CN117531438A - Control method and system for improving stability of polyoxymethylene product - Google Patents
Control method and system for improving stability of polyoxymethylene product Download PDFInfo
- Publication number
- CN117531438A CN117531438A CN202311470435.8A CN202311470435A CN117531438A CN 117531438 A CN117531438 A CN 117531438A CN 202311470435 A CN202311470435 A CN 202311470435A CN 117531438 A CN117531438 A CN 117531438A
- Authority
- CN
- China
- Prior art keywords
- weight
- time sequence
- liquid level
- feature
- sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 229920006324 polyoxymethylene Polymers 0.000 title claims abstract description 58
- 229930040373 Paraformaldehyde Natural products 0.000 title claims abstract description 56
- -1 polyoxymethylene Polymers 0.000 title claims abstract description 56
- 238000000034 method Methods 0.000 title claims abstract description 50
- 239000007788 liquid Substances 0.000 claims abstract description 157
- 239000013598 vector Substances 0.000 claims description 179
- 230000004927 fusion Effects 0.000 claims description 81
- 238000012549 training Methods 0.000 claims description 55
- 238000004458 analytical method Methods 0.000 claims description 15
- 230000003247 decreasing effect Effects 0.000 claims description 11
- 238000005457 optimization Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 5
- 238000003062 neural network model Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 abstract description 32
- 239000000654 additive Substances 0.000 abstract description 7
- 230000006870 function Effects 0.000 description 17
- 238000006116 polymerization reaction Methods 0.000 description 14
- 238000009826 distribution Methods 0.000 description 13
- 239000012752 auxiliary agent Substances 0.000 description 12
- 239000000463 material Substances 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 239000000843 powder Substances 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 238000005303 weighing Methods 0.000 description 8
- 238000004519 manufacturing process Methods 0.000 description 6
- 230000000737 periodic effect Effects 0.000 description 6
- 230000004913 activation Effects 0.000 description 5
- 239000003999 initiator Substances 0.000 description 5
- 230000001105 regulatory effect Effects 0.000 description 5
- 230000001276 controlling effect Effects 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000001125 extrusion Methods 0.000 description 3
- 239000012847 fine chemical Substances 0.000 description 3
- 239000012530 fluid Substances 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 230000004580 weight loss Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 239000003054 catalyst Substances 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 239000002994 raw material Substances 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 241001391944 Commicarpus scandens Species 0.000 description 1
- 108010076504 Protein Sorting Signals Proteins 0.000 description 1
- 239000006096 absorbing agent Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000003483 aging Methods 0.000 description 1
- 239000003963 antioxidant agent Substances 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 238000003889 chemical engineering Methods 0.000 description 1
- 238000001311 chemical methods and process Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 229920006351 engineering plastic Polymers 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011049 filling Methods 0.000 description 1
- 239000008187 granular material Substances 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000000178 monomer Substances 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000010525 oxidative degradation reaction Methods 0.000 description 1
- 239000008188 pellet Substances 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 239000002002 slurry Substances 0.000 description 1
- 239000003381 stabilizer Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000010977 unit operation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D9/00—Level control, e.g. controlling quantity of material stored in vessel
- G05D9/12—Level control, e.g. controlling quantity of material stored in vessel characterised by the use of electric means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01J—CHEMICAL OR PHYSICAL PROCESSES, e.g. CATALYSIS OR COLLOID CHEMISTRY; THEIR RELEVANT APPARATUS
- B01J4/00—Feed or outlet devices; Feed or outlet control devices
- B01J4/008—Feed or outlet control devices
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C48/00—Extrusion moulding, i.e. expressing the moulding material through a die or nozzle which imparts the desired form; Apparatus therefor
- B29C48/25—Component parts, details or accessories; Auxiliary operations
- B29C48/92—Measuring, controlling or regulating
-
- C—CHEMISTRY; METALLURGY
- C08—ORGANIC MACROMOLECULAR COMPOUNDS; THEIR PREPARATION OR CHEMICAL WORKING-UP; COMPOSITIONS BASED THEREON
- C08G—MACROMOLECULAR COMPOUNDS OBTAINED OTHERWISE THAN BY REACTIONS ONLY INVOLVING UNSATURATED CARBON-TO-CARBON BONDS
- C08G6/00—Condensation polymers of aldehydes or ketones only
Landscapes
- Chemical & Material Sciences (AREA)
- Engineering & Computer Science (AREA)
- Organic Chemistry (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Mechanical Engineering (AREA)
- Health & Medical Sciences (AREA)
- Medicinal Chemistry (AREA)
- Polymers & Plastics (AREA)
- Feedback Control In General (AREA)
Abstract
The application discloses a control method and a control system for improving stability of a polyoxymethylene product, wherein a weight time sequence change signal is used as a feedback signal for a controller by adding feedforward-feedback control, and a liquid level value signal is used as a feedforward signal for the controller, so that interference caused by liquid level change is synchronously compensated, and system fluctuation caused by liquid level fluctuation is avoided. Thus, the precise control of the additives in the polymerizer and the extruder can be realized, and the stability of the polyoxymethylene product is effectively improved.
Description
Technical Field
The present application relates to the field of intelligent control, and more particularly, to a control method and system for improving the stability of polyoxymethylene products.
Background
Polyoxymethylene is an important engineering plastic with excellent physical, chemical and mechanical properties, and is widely applied to the fields of automobiles, electronics, machinery and the like. However, polyoxymethylene also has some disadvantages such as poor thermal stability, susceptibility to thermal degradation and oxidative degradation, resulting in reduced product properties and poor color. In order to increase the stability of polyoxymethylene products, it is generally necessary to add certain amounts of initiators and auxiliaries, such as stabilizers, antioxidants, UV absorbers, etc., in the polymerization and extrusion processes, which are added in a certain order, which would otherwise cause abnormal reactions. If the catalyst is added too little, insufficient reaction can be caused, and slurry appears; if the catalyst is excessively added, the reaction is severe, so that the screw of the polymerization machine is unevenly stressed, vibration occurs, and equipment is damaged. In the polymerization process, various auxiliary agents are in ppm level, fluctuation of feeding of each auxiliary agent can cause fluctuation of product quality, and the product can be seriously exploded or not polymerized, so that how to stably, continuously and accurately control the feeding of various auxiliary agents is very important.
