CN115130671B - Training method of equipment comprehensive efficiency prediction model, storage medium and electronic equipment - Google Patents
Training method of equipment comprehensive efficiency prediction model, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN115130671B CN115130671B CN202211052653.5A CN202211052653A CN115130671B CN 115130671 B CN115130671 B CN 115130671B CN 202211052653 A CN202211052653 A CN 202211052653A CN 115130671 B CN115130671 B CN 115130671B
- Authority
- CN
- China
- Prior art keywords
- data
- time
- equipment
- prediction model
- training
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Biomedical Technology (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Operations Research (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a training method of an equipment comprehensive efficiency prediction model, a storage medium and electronic equipment. The training method comprises the following steps: acquiring a data set, wherein the data set comprises a training set and a testing set, and each group of data in the data set comprises two time series of data and a label; constructing an equipment comprehensive efficiency prediction model based on a time convolution network, wherein the equipment comprehensive efficiency prediction model is provided with two input ends and an output end, the two input ends are respectively used for inputting first time sequence data and second time sequence data, and the output end is used for outputting an OEE predicted value; and training the comprehensive efficiency prediction model of the equipment by using each group of data in the training set to minimize the loss value, and testing the trained comprehensive efficiency prediction model of the equipment by using each group of data in the testing set to obtain a final comprehensive efficiency prediction model of the equipment. And predicting the OEE data at the future moment by using the equipment comprehensive efficiency prediction model obtained by the training method.
Description
Technical Field
The invention relates to the technical field of production equipment, in particular to a training method of an equipment comprehensive efficiency prediction model, a storage medium and electronic equipment.
Background
OEE (Overall Equipment efficiency) is a simple and independent production management measuring tool, can effectively measure the utilization efficiency of a single Equipment, is used for expressing the ratio of actual production capacity to theoretical capacity, and consists of three key elements of availability, expressiveness and quality index. OEE has great significance in the actual production environment, with one important objective being to help managers discover and reduce the six losses that exist in the manufacturing industry: the method comprises the following steps of shutdown loss, reloading and debugging loss, temporary shutdown loss, deceleration loss, defective loss in the starting process and defective loss generated during normal operation of production.
With the advent of the big data era, machine learning and artificial intelligence bring new changes to industrial production, and the artificial intelligence can learn in data and even accurately infer future results by using historical information. For a large amount of equipment condition information and OEE performance data recorded every day, the change rule of the data can be learned by means of an artificial intelligent algorithm such as deep learning, and the OEE index of the equipment in a future period can be predicted. The manufacturer can master the initiative of equipment maintenance, which greatly reduces the frequency of equipment failure and improves the production benefit of the equipment. In the related technology, the OEE index of the equipment in a future period is calculated and predicted to be inaccurate by utilizing artificial intelligent algorithms such as deep learning and the like, so that the production benefit of the equipment cannot be maximized.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art. Therefore, one objective of the present invention is to provide a training method for an equipment comprehensive efficiency prediction model, which uses the trained equipment comprehensive efficiency prediction model to realize prediction of future-time OEE data.
The second purpose of the invention is to provide a method for predicting the comprehensive efficiency of the equipment.
The third purpose of the invention is to provide a device for predicting the comprehensive efficiency of the equipment.
A fourth object of the invention is to propose a computer-readable storage medium.
A fifth object of the invention is to propose an electronic device.
In order to achieve the above object, an embodiment of a first aspect of the present invention provides a training method for a device comprehensive efficiency prediction model, where the training method includes: acquiring a data set, wherein the data set comprises a training set and a testing set, each group of data in the data set comprises two time series of data and a label, the label is an equipment comprehensive efficiency OEE true value, the two time series of data are respectively marked as first time series of data and second time series of data, the first time series of data comprises equipment working data and OEE corresponding to the equipment working data, the second time series of data comprises equipment working data, and the corresponding time period is later than the time period corresponding to the first time series of data; constructing an equipment comprehensive efficiency prediction model based on a time convolution network, wherein the equipment comprehensive efficiency prediction model is provided with two input ends and an output end, the two input ends are respectively used for inputting first time sequence data and second time sequence data, and the output end is used for outputting an OEE predicted value; and training the comprehensive efficiency prediction model of the equipment by utilizing each group of data in the training set, calculating a loss value by utilizing a loss function to minimize the loss value, and testing the trained comprehensive efficiency prediction model of the equipment by utilizing each group of data in the testing set to obtain a final comprehensive efficiency prediction model of the equipment, wherein the loss value is calculated by utilizing the loss function according to an OEE predicted value and an OEE real value corresponding to the OEE predicted value.
According to the training method of the equipment comprehensive efficiency prediction model, historical known equipment working data corresponding to each moment in each data set, namely first time sequence data, and known equipment working data corresponding to each moment in a future period, namely second time sequence data are input into the equipment comprehensive efficiency prediction model with two inputs and one output, and the equipment comprehensive efficiency prediction model is trained to obtain the equipment comprehensive efficiency prediction model capable of predicting the OEE index of the future moment according to the historical known equipment working data, the OEE true value corresponding to the historical known equipment working data and the partial known equipment working data of the future moment. The device comprehensive efficiency prediction model obtained by the training method can realize prediction of OEE data at a future moment.
In addition, the training method of the device comprehensive efficiency prediction model provided by the above embodiment of the present invention may further have the following additional technical features:
according to an embodiment of the invention, the acquiring a data set comprises: acquiring equipment working data of equipment in a historical preset time period and an OEE corresponding to the equipment working data; dividing the equipment working data in the preset time period and the OEE corresponding to the equipment working data into a plurality of batches of data according to a first preset time length and a preset step length; aiming at each batch of data, dividing the batch of data into a plurality of groups of training data according to a second preset time length, wherein each group of training data comprises first time sequence data and second time sequence data, and acquiring a corresponding OEE real value; and taking a group of data with the most recent time as a test set, and taking other groups of data as a training set, wherein in each training period, a group of data in the training set is adopted to train the comprehensive efficiency prediction model of the equipment.
According to an embodiment of the present invention, the device comprehensive efficiency prediction model includes two layers of gated cyclic units GRU networks and a feature merging module, where the two layers of gated cyclic units GRU networks are respectively marked as a first GRU network and a second GRU network, and the training of the device comprehensive efficiency prediction model by using each set of data in the training set includes: for each training period, inputting first time sequence data of the corresponding training period into the first GRU network, wherein the first GRU network obtains a first prediction characteristic through calculation and output, and inputting second time sequence data of the corresponding training period into the second GRU network, and the second GRU network obtains a second prediction characteristic through calculation and output; and the first prediction characteristic and the second prediction characteristic are combined through the characteristic combination module to obtain a third prediction characteristic.
According to an embodiment of the present invention, the device comprehensive efficiency prediction model further includes a multivariate time convolution network and a backend network connected in sequence, and the training of the device comprehensive efficiency prediction model by using each set of data in the training set further includes: inputting the third prediction characteristic into the multivariate time convolution network, and outputting the multivariate time convolution network to obtain a time convolution prediction characteristic; and inputting the time convolution prediction characteristics into the back-end network to obtain the OEE prediction value of the corresponding training period.
According to an embodiment of the present invention, the multi-time convolution network includes a first time convolution network, a second time convolution network, and a third time convolution network, the number of convolution kernels of the first time convolution network, the second time convolution network, and the third time convolution network is the same, the sizes of the convolution kernels are different, and the third prediction feature is input to the first time convolution network, the second time convolution network, and the third time convolution network at the same time, so as to obtain a first time convolution feature, a second time convolution feature, and a third time convolution feature; splicing the first time convolution characteristic, the second time convolution characteristic and the third time convolution characteristic to obtain a time convolution prediction characteristic; the back-end network is formed by connecting a full connection layer and a softmax layer in sequence.
According to one embodiment of the invention, the loss function is a mean square error loss function.
In order to achieve the above object, a second aspect of the present invention provides a method for predicting comprehensive efficiency of a device, where the method includes: acquiring data to be detected, wherein the data to be detected comprises first time sequence data and second time sequence data; and inputting the data to be tested into a trained equipment comprehensive efficiency prediction model to obtain a corresponding OEE prediction value, wherein the equipment comprehensive efficiency prediction model is obtained by utilizing the training method of the equipment comprehensive efficiency prediction model provided by the embodiment of the first aspect of the invention.
In order to achieve the above object, a third aspect of the present invention provides an apparatus for predicting comprehensive efficiency of a device, the apparatus including: the device comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring data to be detected, and the data to be detected comprises first time sequence data and second time sequence data; and the prediction module is used for inputting the data to be tested into the trained equipment comprehensive efficiency prediction model to obtain a corresponding OEE prediction value, wherein the equipment comprehensive efficiency prediction model is obtained by utilizing the training method of the equipment comprehensive efficiency prediction model provided by the embodiment of the first aspect of the invention.
To achieve the above object, a fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements a method for training a plant comprehensive efficiency prediction model as set forth in the first aspect of the present invention or a method for predicting plant comprehensive efficiency as set forth in the second aspect of the present invention.
To achieve the above object, a fifth embodiment of the present invention provides an electronic device, which includes a memory and a processor, where the memory stores a computer program, and the computer program, when executed by the processor, implements a method for training a device comprehensive efficiency prediction model as set forth in the first embodiment of the present invention or a method for predicting device comprehensive efficiency as set forth in the second embodiment of the present invention.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
FIG. 1 is a flow chart of a method of training a plant complex efficiency prediction model according to one embodiment of the present invention;
FIG. 2 is a flow diagram of acquiring a data set according to one embodiment of the invention;
FIG. 3 is a schematic diagram of the partitioning of a data set according to one embodiment of the present invention;
FIG. 4 is a flow chart of a method of training a plant complex efficiency prediction model according to one embodiment of the present invention;
FIG. 5 is a flow chart of obtaining a third predictive feature according to one embodiment of the invention;
FIG. 6 is a flow chart of obtaining OEE predicted values according to one embodiment of the present invention;
FIG. 7 is a schematic diagram of a multivariate time convolution network of one embodiment of the present invention;
FIG. 8 is a schematic diagram of a time convolutional network of one embodiment of the present invention;
FIG. 9 is a flow chart of a method for plant integrated efficiency prediction according to one embodiment of the present invention;
fig. 10 is a schematic diagram of an apparatus comprehensive efficiency predicting apparatus according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
The method for training the device comprehensive efficiency prediction model, the storage medium and the electronic device according to the embodiment of the present invention will be described in detail with reference to fig. 1 to 10 and the detailed description.
Before specifically describing the embodiments of the present invention, for ease of understanding, OEE (Overall Equipment efficiency) is first introduced:
OEE is called overall equipment efficiency, is a simple and independent production management measuring tool, can effectively measure the utilization efficiency of single equipment, is used for expressing the ratio of actual production capacity to theoretical capacity, and consists of three key factors of availability, expressiveness and quality index. OEE has great significance in a real production environment, where one important objective is to help managers discover and reduce the six losses that exist in the manufacturing industry: the system comprises a machine halt loss, a replacement and debugging loss, a halt loss, a deceleration loss, a defective product loss in the starting process and a defective product loss generated in normal operation of production.
FIG. 1 is a flow chart of a method for training a plant synthetic efficiency prediction model according to an embodiment of the present invention. As shown in fig. 1, the method for training the plant comprehensive efficiency prediction model may include:
the method comprises the following steps of S1, acquiring a data set, wherein the data set comprises a training set and a testing set, each group of data in the data set comprises two time sequence data and a label, the label is an OEE real value of the comprehensive efficiency of the equipment, the two time sequence data are respectively marked as a first time sequence data and a second time sequence data, the first time sequence data comprise equipment working data and OEE corresponding to the equipment working data, the second time sequence data comprise equipment working data, and the corresponding time period is later than the time period corresponding to the first time sequence data.
Specifically, a data set is obtained, the data set comprises a large number of data groups, the data groups in the data set can be divided into a training set and a testing set according to time points, the data groups in the training set are used for training the comprehensive efficiency prediction model of the equipment, the data groups in the testing set are used for predicting the comprehensive efficiency prediction model of the equipment, and the training effect of the comprehensive efficiency prediction model of the equipment is detected.
It should be noted that each set of data includes first time series data, second time series data, and a tag. The first time sequence data are historical known equipment working data at each moment, such as production rate and running duration corresponding to each moment history. The second time sequence data is partially known equipment working data in a period of time in the future at each moment, such as scheduled working time corresponding to each moment history. The label is an OEE real value corresponding to the first time sequence data and the second time sequence data. By acquiring historical known equipment working data at each moment, partially known equipment working data within a period of time in the future at each moment and OEE real values corresponding to each moment, and training the comprehensive efficiency prediction model by using the data, the past data characteristics of OEE indexes and the future data characteristics can be associated, so that OEE predicted values output by the trained comprehensive efficiency prediction model of the equipment are more accurate.
In an embodiment of the present invention, as shown in FIG. 2, acquiring a data set may include:
s11, collecting the equipment working data of the equipment in a historical preset time period and the OEE corresponding to the equipment working data.
Specifically, the method collects the equipment working data of the equipment in a historical preset time period, namely collects data related to OEE such as operating equipment parameters, operating duration, temperature, production rate, equipment working duration, predicted time points, operating equipment parameters and the like in a historical certain time period.
And S12, dividing the equipment working data in the preset time period and the OEE corresponding to the equipment working data into a plurality of batches of data according to the first preset time length and the preset step length.
Specifically, the device working data and the corresponding OEE in the preset time period are divided into a plurality of batches of data according to a first preset time length and a preset step length. The first preset time length is the time length of each batch of data.
And S13, aiming at each batch of data, dividing the batch of data into a plurality of groups of training data according to a second preset time length, wherein each group of training data comprises first time sequence data and second time sequence data, and acquiring a corresponding OEE real value.
Specifically, each set of training data is divided into first time sequence data and second time sequence data according to a second preset time length. The first time sequence data includes an OEE real value, the second time sequence data does not include an OEE real value, and the first time sequence data is later than the second time sequence data, as shown in fig. 3.
It should be noted that the first time series data in each set of data is historical known operating equipment operating data corresponding to each time, and since the second time series data is taken from continuous time series data later than the first time series data, the second time series data can be regarded as part of known equipment operating data in a future period of time relative to the first time series data. The time points, the operating equipment parameters, the operating duration, the temperature, the production rate and the like corresponding to each time point are historical known data which can be collected at each time point. The working time of the equipment, the predicted time point, the parameters of the running equipment and the like are future known data which can be collected at each time point. The data set is divided in such a way, so that the model can learn the data conversion characteristics from the past time period to the future time period, and the prediction capability of the model is enhanced.
And S14, taking a group of data with the latest time as a test set, and taking other groups of data as a training set, wherein in each training period, a group of data in the training set is adopted to train the comprehensive efficiency prediction model of the equipment.
Specifically, according to the time point, a group of data with the latest time is used as a test set, and other groups of data are used as a training set, so as to test the training effect of the device comprehensive efficiency prediction model obtained through training.
And S2, constructing an equipment comprehensive efficiency prediction model based on the time convolution network, wherein the equipment comprehensive efficiency prediction model is provided with two input ends and an output end, the two input ends are respectively used for inputting the first time sequence data and the second time sequence data, and the output end is used for outputting an OEE prediction value.
Specifically, the equipment comprehensive efficiency prediction model is a model based on a time convolution network and is provided with two input ends (a first input end and a second input end) and an output end, first time sequence data in each group of data are input into the first input end, second time sequence data in each group of data are input into the second input end, the equipment comprehensive efficiency prediction model predicts OEE data at a future moment by extracting time sequence information characteristics in the first time sequence data and the second time sequence data, and the output end outputs OEE predicted values corresponding to each group of data.
And S3, training the built equipment comprehensive efficiency prediction model by utilizing each group of data in the training set, calculating a loss value by utilizing a loss function, enabling the loss value to reach the minimum value, testing the trained equipment comprehensive efficiency prediction model by utilizing each group of data in the testing set, and obtaining a final equipment comprehensive efficiency prediction model, wherein the loss value is calculated by utilizing the loss function according to the OEE predicted value and the corresponding OEE true value.
In an embodiment of the present invention, the loss function may be a mean square error loss function. Wherein the mean square error loss function
The expression of (a) is:
wherein, N is the number of training data contained in each batch;
representing a real label value corresponding to the ith training data, namely a real OEE value;
and the model OEE predicted value is output by the output end of the model OEE predicted value which represents the calculation of the comprehensive efficiency prediction model of the equipment.
Specifically, the loss value can be calculated according to the OEE predicted value and the OEE true value corresponding to the OEE predicted value by using the above-mentioned mean square error loss function.
Specifically, each group of data in the training set is used for training the comprehensive efficiency prediction model of the equipment, and in the training process, the OEE predicted value of each group of data output by the comprehensive efficiency prediction model of the equipment and the OEE true value corresponding to each group of data are used for calculating loss. And inputting each group of data in the training set into the equipment comprehensive efficiency prediction model and calculating a loss value, which is called a round of model training. And when the number of times of model training is lower than the preset number of times of training rounds, continuously sending each group of data in the training set into the model for training, and striving for the convergence of the loss value to the minimum. And when each training round is finished, testing the trained equipment comprehensive efficiency prediction model by using each group of data in the test set. And after all the training rounds are finished, the obtained equipment comprehensive efficiency prediction model is the finally trained equipment comprehensive efficiency prediction model. During the training process, the model parameters may be adjusted by an optimization algorithm such that the loss values converge to a minimum. Illustratively, the optimization model parameters are updated using the Adam optimization algorithm.
In an embodiment of the present invention, as shown in fig. 4 and 5, the device comprehensive efficiency prediction model may include two layers of gated loop Unit GRU (Gate recovery Unit) networks and a feature merging module. The two layers of gated cyclic unit GRU networks are respectively marked as a first GRU network and a second GRU network, and the equipment comprehensive efficiency prediction model is trained by using each group of data in a training set, which may include:
s31, aiming at each training period, inputting first time sequence data of the corresponding training period into a first GRU network, calculating and outputting the first time sequence data by the first GRU network to obtain a first prediction characteristic, inputting second time sequence data of the corresponding training period into a second GRU network, and calculating and outputting the second time sequence data by the second GRU network to obtain a second prediction characteristic;
and S32, merging the first prediction characteristic and the second prediction characteristic through the characteristic merging module to obtain a third prediction characteristic.
Specifically, the first time series data is input into a first GRU network, and the first GRU network extracts time series information in the first time series data to obtain a first prediction characteristic (T1, D1). Where T1 represents the first predicted feature sequence length, and D1 represents the vector dimension of the first predicted feature. And inputting the second time sequence data into a second GRU network, and extracting the time sequence information in the second time sequence data by the second GRU network to obtain a second prediction characteristic (T2, D2). Wherein T2 represents the length of the second predicted feature sequence, and D2 represents the vector dimension of the second predicted feature. After the first GRU network and the second GRU network respectively extract the time sequence information of the first time sequence data and the second time sequence data, the first prediction feature and the second prediction feature are converted into the same dimension D, and the converted two are combined in the time dimension.
More specifically, the obtained first prediction features (T1, D1) and the second prediction features (T2, D2) are input into a merging module, the merging module converts the first prediction features (T1, D1) and the second prediction features (T2, D2) into the same dimension D, and then performs feature vector merging on a time dimension to obtain third prediction features (T1 + T2, D).
In the embodiment of the present invention, as shown in fig. 4 and 6, the device comprehensive efficiency prediction model further includes a multivariate time convolution network and a back-end network connected in sequence, and the training of the device comprehensive efficiency prediction model by using each group of data in the training set may further include:
s33, inputting the third prediction characteristic into a multivariate time convolution network to obtain an OEE prediction characteristic;
and S34, inputting the OEE prediction characteristics into a rear-end network to obtain the OEE prediction value of the corresponding training period.
Specifically, the third predictive feature (T1 + T2, D) is input into a Multivariate time Convolutional Network (M-TCN). The multi-element time convolution network comprises a plurality of time convolution networks with different scales, the time convolution networks with different scales are used for processing third prediction features (T1 + T2, D) in parallel, three processed time prediction features are spliced, and a time convolution prediction feature is obtained through output, wherein the vector dimension of the time convolution prediction feature is the same as that of the third prediction features (T1 + T2, D). And finally, inputting the time convolution prediction characteristics into a back-end network to obtain the OEE prediction value of the corresponding training period.
As a specific embodiment, as shown in fig. 7, the multivariate time convolution network M-TCN includes a first time convolution network TCN1, a second time convolution network TCN2, and a third time convolution network TCN3, where the numbers of convolution kernels of the first time convolution network TCN1, the second time convolution network TCN2, and the third time convolution network TCN3 are the same, and the sizes of the convolution kernels are different, and the third prediction feature is simultaneously input to the first time convolution network TCN1, the second time convolution network TCN2, and the third time convolution network TCN3, respectively, to obtain a first time convolution feature, a second time convolution feature, and a third time convolution feature; splicing the first time convolution characteristic, the second time convolution characteristic and the third time convolution characteristic to obtain a time convolution prediction characteristic;
the back-end network is formed by connecting a full connection layer and a softmax layer in sequence.
Specifically, referring to fig. 7, the size of the convolution kernel of the first time convolution network TCN1 is 3, the size of the convolution kernel of the second time convolution network TCN2 is 5, and the size of the convolution kernel of the third time convolution network TCN3 is 7, and time convolution networks with different convolution kernel sizes are adopted to capture information of different time scales in the third prediction feature so as to adapt to different correlations of future OEE data to equipment working data at a past time.
It should be noted that a Time Convolutional Network (TCN) is a special Convolutional neural Network, and its structure is shown in fig. 8, and the time Convolutional Network includes three main structures: causal convolution, dilated convolution, and residual concatenation.
In causal convolution, the output value at the time t of the previous layer depends only on the input values at the time t and before the time t of the next layer. The difference with the traditional convolution neural network is that the causal convolution can not see future data, and the causal convolution is a one-way structure, is not two-way and is a strict time constraint model.
The simple causal convolution also has the problem of the traditional convolutional neural network that the modeling length of time is limited by the size of the convolution kernel, and if a longer dependency is to be caught, a large number of layers need to be linearly stacked. The dilation convolution allows for an interval sampling of the input at the time of convolution, the sampling rate being controlled by a controlled variable d. D =1 in the lowest layer indicates that each point is sampled at the time of input, and d =2 in the middle layer indicates that each 2 points are sampled at the time of input as an input value. Generally higher levels use larger d sizes. Therefore, the dilation convolution causes the size of the effective window to grow exponentially with the number of layers. Thus, the convolution network can obtain a large reception field by using a small number of layers.
Residual concatenation has proven to be an effective method of training deep networks, allowing the network to deliver information in a cross-layer manner. Therefore, the outputs of each two layers of convolutional layers in the time convolutional network are connected by a residual connection, so that the information of the shallow network can be directly transmitted to the deep network.
According to the training method of the equipment comprehensive efficiency prediction model, the time sequence change information of each group of data in the training set is learned by utilizing the recurrent neural network, the multi-scale time convolution network is used as the prediction model of the OEE index, the multi-scale time convolution network adopts a plurality of convolution kernels with different sizes, information with different time scales can be captured, and the accuracy of the output OEE prediction value is improved. The comprehensive efficiency prediction model of the equipment obtained by the training method can realize prediction of OEE data at a future moment so as to master initiative for equipment maintenance, greatly reduce frequency of equipment failure and improve production benefit of the equipment.
The invention also provides a device comprehensive efficiency prediction method.
FIG. 9 is a flow chart of a method for predicting the integrated efficiency of a plant according to an embodiment of the present invention. As shown in fig. 9, the method for predicting the integrated efficiency of the device may include:
s21, acquiring data to be detected, wherein the data to be detected comprises first time sequence data and second time sequence data;
and S22, inputting the data to be tested into the trained equipment comprehensive efficiency prediction model to obtain a corresponding OEE prediction value, wherein the equipment comprehensive efficiency prediction model is obtained by utilizing the training method of the equipment comprehensive efficiency prediction model.
It should be noted that, for other specific implementations of the method for predicting comprehensive efficiency of equipment according to the embodiment of the present invention, reference may be made to the specific implementation of the method for training the comprehensive efficiency prediction model of equipment according to the above embodiment of the present invention.
According to the equipment comprehensive efficiency prediction method, the trained equipment comprehensive efficiency prediction model is utilized to predict the data to be measured of the equipment, so that the OEE prediction value of the equipment is obtained, the initiative for equipment maintenance is mastered, the frequency of equipment failure occurrence is greatly reduced, and the production benefit of the equipment is improved.
The invention also provides a device for predicting the comprehensive efficiency of the equipment.
Fig. 10 is a schematic diagram of an apparatus comprehensive efficiency prediction device according to an embodiment of the present invention. As shown in fig. 10, the plant integrated efficiency prediction apparatus 100 may include an acquisition module 10 and a prediction module 20.
The acquiring module 10 is configured to acquire data to be detected, where the data to be detected includes first time sequence data and second time sequence data; the prediction module 20 is configured to input data to be tested into the trained device comprehensive efficiency prediction model to obtain a corresponding OEE prediction value, where the device comprehensive efficiency prediction model is obtained by using the training method of the device comprehensive efficiency prediction model.
It should be noted that, for other specific implementations of the device comprehensive efficiency prediction apparatus according to the embodiment of the present invention, reference may be made to the specific implementation of the training method of the device comprehensive efficiency prediction model according to the above-mentioned embodiment of the present invention.
In the device comprehensive efficiency prediction apparatus of the embodiment of the present invention, the prediction module 20 predicts the device to-be-measured data acquired by the acquisition module 10 by using the trained device comprehensive efficiency prediction model, so as to obtain the OEE prediction value of the device, master the initiative for device maintenance, greatly reduce the frequency of device faults, and improve the production benefit of the device.
The invention also provides a computer readable storage medium.
In this embodiment, a computer program is stored on a computer-readable storage medium, and the computer program corresponds to the above-mentioned training method for the plant total efficiency prediction model, or the above-mentioned plant total efficiency prediction method, when executed by a processor, implements the above-mentioned training method for the plant total efficiency prediction model, or the above-mentioned plant total efficiency prediction method.
The invention also provides the electronic equipment.
In this embodiment, the electronic device includes a processor, a memory, and a computer program stored in the memory, and when the processor executes the computer program, the method for training the device comprehensive efficiency prediction model or the method for predicting the device comprehensive efficiency is implemented.
According to the storage medium and the electronic equipment, the trained equipment comprehensive efficiency prediction model is utilized to predict the data to be measured of the equipment, so that a manufacturer can master the initiative of equipment maintenance, the frequency of equipment failure is greatly reduced, and the production benefit of the equipment is improved.
It should be noted that the logic and/or steps represented in the flowcharts or otherwise described herein, such as an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Further, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following technologies, which are well known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
In the description of the present invention, it is to be understood that the terms "central," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," "circumferential," and the like are used in the orientations and positional relationships indicated in the drawings for convenience in describing the invention and to simplify the description, and are not intended to indicate or imply that the referenced devices or elements must have a particular orientation, be constructed and operated in a particular orientation, and are therefore not to be considered limiting of the invention.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
In the present invention, unless otherwise expressly stated or limited, the terms "mounted," "connected," "secured," and the like are to be construed broadly and can, for example, be fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; they may be directly connected or indirectly connected through intervening media, or they may be connected internally or in any other suitable relationship, unless expressly stated otherwise. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.
In the present invention, unless expressly stated or limited otherwise, the first feature "on" or "under" the second feature may be directly contacting the second feature or the first and second features may be indirectly contacting each other through intervening media. Also, a first feature "on," "over," and "above" a second feature may be directly or diagonally above the second feature, or may simply indicate that the first feature is at a higher level than the second feature. A first feature being "under," "below," and "beneath" a second feature may be directly under or obliquely under the first feature, or may simply mean that the first feature is at a lesser elevation than the second feature.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. A training method of a device comprehensive efficiency prediction model is characterized by comprising the following steps:
acquiring a data set, wherein the data set comprises a training set and a testing set, each group of data in the data set comprises two time series of data and a label, the label is an equipment comprehensive efficiency OEE true value, the two time series of data are respectively marked as first time series of data and second time series of data, the first time series of data comprises equipment working data and OEE corresponding to the equipment working data, the second time series of data comprises equipment working data, and the corresponding time period is later than the time period corresponding to the first time series of data;
constructing an equipment comprehensive efficiency prediction model based on a time convolution network, wherein the equipment comprehensive efficiency prediction model is provided with two input ends and an output end, the two input ends are respectively used for inputting first time sequence data and second time sequence data, and the output end is used for outputting an OEE predicted value;
and training the comprehensive efficiency prediction model of the equipment by utilizing each group of data in the training set, calculating a loss value by utilizing a loss function to minimize the loss value, and testing the trained comprehensive efficiency prediction model of the equipment by utilizing each group of data in the testing set to obtain a final comprehensive efficiency prediction model of the equipment, wherein the loss value is calculated by utilizing the loss function according to an OEE predicted value and an OEE real value corresponding to the OEE predicted value.
2. The method for training the plant integrated efficiency prediction model according to claim 1, wherein the obtaining the data set comprises:
acquiring equipment working data of equipment in a historical preset time period and an OEE corresponding to the equipment working data;
dividing the equipment working data and the corresponding OEE in the preset time period into a plurality of batches of data according to a first preset time length and a preset step length;
aiming at each batch of data, dividing the batch of data into a plurality of groups of training data according to a second preset time length, wherein each group of training data comprises first time sequence data and second time sequence data, and acquiring a corresponding OEE real value;
and taking a group of data with the latest time as a test set, and taking other groups of data as a training set, wherein in each training period, a group of data in the training set is adopted to train the comprehensive efficiency prediction model of the equipment.
3. The method for training the plant integrated efficiency prediction model according to claim 2, wherein the plant integrated efficiency prediction model comprises two layers of gated cyclic units GRU networks and a feature merging module, wherein the two layers of gated cyclic units GRU networks are respectively denoted as a first GRU network and a second GRU network, and the training of the plant integrated efficiency prediction model using each set of data in the training set comprises:
for each training period, inputting first time sequence data of the corresponding training period into the first GRU network, wherein the first GRU network obtains a first prediction characteristic through calculation and outputting, and inputting second time sequence data of the corresponding training period into the second GRU network, and the second GRU network obtains a second prediction characteristic through calculation and outputting;
and combining the first prediction characteristic and the second prediction characteristic through the characteristic combining module to obtain a third prediction characteristic.
4. The method for training the plant integrated efficiency prediction model according to claim 3, wherein the plant integrated efficiency prediction model further includes a multivariate time convolution network and a backend network connected in sequence, and the training of the plant integrated efficiency prediction model using each set of data in the training set further includes:
inputting the third prediction characteristic into the multivariate time convolution network, and outputting the multivariate time convolution network to obtain a time convolution prediction characteristic;
and inputting the time convolution prediction characteristics into the back-end network to obtain the OEE prediction value of the corresponding training period.
5. The method of training an integrated plant efficiency prediction model according to claim 4,
the multivariate time convolution network comprises a first time convolution network, a second time convolution network and a third time convolution network, the number of convolution kernels of the first time convolution network, the number of convolution kernels of the second time convolution network and the number of convolution kernels of the third time convolution network are the same, the sizes of the convolution kernels are different, and the third prediction characteristic is input into the first time convolution network, the second time convolution network and the third time convolution network respectively and simultaneously to obtain a first time convolution characteristic, a second time convolution characteristic and a third time convolution characteristic; splicing the first time convolution characteristic, the second time convolution characteristic and the third time convolution characteristic to obtain a time convolution prediction characteristic;
the back-end network is formed by connecting a full connection layer and a softmax layer in sequence.
6. The method of claim 1, wherein the loss function is a mean square error loss function.
7. A method for predicting the comprehensive efficiency of equipment is characterized by comprising the following steps:
acquiring data to be detected, wherein the data to be detected comprises first time sequence data and second time sequence data;
inputting the data to be tested into a trained equipment comprehensive efficiency prediction model to obtain a corresponding OEE prediction value, wherein the equipment comprehensive efficiency prediction model is obtained by utilizing the training method of the equipment comprehensive efficiency prediction model according to any one of claims 1-6.
8. An apparatus for predicting plant integrated efficiency, the apparatus comprising:
the device comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring data to be detected, and the data to be detected comprises first time sequence data and second time sequence data;
the prediction module is used for inputting the data to be tested into a trained equipment comprehensive efficiency prediction model to obtain a corresponding OEE prediction value, wherein the equipment comprehensive efficiency prediction model is obtained by utilizing the training method of the equipment comprehensive efficiency prediction model according to any one of claims 1-6.
9. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a method of training a plant integrated efficiency prediction model according to any one of claims 1 to 6 or a method of plant integrated efficiency prediction according to claim 7.
