Disclosure of Invention
The embodiment of the application provides a system, a method, computer equipment and a computer readable storage medium for processing production process parameters, so as to at least solve the problems that the overfitting problem and the low accuracy often occur when the parameter curve speculation is performed through a machine learning algorithm in the related art.
In a first aspect, an embodiment of the present application provides a system for processing production process parameters, where the system includes: the system comprises a meter control device and a central control center, wherein the meter control device records the measurement value of a digital meter and generates a measurement curve set:
the central control center acquires a golden curve according to the production process parameters of a plurality of batches, and calculates a matching path between the golden curve and each batch of measurement curves;
the central control center trains the quality index, the time sequence in the measurement curve and the instrument parameter according to the matching path to obtain a parameter optimization model;
and the central control center obtains an optimized parameter curve corresponding to the target quality index through the parameter optimization model according to the target quality index.
In some of these embodiments, the central control center is further configured to:
aligning the measured values in the golden curve and the measuring curve according to a time sequence by using a dynamic time warping algorithm (DTW), and using the aligned measured values as input parameters of a Support Vector Regression (SVR), wherein the measured values comprise the instrument parameters;
and taking the quality index as an output parameter of the SVR, and training the input parameter and the output parameter through the SVR to obtain the parameter optimization model.
In some of these embodiments, the central control center is further configured to:
and based on the parameter optimization model, performing reverse deduction according to the target quality index through a nonlinear optimization function to obtain the optimized parameter curve.
In a second aspect, embodiments of the present application provide a method for processing production process parameters, the method including:
acquiring golden curves according to production process parameters of a plurality of batches, and calculating a matching path between the golden curves and each batch of measurement curves;
according to the matching path, training a quality index, a time sequence in the measurement curve and instrument parameters to obtain a parameter optimization model;
and obtaining an optimized parameter curve corresponding to the target quality index through the parameter optimization model according to the target quality index.
In some embodiments, the obtaining a parameter optimization model by training the quality indicator, the time series in the measurement curve, and the instrument parameter according to the matching path includes:
aligning the measured values in the golden curve and the measuring curve according to a time sequence by using a dynamic time warping algorithm (DTW), and using the aligned measured values as input parameters of a Support Vector Regression (SVR), wherein the measured values comprise the instrument parameters;
and taking the quality index as an output parameter of the SVR, and training the input parameter and the output parameter through the SVR to obtain the parameter optimization model.
In some embodiments, the obtaining, according to the target quality index and through the parameter optimization model, an optimized parameter curve corresponding to the target quality index includes:
and based on the parameter optimization model, performing reverse deduction according to the target quality index through a nonlinear optimization function to obtain the optimized parameter curve.
In some embodiments, the obtaining the golden curve according to the production process parameters of the plurality of batches includes:
determining a measurement curve set according to the production process parameters, randomly selecting a first measurement curve and a second measurement curve from the measurement curve set, calculating the sum of first distances between the first measurement curve and all curves in the measurement curve set, and acquiring a matching path between the first measurement curve and the second measurement curve;
obtaining the golden curve according to the matching path and a weight parameter sequence, and calculating a second distance sum of the golden curve and all curves in the measurement curve set, wherein the weight parameter sequence comprises weight parameters of the first measurement curve and the second measurement curve;
and outputting the golden curve under the condition that the difference value of the second distance sum and the first distance sum is smaller than a preset threshold value.
In some embodiments, the matching path, the sum of distances, and the golden curve are derived by DTW.
In a third aspect, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements any one of the above methods when executing the computer program.
In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement any of the above methods.
Compared with the related art, the method for processing the production process parameters, provided by the embodiment of the application, obtains the golden curve according to the production process parameters of a plurality of batches, calculates the matching path between the golden curve and each batch of the measurement curves, obtains the parameter optimization model according to the matching path by training the quality index, the time sequence in the measurement curves and the instrument parameters, obtains the optimized parameter curve corresponding to the target quality index according to the target quality index and through the parameter optimization model, solves the problem that the parameter curve conjecture through a machine learning algorithm is often over-fitted, and improves the accuracy of predicting the production process parameters.
The details of one or more embodiments of the application are set forth in the accompanying drawings and the description below to provide a more thorough understanding of the application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application.
It is obvious that the drawings in the following description are only examples or embodiments of the present application, and that it is also possible for a person skilled in the art to apply the present application to other similar contexts on the basis of these drawings without inventive effort. Moreover, it should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.
Unless defined otherwise, technical or scientific terms referred to herein shall have the ordinary meaning as understood by those of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The present application is directed to the use of the terms "including," "comprising," "having," and any variations thereof, which are intended to cover non-exclusive inclusions; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. Reference to "connected," "coupled," and the like in this application is not intended to be limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The term "plurality" as referred to herein means two or more. "and/or" describes an association relationship of associated objects, meaning that three relationships may exist, for example, "A and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. Reference herein to the terms "first," "second," "third," and the like, are merely to distinguish similar objects and do not denote a particular ordering for the objects.
The method for processing production process parameters provided by the present application can be applied to the application environment shown in fig. 1, and fig. 1 is a schematic application environment diagram of the method for processing production process parameters according to the embodiment of the present application, as shown in fig. 1. Wherein, the meter 102 and the central control center 104 perform data transmission through a network. The meter 102 obtains measurements from a manufacturing process, and the central control center 104 obtains a golden curve from the measurements from a plurality of batches and calculates a matching path between the golden curve and each batch of measurements. The central control center 104 obtains a parameter optimization model by training the quality index, the time series in the measurement curve and the instrument parameter according to the matching path, and obtains an optimized parameter curve corresponding to the target quality index through the parameter optimization model according to the target quality index. The central control center 104 may be implemented by a separate server or a server cluster composed of a plurality of servers.
The embodiment provides a method for processing production process parameters. FIG. 2 is a flow chart of a method of processing production process parameters according to an embodiment of the present application, as shown in FIG. 2, the method comprising the steps of:
step S201, obtaining a golden curve according to the production process parameters of a plurality of batches, and calculating a matching path between the golden curve and each batch of measurement curves.
In this embodiment, the golden curve is an average value of measurement curves obtained according to a historical production process, and is used for guiding the production process so as to improve the product quality.
The production process refers to a work, a method or a technology that a production worker processes or treats various raw materials, materials and semi-finished products by using production tools and equipment to finally form finished products, and the production processes are generally realized on production devices such as a reaction kettle and the like. In a set of production equipment, there are various digital instruments, such as weighing module, thermometer, pressure gauge, flowmeter, pneumatic valve, stirrer, etc., the measured value of each digital instrument is recorded in real-time database, and the measured value in the database is processed to obtain the measured curve for recording production process. Each measurement curve records the variation of the measurement value of one batch in the production process.
Because the acquisition time of each instrument measurement value is inconsistent in different measurement curves, the measurement values of the instruments need to be corresponded and matched according to the production flow, and then a matching path between the two measurement curves is obtained.
And step S202, according to the matching path, training the quality index, the time sequence in the measurement curve and the instrument parameter to obtain a parameter optimization model.
In this embodiment, the quality index is used as an output of the parameter optimization model, and represents a quality parameter corresponding to a quality that the production process is intended to achieve, the time series is a series formed at time points corresponding to different process measurement values in each measurement curve, and the instrument parameter includes various parameters related to the production process in the measurement process, such as temperature, pressure, weight, and the like. And training the historical process parameters to obtain a parameter optimization model.
And step S203, obtaining an optimized parameter curve corresponding to the target quality index through the parameter optimization model according to the target quality index.
In this embodiment, the target quality index is an expected quality index of the product, the performance of the product is optimal under the target quality index, the optimized parameter curve corresponds to the target quality index, and is a measurement curve formed by parameter values to be matched during the production process of the product, including temperature, air pressure, and the like. Under the guidance of the optimized parameter curve, the production process can optimize the performance of the product.
Through the steps S201 to S203, in this embodiment, on the basis of the golden curve, a parameter optimization model is obtained by training according to a matching path between the golden curve and another measured curve, and an optimized parameter curve corresponding to the output is obtained by setting the output to the parameter optimization model, so that the problem of overfitting often occurs when a machine learning algorithm is used for parameter curve prediction is solved, and the accuracy of predicting production process parameters is improved.
In some embodiments, fig. 3 is a flowchart of a method for training a parameter optimization model according to an embodiment of the present application, and as shown in fig. 3, the method includes the following steps:
step S301, aligning the measured values in the golden curve and the measuring curve according to the time sequence by using a dynamic time warping algorithm DTW, and using the aligned measured values as input parameters of a Support Vector Regression (SVR), wherein the measured values comprise the instrument parameters.
In the Dynamic Time Warping (DTW) algorithm used in this embodiment, local stretching or scaling of a track can be realized by copying track points, and then similarity comparison is performed on two unaligned sequences, and data points are aligned by a distance between data points, so that an optimal matching path between the two sequences is calculated.
The measurement curve of each batch can be represented by the following formula 1:
m (t), t ═ 0., Lm equation 1
In equation 1, the measurement curve m is a multidimensional vector at time point t, the time point t is a point in a time series, Lm is the length of the curve m, which can be the number of time points, and each time point corresponds to a measurement value of a measurement index.
The golden curve can be obtained from the following equation 2:
g (t), t ═ 0., Lg equation 2
In equation 2, g represents the golden curve, Lg represents the length of the golden curve, and the dimensions of g and m are the same.
In this embodiment, an optimal matching path is found based on the maximum lengths of m and g, which can be obtained from the following formula 3:
lmax ═ max (Lg, Lm) equation 3
In formula 3, max () represents taking the larger value in parentheses, and Lmax is the maximum length. The best matching path of the measurement curve m can be represented by the following equation 4:
m (tm (t)), t ═ 0., Lmax equation 4
In formula 4, Tm represents a value range of a time sequence of m, Lmax is a maximum length, values of formula 4 at each time point in the time sequence and in each measurement dimension are input parameters of a measurement curve in a parameter optimization model, and a quality index corresponding to the measurement curve is an output parameter of the parameter optimization model.
Step S302, the quality index is used as an output parameter of the SVR, and the input parameter and the output parameter are trained through the SVR to obtain the parameter optimization model.
In this embodiment, a Support Vector regression (SVM) is used for model training, the SVM is a generalized linear classifier (generalized linear classifier) that performs binary classification on data in a supervised learning (supervised learning) manner, and has sparsity and stability, and a decision boundary of the SVM is a maximum margin hyperplane for solving a learning sample.
Through the steps S301 and S302, each measuring curve is matched with the golden curve through the DTW algorithm to obtain the optimal matching path, the aligned measured value is used as the input of the SVR model based on the optimal matching path, the quality index is used as the output of the SVR model for model training, and finally the parameter optimization model is obtained, so that the overfitting problem caused by model training through deep learning is avoided, the accuracy of the corresponding relation between the measuring curve and the golden curve is improved through the DTW algorithm, and the accuracy of the parameter optimization model is further improved.
In some embodiments, obtaining, by the parameter optimization model, an optimized parameter curve corresponding to the target quality index according to the target quality index includes: and based on the parameter optimization model, obtaining an optimized parameter curve by reverse deduction according to the target quality index through a nonlinear optimization function. The main idea of the nonlinear optimization method is to find an extreme value in a constraint condition to obtain an optimal solution of an objective function, and the specific nonlinear optimization method includes a gradient descent method, a newton method and the like. In the embodiment, input variables including time series and instrument parameters are reversely deduced through a parameter optimization model based on a nonlinear optimization method under the condition that a preset target quality index is taken as output, and the input variables form an optimized parameter curve for guiding the process production process and improving the product quality.
In some embodiments, fig. 4 is a flowchart of a method of calculating a golden curve according to an embodiment of the present application, as shown in fig. 4, the method comprising the steps of:
step S401, determining a measurement curve set according to production process parameters, randomly selecting a first measurement curve and a second measurement curve from the measurement curve set, calculating a first distance sum of the first measurement curve and all curves in the measurement curve set, and obtaining a matching path between the first measurement curve and the second measurement curve.
In the production process, more than one instrument is generally corresponding to the key process parameters, so that the measurement curve is a multi-dimensional vector curve. Each measurement curve records the change of the measurement value of one batch in the production process, the measurement curves of a plurality of batches form a measurement curve set, and the first measurement curve and the second measurement curve are the measurement curves in the measurement curve set.
Since the specific processes of each batch production are difficult to be completely consistent, for example, the time length of each batch production, the start time or end time of the same step, the temperature and other process parameters are difficult to be completely identical, there is a difference between the measured values of the measurement curves, and in two measurement curves, the difference between the same measured value and the same measured value in the values is calculated for the distances of the measured value on different measurement curves, and the distance between the two measurement curves can be obtained by calculating the sum of the distances of all the measured values. In this embodiment, the sum of the distances between the first measurement curve and all other curves in the measurement curve set needs to be calculated.
Step S402, obtaining a golden curve according to the matching path and the weight parameter sequence, and calculating a second distance sum of the golden curve and all curves in the measurement curve set, wherein the weight parameter sequence includes weight parameters of the first measurement curve and the second measurement curve.
The weighting parameter sequence in this embodiment includes that, in the process of calculating the golden curve, the weighting parameter of the first measurement curve and the weighting parameter of the second measurement curve are weighted and summed to obtain the golden curve.
In step S403, the golden curve is output when the difference between the second distance sum and the first distance sum is smaller than a predetermined threshold.
The user can set the preset threshold value according to the requirement, and the golden curve corresponding to the sum of the second distances is a reference measurement curve meeting the requirement under the condition that the difference value is smaller than the preset threshold value, so that the golden curve can be used for guiding the setting of the instrument parameters in the production process.
In other embodiments, in the case that the difference value is greater than or equal to the preset threshold, the measurement curve set may be rearranged, and in the rearranged measurement curve set, the golden curve may be iteratively calculated until the difference value is less than the preset threshold.
In other embodiments, after obtaining the golden curve, selecting other measurement curves from the set of measurement curves, performing a weighted summation of the measurement curves and the golden curve until all the curves in the set of measurement curves have been calculated, implementing an adjustment of the golden curve, calculating a second distance sum for the golden curve obtained after the adjustment, and comparing the second distance sum with the first distance sum.
Through the steps S401 to S403, a golden curve is obtained based on the measured value of the historical production process, the golden curve is closer to the actual situation of the production process, and the accuracy of prediction of the parameter optimization model can be improved by training the parameter optimization model on the basis of the golden curve.
In some embodiments, the matching path, the sum of distances and the golden curve are obtained by DTW. In this embodiment, a DTW algorithm is used to calculate a golden curve, and the DTW algorithm realizes local stretching or scaling of a track by copying track points, so as to compare similarity of two unaligned sequences, and calculate an optimal matching path between the two sequences by a distance between data points. Through the DTW algorithm, based on the idea of dynamic programming, the accuracy of the calculation of the matching path can be improved.
It should be noted that the steps illustrated in the above-described flow diagrams or in the flow diagrams of the figures may be performed in a computer system, such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flow diagrams, in some cases, the steps illustrated or described may be performed in an order different than here.
The present embodiment further provides a device for processing parameters of a production process, which is used to implement the foregoing embodiments and preferred embodiments, and the description of the device is omitted here. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
In some embodiments, fig. 5 is a block diagram of a system for processing production process parameters according to an embodiment of the present application, and as shown in fig. 5, the system includes a meter control device 51 and a central control center 52, the meter control device 51 records measurement values of a digital meter and generates a set of measurement curves, the central control center 52 obtains a golden curve according to production process parameters of a plurality of batches, and calculates a matching path between the golden curve and each batch of measurement curves. The central control center 52 also obtains a parameter optimization model by training the quality index, the time series in the measurement curve and the instrument parameter according to the matching path, and the central control center 52 obtains an optimized parameter curve corresponding to the target quality index through the parameter optimization model according to the target quality index. In this embodiment, the central control center 52 trains the golden curve to obtain the parameter optimization model according to the matching path between the golden curve and other measured curves, and sets an output for the parameter optimization model to obtain an optimized parameter curve corresponding to the output, thereby solving the problem of overfitting when a machine learning algorithm is used to predict the parameter curve, and improving the accuracy of predicting the production process parameters.
In some embodiments, the central control center 52 is further configured to align the measured values in the golden curve and the measured curve according to a time sequence through a dynamic time warping algorithm DTW, use the aligned measured values as input parameters of a support vector regression SVR, use the quality index as an output parameter of the SVR, and train the input parameters and the output parameters through the SVR to obtain the parameter optimization model, where the measured values include the meter parameters. In the embodiment, through the DTW algorithm, the central control center 52 matches each measurement curve with the golden curve to obtain an optimal matching path, based on the optimal matching path, the aligned measurement value is used as the input of the SVR model, the quality index is used as the output of the SVR for model training, and finally, a parameter optimization model is obtained, so that the over-fitting problem caused by model training through deep learning is avoided, the accuracy of the corresponding relationship between the measurement curve and the golden curve is improved through the DTW algorithm, and the accuracy of the parameter optimization model is further improved.
In some embodiments, the central control center 52 is further configured to obtain the optimized parameter curve by performing a non-linear optimization function based on the parameter optimization model and performing a back-stepping operation according to the target quality index. In this embodiment, the central control center 52 reversely deduces input variables including time series and instrument parameters through a parameter optimization model based on a nonlinear optimization method under the condition that a preset target quality index is taken as an output, and the input variables form an optimized parameter curve for guiding a process production process and improving the product quality.
The meter control device 51 and the central control center 52 may be functional modules or program modules, and may be implemented by software or hardware. For a module implemented by hardware, the modules may be located in the same processor; or the modules can be respectively positioned in different processors in any combination.
In an embodiment, fig. 6 is a schematic diagram of parameter curve comparison according to an embodiment of the present application, as shown in fig. 6, in a measurement curve set having 26 measurement curve samples, a sample corresponding to 22 curves is a good product, a sample corresponding to 4 curves is a bad product, a dotted line in fig. 6 is a golden curve obtained according to the measurement curve set, a solid line is an optimized parameter curve, the optimized parameter curve is obtained by performing path matching through DTW on the basis of the golden curve to obtain an input variable, training with an SVR algorithm to obtain a parameter optimization model, and then performing inverse extrapolation with a general nonlinear optimization function. Wherein X-1, 13A, water addition 1, water addition 2, water addition 3, water addition 4, water addition 5, water addition 6 and feeding 20# of the horizontal axis are different steps in the production process, and the vertical axis represents parameter values corresponding to the steps in the horizontal axis.
In one embodiment, a computer device is provided, which may be a terminal. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of processing a production process parameter. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
In one embodiment, fig. 7 is a schematic diagram of an internal structure of an electronic device according to an embodiment of the present application, and as shown in fig. 7, there is provided an electronic device, which may be a server, and an internal structure diagram of which may be as shown in fig. 7. The electronic device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic equipment comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the electronic device is used for storing data. The network interface of the electronic device is used for connecting and communicating with an external terminal through a network. The computer program is executed by a processor to implement a method of processing a production process parameter.
Those skilled in the art will appreciate that the architecture shown in fig. 7 is a block diagram of only a portion of the architecture associated with the subject application, and does not constitute a limitation on the electronic devices to which the subject application may be applied, and that a particular electronic device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the steps in the method for processing production process parameters provided in the above embodiments are implemented.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps in the method of production process parameter processing provided by the various embodiments described above.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.