Currently, high-precision diaphragm metering pumps are commonly used in the industry to control the amount of additive added. The operating characteristics of diaphragm metering pumps result in uncontrolled instantaneous flow per stroke, and serious persons may cause shutdowns, increasing the effort and production costs of the enterprise. The addition of the damper to the pump outlet can slow down the pressure fluctuation caused by the pulse, but the problem of unstable flow of the auxiliary agent can not be fundamentally solved. Meanwhile, the diaphragm metering pump has high requirement on the type selection of the instrument, the investment is large, and the instrument is easy to break. Therefore, in normal production, the phenomenon of uneven product quality and unstable performance caused by unstable addition amount of the auxiliary agent which leads to uneven normal distribution of the molecular weight of the product often occurs.
Accordingly, an optimized control scheme for improving the stability of polyoxymethylene products is desired.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides a control method and a control system for improving the stability of a polyoxymethylene product, which are characterized in that a sensor is used for feeding a weight time sequence change signal to a controller to serve as a feedback signal by adding feedforward-feedback control, and meanwhile, a liquid level value signal is fed to the controller to serve as a feedforward signal, so that interference caused by liquid level change is synchronously compensated, and system fluctuation caused by liquid level fluctuation is avoided. Thus, the precise control of the additives in the polymerizer and the extruder can be realized, and the stability of the polyoxymethylene product is effectively improved.
According to one aspect of the present application, there is provided a control method for improving the stability of a polyoxymethylene product, comprising:
acquiring weight signals and liquid level signals at a plurality of preset time points in a preset time period;
arranging the weight signals and the liquid level signals of the plurality of preset time points into weight signal time sequence input vectors and liquid level signal time sequence input vectors according to a time dimension;
carrying out local time sequence feature analysis on the weight signal time sequence input vector to obtain a sequence of weight signal local time sequence feature vectors;
embedding feedback fusion is carried out on the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector so as to obtain a weight-liquid level time sequence embedding feedback fusion feature; and
and based on the weight-liquid level time sequence embedded feedback fusion characteristic, determining that the rotating speed value of the variable frequency motor at the current time point should be increased, decreased or maintained.
According to another aspect of the present application, there is provided a control system for improving the stability of a polyoxymethylene product, comprising:
the signal acquisition module is used for acquiring weight signals and liquid level signals at a plurality of preset time points in a preset time period;
the arrangement module is used for arranging the weight signals and the liquid level signals of the plurality of preset time points into weight signal time sequence input vectors and liquid level signal time sequence input vectors according to the time dimension;
The local time sequence feature analysis module is used for carrying out local time sequence feature analysis on the weight signal time sequence input vector so as to obtain a sequence of weight signal local time sequence feature vectors;
the embedded feedback fusion module is used for carrying out embedded feedback fusion on the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector to obtain a weight-liquid level time sequence embedded feedback fusion feature; and
and the control result generation module is used for determining that the rotating speed value of the variable frequency motor at the current time point should be increased, decreased or kept based on the weight-liquid level time sequence embedded feedback fusion characteristic.
Compared with the prior art, the control method and the control system for improving the stability of the polyoxymethylene product are provided, by adding feedforward-feedback control, a weight time sequence change signal is used for a controller to serve as a feedback signal by a sensor, and meanwhile, a liquid level value signal is used for the controller to serve as a feedforward signal, so that interference caused by liquid level change is synchronously compensated, and system fluctuation caused by liquid level fluctuation is avoided. Thus, the precise control of the additives in the polymerizer and the extruder can be realized, and the stability of the polyoxymethylene product is effectively improved.
Drawings
The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
FIG. 1 is a flow chart of a control method for improving the stability of a polyoxymethylene product according to an embodiment of the present application;
FIG. 2 is a system architecture diagram of a control method for improving the stability of polyoxymethylene products according to an embodiment of the present application;
FIG. 3 is a flow chart of a training phase of a control method for improving the stability of a polyoxymethylene product according to an embodiment of the present application;
FIG. 4 is a flowchart of sub-step S3 of a control method for improving the stability of a polyoxymethylene product according to an embodiment of the present application;
FIG. 5 is a block diagram of a control system for improving the stability of a polyoxymethylene product according to an embodiment of the present application;
FIG. 6 is a schematic diagram of the operation of the weightless system;
FIG. 7 is a schematic diagram of an improved configuration system;
FIG. 8 is a schematic diagram of the operation of the ingredients;
FIG. 9 is a flow chart of a powder delivery system and a sequence control program as shown in the figure;
fig. 10 is a schematic diagram of a control system.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
As used in this application and in the claims, the terms "a," "an," "the," and/or "the" are not specific to the singular, but may include the plural, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.
Flowcharts are used in this application to describe the operations performed by systems according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
It should be understood that the weightless feeding is a feeding device commonly used in fine chemical engineering in recent years, and is mostly used in occasions with high feeding precision requirement and strict material proportion. In a polyoxymethylene polymerization device, the procedures of adding an initiator into a polymerization machine and adding various auxiliary agents in the extrusion process are all high in precision requirement, and in the occasion with strict proportioning requirement, the conventional flow meter is difficult to meet the actual requirement, and the problems can be well solved by adopting a weightless weighing rule. The weight loss method is to flow out materials from a container, and measure the change value of weight in unit time, namely 'lost weight', by a weighing mode, so as to obtain the instantaneous flow of fluid. The weighing sensor sends a weight signal to the Flow Controller (FC), the controller calculates the flow, compares the flow with a given value (SP) to obtain deviation, calculates the deviation and then outputs a control signal to the frequency converter (f), and further adjusts the rotating speed of the variable frequency motor to achieve the purpose of controlling the flow. From the above, the sensor, the controller, the frequency converter and the variable frequency motor form a closed loop regulating circuit, so that the flow can be stabilized. The requirements of common application occasions can be met, but the requirements of raw material proportion are high for polymerization reaction, so that the adding flow is more stable, and the object characteristics are also needed to be considered. The periodic variation of the liquid level in the apparatus can be divided into 2 phases:
(1) When the liquid level reaches high, closing a charging valve, and regulating the flow according to a weighing signal, namely a weight metering mode; (2) When the liquid level reaches low, a charging valve is opened, and the motor keeps the last moment of rotating speed, which is called a volume metering mode; accordingly, considering that the periodic variation of the liquid level in the apparatus causes a non-negligible disturbance to the flow rate fed into the polymerizer, both in weight mode and in volume mode, in particular equivalent to open loop control, it is necessary to compensate correspondingly the flow rate fluctuations caused by the variation of the liquid level in order to improve the flow stability. Particularly, the linear relation between the weight of the material in the storage tank and the flow fluctuation caused by the liquid level change is considered, so that the weight time sequence change signal can be used for making a feedback signal to the controller by adding feedforward-feedback control, and meanwhile, the liquid level value signal can be used for making a feedforward signal to the controller, so that the disturbance caused by the liquid level change can be synchronously compensated, and the system fluctuation caused by the liquid level fluctuation can be avoided. Thus, the precise control of the additives in the polymerizer and the extruder can be realized, and the stability of the polyoxymethylene product is effectively improved.
In the technical scheme of the application, a control method for improving the stability of a polyoxymethylene product is provided. Fig. 1 is a flow chart of a control method for improving the stability of a polyoxymethylene product according to an embodiment of the present application. Fig. 2 is a system architecture diagram of a control method for improving the stability of a polyoxymethylene product according to an embodiment of the present application. As shown in fig. 1 and 2, a control method for improving stability of a polyoxymethylene product according to an embodiment of the present application includes the steps of: s1, acquiring weight signals and liquid level signals of a plurality of preset time points in a preset time period; s2, arranging the weight signals and the liquid level signals at the plurality of preset time points into weight signal time sequence input vectors and liquid level signal time sequence input vectors according to a time dimension; s3, carrying out local time sequence feature analysis on the weight signal time sequence input vector to obtain a sequence of weight signal local time sequence feature vectors; s4, embedding feedback fusion is carried out on the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector so as to obtain a weight-liquid level time sequence embedded feedback fusion feature; and S5, based on the weight-liquid level time sequence embedded feedback fusion characteristic, determining that the rotating speed value of the variable frequency motor at the current time point should be increased, decreased or maintained.
In particular, the step S1 is to acquire weight signals and liquid level signals at a plurality of predetermined time points within a predetermined period. That is, in one specific example of the present application, weight signals at a plurality of predetermined time points within a predetermined period of time are acquired by a weight signal sensor; and acquiring liquid level signals at a plurality of predetermined time points within a predetermined period of time by a liquid level signal sensor. It is worth mentioning that the weight signal sensor is a sensor for measuring the weight or load of an object. They typically use the principle of force or pressure sensors to effect measurements. The liquid level signal sensor is a sensor for measuring the liquid level of a liquid or material.
In particular, the step S2 is to arrange the weight signal and the liquid level signal at the plurality of predetermined time points into a weight signal time sequence input vector and a liquid level signal time sequence input vector according to a time dimension. Considering that the weight signal and the liquid level signal both have dynamic change rules in the time dimension, the change and fluctuation conditions of the flow can be reflected, so that in order to make corresponding compensation for the flow fluctuation caused by the liquid level change to improve the flow stability, the time sequence change characteristics of the weight signal need to be captured and described. Based on this, in the technical solution of the present application, the weight signals and the liquid level signals at the plurality of predetermined time points are arranged into a weight signal time sequence input vector and a liquid level signal time sequence input vector according to a time dimension, so as to integrate distribution information of the weight signals and the liquid level signals in time sequence, respectively.
In particular, the step S3 performs a local time sequence feature analysis on the weight signal time sequence input vector to obtain a sequence of weight signal local time sequence feature vectors. In particular, in one specific example of the present application, as shown in fig. 4, the S3 includes: s31, vector segmentation is carried out on the weight signal time sequence input vector so as to obtain a sequence of weight signal local time sequence input vectors; and S32, passing the sequence of the weight signal local time sequence input vectors through a weight time sequence feature extractor based on a one-dimensional convolution layer to obtain the sequence of the weight signal local time sequence feature vectors.
Specifically, in S31, the weight signal time sequence input vector is subjected to vector slicing to obtain a sequence of weight signal local time sequence input vectors. The change in weight signal is generally continuous during polyoxymethylene production and there is some dynamic and periodic nature. Therefore, in order to better capture the time sequence characteristics of the weight signal, in the technical scheme of the application, vector segmentation is further performed on the time sequence input vector of the weight signal so as to obtain a sequence of local time sequence input vectors of the weight signal. By vector slicing the weight signal time sequence input vector, a continuous signal sequence can be divided into a plurality of sequences of local time sequence input vectors, so that weight change characteristics in different time periods, such as instantaneous flow change, periodic fluctuation and the like in the feeding process, are captured. Such local features can provide more information for subsequent feature extraction and classification, thereby better controlling the variable frequency motor speed to achieve control over polyoxymethylene product stability.
Specifically, the step S32 is to pass the sequence of the weight signal local time sequence input vectors through a weight time sequence feature extractor based on a one-dimensional convolution layer to obtain the sequence of the weight signal local time sequence feature vectors. It should be understood that, in order to extract local features in each local time sequence input vector to better reflect the variation trend and rule of the weight signal, the sequence of the local time sequence input vectors of the weight signal needs to be further extracted by feature mining in a weight time sequence feature extractor based on a one-dimensional convolution layer, so as to extract time sequence feature information of the weight signal in the time dimension, thereby obtaining the sequence of the local time sequence feature vectors of the weight signal. By extracting and analyzing the characteristics of the local time sequence input vectors of each weight signal, more detailed and accurate time sequence change and fluctuation characteristic information of the weight signal can be obtained, so that the accuracy of estimating and controlling the instantaneous flow of the fluid is improved. More specifically, each layer using the weight timing feature extractor based on one-dimensional convolution layer performs on input data in forward transfer of the layer: carrying out convolution processing on input data to obtain a convolution characteristic diagram; pooling the convolution feature images based on a feature matrix to obtain pooled feature images; performing nonlinear activation on the pooled feature map to obtain an activated feature map; the output of the last layer of the weight time sequence feature extractor based on the one-dimensional convolution layer is the sequence of the weight signal local time sequence feature vectors, and the input of the first layer of the weight time sequence feature extractor based on the one-dimensional convolution layer is the sequence of the weight signal local time sequence input vectors.
Notably, the one-dimensional convolutional layer (1D Convolutional Layer) is a base layer type in convolutional neural networks for processing one-dimensional sequence data. It applies a one-dimensional convolution operation on the input data, extracting local features of the input sequence by learning the weights of the convolution kernel (also called a filter). The input to the one-dimensional convolution layer may be one-dimensional time series data, word sequences in text data, or other continuous one-dimensional data. The main purpose of this layer is to capture the local patterns and features in the input sequence by convolution operations. In the one-dimensional convolution layer, the input sequence is subject to a sliding convolution operation by one or more convolution kernels. Each convolution kernel is a small window that slides over the input sequence and performs element-by-element multiplication and summation with the data within the window to yield an element of the convolution output. By sliding the window, the convolution operation can extract local features over the entire input sequence. The one-dimensional convolution layer also typically includes an activation function and a pooling operation. The activation function introduces a nonlinear characteristic that increases the expressive power of the network. The pooling operation is used to reduce the size of the feature map, reduce the computational complexity, and extract more salient features. Parameters of the one-dimensional convolution layer include the number, size, and stride of the convolution kernels, the selection of the activation function, and the like. These parameters may be adjusted and optimized according to specific tasks and data characteristics. By stacking a plurality of one-dimensional convolution layers, a deeper model can be constructed, and higher-level abstract features can be extracted. One-dimensional convolutional layers are widely used in many areas including speech recognition, natural language processing, bioinformatics, and the like.
It should be noted that, in other specific examples of the present application, the local time sequence feature analysis may be performed on the weight signal time sequence input vector in other manners to obtain a sequence of weight signal local time sequence feature vectors, for example: dividing the weight signal timing data into a series of successive timing input vectors; for each timing input vector, defining a local timing signature analysis window for signature analysis within the window: extracting local time sequence features from the weight signals within each local time sequence feature analysis window using an appropriate feature extraction method; combining the local time sequence features extracted from each window into a local time sequence feature vector; for each time sequence input vector, extracting local time sequence characteristics and constructing local time sequence characteristic vectors; and combining the local time sequence characteristic vectors corresponding to each time sequence input vector into a characteristic vector sequence. This sequence will preserve the timing information of the original weight signal and each feature vector represents a local timing feature of the corresponding timing input vector.
And S4, performing embedded feedback fusion on the sequence of the weight signal local time sequence characteristic vector and the liquid level signal time sequence input vector to obtain a weight-liquid level time sequence embedded feedback fusion characteristic. It should be understood that the amount of flow fluctuation caused by the change in the weight and level of the material in the tank is linear during the polyoxymethylene production process. By fusing the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector, the information of the two signals can be mutually complemented, and richer and comprehensive feature representation is provided. Specifically, in the technical scheme of the application, by adding feedforward-feedback control, the time sequence characteristic of the weight signal is used as a feedback signal, and meanwhile, the time sequence distribution information of the liquid level signal is used as a feedforward signal, so that the interference caused by liquid level change is synchronously compensated, and the system fluctuation caused by liquid level fluctuation is avoided. Based on the above, in the technical scheme of the application, the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector are further processed through a feature embedding feedback fusion module to obtain a weight-liquid level time sequence embedding feedback fusion feature vector. It should be understood that, by the processing of the feature embedding feedback fusion module, feature embedding feedback can be performed on the time sequence distribution information of the liquid level signal based on the local time sequence features of all weight signals as references, and thus the obtained time sequence information, dynamic variation trend and relationship between the weight and liquid level signals can be contained in the weight-liquid level time sequence embedding feedback fusion feature vector, so that the comprehensive state of the weight and liquid level in the polyoxymethylene production process can be reflected better, and more accurate input can be provided for subsequent classification and control. Specifically, the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector are processed through a feature embedding feedback fusion module to obtain the weight-liquid level time sequence embedding feedback fusion feature vector as the weight-liquid level time sequence embedding feedback fusion feature, and the method comprises the following steps: the liquid level signal time sequence input vector passes through a liquid level signal time sequence feature extractor based on a full convolution neural network model to obtain a liquid level signal full convolution time sequence feature vector; arranging the sequence of the weight signal local time sequence feature vectors into a weight signal global time sequence feature vector; fusing the liquid level signal full convolution time sequence feature vector and the weight signal global time sequence feature vector to obtain a liquid level-weight fusion feature vector; carrying out semantic association coding on the sequence of the weight signal local time sequence feature vectors to obtain weight global semantic association feature vectors; and fusing the liquid level-weight fusion feature vector and the weight global semantic association feature vector to obtain the weight-liquid level time sequence embedded feedback fusion feature vector.
In particular, the step S5 is to determine that the rotation speed value of the variable frequency motor at the current time point should be increased, decreased or maintained based on the weight-liquid level time sequence embedded feedback fusion feature. That is, in the technical solution of the present application, the weight-liquid level time sequence is embedded into the feedback fusion feature vector to pass through the classifier to obtain a classification result, where the classification result is used to indicate that the rotation speed value of the variable frequency motor at the current time point should be increased, decreased or maintained. The method is characterized in that the weight signal time sequence characteristic is embedded into the fusion characteristic information of the feedback liquid level time sequence information to conduct classification processing, so that the rotating speed of the variable frequency motor is adaptively controlled. Therefore, the flow fluctuation caused by liquid level change can be correspondingly compensated, the system fluctuation caused by the liquid level fluctuation is avoided, and the flow and the system stability are improved. Specifically, using a plurality of full-connection layers of the classifier to perform full-connection coding on the weight-liquid level time sequence embedded feedback fusion feature vector so as to obtain a coding classification feature vector; and passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
A classifier refers to a machine learning model or algorithm that is used to classify input data into different categories or labels. The classifier is part of supervised learning, which performs classification tasks by learning mappings from input data to output categories.
Fully connected layers are one type of layer commonly found in neural networks. In the fully connected layer, each neuron is connected to all neurons of the upper layer, and each connection has a weight. This means that each neuron in the fully connected layer receives inputs from all neurons in the upper layer, and weights these inputs together, and then passes the result to the next layer.
The Softmax classification function is a commonly used activation function for multi-classification problems. It converts each element of the input vector into a probability value between 0 and 1, and the sum of these probability values equals 1. The Softmax function is commonly used at the output layer of a neural network, and is particularly suited for multi-classification problems, because it can map the network output into probability distributions for individual classes. During the training process, the output of the Softmax function may be used to calculate the loss function and update the network parameters through a back propagation algorithm. Notably, the output of the Softmax function does not change the relative magnitude relationship between elements, but rather normalizes them. Thus, the Softmax function does not change the characteristics of the input vector, but simply converts it into a probability distribution form.
It should be appreciated that the one-dimensional convolutional layer-based weight timing feature extractor, the feature embedding feedback fusion module, and the classifier need to be trained prior to the inference using the neural network model described above. That is, the control method for improving the stability of the polyoxymethylene product further comprises a training stage for training the weight time sequence feature extractor based on the one-dimensional convolution layer, the feature embedding feedback fusion module and the classifier.
Fig. 3 is a flow chart of a training phase of a control method for improving the stability of a polyoxymethylene product according to an embodiment of the present application. As shown in fig. 3, a control method for improving stability of a polyoxymethylene product according to an embodiment of the present application includes: a training phase comprising: s110, training data are acquired, wherein the training data comprise training weight signals and training liquid level signals at a plurality of preset time points in a preset time period, and the rotating speed value of the variable frequency motor at the current time point is required to be increased, reduced or maintained; s120, arranging the training weight signals and the training liquid level signals at a plurality of preset time points into training weight signal time sequence input vectors and training liquid level signal time sequence input vectors according to a time dimension; s130, vector segmentation is carried out on the training weight signal time sequence input vector so as to obtain a sequence of training weight signal local time sequence input vector; s140, passing the sequence of training weight signal local time sequence input vectors through the weight time sequence feature extractor based on the one-dimensional convolution layer to obtain a sequence of training weight signal local time sequence feature vectors; s150, passing the sequence of the training weight signal local time sequence feature vector and the training liquid level signal time sequence input vector through the feature embedding feedback fusion module to obtain a training weight-liquid level time sequence embedding feedback fusion feature vector; s160, optimizing the training weight-liquid level time sequence embedded feedback fusion feature vector position by position to obtain an optimized training weight-liquid level time sequence embedded feedback fusion feature vector; s170, embedding the optimized training weight-liquid level time sequence into a feedback fusion feature vector to pass through the classifier so as to obtain a classification loss function value; and S180, training the weight time sequence feature extractor based on the one-dimensional convolution layer, the feature embedding feedback fusion module and the classifier based on the classification loss function value.
Wherein, insert feedback fusion eigenvector through said classifier with said optimization training weight-liquid level time sequence to get the classification loss function value, include: and calculating a cross entropy loss function value between the training classification result and a true value which is required to be increased, reduced or maintained by the rotating speed value of the variable frequency motor at the current time point as the classification loss function value.
In particular, in the technical solution of the present application, each weight signal local time sequence feature vector in the sequence of weight signal local time sequence feature vectors expresses a time sequence correlation feature in a local time domain determined by global time domain under vector slicing, and the liquid level signal time sequence input vector expresses a global time domain source time sequence distribution of a liquid level signal, so that after the sequence of weight signal local time sequence feature vectors and the liquid level signal time sequence input vector pass through a feature embedding feedback fusion module, the time sequence correlation feature distribution in the local time domain can be constrained by the source time sequence distribution in the global time domain, so that the weight-liquid level time sequence embedding feedback fusion feature vector has a cross-source domain-feature domain time sequence correlation distribution in a global-local mixed time domain space scale. That is, since the weight-liquid level time sequence embedded feedback fusion feature vector has a multi-scale cross-domain time sequence associated feature distribution property as a whole, it is required to improve the classification regression efficiency when the weight-liquid level time sequence embedded feedback fusion feature vector is classified and regressed by a classifier. Therefore, when the weight-liquid level time sequence embedded feedback fusion feature vector carries out classification regression through a classifier, the applicant of the application carries out position-by-position optimization on the weight-liquid level time sequence embedded feedback fusion feature vector, which is specifically expressed as follows:
Wherein v is i Is the characteristic value of the ith position of the weight-liquid level time sequence embedded feedback fusion characteristic vector,is the weight-liquid level time sequence embedded feedback meltGlobal average of all eigenvalues of the sum eigenvector and v max Is the maximum eigenvalue of the weight-liquid level time sequence embedded feedback fusion eigenvector, exp () represents exponential operation, v i ' is the optimal training weight-level time sequence embedded feedback fusion feature vector. That is, through the regularization function-like concept of global distribution parameters, the optimization simulates a cost function with the regular expression of regression probability based on the parameter vector representation of global distribution of the weight-liquid level time sequence embedded feedback fusion feature vector, so that the feature manifold representation of the weight-liquid level time sequence embedded feedback fusion feature vector in a high-dimensional feature space models the point-by-point regression characteristic of a weight matrix based on a classifier under a quasi-regression probability to capture a parameter smooth optimization track of the weight-liquid level time sequence embedded feedback fusion feature vector to be classified under the scene geometry of the high-dimensional feature manifold through the parameter space of a classifier model, and the training efficiency of the weight-liquid level time sequence embedded feedback fusion feature vector under the classification probability regression of the classifier is improved. Therefore, the flow fluctuation caused by liquid level change can be correspondingly compensated to improve the flow stability, and by adopting the mode, the accurate control of additives in a polymerizer and an extruder can be realized, and the stability of a polyoxymethylene product is effectively improved.
In summary, the control method for improving the stability of the polyoxymethylene product according to the embodiment of the present application is clarified by adding feedforward-feedback control, using a sensor to send a weight time sequence change signal to a controller as a feedback signal, and simultaneously sending a liquid level value signal to the controller as a feedforward signal, so that the disturbance caused by the liquid level change is synchronously compensated, and the system fluctuation caused by the liquid level fluctuation is avoided. Thus, the precise control of the additives in the polymerizer and the extruder can be realized, and the stability of the polyoxymethylene product is effectively improved.
Further, a control system for improving the stability of a polyoxymethylene product is also provided.
Fig. 5 is a block diagram of a control system for improving the stability of a polyoxymethylene product according to an embodiment of the present application. As shown in fig. 5, a control system 300 for improving the stability of a polyoxymethylene product according to an embodiment of the present application includes: a signal acquisition module 310 for acquiring weight signals and liquid level signals at a plurality of predetermined time points within a predetermined period of time; an arrangement module 320, configured to arrange the weight signals and the liquid level signals at the plurality of predetermined time points into a weight signal time sequence input vector and a liquid level signal time sequence input vector according to a time dimension; the local time sequence feature analysis module 330 is configured to perform local time sequence feature analysis on the weight signal time sequence input vector to obtain a sequence of weight signal local time sequence feature vectors; the embedded feedback fusion module 340 is configured to perform embedded feedback fusion on the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector to obtain a weight-liquid level time sequence embedded feedback fusion feature; and a control result generating module 350, configured to determine, based on the weight-liquid level time sequence embedded feedback fusion feature, that the rotation speed value of the variable frequency motor at the current time point should be increased, decreased or maintained.
As described above, the control system 300 for improving the stability of a polyoxymethylene product according to an embodiment of the present application may be implemented in various wireless terminals, such as a server or the like having a control algorithm for improving the stability of a polyoxymethylene product. In one possible implementation, the control system 300 for improving the stability of a polyoxymethylene product according to an embodiment of the present application may be integrated into a wireless terminal as one software module and/or hardware module. For example, the control system 300 for improving the stability of the polyoxymethylene product may be a software module in the operating system of the wireless terminal, or may be an application developed for the wireless terminal; of course, the control system 300 for improving the stability of the polyoxymethylene product may also be one of a number of hardware modules of the wireless terminal.
Alternatively, in another example, the control system 300 for improving stability of a polyoxymethylene product and the wireless terminal may be separate devices, and the control system 300 for improving stability of a polyoxymethylene product may be connected to the wireless terminal through a wired and/or wireless network and transmit interactive information in a agreed data format.
In one example, the post-aggregation processing runs stably for a long period, and optimization compatibility and overall control are required among all logic modules.
1. Adding logic sequential control program
1. The polyformaldehyde has strict requirements on the feeding sequence and interval of the polymerized monomers and the auxiliary agents, and the abnormal feeding sequence or overlong feeding interval of different materials can cause the polymerization reaction to not normally occur; according to the method, the one-key control feeding is realized by adding the logic sequential control program, so that the feeding sequence and interval of various materials are more accurate.
2. Weightless feeding control system
The weightless feeding device is a feeding device commonly used in fine chemical industry in recent years, and is mostly used in occasions with high feeding precision requirements and strict material proportion. In a polyoxymethylene polymerization device, the procedures of adding an initiator into a polymerization machine and adding various auxiliary agents in the extrusion process are all high in precision requirement, and in the occasion with strict proportioning requirement, the conventional flow meter is difficult to meet the actual requirement, and the problems can be well solved by adopting a weightless weighing rule.
1. Working principle of weightlessness system
The weight loss method is to flow out materials from a container, and measure the change value of weight in unit time, namely 'lost weight', by a weighing mode, so as to obtain the instantaneous flow of fluid.
As shown in fig. 6, the load cell sends the weight signal to the flow controller "FC", the controller calculates the flow, compares the flow with the set value "SP" to obtain a deviation, calculates the deviation, and outputs a control signal to the frequency converter f. And then the rotating speed of the variable frequency motor is regulated to achieve the purpose of controlling the flow.
From the above, the sensor, the controller, the frequency converter and the variable frequency motor form a closed loop regulating circuit, so that the flow can be stabilized. The requirements of common application occasions can be met, but the requirements of raw material proportion are high for polymerization reaction, so that the adding flow is more stable, and the object characteristics are also needed to be considered.
The periodic variation of the liquid level in the apparatus can be divided into 2 phases:
(1) When the liquid level reaches high, the charging valve is closed, and the flow is regulated according to the weighing signal, which is called a weight metering mode.
(2) When reaching low liquid level, the charging valve is opened, and the motor keeps the last moment of rotating speed, which is called a volume metering mode.
2. Improvement measure
Whether metering in gravimetric or volumetric mode, the periodic variation of the liquid level in the apparatus produces a non-negligible disturbance to the flow rate fed to the polymerizer; in particular, the method is equivalent to open-loop control in a volume mode, so that flow fluctuation caused by liquid level change is correspondingly compensated, and flow stability is improved. Through simulation, the flow fluctuation caused by the weight of the material in the storage tank and the liquid level change is calculated to be in a linear relation, and the sensor can be used for sending a weight change rate signal to the controller to be used as a feedback signal by adding feedforward-feedback control, and sending a weight value signal to the controller to be used as a feedforward signal, so that the interference caused by the liquid level change is synchronously compensated, and the system fluctuation caused by the liquid level fluctuation is avoided, as shown in fig. 7 and 8.
3. Optimizing the location
(1) Charging system for initiator of polymerization machine
The weight loss control system is adopted to replace the traditional diaphragm metering pump filling system, so that the stability of the addition flow of auxiliary agents such as a system initiator and the like is effectively controlled, and the uniformity of products is fundamentally improved.
(2) Extruder auxiliary agent configuration system
a. The manual configuration is changed into automatic feeding, so that the contact between a person and powder is avoided, the working condition is improved, the labor intensity is reduced, and the powder utilization rate is improved;
b. the whole process is automatically produced, so that the production efficiency is improved, the weighing precision is controlled, and the stability and uniformity of the product are further improved;
c. the proportion control of the auxiliary agent and the powder is increased, so that the batch is changed from intermittent operation to continuous operation.
3. Sequence control program
The polymerization and post-treatment procedures in the preparation process of the polyformaldehyde are the cooperation of unit operation and time control, and the automatic control is difficult to realize by adopting control procedures such as pressure, liquid level, cascade and the like in the traditional fine chemical process; at present, most similar enterprises adopt manual operation, so that the workload of central control monitoring disc personnel is large, the control time is unequal due to factors such as fatigue and the like, and the overall working condition fluctuates.
At present, by powder conveying in the polymerization and post-treatment processes, mixing, ageing and conveying of granules, and adding sequential control procedures in powder configuration, the accurate control of time and the stability of the whole working condition are improved, and meanwhile, the workload of central control personnel is reduced.
1. Powder conveying system
The flow chart of the powder conveying system and the sequence control procedure are shown in fig. 9.
Sequential control logic:
(1) Selecting an automatic state
(2) Selection target stock bin
(3) Setting the feeding amount of a target bin
(4) The starting sequence is that a feeding valve V3 of the target bin is opened, a valve opening signal is fed back to the DCS, a conveyor is started, a conveyor starting signal is fed back to the system, and a blanking valve 1 is started to start feeding to the target bin after a delay of 10 s; when the feeding amount reaches the feeding amount of the set target bin, the feeding valve V4 of the other bin is automatically opened, and the valve V3 is closed after a valve V4 opening signal is fed back to the background.
2. Pellet delivery system
4. Advanced control technique
At present, enterprises such as petroleum, chemical industry and the like mostly adopt a traditional DCS control system, a control loop adopts a control mode of a valve, except that a small part of loops such as pressure control, liquid level control and the like adopt a cascade control mode, the rest adopts a manual control mode, the single variable control mode starts feedback adjustment when working conditions change, and adjustment is lagged, and because the single variable control mode is variable adjustment, the workload of operators is large, and system fluctuation is very easy to cause.
The advanced control system calculates the optimal control parameters through simulation by collecting big data parameters of stable operation of working conditions, replaces manual prejudgment by the DCS control system during system operation, performs advanced control and multi-variable integral adjustment to ensure the stable operation of the system, and greatly reduces the operation intensity of staff.
The control system architecture is shown in fig. 10.
The polymerization post-treatment is integrally controlled by adopting an advanced control system, so that the purposes of stable operation of the device, more accurate addition of the auxiliary agent, labor intensity reduction, efficiency improvement, energy conservation and consumption reduction are realized.
The foregoing description of the embodiments of the present disclosure has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (9)
1. A control method for improving the stability of a polyoxymethylene product, comprising:
acquiring weight signals and liquid level signals at a plurality of preset time points in a preset time period;
arranging the weight signals and the liquid level signals of the plurality of preset time points into weight signal time sequence input vectors and liquid level signal time sequence input vectors according to a time dimension;
Carrying out local time sequence feature analysis on the weight signal time sequence input vector to obtain a sequence of weight signal local time sequence feature vectors;
embedding feedback fusion is carried out on the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector so as to obtain a weight-liquid level time sequence embedding feedback fusion feature; and
and based on the weight-liquid level time sequence embedded feedback fusion characteristic, determining that the rotating speed value of the variable frequency motor at the current time point should be increased, decreased or maintained.
2. The control method for improving the stability of a polyoxymethylene product as set forth in claim 1, wherein performing a local time series feature analysis on the weight signal time series input vector to obtain a sequence of weight signal local time series feature vectors, comprises:
vector segmentation is carried out on the weight signal time sequence input vector so as to obtain a sequence of weight signal local time sequence input vectors; and
the sequence of weight signal local time sequence input vectors is passed through a weight time sequence feature extractor based on a one-dimensional convolution layer to obtain the sequence of weight signal local time sequence feature vectors.
3. The control method for improving the stability of a polyoxymethylene product according to claim 2, wherein performing an embedded feedback fusion of the sequence of weight signal local timing feature vectors and the liquid level signal timing input vector to obtain a weight-liquid level timing embedded feedback fusion feature, comprises: and the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector pass through a feature embedding feedback fusion module to obtain the weight-liquid level time sequence embedding feedback fusion feature vector as the weight-liquid level time sequence embedding feedback fusion feature.
4. The control method for improving the stability of a polyoxymethylene product according to claim 3, wherein passing the sequence of weight signal local timing feature vectors and the liquid level signal timing input vector through a feature embedding feedback fusion module to obtain the weight-liquid level timing embedding feedback fusion feature vector as the weight-liquid level timing embedding feedback fusion feature, comprises:
the liquid level signal time sequence input vector passes through a liquid level signal time sequence feature extractor based on a full convolution neural network model to obtain a liquid level signal full convolution time sequence feature vector;
arranging the sequence of the weight signal local time sequence feature vectors into a weight signal global time sequence feature vector;
fusing the liquid level signal full convolution time sequence feature vector and the weight signal global time sequence feature vector to obtain a liquid level-weight fusion feature vector;
carrying out semantic association coding on the sequence of the weight signal local time sequence feature vectors to obtain weight global semantic association feature vectors;
and fusing the liquid level-weight fusion feature vector and the weight global semantic association feature vector to obtain the weight-liquid level time sequence embedded feedback fusion feature vector.
5. The control method for improving the stability of a polyoxymethylene product as set forth in claim 4, wherein determining that a value of a rotational speed of the variable frequency motor at a current point in time should be increased, decreased or maintained based on the weight-level time series embedded feedback fusion feature, comprises: and embedding the weight-liquid level time sequence into a feedback fusion feature vector to obtain a classification result through a classifier, wherein the classification result is used for indicating that the rotating speed value of the variable frequency motor at the current time point should be increased, decreased or maintained.
6. The control method for improving the stability of a polyoxymethylene product as set forth in claim 5, further comprising a training step of: the weight time sequence feature extractor based on the one-dimensional convolution layer, the feature embedding feedback fusion module and the classifier are used for training.
7. The control method for improving the stability of a polyoxymethylene product as set forth in claim 6, wherein said training step comprises:
acquiring training data, wherein the training data comprises training weight signals and training liquid level signals at a plurality of preset time points in a preset time period, and a real value which is required to be increased, reduced or maintained by a rotating speed value of a variable frequency motor at the current time point;
Arranging the training weight signals and the training liquid level signals at a plurality of preset time points into training weight signal time sequence input vectors and training liquid level signal time sequence input vectors according to a time dimension;
vector segmentation is carried out on the training weight signal time sequence input vector so as to obtain a sequence of training weight signal local time sequence input vector;
passing the sequence of training weight signal local time sequence input vectors through the weight time sequence feature extractor based on the one-dimensional convolution layer to obtain a sequence of training weight signal local time sequence feature vectors;
the training weight signal local time sequence feature vector sequence and the training liquid level signal time sequence input vector are processed through the feature embedding feedback fusion module to obtain a training weight-liquid level time sequence embedding feedback fusion feature vector;
performing position-by-position optimization on the training weight-liquid level time sequence embedded feedback fusion feature vector to obtain an optimized training weight-liquid level time sequence embedded feedback fusion feature vector;
embedding the optimized training weight-liquid level time sequence into a feedback fusion feature vector to pass through the classifier so as to obtain a classification loss function value; and
and training the weight time sequence feature extractor based on the one-dimensional convolution layer, the feature embedding feedback fusion module and the classifier based on the classification loss function value.
8. The control method for improving stability of a polyoxymethylene product as set forth in claim 7, wherein embedding the optimized training weight-level time series into a feedback fusion feature vector through the classifier to obtain a class loss function value, comprises:
processing the optimized training weight-level time sequence embedded feedback fusion feature vector by using the classifier to obtain training classification results, and
and calculating a cross entropy loss function value between the training classification result and a true value which is to be increased, decreased or maintained by the rotating speed value of the variable frequency motor at the current time point as the classification loss function value.
9. A control system for improving the stability of a polyoxymethylene product, comprising:
the signal acquisition module is used for acquiring weight signals and liquid level signals at a plurality of preset time points in a preset time period;
the arrangement module is used for arranging the weight signals and the liquid level signals of the plurality of preset time points into weight signal time sequence input vectors and liquid level signal time sequence input vectors according to the time dimension;
the local time sequence feature analysis module is used for carrying out local time sequence feature analysis on the weight signal time sequence input vector so as to obtain a sequence of weight signal local time sequence feature vectors;
The embedded feedback fusion module is used for carrying out embedded feedback fusion on the sequence of the weight signal local time sequence feature vector and the liquid level signal time sequence input vector to obtain a weight-liquid level time sequence embedded feedback fusion feature; and
and the control result generation module is used for determining that the rotating speed value of the variable frequency motor at the current time point should be increased, decreased or kept based on the weight-liquid level time sequence embedded feedback fusion characteristic.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311470435.8A CN117531438A (en) | 2023-11-06 | 2023-11-06 | Control method and system for improving stability of polyoxymethylene product |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311470435.8A CN117531438A (en) | 2023-11-06 | 2023-11-06 | Control method and system for improving stability of polyoxymethylene product |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117531438A true CN117531438A (en) | 2024-02-09 |
Family
ID=89793013
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311470435.8A Pending CN117531438A (en) | 2023-11-06 | 2023-11-06 | Control method and system for improving stability of polyoxymethylene product |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117531438A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118011778A (en) * | 2024-03-06 | 2024-05-10 | 湖南省海昆农业科技有限公司 | Temperature self-adaptive regulation and control method of insect protein fresh-keeping system |
-
2023
- 2023-11-06 CN CN202311470435.8A patent/CN117531438A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118011778A (en) * | 2024-03-06 | 2024-05-10 | 湖南省海昆农业科技有限公司 | Temperature self-adaptive regulation and control method of insect protein fresh-keeping system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN117531438A (en) | Control method and system for improving stability of polyoxymethylene product | |
US10466658B2 (en) | Numerical controller and machine learning device | |
CN109460890B (en) | Intelligent self-healing method based on reinforcement learning and control performance monitoring | |
KR20190132571A (en) | Systems and methods for control of polymer reactions and processing using automatic continuous online monitoring | |
Charest et al. | Integration of artificial intelligence in an injection molding process for on-line process parameter adjustment | |
CN1129037A (en) | Method and apparatus for fuzzy logic control with automatic tuning | |
CN113408648A (en) | Unit combination calculation method combined with deep learning | |
CN111815053A (en) | Prediction method and system for industrial time sequence data | |
Qin et al. | An intelligent non-optimality self-recovery method based on reinforcement learning with small data in big data era | |
CN103226728B (en) | High density polyethylene polymerization cascade course of reaction Intelligent Measurement and yield optimization method | |
CN117160072B (en) | Multidirectional dynamic continuous extraction system and method thereof | |
CN100580585C (en) | Soft measuring method of industrial process under condition of small sample | |
CN101285816A (en) | Copper matte air refining procedure parameter soft sensing instrument and its soft sensing method | |
CN114403486A (en) | Intelligent control method of airflow type cut-tobacco drier based on local peak value coding circulation network | |
CN113885310A (en) | Intelligent control system for vacuum dry pump test | |
CN110794676A (en) | CSTR process nonlinear control method based on Hammerstein-Wiener model | |
Hosen et al. | Prediction interval-based modelling of polymerization reactor: A new modelling strategy for chemical reactors | |
Zhang et al. | Optimization control of a fed-batch process using an improved reinforcement learning algorithm | |
Duran-Villalobos et al. | Iterative learning modelling and control of batch fermentation processes | |
Owa | Nonlinear Model Predictive Control Strategy Based on Soft Computing Approaches and Real Time Implementation on a Coupled-Tank System | |
CN106127326B (en) | Chemical material processing melt index forecasting method | |
Garg et al. | Subspace identification and predictive control of batch particulate processes | |
CN118099915B (en) | Constant-current driving method for laser power supply | |
CN118426306A (en) | Offshore platform three-phase separator liquid level control method and system | |
Ichikawa et al. | Dynamical Mechanism of Sampling-Based Probabilistic Inference Under Probabilistic Population Codes |
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 |