A fluorine chemical industry multi-regulating valve automatic control method and system

Abstract

The embodiment of the present application relates to the technical field of automatic control of multiple regulating valves in fluorine chemical industry, and provides a method and system for automatic control of multiple regulating valves in fluorine chemical industry, so as to solve the problems of low efficiency and poor safety of automatic control of multiple regulating valves in fluorine chemical industry in the prior art. Among them, based on the Markov decision process, the optimal operating conditions under the current fluorine chemical production conditions are evaluated and the optimal operating conditions are predicted and processed; based on the process safety limit and the product quality target, an ideal opening adjustment plan is generated, and the pulse width modulation control instruction is sent to the corresponding actuator of the fluorine chemical industry multiple regulating valve through the industrial communication bus to obtain the actual valve control strategy of the fluorine chemical industry multiple regulating valve; the difference between the actual effect and the expected effect is compared to obtain an improved control strategy; the control strategy is continuously optimized based on the experience accumulated from the improved control strategy. The technical solution provided by the embodiment of the present application can improve the efficiency and safety of automatic control of multiple regulating valves in fluorine chemical industry.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of automatic control of multiple regulating valves in fluorine chemical industry, and in particular to a method and system for automatic control of multiple regulating valves in fluorine chemical industry. Background Art

[0002] In the fluorine chemical industry, precise control of multi-control valves is essential to ensure product quality, improve production efficiency and reduce energy consumption. Due to the complex and changing working conditions in the production process, a method is needed to evaluate the optimal operating conditions in real time and dynamically adjust the valve opening according to these conditions. In addition, in order to meet strict process safety limits and product quality targets, the control system must have a high degree of safety and reliability, and it must also be able to adapt to equipment aging and environmental changes.

[0003] The existing fluorine chemical multi-control valve automatic control method usually adopts the following steps: compare the expected behavior of the physical entity with the pre-defined behavior pattern through the fuzzy rule library to identify the behavior deviation area; use fuzzy reasoning rules to quantitatively analyze the degree and type of behavior deviation to reduce uncertainty; use basic machine vision technology to monitor key interaction points, but lack the ability of fine-grained monitoring and space-time fusion algorithm to create a three-dimensional space-time model; correction instructions are directly applied to the actual system without prior simulation execution in the digital twin model, and the feedback mechanism is simple, usually only involving error integration, which is used for subsequent control adjustments.

[0004] Although the above existing solutions have improved the control level to a certain extent, there are still some significant shortcomings: traditional controllers and other static control strategies are difficult to quickly adapt to changes in operating conditions, which may cause fluctuations in the production process and affect product quality; for complex systems involving multiple variables, such as multi-control valve control in fluorine chemical industry, it is difficult for existing solutions to find the global optimal solution, especially when considering safety limits and product quality goals at the same time; existing control systems have poor adaptability to environmental changes and equipment aging, and may not be able to maintain the initial control performance after long-term operation. Summary of the invention

[0005] The embodiments of the present application provide a method and system for automatic control of multiple regulating valves in a fluorine chemical industry, which are used to solve the problems of low efficiency and poor safety of automatic control of multiple regulating valves in a fluorine chemical industry in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a fluorine chemical industry multi-regulating valve automatic control method, comprising: Based on the Markov decision process, the optimal operating conditions under the current fluorine chemical production conditions are evaluated, and the optimal operating conditions are predicted and processed to obtain the optimal process state; According to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search for the opening of the fluorine chemical multi-regulating valve, and an ideal opening adjustment scheme is generated based on the process safety limit and the product quality target, wherein the ideal opening adjustment scheme meets the safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production; Convert the ideal opening adjustment scheme into a pulse width modulation control instruction, send the pulse width modulation control instruction to the corresponding actuator of the fluorine chemical multi-control valve through an industrial communication bus, and use fault prediction and health management technology to identify and handle potential equipment failure risks, so as to obtain an actual valve control strategy for the fluorine chemical multi-control valve; Using an adaptive feedback loop of a deep neural network, the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme is compared, and the actual valve control strategy is dynamically corrected using a back propagation algorithm to obtain an improved control strategy; By utilizing an online learning mechanism, the control strategy of the fluorine chemical multi-regulating valve is continuously optimized based on the experience accumulated during the operation of the improved control strategy.

[0007] Optionally, the ideal opening adjustment scheme is converted into a pulse width modulation control instruction, and the pulse width modulation control instruction is sent to the corresponding actuator of the fluorine chemical multi-control valve through an industrial communication bus, and the potential equipment failure risk is identified and processed by using fault prediction and health management technology to obtain the actual valve control strategy of the fluorine chemical multi-control valve, including: Based on the ideal opening adjustment scheme, converting the opening information in the ideal opening adjustment scheme into a pulse width modulation control instruction of the actuator of the fluorine chemical multi-regulating valve to generate a pulse width modulation signal; Based on the pulse width modulation signal and the industrial communication bus, in combination with the time synchronization protocol, the pulse width modulation control instruction is timestamped, and the pulse width modulation control instruction is optimized and scheduled through distributed control to ensure that the pulse width modulation control instruction is sent to the corresponding actuator of the fluorine chemical multi-regulating valve, and a communication connection is established between the industrial communication bus and the actuator of the fluorine chemical multi-regulating valve; Based on the communication connection, using fault prediction and health management technology, combined with real-time data analysis and pattern recognition algorithms, the operation status of the communication connection is monitored in real time, an abnormality detection algorithm is used to automatically identify abnormal operation modes of the communication connection, and potential risks are evaluated through a machine learning model to generate a potential risk assessment report; Based on the potential risk assessment report, adaptive filtering is used to filter noise of the communication connection, and preventive measures are generated to ensure safe operation of the communication connection; Based on the preventive measures and the pulse width modulation signal, the actuator of the fluorine chemical multi-control valve adjusts the opening information intelligently to generate an actual valve control strategy.

[0008] Optionally, based on the communication connection, using fault prediction and health management technology, combined with real-time data analysis and pattern recognition algorithms, the operation status of the communication connection is monitored in real time, an abnormality detection algorithm is used to automatically identify abnormal operation modes of the communication connection, and potential risks are evaluated through a machine learning model to generate a potential risk assessment report, including: Based on the communication connection, using fault prediction and health management technology, combined with real-time data analysis and pattern recognition algorithms, continuously monitoring the operating status of the communication connection, collecting status information of the communication connection through sensors on the industrial communication bus, and transmitting the status information to the central processing unit of the communication connection to obtain the operating status of the communication connection; According to the operation status of the communication connection, the collected operation status is analyzed and processed using an anomaly detection algorithm to automatically identify abnormal operation modes of the communication connection to generate a potential abnormal operation mode report; Based on the potential abnormal operation mode report, potential risks are evaluated through a machine learning model to generate a potential risk assessment report.

[0009] Optionally, using an adaptive feedback loop of a deep neural network, comparing the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, and dynamically correcting the actual valve control strategy using a back propagation algorithm to obtain an improved control strategy, including: Based on the adaptive feedback loop of the deep neural network, the operating state of the actual valve control strategy is monitored and analyzed to obtain the actual effect of the actual valve control strategy; Based on the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, a comparison is performed to analyze the difference between the actual effect and the expected effect to generate a difference analysis result; According to the difference analysis results, using the back propagation algorithm in the deep neural network, analyzing the specific reasons causing the differences in the difference analysis results, and adjusting the weights of the deep neural network to generate correction suggestions; According to the correction suggestion, the actual valve control strategy is optimized to improve the accuracy and efficiency of the actual valve control strategy and generate an improved control strategy.

[0010] Optionally, according to the difference analysis results, using a back propagation algorithm in a deep neural network, analyzing specific causes of the differences in the difference analysis results, and adjusting the weights of the deep neural network to generate correction suggestions, including: According to the difference analysis result, the gap between the actual effect obtained and the expected effect is quantified to obtain an initial error value, the initial error value is transmitted layer by layer through the back propagation algorithm in the deep neural network, and the specific cause of the gap is identified to generate a preliminary recognition result; Based on the preliminary identification results, the key influencing factors in the preliminary identification results are extracted by principal component analysis, and the key influencing factors are refined in combination with sensitivity analysis to obtain a list of key influencing factors; According to the list of key influencing factors, using adaptive moment estimation, adjusting the key influencing factors to reduce the difference between the actual effect and the expected effect, so as to obtain preliminary correction suggestions; Based on the preliminary correction suggestions, the weights of the deep neural network are adjusted, and the obtained control strategy is improved to generate correction suggestions.

[0011] Optionally, according to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search on the opening of the fluorine chemical multi-control valve, and based on the process safety limit and product quality target, an ideal opening adjustment plan is generated, including: Analyzing the opening settings of the fluorine chemical multi-regulating valve according to the optimal operating conditions and the optimal process state to generate different opening setting combinations; Based on the opening setting combination, a group of the opening setting combinations representing different opening settings are initialized as initial solutions, and according to the current optimal operating conditions and the optimal process state, the energy efficiency ratio, safety and product quality performance of each of the opening setting combinations are evaluated to obtain candidate solutions composed of the opening setting combinations with excellent performance; Based on the candidate solutions, using the selection, crossover and mutation steps in the genetic algorithm, exploring the adjustment solutions of the opening setting combinations in the candidate solutions to generate a preliminary ideal opening adjustment solution; Based on the preliminary ideal opening adjustment plan, the iterative optimization is continued, and an ideal opening adjustment plan is generated according to the process safety limit and product quality goals. The ideal opening adjustment plan achieves the highest energy efficiency ratio of the fluorine chemical production.

[0012] Optionally, it is characterized in that, using an online learning mechanism, based on the experience accumulated during the operation of the improved control strategy, the control strategy of the fluorine chemical multi-regulating valve is continuously optimized, including: Based on the online learning mechanism, the operating status of the improved control strategy is monitored to obtain an information report reflecting the actual operating effect of the improved control strategy; Based on the information report, according to the information of the operating status of the improved control strategy collected in real time by the distributed sensor network, the actual operating effect of the improved control strategy is analyzed, the successful operating mode of the improved control strategy is identified, and the operating mode is accumulated as experience to generate an experience library; Based on the experience database, the control strategy of the fluorine chemical multi-control valve is updated by using machine learning, the weight in the control strategy is adjusted, the control strategy is ensured to adapt to the production needs of the fluorine chemical industry, and an updated control strategy is generated; The updated control strategy is processed with optimization suggestions to further improve the control accuracy and control efficiency of the fluorine chemical multi-control valve, so as to generate optimization suggestions, and the optimization suggestions are applied to actual control to ensure that the control strategy of the fluorine chemical multi-control valve is continuously optimized during operation.

[0013] Optionally, the fault prediction and health management technology is used to identify and handle potential equipment failure risks to obtain an actual valve control strategy for the fluorine chemical multi-regulating valve, including: Continuously monitor the health status of equipment through fault prediction and health management technology to identify and address potential equipment failure risks ; Potential equipment failure risk The calculation formula is: ; in, is the potential equipment failure risk, ranging from no risk to high risk, with a value range of between, The health index representing the health management technology reflects the health of the device at a future point in time. health status, represents the preset safety threshold, that is, the safety upper limit of the health index. When it is less than the safety threshold, it indicates that there is a potential risk of equipment failure. Represents the urgency of the expected failure time, indicating the distance from the expected time of equipment failure to the current time. Represents the maximum expected failure time urgency, used for standardization The maximum value of The severity of the equipment failure indicates the degree of impact on the fluorine chemical production process when the equipment failure occurs. is the maximum device fault severity, used to normalize The maximum value of is the repair cost, which represents the cost or resource consumption required to repair the equipment failure. is the maximum repair cost, used for normalization The maximum value of is the weight coefficient used to adjust the importance of the health index part, is the nonlinear influence index of the health index, used to capture the nonlinear trend of the health index changing over time, is the weight coefficient used to adjust the importance of the urgency of the estimated failure time part, is the weight coefficient used to adjust the importance of the device failure severity part, is a weight coefficient used to adjust the importance of the repair cost component; The potential equipment failure risk is analyzed and processed to obtain the actual valve control strategy of the fluorine chemical multi-regulating valve.

[0014] Optionally, according to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search on the opening of the fluorine chemical multi-regulating valve, and an ideal opening adjustment scheme is generated based on the process safety limit and the product quality target, wherein the ideal opening adjustment scheme meets the safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production, including: According to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search on the opening of the fluorine chemical multi-regulating valve; Global search calculation formula: ; in, is the global search calculation formula, i.e., the best opening of the fluorine chemical multi-regulating valve found, is the fitness function, which is used to evaluate the quality of each opening setting. Represents the opening setting space of all the fluorine chemical multi-control valves, fitness function , Represents the opening setting The energy efficiency ratio of fluorine chemical production under represents the safety score, is the product quality rating, , , , , , are the corresponding weight coefficient and nonlinear index, respectively, reflecting the importance of the safety score, the product quality score and the energy efficiency ratio of the fluorine chemical production and their trends over time or conditions; Based on the calculation results of the global search, an ideal opening adjustment plan is generated according to the process safety limit and the product quality target; Calculation formula for ideal opening adjustment scheme: ; in, is the ideal opening adjustment scheme, X represents the opening setting of the fluorine chemical multi-regulating valve, is a penalty function used to deal with individuals that exceed the process safety limit and the product quality target; Penalty Function : ; in, represents the process safety limit, Represents the product quality target, Represents the actual process safety value under the current opening setting. is the actual product quality value under the current opening setting, and are the corresponding safety and quality penalty coefficients, which are used to adjust the severity of constraint violations; Fitness function: , Represents the opening setting The energy efficiency ratio of fluorine chemical production under represents the safety score, is the product quality rating, , , , , , are the corresponding weight coefficient and nonlinear index, respectively, reflecting the importance of the safety score, the product quality score and the energy efficiency ratio of the fluorine chemical production and their trends over time or conditions.

[0015] In a second aspect, the present application provides a fluorine chemical multi-regulating valve automatic control system, including: A prediction module, which evaluates the optimal operating conditions under the current fluorine chemical production conditions based on the Markov decision process, and performs prediction processing on the optimal operating conditions to obtain the optimal process state; A search module, which uses a genetic algorithm to perform a global search for the opening of the fluorine chemical multi-regulating valve according to the optimal operating conditions and the optimal process state, and generates an ideal opening adjustment plan based on process safety limits and product quality targets, wherein the ideal opening adjustment plan meets safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production; An identification module converts the ideal opening adjustment scheme into a pulse width modulation control instruction, sends the pulse width modulation control instruction to the corresponding actuator of the fluorine chemical multi-control valve through an industrial communication bus, and uses fault prediction and health management technology to identify and handle potential equipment failure risks to obtain an actual valve control strategy for the fluorine chemical multi-control valve; A correction module, using an adaptive feedback loop of a deep neural network, compares the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, and dynamically corrects the actual valve control strategy using a back propagation algorithm to obtain an improved control strategy; The optimization module uses an online learning mechanism to continuously optimize the control strategy of the fluorine chemical multi-control valve based on the experience accumulated during the operation of the improved control strategy.

[0016] In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a fluorine chemical multi-regulating valve automatic control method as described in the first aspect above.

[0017] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a fluorine chemical multi-regulating valve automatic control method as described in the first aspect.

[0018] In an embodiment of the present application, the optimal operating conditions under the current fluorine chemical production conditions are evaluated based on a Markov decision process, and the optimal operating conditions are predicted and processed to obtain an optimal process state; according to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search for the opening of the fluorine chemical multi-control valve, and based on the process safety limit and the product quality target, an ideal opening adjustment scheme is generated, wherein the ideal opening adjustment scheme meets the safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production; the ideal opening adjustment scheme is converted into a pulse width modulation control instruction, and the pulse width modulation is transmitted through an industrial communication bus. The control instructions are sent to the corresponding actuators of the fluorine chemical multi-control valves, and the fault prediction and health management technology is used to identify and deal with potential equipment failure risks to obtain the actual valve control strategy of the fluorine chemical multi-control valves; the adaptive feedback loop of the deep neural network is used to compare the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, and the back propagation algorithm is used to dynamically correct the actual valve control strategy to obtain an improved control strategy; the online learning mechanism is used to continuously optimize the control strategy of the fluorine chemical multi-control valve according to the experience accumulated during the operation of the improved control strategy.

[0019] The technical solution of this application has the following beneficial effects: This application evaluates the optimal operating conditions under the current working conditions based on the Markov decision process, and performs predictive processing to obtain the optimal process state, thereby ensuring the high efficiency and stability of production; converts the ideal opening adjustment plan into a control instruction, and sends it to the actuator through the industrial communication bus, while using technology to identify and deal with potential equipment failure risks, ensuring the practical feasibility and safety of the valve control strategy; uses the adaptive feedback loop of the deep neural network to compare the difference between the actual effect and the expected effect, dynamically corrects the control strategy through the back-propagation algorithm, and continuously optimizes the control strategy in combination with the online learning mechanism to ensure that the system can self-learn and improve.

[0020] Furthermore, time synchronization protocols and distributed control are used to optimize the scheduling of pulse width modulation control instructions to ensure the timeliness and accuracy of instruction transmission; through real-time data analysis and pattern recognition algorithms, the status of communication connections is continuously monitored, abnormal operating modes are automatically identified, and the reliability of the system is enhanced; preventive measures are taken based on potential risk assessment reports, and the actuator's adjustment of opening information is intelligently adjusted to reduce the occurrence of failures and improve the safety of system operation.

[0021] Furthermore, sensors on the industrial communication bus continuously monitor the operating status of the communication connection to ensure that all relevant data is collected and transmitted to the central processing unit in a timely manner, providing comprehensive data support; anomaly detection algorithms are used to analyze the operating status, automatically identify abnormal operating modes, and generate detailed reports on potential abnormal operating modes to ensure rapid problem location; potential risks are evaluated based on machine learning models, risk assessment reports are generated, and effective preventive measures are guided, which improves the system's predictability and response speed, greatly enhances the system's reliability and robustness in the face of complex environmental changes, and ensures the safe and stable operation of the fluorine chemical production process.

[0022] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 A flow chart of a fluorine chemical industry multi-regulating valve automatic control method provided by the present application is shown; Figure 2 The schematic diagram of the structure of a fluorine chemical multi-regulating valve automatic control system provided by the present application is shown; Figure 3 A schematic diagram of the structure of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0026] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0028] Figure 1 A flowchart of a fluorine chemical multi-regulating valve automatic control method is provided for the embodiment of the present application, such as Figure 1 As shown, the method includes: 101. Evaluate the optimal operating conditions under the current fluorine chemical production conditions based on the Markov decision process, and predict and process the optimal operating conditions to obtain the optimal process state; In this step, the Markov decision process is a mathematical framework for modeling decision problems with randomness, which includes states, actions, transition probabilities, and reward functions. In fluorine chemical production, the Markov decision process is used to evaluate the optimal operating conditions under the current working conditions. By collecting real-time data such as temperature, pressure, flow, and historical operation records, a dynamic model is established to predict the results under different operating conditions. This information is used to determine the optimal process state, that is, the conditions that are most likely to achieve efficient, safe, and high-quality production.

[0029] In an application example, in an application scenario of intelligent prediction and preventive maintenance, it is assumed that a fluorine chemical plant faces the problem of raw material supply fluctuations. By deploying a wide network of sensors, engineers can obtain detailed production data streams. Using the Markov decision process model, the system can not only identify the current optimal operating settings, but also foresee upcoming challenges in advance, such as changes in reaction rates due to changes in raw material properties. This allows the factory to adjust production strategies before problems occur and avoid potential risks.

[0030] 102. According to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search for the opening of the fluorine chemical multi-control valve, and an ideal opening adjustment plan is generated based on the process safety limit and the product quality target, wherein the ideal opening adjustment plan meets the safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production; In this step, the genetic algorithm is a type of optimization algorithm that simulates natural selection and genetic mechanisms and is suitable for solving complex multivariable optimization problems. In this step, the genetic algorithm is used to explore all possible valve opening combinations to find the optimal solution that meets both safety requirements and product quality. The genetic algorithm iteratively improves the quality of the solution through operations such as encoding candidate solutions, evaluating fitness, selection, crossover, and mutation, and finally converges to one or more approximate optimal solutions.

[0031] In the application example, continuing the scenario in the above example, after determining the optimal process state, the factory needs to be further refined to a specific control level. At this point, the genetic algorithm comes into play. By exploring a large number of valve opening combinations, it finds a set of ideal settings that can maintain high production while ensuring that product quality is not affected. For example, some valves can significantly improve energy efficiency under certain conditions with only slight adjustments without sacrificing product purity or safety.

[0032] 103. Convert the ideal opening adjustment scheme into a pulse width modulation control instruction, send the pulse width modulation control instruction to the corresponding actuator of the fluorine chemical multi-control valve through the industrial communication bus, and use fault prediction and health management technology to identify and handle potential equipment failure risks, so as to obtain an actual valve control strategy for the fluorine chemical multi-control valve; In this step, pulse width modulation is a technology used to control power electronic devices, which adjusts power delivery by changing the duty cycle of the output signal. In this step, the ideal opening adjustment scheme is converted into pulse width modulation control instructions to accurately drive the actuator of the control valve. In addition, fault prediction and health management technology is used to monitor the health of the equipment, detect and deal with possible problems in a timely manner, and thus ensure the reliable operation of the system.

[0033] In the application examples, based on the 101 and 102 examples, the factory has obtained an ideal valve opening adjustment scheme. Now, this scheme is converted into pulse width modulation instructions and sent to each control valve through the industrial communication bus inside the factory. At the same time, the fault prediction and health management system is silently guarding the safety of the entire process in the background. For example, when a control valve shows an abnormal vibration pattern, the fault prediction and health management responds quickly, adjusts the working mode of the valve, and even recommends that the maintenance team check in advance to ensure the continuity and stability of production.

[0034] 104. Using an adaptive feedback loop of a deep neural network, compare the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, and dynamically modify the actual valve control strategy using a back propagation algorithm to obtain an improved control strategy; In this step, deep neural networks are a form of machine learning that is particularly good at processing data with complex nonlinear relationships. Adaptive feedback loops refer to the ability of a control system to continuously adjust its behavior based on real-time feedback. In this step, deep neural networks are used to compare the effects of the actual valve control strategy with the expectations of the ideal solution, learn the gap between the two through the back-propagation algorithm, and adjust the control logic accordingly to achieve better control accuracy and adaptability.

[0035] In the application embodiment, continuing the scenario of the above embodiment, the factory introduced a deep neural network adaptive feedback loop to enhance the flexibility of the control system. Every time the valve works according to the newly set opening, the deep neural network will compare the difference between the actual performance and the ideal effect. For example, if a valve fails to achieve the expected flow control at a specific opening, the deep neural network will fine-tune its control logic through the back-propagation algorithm to ensure that it can better match expectations under the same conditions next time. Over time, the system gradually learns how to respond to various working conditions more accurately.

[0036] 105. Utilize an online learning mechanism to continuously optimize the control strategy of the fluorine chemical multi-control valve based on the experience accumulated during the operation of the improved control strategy.

[0037] In this step, online learning means that the system can learn from the latest data and update its knowledge base without stopping. For fluorine chemical production, this means that the control system can instantly adjust its behavior mode according to the results of each operation without waiting for regular maintenance or restart. This approach allows the system to quickly adapt to environmental changes and technological advances, and always operate in the best state.

[0038] In the application embodiment, based on the previous steps, the factory has achieved true intelligent production. Each time the valve operates according to the latest adjusted strategy, the system will record the results and compare them with the previous situation. If a better control method is found, the system will automatically adopt it and apply it in subsequent operations. In this way, even in the face of unprecedented challenges, such as a surge in demand caused by sudden market changes, the factory can rely on its highly adaptive control system to quickly adjust the production plan to ensure that it is always in the best operating state.

[0039] In summary, through steps 101 to 105, the present invention realizes intelligent management of fluorine chemical production by integrating Markov decision process to evaluate optimal operating conditions, genetic algorithm to optimize valve opening, pulse width modulation to accurately control actuators, deep neural network to dynamically correct control strategies, and online learning mechanism to continuously optimize. It not only improves production efficiency and product quality, ensures the maximization of process safety and energy efficiency, but also enhances the reliability and adaptability of the system through fault prediction and health management technology, and finally achieves the advanced level of intelligent prediction and preventive maintenance.

[0040] Optionally, in step 103, the ideal opening adjustment scheme is converted into a pulse width modulation control instruction, and the pulse width modulation control instruction is sent to the corresponding actuator of the fluorine chemical multi-control valve through the industrial communication bus, and the potential equipment failure risk is identified and processed by using fault prediction and health management technology to obtain the actual valve control strategy of the fluorine chemical multi-control valve, including: Based on the ideal opening adjustment scheme, the opening information in the ideal opening adjustment scheme is converted into a pulse width modulation control instruction of the actuator of the fluorine chemical multi-control valve to generate a pulse width modulation signal; based on the pulse width modulation signal and the industrial communication bus, in combination with the time synchronization protocol, the pulse width modulation control instruction is timestamped, and the pulse width modulation control instruction is optimized and scheduled through distributed control to ensure that the pulse width modulation control instruction is sent to the corresponding actuator of the fluorine chemical multi-control valve, and a communication connection is established between the industrial communication bus and the actuator of the fluorine chemical multi-control valve; based on the communication The communication connection is connected by using fault prediction and health management technology, combined with real-time data analysis and pattern recognition algorithms, to monitor the operating status of the communication connection in real time, and automatically identify abnormal operating modes of the communication connection by using anomaly detection algorithms, and evaluate potential risks through machine learning models to generate a potential risk assessment report; based on the potential risk assessment report, adaptive filtering is used to filter the noise of the communication connection to generate preventive measures to ensure the safe operation of the communication connection; based on the preventive measures and the pulse width modulation signal, the actuator of the fluorine chemical multi-control valve is intelligently adjusted to adjust the opening information to generate an actual valve control strategy.

[0041] Optionally, based on the communication connection, using fault prediction and health management technology, combined with real-time data analysis and pattern recognition algorithms, the operation status of the communication connection is monitored in real time, an abnormality detection algorithm is used to automatically identify abnormal operation modes of the communication connection, and potential risks are evaluated through a machine learning model to generate a potential risk assessment report, including: Based on the communication connection, the fault prediction and health management technology is used in combination with real-time data analysis and pattern recognition algorithms to continuously monitor the operating status of the communication connection, collect status information of the communication connection through sensors on the industrial communication bus, and transmit the status information to the central processing unit of the communication connection to obtain the operating status of the communication connection; according to the operating status of the communication connection, the collected operating status is analyzed and processed using an anomaly detection algorithm, and abnormal operating modes of the communication connection are automatically identified to generate a potential abnormal operating mode report; based on the potential abnormal operating mode report, potential risks are evaluated through a machine learning model to generate a potential risk assessment report.

[0042] In this embodiment, the ideal opening adjustment scheme is converted into a pulse width modulation control instruction and sent to the actuator of the control valve through the industrial communication bus. Pulse width modulation signal is a technology used to accurately control power electronic devices. It adjusts power transfer by changing the duty cycle of the output signal to ensure that the valve can operate according to the preset opening. The time synchronization protocol and distributed control system ensure that the pulse width modulation instructions are accurately transmitted to each actuator and a stable communication connection is established. Fault prediction and health management technology combines real-time data analysis, pattern recognition algorithms and machine learning models to continuously monitor the status of communication connections, automatically identify and evaluate potential risks, such as abnormal operating modes or noise interference, so as to generate preventive measures to ensure the stability and safety of the system.

[0043] In the embodiment of the present application, the opening information in the ideal opening adjustment scheme is first converted into a pulse width modulation control instruction suitable for the actuator to understand, and an accurate pulse width modulation signal is generated. Then, the pulse width modulation instruction is timestamped using a time synchronization protocol, and the distributed control system optimizes scheduling to ensure that the instruction reaches the target device in a timely and accurate manner, while establishing a stable communication connection between the industrial communication bus and the actuator. Then, the system uses fault prediction and health management technology based on the communication connection, combines the status information collected by the sensor, analyzes the operating status through anomaly detection algorithms, identifies abnormal operating modes, and generates a risk assessment report. Finally, preventive measures such as adaptive filtering are taken according to the risk assessment results, and the opening setting of the actuator is intelligently adjusted to form the final actual valve control strategy to achieve efficient and safe operation.

[0044] In an application scenario of intelligent predictive and preventive maintenance, engineers at a fluorine chemical plant deployed an intelligent control system that integrates Markov decision processes, genetic algorithms, and pulse width modulation. When the ideal opening adjustment plan is converted into a pulse width modulation instruction and sent through the industrial communication bus, the system uses a time synchronization protocol to ensure that the instruction reaches the control valve on time. During this period, the fault prediction and health management technology continuously monitors the communication connection status. For example, a certain control valve experienced a slight data transmission delay. The fault prediction and health management system quickly detects anomalies through real-time data analysis and pattern recognition algorithms, and uses machine learning models to assess risks and generate preventive measures. The system automatically enables adaptive filters to reduce noise and optimizes data transmission priorities to avoid potential network congestion. At the same time, it is recommended that the maintenance team check the hardware to prevent failures in advance, ensure production continuity and safety, and significantly improve energy efficiency and product quality.

[0045] Optionally, in step 104, an adaptive feedback loop of a deep neural network is used to compare the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, and a back propagation algorithm is used to dynamically correct the actual valve control strategy to obtain an improved control strategy, including: Based on an adaptive feedback loop of a deep neural network, the operating status of the actual valve control strategy is monitored and analyzed to obtain the actual effect of the actual valve control strategy; based on the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, the difference between the actual effect and the expected effect is compared and analyzed to generate a difference analysis result; based on the difference analysis result, the back propagation algorithm in the deep neural network is used to analyze the specific reasons causing the difference in the difference analysis result, and the weights of the deep neural network are adjusted to generate correction suggestions; based on the correction suggestions, the actual valve control strategy is optimized to improve the accuracy and efficiency of the actual valve control strategy and generate an improved control strategy.

[0046] Optionally, according to the difference analysis results, using a back propagation algorithm in a deep neural network, analyzing specific causes of the differences in the difference analysis results, and adjusting the weights of the deep neural network to generate correction suggestions, including: According to the difference analysis results, the gap between the actual effect obtained and the expected effect is quantified to obtain an initial error value, and the initial error value is transferred layer by layer through the back propagation algorithm in the deep neural network, and the specific cause of the gap is identified to generate a preliminary identification result; based on the preliminary identification result, the key influencing factors in the preliminary identification result are extracted using principal component analysis, and the key influencing factors are refined in combination with sensitivity analysis to obtain a list of key influencing factors; according to the list of key influencing factors, the key influencing factors are adjusted using adaptive moment estimation to narrow the difference between the actual effect and the expected effect to obtain preliminary correction suggestions; based on the preliminary correction suggestions, the weights of the deep neural network are adjusted, and the obtained control strategy is improved to generate correction suggestions.

[0047] In this embodiment, the adaptive feedback loop of the deep neural network is an intelligent control system for monitoring and analyzing the operating status of the actual valve control strategy. It generates a difference analysis result by comparing the difference between the actual effect and the expected effect of the ideal opening adjustment scheme. The back propagation algorithm passes the initial error value layer by layer in the deep neural network, identifies the specific cause of the difference, and adjusts the network weights to optimize the control strategy. Principal component analysis and sensitivity analysis help extract key influencing factors, while adaptive moment estimation is used to fine-tune these factors to ensure that the control strategy is more accurate and efficient.

[0048] In the embodiment of the present application, the scheme first uses the adaptive feedback loop of the deep neural network to monitor the effect of the actual valve control strategy, quantify the gap between the actual effect and the expected effect, and generate an initial error value. Then, the error is transmitted layer by layer through the back propagation algorithm to identify the specific cause and generate a preliminary identification result. Then, the key influencing factors are extracted by principal component analysis, and the sensitivity analysis is combined with the refinement processing to form a list of key influencing factors. Finally, based on this list, adaptive moment estimation is used to adjust the key factors, optimize the deep neural network weights, and improve the control strategy, thereby improving its accuracy and efficiency, and finally generating an improved control strategy.

[0049] In order to further improve the efficiency of intelligent prediction and preventive maintenance, in a fluorine chemical plant, based on the communication connection and real-time monitoring mechanism established in step 103, the system detected that the actual flow of a control valve did not meet expectations. Through quantitative processing, the deep neural network identified that temperature fluctuations were the main cause. Using the back propagation algorithm, the system adjusted the relevant parameters, and combined with principal component analysis and sensitivity analysis to determine the degree of influence of temperature on flow. Subsequently, the temperature control logic was fine-tuned through adaptive moment estimation, which significantly narrowed the gap between actual and expected results, improved the accuracy and energy efficiency of valve operation, and ensured the safe and stable operation of production.

[0050] Optionally, in step 102, according to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search for the opening of the fluorine chemical multi-control valve, and based on the process safety limit and the product quality target, an ideal opening adjustment plan is generated, including: According to the optimal operating conditions and the optimal process state, the opening settings of the fluorine chemical multi-control valve are analyzed to generate different opening setting combinations; based on the opening setting combinations, a group of opening setting combinations representing different opening settings are initialized as an initial scheme, and according to the current optimal operating conditions and the optimal process state, the energy efficiency ratio, safety and product quality performance of each opening setting combination are evaluated to obtain candidate schemes composed of the opening setting combinations with excellent performance; based on the candidate schemes, the selection, crossover and mutation steps in the genetic algorithm are used to explore the adjustment schemes of the opening setting combinations in the candidate schemes to generate a preliminary ideal opening adjustment scheme; based on the preliminary ideal opening adjustment scheme, iterative optimization is continued, and an ideal opening adjustment scheme is generated according to the process safety limit and the product quality target, and the ideal opening adjustment scheme achieves the highest energy efficiency ratio of the fluorine chemical production.

[0051] In this embodiment, a genetic algorithm is an optimization method that simulates natural selection and genetic mechanisms and is used to solve complex multivariable optimization problems. In this step, the opening settings of the fluorine chemical multi-control valve are analyzed according to the optimal operating conditions and the optimal process state, and different combinations of opening settings are generated. These combinations are initialized as a set of initial schemes representing different opening settings. The system evaluates the energy efficiency ratio, safety and product quality performance of each opening setting combination, and selects the combinations with excellent performance to form candidate solutions. The selection, crossover and mutation steps in the genetic algorithm are used to explore the adjustment schemes in the candidate schemes, and finally generate the ideal opening adjustment scheme.

[0052] In the embodiment of the present application, the scheme first analyzes the opening setting of the fluorine chemical multi-control valve based on the optimal operating conditions and the optimal process state, and generates a variety of possible opening setting combinations. Then, a set of opening setting combinations is initialized as the initial scheme, and the performance of each combination is evaluated according to the current optimal operating conditions and the optimal process state, and the candidate schemes with excellent performance are screened out. Next, the selection, crossover and mutation steps in the genetic algorithm are used to explore the adjustment schemes in the candidate schemes and generate a preliminary ideal opening adjustment scheme. Finally, through iterative optimization, combined with the process safety limit and product quality goals, the final ideal opening adjustment scheme is generated to ensure the highest energy efficiency ratio.

[0053] In a fluorine chemical plant, based on the optimal operating conditions and optimal process conditions determined in step 102, the system generates multiple combinations of opening settings and selects candidate solutions with excellent performance. Subsequently, the selection, crossover and mutation steps of the genetic algorithm are used to continuously explore better opening settings. In step 103, these ideal opening adjustment schemes are converted into pulse width modulation control instructions and sent to the actuator, while the health of the communication connection is monitored. Further, in step 104, the adaptive feedback loop of the deep neural network continuously compares the difference between the actual effect and the expected effect, dynamically corrects the control strategy, and ensures that the valve operation is accurate and efficient. This series of measures has significantly improved production efficiency and product quality, and ensured the continuity and safety of production.

[0054] Optionally, in step 105, an online learning mechanism is used to continuously optimize the control strategy of the fluorine chemical multi-control valve according to the experience accumulated during the operation of the improved control strategy, including: Based on the online learning mechanism, the operating status of the improved control strategy is monitored to obtain an information report reflecting the actual operating effect of the improved control strategy; based on the information report, according to the information on the operating status of the improved control strategy collected in real time by the distributed sensor network, the actual operating effect of the improved control strategy is analyzed, the successful operation mode of the improved control strategy is identified, and the operation mode is accumulated as experience to generate an experience library; based on the experience library, the control strategy of the fluorine chemical multi-control valve is updated by machine learning, the weights in the control strategy are adjusted to ensure that the control strategy adapts to the production needs of the fluorine chemical industry, and an updated control strategy is generated; the updated control strategy is optimized and recommended to further improve the control accuracy and control efficiency of the fluorine chemical multi-control valve to generate optimization suggestions, and the optimization suggestions are applied to actual control to ensure that the control strategy of the fluorine chemical multi-control valve is continuously optimized during operation.

[0055] In this embodiment, the online learning mechanism means that the system can continuously learn from the latest data during operation and update its own knowledge base to adapt to the changing environment and technological progress. In this step, the system monitors the actual operating effect of the improved control strategy and generates an information report reflecting the actual operating effect. The distributed sensor network collects the operating status information of the improved control strategy in real time, which is used to analyze the actual operating effect, identify successful operating modes, and accumulate these modes into an experience library. The machine learning algorithm uses the experience library to update the control strategy of the fluorine chemical multi-control valve, adjust the weights in the control strategy, ensure that it adapts to production needs, and thus generate an updated control strategy.

[0056] In the embodiment of the present application, the scheme first monitors the operating status of the improved control strategy based on the online learning mechanism, collects data reflecting the actual operating effect, and generates an information report. Then, according to the operating status information collected in real time by the distributed sensor network, the actual operating effect of the improved control strategy is analyzed, the successful operation mode is identified, and it is accumulated as experience to form an experience library. Next, the control strategy of the fluorine chemical multi-control valve is updated using machine learning technology, the weights in the control strategy are adjusted to ensure that it adapts to production needs, and an updated control strategy is generated. Finally, the updated control strategy is optimized and recommended to further improve the control accuracy and efficiency, and the optimization suggestions are applied to the actual control to ensure that the control strategy is continuously optimized during operation.

[0057] In order to further improve the efficiency of intelligent prediction and preventive maintenance, in a fluorine chemical plant, based on the ideal opening adjustment scheme generated in step 102, in step 103, the system converts these schemes into pulse width modulation control instructions and sends them to the actuator, while using fault prediction and health management technology to monitor the health of the communication connection. In step 104, the adaptive feedback loop of the deep neural network dynamically corrects the control strategy to ensure accurate and efficient valve operation. On this basis, through the online learning mechanism, the system monitors the actual operation effect of the improved control strategy, generates information reports, and combines the data of the distributed sensor network to identify successful operation modes and accumulate them into an experience library. Subsequently, the control strategy is updated using a machine learning algorithm, and the weights are adjusted to ensure that it adapts to production needs. Finally, by optimizing the updated control strategy, the control accuracy and efficiency are further improved, ensuring that the control strategy of the multi-control valve is continuously optimized during operation, and significantly improving the continuity and stability of production.

[0058] This application takes into account that in fluorine chemical production, ensuring the stable operation of multi-control valves is crucial to maintaining process safety and product quality. Traditional fault detection methods are inefficient and difficult to deal with emergencies. Fault prediction and health management technology has gradually become a key means to ensure equipment health and improve production continuity. The development of potential equipment failure risk calculation formula aims to more accurately evaluate the health status of equipment, identify and deal with potential failure risks in advance, realize intelligent prediction and preventive maintenance, and ensure production efficiency and safety. Therefore, a new optional solution is proposed, which includes: Optionally, the fault prediction and health management technology is used to identify and handle potential equipment failure risks to obtain an actual valve control strategy for the fluorine chemical multi-regulating valve, including: Continuously monitor the health status of equipment through fault prediction and health management technology to identify and address potential equipment failure risks ; Potential equipment failure risk The calculation formula is: ; in, is the potential equipment failure risk, ranging from no risk to high risk, with a value range of between, The health index representing the health management technology reflects the health of the device at a future point in time. health status, represents the preset safety threshold, that is, the safety upper limit of the health index. When it is less than the safety threshold, it indicates that there is a potential risk of equipment failure. Represents the urgency of the expected failure time, indicating the distance from the expected time of equipment failure to the current time. Represents the maximum expected failure time urgency, used for standardization The maximum value of The severity of the equipment failure indicates the degree of impact on the fluorine chemical production process when the equipment failure occurs. is the maximum device fault severity, used to normalize The maximum value of is the repair cost, which represents the cost or resource consumption required to repair the equipment failure. is the maximum repair cost, used for normalization The maximum value of is the weight coefficient used to adjust the importance of the health index part, is the nonlinear influence index of the health index, used to capture the nonlinear trend of the health index changing over time, is the weight coefficient used to adjust the importance of the urgency of the estimated failure time part, is the weight coefficient used to adjust the importance of the device failure severity component, is a weight coefficient used to adjust the importance of the repair cost component; The potential equipment failure risk is analyzed and processed to obtain the actual valve control strategy of the fluorine chemical multi-regulating valve.

[0059] The formula comprehensively evaluates the potential failure risk of fluorine chemical multi-control valves, and provides a comprehensive risk assessment by quantifying factors such as health index, expected failure time urgency, equipment failure severity and repair cost. The sum of each sub-item ensures multi-dimensional consideration, and the weight coefficient adjusts the importance of each factor, making the risk assessment more scientific and accurate, helping to identify and deal with potential failures in advance, and ensuring production safety and efficiency.

[0060] The following is a brief explanation of the design reasons for each item in the formula: Potential equipment failure risk Reasons for designing the calculation formula: The formula comprehensively and accurately assesses the risk of potential equipment failure, helping decision makers take preventive measures in a timely manner, reducing unplanned downtime and maintenance costs, and improving production continuity and safety.

[0061] Formula expression: Potential equipment failure risk : Indicates the degree from no risk to high risk, with a value range of between. The closer it is to 1, the higher the potential failure risk; Health Index : reflects the device at a future time point The health status of the device, which is usually calculated from sensor data and historical records, and the full score may be 100 or other standardized values; Health threshold : The safety upper limit of the device health index, when When it is less than this threshold, it indicates that there is a potential risk of equipment failure; Estimated time to failure : Indicates the distance between the estimated time when the equipment failure occurs and the current time, that is, the time point when the failure may occur, which is used to assess the urgency of the failure; Maximum expected failure time urgency : Used for standardization The maximum value ensures Between [0, 1]; Maximum expected failure time urgency : Used for standardization The maximum value ensures Between [0, 1]; Equipment failure severity : Indicates the degree of impact of equipment failure on the fluorine chemical production process, such as downtime, product quality decline, etc. The full score may be 100 or other standardized values; Maximum equipment failure severity : Used for standardization The maximum value ensures Between [0, 1]; Repair costs : Indicates the cost or resource consumption required to repair equipment failure, such as material costs, labor costs, etc.; Maximum repair cost : Used for standardization The maximum value ensures Between [0, 1]; Weight coefficient , , , , : Used to adjust the importance of each sub-item, for example, and The importance of the health index component and its nonlinear effect are adjusted separately, while , , Then adjust the importance of the estimated failure time urgency, equipment failure severity and repair cost components respectively; :This part mainly considers the changing trend of the equipment health index. As the health index decreases, the potential failure risk increases. Capturing the nonlinear trend of the health index over time, the formula can more accurately reflect the speed of health deterioration, and the weight coefficient Used to adjust the importance of this factor in the overall risk assessment; :This part evaluates the urgency of the expected failure time. The closer to the failure time, the higher the risk. Normalized to , making sure its value is between, which is convenient for comparison with other factors, and the weight coefficient Used to adjust the importance of this factor in the overall risk assessment; :This part evaluates the impact of equipment failure on the production process. If the failure seriously affects production, the risk is naturally higher. Normalized to , making sure its value is between, which is convenient for comparison with other factors. Used to adjust the importance of this factor in the overall risk assessment; :This part evaluates the cost of repairing the fault. High repair cost means higher risk because the repair cost may exceed the budget or cause other resources to be strained. Normalized to , making sure its value is between, which is convenient for comparison with other factors, and the weight coefficient Used to adjust the importance of this factor in the overall risk assessment; The reason for adding up the sub-items is that each factor independently reflects the potential failure risk in different aspects. By adding them together, the influence of multiple factors can be comprehensively considered to form a comprehensive risk assessment index.

[0062] The following is a brief explanation of how to obtain the parameters of the formula: Health Index :The sensor network deployed on the equipment collects data in real time, such as temperature, pressure, vibration, etc., combines historical maintenance records and performance indicators, and uses machine learning or statistical models to predict the equipment at a future time point The health status of or other scope; Health threshold :According to industry standards, equipment manufacturer recommendations or historical failure data analysis, this threshold represents the safe upper limit of the equipment health status. When the health index is lower than this value, it is considered that there is a potential failure risk; Estimated time to failure : Use the fault prediction model to predict the time from when the equipment may fail, based on historical fault data, current operating conditions and trend analysis; Maximum expected failure time urgency : Set as the longest possible fault warning period for the equipment, usually determined based on historical data or expert experience, and used for standardization , ensure that its value is between [0,1]; Equipment failure severity :Evaluate the impact of equipment failure on the production process, including downtime, product quality degradation and other factors, through expert evaluation, simulation analysis or historical failure case summary, usually standardized as or other scope; Maximum equipment failure severity : Used for standardization based on the worst failure cases in history or the theoretical maximum impact , making sure its value is between; Repair costs :Estimate the cost or resource consumption required to repair equipment failure, including material costs, labor costs, downtime losses, etc., calculated through cost estimation models or by referring to the actual repair costs of similar failures; Maximum repair cost : Based on the highest historical repair cost case or theoretical maximum repair cost setting, used for standardization , making sure its value is between; Weight coefficient , , , , :The weight coefficients are set by domain experts based on experience and actual needs, or automatically adjusted through sensitivity analysis and optimization algorithms. These coefficients reflect the relative importance of each factor in risk assessment, ensuring that the formula can accurately reflect the actual situation.

[0063] In the embodiment of this application, in a modern fluorine chemical plant, engineers deployed an advanced fault prediction and health management system to monitor the health status of multiple control valves. Assume that there is a control valve that needs to assess its potential equipment failure risk. The following are the specific implementation steps: Health Index :Through the data collected by the sensor network, it is found that the valve will The health index is 75 (out of 100), and the preset safety threshold is 85. This means that the health of the valve is close to the critical point; Estimated time to failure : Based on historical data and trend analysis, it is predicted that the valve will fail within the next 6 months, i.e. =6 months, maximum expected failure time urgency Set to 12 months; Equipment failure severity :The evaluation shows that if the valve fails, it will have a serious impact on the production line, so the equipment failure severity is set to 90 (out of 100), and the maximum equipment failure severity is Also 100; Repair costs : The estimated cost of repairing the valve failure is RMB 20,000, and the maximum repair cost is Set at RMB 50,000; Weight coefficient , , , , : Based on expert opinions and experience, they are set to 0.4, 2, 0.2, 0.2 and 0.2 respectively; Substitute the numerical values ​​to calculate the potential equipment failure risk R:

[0064] The calculation results of the formula show that the potential equipment failure risk of the control valve is The value is 0.36552, indicating a moderate risk. Specifically, the health index contributes little because, although it is below the safety threshold, the gap is not large; the urgency of the expected failure time is medium, which means that the failure may occur soon, but not imminently; the severity of equipment failure is high, indicating that once a failure occurs, it will have a greater impact on the production process; the repair cost is relatively low and will not put too much pressure on the budget.

[0065] Based on this conclusion, the engineer decided to take preventive measures, such as arranging maintenance or replacing the valve in advance, and adjusting the production plan to reduce the impact of potential failures. At the same time, the status of the valve will continue to be closely monitored to ensure that any abnormalities can be discovered and handled in a timely manner, thereby ensuring the continuity and safety of production.

[0066] This application takes into account that in fluorine chemical production, it is crucial to ensure that the opening setting of the multi-control valve meets both safety and quality requirements and achieves the highest energy efficiency ratio. Traditional methods are difficult to cope with complex working conditions. Genetic algorithms optimize the opening setting through global search, combine process safety and quality goals, generate ideal adjustment plans, improve production efficiency and safety, and meet the needs of intelligent prediction and preventive maintenance. Therefore, a new optional solution is proposed, which includes: Optionally, according to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search on the opening of the fluorine chemical multi-regulating valve, and an ideal opening adjustment scheme is generated based on the process safety limit and the product quality target, wherein the ideal opening adjustment scheme meets the safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production, including: According to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search on the opening of the fluorine chemical multi-regulating valve; Global search calculation formula: ; in, is the global search calculation formula, i.e., the best opening of the fluorine chemical multi-regulating valve found, is the fitness function, which is used to evaluate the quality of each opening setting. Represents the opening setting space of all the fluorine chemical multi-control valves, fitness function , Represents the opening setting The energy efficiency ratio of fluorine chemical production under represents the safety score, is the product quality rating, , , , , , are the corresponding weight coefficient and nonlinear index, respectively, reflecting the importance of the safety score, the product quality score and the energy efficiency ratio of the fluorine chemical production and their trends over time or conditions; Based on the calculation results of the global search, an ideal opening adjustment plan is generated according to the process safety limit and the product quality target; Calculation formula for ideal opening adjustment scheme: ; in, is the ideal opening adjustment scheme, Represents the opening setting of the fluorine chemical multi-regulating valve, is a penalty function used to deal with individuals that exceed the process safety limit and the product quality target; Penalty Function : ; in, represents the process safety limit, Represents the product quality target, Represents the actual process safety value under the current opening setting. is the actual product quality value under the current opening setting, and are the corresponding safety and quality penalty coefficients, which are used to adjust the severity of constraint violations; Fitness function: , Represents the opening setting The energy efficiency ratio of fluorine chemical production under represents the safety score, is the product quality rating, , , , , , are the corresponding weight coefficient and nonlinear index, respectively, reflecting the importance of the safety score, the product quality score and the energy efficiency ratio of the fluorine chemical production and their trends over time or conditions.

[0067] The above formula is designed to perform a global search for the opening of the fluorine chemical multi-control valve through a genetic algorithm, and generate an ideal opening adjustment plan in combination with the process safety limit and product quality target. This method not only improves the flexibility and efficiency of production, but also ensures the safety of operation and product quality, and meets the needs of modern chemical production for intelligent prediction and preventive maintenance.

[0068] The following is a brief explanation of the design reasons for each item in the formula: The reason for designing the global search calculation formula is to find the optimal solution in all possible opening setting spaces 𝑋 through genetic algorithms. The fitness function 𝑓(𝑥) comprehensively evaluates energy efficiency, safety, and product quality to ensure that the selected opening setting is not only efficient, but also meets safety and quality requirements, thereby optimizing the production process.

[0069] Formula expression: It represents the optimal opening setting of the fluorine chemical multi-control valve found by global search, that is, in all possible opening setting spaces 𝑋, the fitness function The opening setting that reaches the maximum value.

[0070] : Fitness function, used to evaluate each opening setting The comprehensive advantages and disadvantages of the product are comprehensively considered, which includes three key factors: energy efficiency ratio, safety and product quality, to ensure that the selected opening setting is not only efficient, but also meets safety and quality requirements; : Weight coefficients, which adjust the relative importance of energy efficiency ratio, safety and product quality in the fitness function. These coefficients can be set according to specific production needs and priorities to reflect the importance of different factors; : Non-linear index, used to capture the trend of factors changing over time or conditions. For example, non-linear index can be used to show how the influence of a factor increases or decreases as the opening setting changes. This allows the formula to more accurately reflect the actual situation; : Energy efficiency rating, indicating the energy efficiency rating at the opening setting. The energy efficiency ratio of fluorine chemical production is usually standardized as or other ranges, where 1 represents the highest energy efficiency ratio, which reflects the energy efficiency at a specific opening setting; : Safety score, indicating the opening setting The safety level of equipment operation is also standardized as or other ranges, where 1 represents the safest operating state, and the score reflects the level of safety risk at a specific opening setting; : Product quality rating, indicating the opening setting The quality of the following products is standardized or other ranges, where 1 indicates the highest product quality, and the score reflects the consistency and quality standard of the product at a particular opening setting; :This part mainly considers the energy efficiency ratio of fluorine chemical production and the weight coefficient Used to adjust the importance of energy efficiency ratio in the entire fitness function, rather than the linear exponent Capture the nonlinear trend of energy efficiency ratio changing with the opening setting. This ensures that the formula can more accurately reflect the actual impact of different opening settings on energy efficiency ratio; :This part evaluates the safety of equipment operation, weight coefficient Used to adjust the importance of security in the entire fitness function, rather than a linear exponent Capturing the nonlinear trend of safety changing with the opening setting, which ensures that the formula can fully consider the safety risks under different opening settings; :This part evaluates product quality, weight coefficient Used to adjust the importance of product quality in the entire fitness function, rather than a linear exponent Capturing the nonlinear trend of product quality as the opening setting changes, which ensures that the formula can accurately reflect the impact of different opening settings on product quality; The reason for adding up the sub-items is that each sub-item independently reflects the performance indicators of different aspects. By adding them together, the influence of multiple factors can be comprehensively considered to form a comprehensive risk assessment indicator.

[0071] Reasons for designing the ideal opening adjustment scheme: Based on the global search, the penalty function 𝑃(𝑥) is introduced to ensure that the final opening setting meets the process safety limit and product quality target. Below safety limit , or actual product quality value Below quality target When , the penalty function will add additional costs or reduce the fitness score. The fitness function 𝑓(𝑥) can comprehensively and accurately evaluate the comprehensive advantages and disadvantages of each opening setting, helping decision makers find the best opening setting that can improve production efficiency while ensuring safety and quality. This ensures that the final opening setting not only has the best performance, but also strictly abides by safety and quality constraints to ensure the continuity and reliability of production.

[0072] Formula expression: : Ideal opening adjustment scheme, indicating the ideal opening adjustment scheme among all possible opening settings In the above example, the fitness function Subtract the penalty function The opening setting corresponding to the maximum value after the calculation ensures that the final opening setting not only has the best performance but also strictly complies with safety and quality constraints; :All possible opening setting spaces of fluorine chemical multi-control valves, representing all feasible opening configuration options; : Fitness function, used to evaluate each opening setting The comprehensive advantages and disadvantages of the product are considered comprehensively, taking into account three key factors: energy efficiency, safety and product quality; : Penalty function, used to deal with situations that exceed process safety limits and product quality targets. If an opening setting violates these limits, its fitness score is reduced through the penalty function; : Process safety limit, which indicates the safety threshold of equipment operation. Exceeding this value means there is a safety hazard; : Product quality target, indicating the expected product quality standard. A value lower than this means that the product quality does not meet the standard; : The actual process safety value under the current opening setting, reflecting whether the current setting meets the safety requirements; : The actual product quality value under the current opening setting, reflecting whether the current setting meets the quality requirements; and : Safety and quality penalty coefficients, used to adjust the severity of constraint violations. A larger coefficient means a more severe penalty for violating the constraints. , , , , , : Weight coefficient and nonlinear index, which respectively adjust the relative importance of energy efficiency ratio, safety and product quality in the fitness function and their trends over time or conditions; :This part mainly considers the energy efficiency ratio of fluorine chemical production and the weight coefficient Used to adjust the importance of energy efficiency ratio in the entire fitness function, rather than the linear exponent Capturing the nonlinear trend of the energy efficiency ratio changing with the opening setting ensures that the formula can more accurately reflect the actual impact of different opening settings on the energy efficiency ratio.

[0073] :This part evaluates the safety of equipment operation. Weighting factor Used to adjust the importance of security in the entire fitness function, rather than a linear exponent Capturing the nonlinear trend of safety changing with the opening setting, which ensures that the formula can fully consider the safety risks under different opening settings; :This part evaluates product quality, weight coefficient Used to adjust the importance of product quality in the entire fitness function, rather than a linear exponent Capturing the nonlinear trend of product quality as the opening setting changes, which ensures that the formula can accurately reflect the impact of different opening settings on product quality; By subtracting the penalty function from the fitness function, we ensure that the final selection is not only excellent in performance, but also strictly complies with safety and quality constraints. This approach enables the optimization process to pursue efficiency without sacrificing safety and quality, thereby achieving the optimization of the production process.

[0074] The following is a brief explanation of how to obtain the parameters of the formula: : Through genetic algorithms or other optimization methods, in all possible opening setting spaces Search in and find the optimal opening setting that makes the fitness function minus the penalty function reach the maximum value; :The opening setting space is defined by engineers based on the equipment operating range and historical data, and usually includes a series of discrete or continuous opening values ​​to ensure that all feasible adjustment ranges are covered; :The fitness function comprehensively evaluates the energy efficiency ratio, safety and product quality of each opening setting. The specific value is calculated through simulation model, historical data and real-time monitoring data; :The penalty function is used to deal with situations that exceed the process safety limit and product quality target. The specific value is calculated based on the safety and quality monitoring data in actual operation; and :The safety and quality penalty coefficients are set by domain experts based on experience and industry standards, or determined through sensitivity analysis. These coefficients reflect the severity of violations of safety and quality constraints; :Process safety limits are provided by equipment manufacturers or industry specifications, indicating the safety threshold of equipment operation, for example, the maximum allowable values ​​of parameters such as temperature and pressure; :Product quality targets are set by production standards or customer requirements, indicating the expected product quality standards, such as purity, impurity content and other indicators; : The actual process safety value under the current opening setting is obtained through real-time sensor data (such as temperature, pressure, vibration, etc.) and historical data analysis; : The actual product quality value at the current opening setting, obtained through online testing instruments, laboratory analysis or historical quality records; , , , , , :The weight coefficients and nonlinear indexes are set by domain experts based on experience and specific needs, or automatically adjusted through machine learning methods. These coefficients determine the relative importance of each factor in the fitness function and its trend over time or conditions; :The energy efficiency ratio score is obtained through simulation models or historical data predictions, reflecting the energy utilization efficiency under specific opening settings, for example, the ratio of energy consumption to output; :The safety score is obtained through the evaluation of fault prediction and health management technology, combining real-time sensor data and historical fault data to predict the safety status of the equipment at a certain point in the future; : Product quality scores are obtained through online testing instruments, laboratory analysis or historical quality records, reflecting the quality level of the product at a specific opening setting.

[0075] In the present embodiment, in a modern fluorine chemical plant, engineers use genetic algorithms to optimize the opening settings of multiple regulating valves. , in order to achieve the highest energy efficiency ratio, best safety and product quality. The following are the specific implementation steps: Define the fitness function : The fitness function takes into account the energy efficiency ratio , safety score and product quality rating . Weight coefficient , , , , , They are set to 0.4, 2, 0.3, 1.5, 0.3 and 1.2 respectively, reflecting the importance of each factor and its nonlinear influence.

[0076] According to the optimal operating conditions and the best process state, a set of random opening settings is generated as the initial population , each individual represents a possible opening setting.

[0077] Evaluate fitness: For each opening setting Calculate the fitness function , (Full score: 1) (Full score: 1) (Full score: 1) Global search calculation: Use genetic algorithm selection, crossover and mutation operations to iteratively find the optimal solution . Assume that the optimal opening setting is found after multiple rounds of iterations The fitness value of is:

[0078] Introducing penalty function : examine Whether the process safety limit is met and product quality goals , (Below the safety limit ), actual product quality value (Higher than quality target ); then the penalty function for:

[0079] Generate ideal opening adjustment plan : The final ideal opening adjustment plan is the maximum fitness value after considering the penalty:

[0080] The calculation results show that the optimal opening setting obtained by global search of genetic algorithm After considering safety and quality constraints, the fitness value is 0.8197. The specific conclusions are as follows: Energy efficiency ratio: Although It has reached a relatively high level (0.85), but it is not the highest because safety and quality must also be taken into account.

[0081] Safety: Although Below safety limit , but due to the small gap, the penalty term The impact is not significant.

[0082] Product quality: Exceeded quality goals , indicating that the product quality is well guaranteed.

[0083] Based on this result, the engineers decided to adopt As the final opening adjustment solution, it ensures the safety of the production process and product quality, while achieving a high energy efficiency ratio. This solution not only improves the continuity and stability of production, but also reduces the risk of potential failures, providing strong support for intelligent prediction and preventive maintenance.

[0084] Figure 2 A schematic diagram of a fluorine chemical multi-regulating valve automatic control system is provided for the present application embodiment. Figure 2 As shown, the device comprises: The prediction module 21 evaluates the optimal operating conditions under the current fluorine chemical production conditions based on the Markov decision process, and performs prediction processing on the optimal operating conditions to obtain the optimal process state; A search module 22, based on the optimal operating conditions and the optimal process state, uses a genetic algorithm to perform a global search for the opening of the fluorine chemical multi-control valve, and generates an ideal opening adjustment plan based on the process safety limit and the product quality target, wherein the ideal opening adjustment plan meets the safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production; The identification module 23 converts the ideal opening adjustment scheme into a pulse width modulation control instruction, sends the pulse width modulation control instruction to the corresponding actuator of the fluorine chemical multi-control valve through the industrial communication bus, and uses the fault prediction and health management technology to identify and handle potential equipment failure risks, so as to obtain the actual valve control strategy of the fluorine chemical multi-control valve; A correction module 24, using an adaptive feedback loop of a deep neural network, compares the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, and dynamically corrects the actual valve control strategy using a back propagation algorithm to obtain an improved control strategy; The optimization module 25 continuously optimizes the control strategy of the fluorine chemical multi-control valve according to the experience accumulated during the operation of the improved control strategy by using an online learning mechanism.

[0085] Figure 2 The fluorine chemical multi-regulating valve automatic control system can execute Figure 1 The implementation principle and technical effect of the fluorine chemical multi-regulating valve automatic control method described in the embodiment are not described in detail. The specific way in which each module and unit performs operations in the fluorine chemical multi-regulating valve automatic control system in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.

[0086] In one possible design, Figure 2 The fluorine chemical multi-regulating valve automatic control system of the embodiment shown can be implemented as a computing device, such as Figure 3As shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0087] The processing component 32 is used for the above Figure 1 The embodiment provides an automatic control method for multiple regulating valves in a fluorine chemical industry.

[0088] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0089] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0090] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0091] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0092] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0093] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0094] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The embodiment shown is an automatic control method for multiple regulating valves in a fluorine chemical industry.

[0095] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0096] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0097] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A fluorine chemical multi-regulating valve automatic control method, characterized in that: include: Based on the Markov decision process, the optimal operating conditions under the current fluorine chemical production conditions are evaluated, and the optimal operating conditions are predicted and processed to obtain the optimal process state; According to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search for the opening of the fluorine chemical multi-regulating valve, and an ideal opening adjustment scheme is generated based on the process safety limit and the product quality target, wherein the ideal opening adjustment scheme meets the safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production; Convert the ideal opening adjustment scheme into a pulse width modulation control instruction, send the pulse width modulation control instruction to the corresponding actuator of the fluorine chemical multi-control valve through an industrial communication bus, and use fault prediction and health management technology to identify and handle potential equipment failure risks, so as to obtain an actual valve control strategy for the fluorine chemical multi-control valve; Using an adaptive feedback loop of a deep neural network, the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme is compared, and the actual valve control strategy is dynamically corrected using a back propagation algorithm to obtain an improved control strategy; By utilizing an online learning mechanism, the control strategy of the fluorine chemical multi-regulating valve is continuously optimized based on the experience accumulated during the operation of the improved control strategy.

2. The method according to claim 1, characterized in that The ideal opening adjustment scheme is converted into a pulse width modulation control instruction, and the pulse width modulation control instruction is sent to the corresponding actuator of the fluorine chemical multi-control valve through an industrial communication bus, and the potential equipment failure risk is identified and processed by using fault prediction and health management technology to obtain the actual valve control strategy of the fluorine chemical multi-control valve, including: Based on the ideal opening adjustment scheme, converting the opening information in the ideal opening adjustment scheme into a pulse width modulation control instruction of the actuator of the fluorine chemical multi-regulating valve to generate a pulse width modulation signal; Based on the pulse width modulation signal and the industrial communication bus, in combination with the time synchronization protocol, the pulse width modulation control instruction is timestamped, and the pulse width modulation control instruction is optimized and scheduled through distributed control to ensure that the pulse width modulation control instruction is sent to the corresponding actuator of the fluorine chemical multi-regulating valve, and a communication connection is established between the industrial communication bus and the actuator of the fluorine chemical multi-regulating valve; Based on the communication connection, using fault prediction and health management technology, combined with real-time data analysis and pattern recognition algorithms, the operation status of the communication connection is monitored in real time, an abnormality detection algorithm is used to automatically identify abnormal operation modes of the communication connection, and potential risks are evaluated through a machine learning model to generate a potential risk assessment report; Based on the potential risk assessment report, adaptive filtering is used to filter noise of the communication connection, and preventive measures are generated to ensure safe operation of the communication connection; Based on the preventive measures and the pulse width modulation signal, the actuator of the fluorine chemical multi-control valve adjusts the opening information intelligently to generate an actual valve control strategy.

3. The method according to claim 2, characterized in that Based on the communication connection, the operation status of the communication connection is monitored in real time by using fault prediction and health management technology, combined with real-time data analysis and pattern recognition algorithms, and abnormal operation modes of the communication connection are automatically identified by using anomaly detection algorithms. Potential risks are evaluated by machine learning models to generate a potential risk assessment report, including: Based on the communication connection, using fault prediction and health management technology, combined with real-time data analysis and pattern recognition algorithms, continuously monitoring the operating status of the communication connection, collecting status information of the communication connection through sensors on the industrial communication bus, and transmitting the status information to the central processing unit of the communication connection to obtain the operating status of the communication connection; According to the operation status of the communication connection, the collected operation status is analyzed and processed using an anomaly detection algorithm to automatically identify abnormal operation modes of the communication connection to generate a potential abnormal operation mode report; Based on the potential abnormal operation mode report, potential risks are evaluated through a machine learning model to generate a potential risk assessment report.

4. The method according to claim 1, characterized in that: Using an adaptive feedback loop of a deep neural network, the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme is compared, and the actual valve control strategy is dynamically corrected using a back propagation algorithm to obtain an improved control strategy, including: Based on the adaptive feedback loop of the deep neural network, the operating state of the actual valve control strategy is monitored and analyzed to obtain the actual effect of the actual valve control strategy; Based on the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, a comparison is performed to analyze the difference between the actual effect and the expected effect to generate a difference analysis result; According to the difference analysis results, using the back propagation algorithm in the deep neural network, analyzing the specific reasons causing the differences in the difference analysis results, and adjusting the weights of the deep neural network to generate correction suggestions; According to the correction suggestion, the actual valve control strategy is optimized to improve the accuracy and efficiency of the actual valve control strategy and generate an improved control strategy.

5. The method according to claim 4, characterized in that According to the difference analysis results, the back propagation algorithm in the deep neural network is used to analyze the specific reasons causing the differences in the difference analysis results, and the weights of the deep neural network are adjusted to generate correction suggestions, including: According to the difference analysis result, the gap between the actual effect obtained and the expected effect is quantified to obtain an initial error value, the initial error value is transmitted layer by layer through the back propagation algorithm in the deep neural network, and the specific cause of the gap is identified to generate a preliminary recognition result; Based on the preliminary identification results, the key influencing factors in the preliminary identification results are extracted by principal component analysis, and the key influencing factors are refined in combination with sensitivity analysis to obtain a list of key influencing factors; According to the list of key influencing factors, using adaptive moment estimation, adjusting the key influencing factors to reduce the difference between the actual effect and the expected effect, so as to obtain preliminary correction suggestions; Based on the preliminary correction suggestions, the weights of the deep neural network are adjusted, and the obtained control strategy is improved to generate correction suggestions.

6. The method according to claim 1, characterized in that According to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search for the opening of the fluorine chemical multi-control valve, and based on the process safety limit and product quality target, an ideal opening adjustment plan is generated, including: Analyzing the opening settings of the fluorine chemical multi-regulating valve according to the optimal operating conditions and the optimal process state to generate different opening setting combinations; Based on the opening setting combination, a group of the opening setting combinations representing different opening settings are initialized as initial solutions, and according to the current optimal operating conditions and the optimal process state, the energy efficiency ratio, safety and product quality performance of each of the opening setting combinations are evaluated to obtain candidate solutions composed of the opening setting combinations with excellent performance; Based on the candidate solutions, using the selection, crossover and mutation steps in the genetic algorithm, exploring the adjustment solutions of the opening setting combinations in the candidate solutions to generate a preliminary ideal opening adjustment solution; Based on the preliminary ideal opening adjustment plan, the iterative optimization is continued, and an ideal opening adjustment plan is generated according to the process safety limit and product quality goals. The ideal opening adjustment plan achieves the highest energy efficiency ratio of the fluorine chemical production.

7. The method according to claim 1, characterized in that By using an online learning mechanism, based on the experience accumulated during the operation of the improved control strategy, the control strategy of the fluorine chemical multi-control valve is continuously optimized, including: Based on the online learning mechanism, the operating status of the improved control strategy is monitored to obtain an information report reflecting the actual operating effect of the improved control strategy; Based on the information report, according to the information of the operating status of the improved control strategy collected in real time by the distributed sensor network, the actual operating effect of the improved control strategy is analyzed, the successful operating mode of the improved control strategy is identified, and the operating mode is accumulated as experience to generate an experience library; Based on the experience database, the control strategy of the fluorine chemical multi-control valve is updated by using machine learning, the weight in the control strategy is adjusted, the control strategy is ensured to adapt to the production needs of the fluorine chemical industry, and an updated control strategy is generated; The updated control strategy is processed with optimization suggestions to further improve the control accuracy and control efficiency of the fluorine chemical multi-control valve, so as to generate optimization suggestions, and the optimization suggestions are applied to actual control to ensure that the control strategy of the fluorine chemical multi-control valve is continuously optimized during operation.

8. The method according to claim 7, characterized in that Fault prediction and health management technology is used to identify and handle potential equipment failure risks to obtain the actual valve control strategy of the fluorine chemical multi-control valve, including: Continuously monitor the health status of equipment through fault prediction and health management technology to identify and address potential equipment failure risks ; Potential equipment failure risk The calculation formula is: ； in, is the potential equipment failure risk, ranging from no risk to high risk, with a value range of between, The health index representing the health management technology reflects the health of the device at a future point in time. health status, represents the preset safety threshold, that is, the safety upper limit of the health index. When it is less than the safety threshold, it indicates that there is a potential risk of equipment failure. Represents the urgency of the expected failure time, indicating the distance from the expected time of equipment failure to the current time. Represents the maximum expected failure time urgency, used for standardization The maximum value of The severity of the equipment failure indicates the degree of impact on the fluorine chemical production process when the equipment failure occurs. is the maximum device fault severity, used to normalize The maximum value of is the repair cost, which represents the cost or resource consumption required to repair the equipment failure. is the maximum repair cost, used for normalization The maximum value of is the weight coefficient used to adjust the importance of the health index part, is the nonlinear influence index of the health index, used to capture the nonlinear trend of the health index changing over time, is the weight coefficient used to adjust the importance of the urgency of the estimated failure time part, is the weight coefficient used to adjust the importance of the device failure severity part, is a weight coefficient used to adjust the importance of the repair cost component; The potential equipment failure risk is analyzed and processed to obtain the actual valve control strategy of the fluorine chemical multi-regulating valve.

9. The method according to claim 8, characterized in that According to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search for the opening of the fluorine chemical multi-regulating valve, and an ideal opening adjustment scheme is generated based on the process safety limit and the product quality target, wherein the ideal opening adjustment scheme meets the safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production, including: According to the optimal operating conditions and the optimal process state, a genetic algorithm is used to perform a global search on the opening of the fluorine chemical multi-regulating valve; Global search calculation formula: ； in, is the global search calculation formula, i.e., the best opening of the fluorine chemical multi-regulating valve found, is the fitness function, which is used to evaluate the quality of each opening setting. Represents the opening setting space of all the fluorine chemical multi-control valves, fitness function , Represents the opening setting The energy efficiency ratio of fluorine chemical production under represents the safety score, is the product quality rating, , , , , , are the corresponding weight coefficient and nonlinear index, respectively, reflecting the importance of the safety score, the product quality score and the energy efficiency ratio of the fluorine chemical production and their trends over time or conditions; Based on the calculation results of the global search, an ideal opening adjustment plan is generated according to the process safety limit and the product quality target; The calculation formula for the ideal opening adjustment scheme is: ； in, is the ideal opening adjustment scheme, X represents the opening setting of the fluorine chemical multi-regulating valve, is a penalty function used to deal with individuals that exceed the process safety limit and the product quality target; Penalty Function : ； in, represents the process safety limit, Represents the product quality target, Represents the actual process safety value under the current opening setting. is the actual product quality value under the current opening setting, and are the corresponding safety and quality penalty coefficients, which are used to adjust the severity of constraint violations; Fitness function: , Represents the opening setting The energy efficiency ratio of fluorine chemical production under represents the safety score, is the product quality rating, , , , , , are the corresponding weight coefficient and nonlinear index, respectively, reflecting the importance of the safety score, the product quality score and the energy efficiency ratio of the fluorine chemical production and their trends over time or conditions.

10. A fluorine chemical multi-regulating valve automatic control system, characterized in that: include: A prediction module, which evaluates the optimal operating conditions under the current fluorine chemical production conditions based on the Markov decision process, and performs prediction processing on the optimal operating conditions to obtain the optimal process state; A search module, which uses a genetic algorithm to perform a global search for the opening of the fluorine chemical multi-regulating valve according to the optimal operating conditions and the optimal process state, and generates an ideal opening adjustment plan based on process safety limits and product quality targets, wherein the ideal opening adjustment plan meets safety and quality requirements and achieves the highest energy efficiency ratio of the fluorine chemical production; An identification module converts the ideal opening adjustment scheme into a pulse width modulation control instruction, sends the pulse width modulation control instruction to the corresponding actuator of the fluorine chemical multi-control valve through an industrial communication bus, and uses fault prediction and health management technology to identify and handle potential equipment failure risks to obtain an actual valve control strategy for the fluorine chemical multi-control valve; A correction module, using an adaptive feedback loop of a deep neural network, compares the difference between the actual effect of the actual valve control strategy and the expected effect of the ideal opening adjustment scheme, and dynamically corrects the actual valve control strategy using a back propagation algorithm to obtain an improved control strategy; The optimization module uses an online learning mechanism to continuously optimize the control strategy of the fluorine chemical multi-control valve based on the experience accumulated during the operation of the improved control strategy.

Images