10. An electronic device comprising a memory and a processor, the memory having stored thereon a computer program, wherein the computer program, when executed by the processor, implements a method of training a device integrated efficiency prediction model as claimed in any one of claims 1 to 6 or a method of device integrated efficiency prediction as claimed in claim 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211052653.5A CN115130671B (en) | 2022-08-31 | 2022-08-31 | Training method of equipment comprehensive efficiency prediction model, storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211052653.5A CN115130671B (en) | 2022-08-31 | 2022-08-31 | Training method of equipment comprehensive efficiency prediction model, storage medium and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115130671A CN115130671A (en) | 2022-09-30 |
| CN115130671B true CN115130671B (en) | 2022-11-29 |
Family
ID=83387247
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211052653.5A Active CN115130671B (en) | 2022-08-31 | 2022-08-31 | Training method of equipment comprehensive efficiency prediction model, storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115130671B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116629454B (en) * | 2023-07-19 | 2023-10-03 | 武汉新威奇科技有限公司 | Method and system for predicting production efficiency of servo screw press based on neural network |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105023194A (en) * | 2014-04-28 | 2015-11-04 | 国家电网公司 | Distribution equipment operation efficiency evaluation method involving reliability and economy |
| CN113435593A (en) * | 2021-06-29 | 2021-09-24 | 广州博通信息技术有限公司 | Refrigeration equipment frosting prediction method based on sensor time sequence data analysis |
| CN113723693A (en) * | 2021-09-03 | 2021-11-30 | 华电莱州发电有限公司 | Method for predicting equipment indexes by multi-factor fuzzy time series algorithm |
| CN114692341A (en) * | 2022-04-22 | 2022-07-01 | 桂林电子科技大学 | Fan fault prediction method and storage medium based on fusion spatiotemporal graph ConvLSTM model |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110263944A (en) * | 2019-05-21 | 2019-09-20 | 中国石油大学(华东) | A multi-variable fault prediction method and device |
| US20220085964A1 (en) * | 2020-09-15 | 2022-03-17 | Qualcomm Incorporated | Enhanced timing advance offset value indications |
| US11450226B2 (en) * | 2020-10-07 | 2022-09-20 | Lghorizon, Llc | Predictive building emergency training and guidance system |
| CN114218867B (en) * | 2021-12-20 | 2022-06-28 | 暨南大学 | Process control method and system for special equipment based on entropy optimization and safety reinforcement learning |
| CN114861533B (en) * | 2022-04-26 | 2024-11-08 | 东南大学 | Ultra-short-term wind power forecasting method based on temporal convolutional network |
-
2022
- 2022-08-31 CN CN202211052653.5A patent/CN115130671B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105023194A (en) * | 2014-04-28 | 2015-11-04 | 国家电网公司 | Distribution equipment operation efficiency evaluation method involving reliability and economy |
| CN113435593A (en) * | 2021-06-29 | 2021-09-24 | 广州博通信息技术有限公司 | Refrigeration equipment frosting prediction method based on sensor time sequence data analysis |
| CN113723693A (en) * | 2021-09-03 | 2021-11-30 | 华电莱州发电有限公司 | Method for predicting equipment indexes by multi-factor fuzzy time series algorithm |
| CN114692341A (en) * | 2022-04-22 | 2022-07-01 | 桂林电子科技大学 | Fan fault prediction method and storage medium based on fusion spatiotemporal graph ConvLSTM model |
Non-Patent Citations (2)
| Title |
|---|
| Recurrent Reinforcement Learning for Predictive Overall Equipment Effectiveness;Da-Yin Liao 等;《2018 e-Manufacturing & Design Collaboration Symposium (eMDC)》;20181101;全文 * |
| 基于深度学习的设备剩余寿命预测算法研究及其应用;程成;《中国优秀硕士学位论文全文数据库 (信息科技辑)》;20210228(第2期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115130671A (en) | 2022-09-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117930815B (en) | Wind turbine generator remote fault diagnosis method and system based on cloud platform | |
| CN111967486A (en) | Complex equipment fault diagnosis method based on multi-sensor fusion | |
| CN118468165B (en) | Complex equipment fault diagnosis system and diagnosis method based on multi-metadata fusion | |
| CN116774086B (en) | A lithium battery health status estimation method based on multi-sensor data fusion | |
| CN118604637B (en) | Management method of energy storage battery of railway machine room | |
| CN117893059A (en) | Energy storage data acquisition and analysis method and system based on sensor | |
| KR102667861B1 (en) | heavy electrical equipment treatment apparatus using deep learning-based prediction model and method therefor | |
| CN114936504A (en) | Equipment residual life prediction method and system based on Bayesian multi-source data fusion | |
| CN115130671B (en) | Training method of equipment comprehensive efficiency prediction model, storage medium and electronic equipment | |
| CN115238573A (en) | Method and system for predicting performance degradation trend of hydroelectric unit considering working parameters | |
| CN118739436B (en) | An algorithm for load output based on automatic adjustment of renewable energy output | |
| CN115310650A (en) | Low-complexity high-precision time sequence multi-step prediction method and system | |
| CN116843080B (en) | Machine learning-based carbon element footprint prediction method and system for urea production | |
| CN118711276B (en) | On-board intelligent diagnosis method and system integrating cloud-edge computing | |
| CN115684835A (en) | Power distribution network fault diagnosis method, device, equipment and storage medium | |
| CN118536090B (en) | Method for generating grain temperature field map based on warehouse temperature and humidity monitoring | |
| CN119513732A (en) | An insulation detection method for energy storage system based on artificial intelligence | |
| CN119125902A (en) | Fuel cell performance prediction method, device and computer program product | |
| CN117804811B (en) | Health diagnosis system and method for deep neural network port machinery equipment of attention mechanism | |
| Soualhi et al. | Explainable RUL estimation of turbofan engines based on prognostic indicators and heterogeneous ensemble machine learning predictors | |
| CN115841580B (en) | A method for predicting remaining useful life of bearings based on marker distribution | |
| CN115169458A (en) | Adaptive fault diagnosis method, device and related medium based on active learning | |
| CN119809277B (en) | Data-driven adaptive optimization method and system for vehicle manufacturing supply chain | |
| CN118710082B (en) | Industry chain disturbance identification method and detection system based on GAT and AE | |
| CN119805239B (en) | Battery defect monitoring and identification and life prediction method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |