it is determined whether the assembly tightness meets a predetermined requirement.
2. The intelligent control method for pneumatic ball valve assembly of claim 1, wherein data preprocessing the force values applied to the valve knob at the plurality of predetermined points in time to obtain an upsampled valve knob force application timing input vector comprises:
arranging the force values applied to the valve knob at the plurality of predetermined time points into valve knob force application time sequence input vectors according to the time dimension; and
and carrying out up-sampling processing on the valve knob force applying time sequence input vector to obtain the up-sampling valve knob force applying time sequence input vector.
3. The intelligent control method for pneumatic ball valve assembly of claim 2, wherein feature extraction of the upsampled valve knob apply force timing input vector to obtain a valve knob apply force timing context correlation feature vector comprises:
extracting local time sequence characteristics of the up-sampling valve knob force application time sequence input vector to obtain a sequence of up-sampling valve knob force application time sequence input sub-vectors; and
and extracting correlation features between the sequences of up-sampling valve knob force application time sequence input sub-vectors to obtain the valve knob force application time sequence context correlation feature vectors.
4. The intelligent control method for pneumatic ball valve assembly of claim 3, wherein extracting the local timing features of the upsampling valve knob apply force timing input vector to obtain the sequence of upsampling valve knob apply force timing input sub-vectors comprises:
vector segmentation is carried out on the up-sampling valve knob force application time sequence input vector so as to obtain a sequence of up-sampling valve knob force application time sequence input sub-vectors; and
and passing the sequence of upsampled valve knob force application timing input sub-vectors through a force application feature extractor having a multi-scale one-dimensional convolution structure to obtain the sequence of multi-scale timing valve knob force application feature vectors.
5. The intelligent control method for pneumatic ball valve assembly of claim 4, wherein extracting correlation features between the sequence of upsampled valve knob apply force timing input sub-vectors to obtain the valve knob apply force timing context correlation feature vector comprises:
the sequence of multi-scale sequential valve knob apply force feature vectors is passed through a transducer-based valve knob apply force inter-correlation feature extractor to obtain the valve knob apply force time sequence context correlation feature vector.
6. The intelligent control method for pneumatic ball valve assembly of claim 5, wherein determining whether the assembly tightness meets a predetermined requirement based on the valve knob apply force timing context associated feature vector comprises:
performing feature distribution optimization on the valve knob force application time sequence context associated feature vector to obtain an optimized valve knob force application time sequence context associated feature vector; and
and passing the optimized valve knob force application time sequence context-associated feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the assembly tightness degree meets the preset requirement.
7. The intelligent control method for pneumatic ball valve assembly of claim 6, wherein feature distribution optimizing the valve knob apply force timing context correlation feature vector to obtain an optimized valve knob apply force timing context correlation feature vector comprises: performing feature distribution optimization on the valve knob force applying time sequence context associated feature vector by using the following optimization formula to obtain an optimized valve knob force applying time sequence context associated feature vector;
wherein, the formula is:wherein (1)>Is the valve knob force application time sequence context associated feature vector,>is the +.o. of the valve knob force application time sequence context associated feature vector>Personal characteristic value->Two norms representing the valve knob force application time sequence context associated feature vector, +.>Is a logarithm based on 2, and +.>Is a weight superparameter,/->Is the +.o. of the optimized valve knob force application time sequence context associated feature vector>And characteristic values.
8. An intelligent control system for pneumatic ball valve assembly, comprising:
a data acquisition module for acquiring force values applied to the valve knob at a plurality of predetermined time points within a predetermined time period acquired by the force sensor;
the data preprocessing module is used for preprocessing the data of the force values applied to the valve knob at a plurality of preset time points to obtain an up-sampling valve knob force application time sequence input vector;
the feature extraction module is used for carrying out feature extraction on the up-sampling valve knob force application time sequence input vector so as to obtain a valve knob force application time sequence context correlation feature vector; and
and the result generation module is used for determining whether the assembly tightness degree reaches a preset requirement or not based on the valve knob force application time sequence context correlation characteristic vector.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311150968.8A CN117052970A (en) | 2023-09-07 | 2023-09-07 | Intelligent control system and method for pneumatic ball valve assembly |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311150968.8A CN117052970A (en) | 2023-09-07 | 2023-09-07 | Intelligent control system and method for pneumatic ball valve assembly |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117052970A true CN117052970A (en) | 2023-11-14 |
Family
ID=88669276
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311150968.8A Pending CN117052970A (en) | 2023-09-07 | 2023-09-07 | Intelligent control system and method for pneumatic ball valve assembly |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117052970A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117649943A (en) * | 2024-01-30 | 2024-03-05 | 吉林大学 | Shaping data intelligent analysis system and method based on machine learning |
-
2023
- 2023-09-07 CN CN202311150968.8A patent/CN117052970A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117649943A (en) * | 2024-01-30 | 2024-03-05 | 吉林大学 | Shaping data intelligent analysis system and method based on machine learning |
| CN117649943B (en) * | 2024-01-30 | 2024-04-30 | 吉林大学 | Shaping data intelligent analysis system and method based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111914883B (en) | Spindle bearing state evaluation method and device based on deep fusion network | |
| EP2327019B1 (en) | Systems and methods for real time classification and performance monitoring of batch processes | |
| CN116453438B (en) | Display screen parameter detection method, device, equipment and storage medium | |
| CN109036466B (en) | Emotion dimension PAD prediction method for emotion voice recognition | |
| CN108780315A (en) | Method and apparatus for optimizing diagnostics of rotating equipment | |
| CN109623489B (en) | Improved machine tool health state evaluation method and numerical control machine tool | |
| CN116552306B (en) | Monitoring system and method for direct current pile | |
| CN117784710B (en) | Remote state monitoring system and method for numerical control machine tool | |
| CN111079861A (en) | Power distribution network voltage abnormity diagnosis method based on image rapid processing technology | |
| CN115659583A (en) | Point switch fault diagnosis method | |
| CN112926644A (en) | Method and system for predicting residual service life of rolling bearing | |
| CN117257302B (en) | Personnel mental health state assessment method and system | |
| CN114430850A (en) | Artificial intelligence scalp image diagnosis and analysis system using big data and product recommendation system using same | |
| CN113919540A (en) | Method for monitoring running state of production process and related equipment | |
| CN117198468B (en) | Intervention scheme intelligent management system based on behavior recognition and data analysis | |
| CN117154263A (en) | Lithium battery cascade utilization charging and discharging system and control method | |
| US20210357727A1 (en) | Abnormality detection device | |
| CN111860447A (en) | Method for accurately identifying telemetry time sequence data mode of random phase shift satellite | |
| CN117052970A (en) | Intelligent control system and method for pneumatic ball valve assembly | |
| CN118506846A (en) | Hard disk testing device, system and method | |
| US9002684B2 (en) | Apparatus and method for processing a process simulation database of a process | |
| CN118500981A (en) | Intelligent quality inspection monitoring system and method for paint production process | |
| CN110320802B (en) | Complex system signal time sequence identification method based on data visualization | |
| CN114139643B (en) | Monoglyceride quality detection method and system based on machine vision | |
| CN115659271A (en) | Sensor abnormality detection method, model training method, system, device, and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |