A recovery control system based on an MVR system

Abstract

The present invention discloses a recovery control system based on an MVR system, which relates to the field of recovery control of the MVR system. The key points of its technical solution include: a data acquisition module for collecting the comprehensive state data of the MVR system, performing fusion processing on the collected data, and obtaining a comprehensive data matrix; a segmentation module for obtaining an initial countermeasure set from the comprehensive data matrix through an anomaly recognition model, performing cluster analysis on the initial countermeasure set, and obtaining k groups of sub-countermeasure sets; a solution output module for respectively optimizing the k groups of sub-countermeasure sets by using an optimization algorithm to obtain a set of candidate solutions, and then optimizing the set of candidate solutions again by using the optimization algorithm to obtain an optimal control solution for the recovery control of the MVP system; each module is connected in a wired and / or wireless manner to achieve stability and optimal energy during the recovery control process based on the MVR system.

Description

Technical Field

[0001] The present invention relates to the field of recovery control of MVR systems, and more specifically, it relates to a recovery control system based on an MVR system. Background Art

[0002] The MVR system, whose full name is Mechanical Vapor Recompression system, is an advanced energy-saving evaporation technology and is widely used in fields such as industrial wastewater treatment, food processing, chemical product concentration, and seawater desalination. The core of MVR technology lies in using the secondary steam generated in the system, increasing its energy and temperature through mechanical compression, and then recycling it as a heat source, thereby reducing the demand for fresh steam and achieving the purpose of energy conservation.

[0003] The MVR system involves multiple technical links, including evaporation, compression, heat exchange, etc. These links need to be precisely controlled to achieve the best effect, and need to be integrated with existing factory equipment and process flows, which may involve complex engineering and coordination work. The operation of each link requires professional and accurate instructions. Improper instructions may lead to a reduction in system efficiency or equipment damage, reducing the stability of the MVR system during operation, and the phenomenon of compressor surging will occur. In addition, although the MVR system is designed for energy conservation, in actual operation, the energy consumption may increase due to various reasons. Therefore, the MVR system needs to adapt to different environmental conditions, such as temperature, humidity, etc., to maintain the best performance. Therefore, it is necessary to design the recovery control of the MVR system. Summary of the Invention

[0004] Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a recovery control system based on an MVR system to achieve the recovery control of the MVR system.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A recovery control system based on an MVR system, the recovery control system based on an MVR system includes:

[0006] A data acquisition module: used to collect the comprehensive state data of the MVR system, and perform fusion processing on the collected data to obtain a comprehensive data matrix;

[0007] A segmentation module: obtain an initial countermeasure set from the comprehensive data matrix through an anomaly recognition model, and perform clustering analysis on the initial countermeasure set to obtain k groups of sub-countermeasure sets;

[0008] A solution output module: use an optimization algorithm to optimize each of the k groups of sub-countermeasure sets to obtain a set of candidates, and then use the optimization algorithm to optimize the set of candidates again to obtain the optimal control scheme for the recovery control of the MVP system;

[0009] Each module is connected in a wired and / or wireless manner.

[0010] Preferably, the data categories of the comprehensive status data include evaporator data, compressor data, heat exchanger data, condenser data, and separator data. The evaporator data includes the liquid level of the evaporator. The compressor data includes the steam pressure, temperature, and flow rate of the compressor. The heat exchanger data includes the temperature of the heat exchanger. The condenser data includes the flow rate, pressure, and temperature of the condenser. The separator data includes the liquid level of the separator.

[0011] Preferably, the method for fusing and processing the collected data includes:

[0012] It is assumed that the comprehensive status data is provided by q sensors. At time point t, the measured values of w data categories are obtained to obtain a measurement vector where, z_t i represents the measured value of the i-th sensor, and i = 1,..., q;

[0013] The compressor characteristic coefficient and the condenser characteristic coefficient are obtained through the compressor data and the condenser data. The measured values corresponding to the steam pressure, temperature, and flow rate of the compressor, and the measured values of the flow rate, pressure, and temperature of the condenser in the measurement vector are deleted. The compressor characteristic coefficient and the condenser characteristic coefficient are filled into the measurement vector, and tags are distributed to obtain a fused characteristic vector;

[0014] Taking time point t as the base point, the characteristic vectors at the previous p_a time points are continuously collected and arranged in chronological order to form a time series matrix, which is denoted as the comprehensive data matrix at time point t.

[0015] Preferably, the method for obtaining the compressor characteristic coefficient and the condenser characteristic coefficient through the compressor data and the condenser data includes:

[0016] The measured values corresponding to the steam pressure, temperature, and flow rate of the compressor, and the measured values of the flow rate, pressure, and temperature of the condenser in the measurement vector are extracted, and curve fitting is performed to obtain the compressor characteristic coefficient and the condenser characteristic coefficient. The curve fitting formula is set as where, W_Y o is used to represent the compressor characteristic coefficient or the condenser characteristic coefficient at time point o, and o = t or 1,..., e,..., r, and t - e represents the e previous time points of time point o; is obtained through the least squares method, x_g y represents the correlation coefficient between the y-th measured value and the fault, represents the measurement error variance of the y-th sensor at time point t - 1, z_t yDenotes the measured value of the y-th sensor, and Y represents the number of measured values.

[0017] Preferably, the correlation coefficient x_g y Is used to evaluate the influence degree of the data category represented by each measured value on the equipment failure. Set the equipment status as Z_u, and Z_u = 1 indicates failure, Z_u = 0 indicates normal. Set the failure correlation model as Where, β0 is the intercept term, and βy is the regression coefficient of the y-th measured value z_t y Corresponding to the data category, and the regression coefficient βy is estimated and calculated by maximizing the likelihood function. x_g y (Z_u = 1|X_y) represents that the regression coefficient corresponding to the data category of the y measured values and the equipment failure is used as the correlation coefficient;

[0018] The correlation coefficient x_g at time point t y Is calculated and obtained by using the measured values at time point t - 1 through the failure correlation model.

[0019] Preferably, the initial countermeasure set is obtained from the comprehensive data matrix through the anomaly recognition model, specifically including:

[0020] Set a sample set, which includes H_k subsets. Each subset includes a comprehensive data matrix, anomaly measured values, and the initial countermeasure set corresponding to the anomaly measured values. The comprehensive data matrix is marked as the input of the anomaly recognition model. The anomaly recognition model matches the anomaly measured values with the fixed countermeasure set and outputs the initial countermeasure set. The anomaly measured values and the initial countermeasure set are marked as the outputs of the anomaly recognition model. Divide the sample set into a training set and a validation set, and input them into the anomaly recognition model in turn;

[0021] The fixed countermeasure set includes all the regulation strategies when the anomaly measured values are regulated to the normal state. Each regulation strategy includes the parameter settings used by the equipment during the adjustment of the MVP system. The anomaly recognition model uses an optimization algorithm to search for z_c optimal regulation strategies in the fixed countermeasure set to form the initial countermeasure set;

[0022] Set the anomaly recognition model. The anomaly recognition model is based on RBF and LSTM, optimizes the model parameters by minimizing the loss function, uses the validation set for cross-validation, adjusts the model hyperparameters, and uses accuracy, recall, and F1-score as performance evaluation indicators. Stop until the expected performance evaluation indicators are reached to obtain the trained anomaly recognition model, and export the trained anomaly recognition model in a deployable format;

[0023] Import the comprehensive data matrix at time point t into the trained anomaly recognition model to obtain the initial countermeasure set at time point t.

[0024] Preferably, the method for clustering and analyzing the initial countermeasure set includes:

[0025] Step S1: Set the number of clustering clusters \(k_{zg}\), and randomly select \(k_{zg}\) control strategies as the initial centroids;

[0026] Step S2: Calculate the distance between each control strategy and each centroid, and assign each control strategy to the cluster represented by the centroid with the closest distance;

[0027] Step S3: Calculate the new centroid for all control strategies in each cluster, and the centroid is the mean of all control strategies in the cluster;

[0028] Step S4: Repeat Steps S2 to S3 until the centroids no longer change in two adjacent iterations;

[0029] Step S5: Output the control strategies in each cluster to obtain \(k\) groups of sub-countermeasure sets.

[0030] Preferably, the method for obtaining the number of clustering clusters \(k_{zg}\) includes:

[0031] Pre-set a cluster number selection data set \(K = \{k_1, \ldots, k_n, \ldots, k_N\}\), where \(k_n\) represents the number of clusters represented by the \(n\)-th value in the cluster number selection data set, and \(n = 1, \ldots, N\);

[0032] Randomly select a \(k_n\) as the pre-set number of clusters, and calculate the average silhouette coefficient when the number of clusters is \(k_n\) where \(z_c\) represents the number of control strategies in the initial countermeasure set;

[0033] Traverse and calculate the average silhouette coefficients of all cluster numbers in the cluster number selection data set, and select the cluster number with the largest silhouette coefficient as the optimal cluster number, and set the optimal cluster number as the number of clustering clusters \(k_{zg}\);

[0034] For any control strategy, calculate its average distance from all other control strategies in the same cluster where \(|C_{sj}|\) represents the number of control strategies in cluster \(C_j\), and \(d(j, l)\) represents the distance between the \(j\)-th control strategy and the \(l\)-th control strategy; and calculate its average distance from the points in all other clusters, and find the closest cluster where cluster \(C_h\) is different from cluster \(C_j\), \(d(j, z)\) is the distance between the \(j\)-th control strategy and the \(z\)-th control strategy in cluster \(C_h\), and \(C_{sh}\) represents the number of control strategies in cluster \(C_h\).

[0035] Preferably, the method for optimizing each of the \(k\) groups of sub-countermeasure sets using an optimization algorithm to obtain the candidate set includes:

[0036] Each regulatory strategy in any set of sub - strategies is regarded as a hunting dog individual. Through mutual cooperation, the hunting dogs continuously adjust their positions and approach the optimal prey;

[0037] Set the position update formula for each hunting dog individual: Where, represents the position of the v - th hunting dog at the o - th iteration, wp v is the local best position of the v - th hunting dog, wp is the global best position, β1, β2 are learning factors, and ε1, ε2 are random factors between 0 and 1;

[0038] Set the fitness Where, and are the stability and energy consumption indicators of the regulatory strategy after the v - th hunting dog is updated respectively. γ1, γ2 are weight coefficients. If the fitness exceeds the current global best fitness, then update the global best position, that is Stop when the change rate of adjacent two fitness values is less than the set threshold in several consecutive iterations;

[0039] Set the constraint conditions as Where, SF represents the number of stability - determining parameters, θ wx represents the weight coefficient of the wx - th parameter, Δh_l represents the absolute difference between the updated regulatory strategy and the standard value, f_op, YU_D are pre - set thresholds, θ1, θ2 represent weight coefficients, Q RE represents the recovered heat, Q IN represents the input heat, Q FR represents the usage amount of fresh steam;

[0040] Output the current global best position wp, which is the corresponding optimal regulatory strategy in the set of sub - strategies.

[0041] Preferably, the method of obtaining the optimal regulatory scheme for the MVP system recovery control by re - optimizing the candidate set through an optimization algorithm includes:

[0042] Collect the optimal regulatory strategies of all sub - strategy sets to form a candidate set, re - optimize the regulatory strategies in the candidate set through an optimization algorithm, output the optimal regulatory strategy in the candidate set, denoted as the optimal regulatory strategy at time point t, and the MVP system adjusts the parameters at the next time point according to the optimal regulatory strategy at this time.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] In the present invention, the compressor characteristic coefficient and the condenser characteristic coefficient are used to represent the compressor data and the condenser data, which not only simplifies the amount of data, but also reflects the comprehensive characteristics of the data, improves the data processing speed, enhances the data mining depth, and realizes the high-quality input of subsequent data. The anomaly recognition model is based on RBF and LSTM to learn the change trends of the measured values at multiple time points in the comprehensive data matrix, analyze and predict the values, and obtain the measured values with abnormal or dangerous states, which is convenient for subsequent analysis of abnormal measured values, reduces the processing of the amount of data, improves the optimized use of data. The optimization algorithm implemented by the present invention realizes the multi-layer optimization of the regulation strategy and outputs the optimal regulation strategy. Through the segmentation of the multi-layer regulation strategy, it is convenient for the centralized analysis and processing of data with the same characteristics. The layer-by-layer optimization not only reduces the amount of data processing, but also the parallel computing improves the operation rate; the adaptive regulation optimization mechanism continuously improves the overall efficiency of intelligent analysis, and the multi-modules cooperate with each other to realize the comprehensive acquisition, processing and analysis of data, and provide real-time and accurate analysis of detection data. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 FIG. is a schematic structural diagram of a recovery control system based on an MVR system proposed by the present invention;

[0046] Figure 2 FIG. is a schematic method diagram of a recovery control system applied to an MVR system in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0048] Embodiment 1

[0049] Refer to Figure 1 and Figure 2 to further illustrate a recovery control system based on an MVR system proposed by the present invention in Embodiment 1.

[0050] The MVR system, full name Mechanical Vapor Recompression system, is an advanced energy-saving evaporation technology widely used in industrial wastewater treatment, food processing, chemical product concentration, seawater desalination and other fields. The core of the MVR technology lies in using the secondary steam generated in the system, increasing its energy and temperature through mechanical compression, and then using it as a heat source for recycling, thereby reducing the demand for fresh steam and achieving the purpose of energy conservation.

[0051] The working principle of the MVR system includes:

[0052] Steam generation: The material is heated in the evaporator to generate steam;

[0053] Steam compression: The generated secondary steam is compressed by a compressor to increase its temperature and pressure;

[0054] Heat energy utilization: The compressed steam returns to the evaporator as a heat source to continue heating the material;

[0055] Evaporation and concentration: The material continues to evaporate under the action of high-temperature steam to achieve the purpose of concentration or crystallization;

[0056] Heat energy recovery: The condensed water and concentrated liquid generated by the system recover heat energy through a heat exchanger to preheat the feed;

[0057] Recycling: During the whole process, the steam is recycled, reducing the demand for fresh steam.

[0058] The MVR system has extremely high economic value in recovering secondary steam, for example:

[0059] A. Energy-saving and efficient: The MVR system reduces the demand for fresh steam or external heat sources by recovering secondary steam, significantly reducing energy consumption.

[0060] B. Reducing operating costs: Reducing dependence on external energy directly reduces the energy procurement cost, thus reducing the overall operating cost.

[0061] C. Improving thermal efficiency: After the recovered secondary steam is compressed by the compressor, its heat energy is reused, improving the thermal efficiency of the whole system.

[0062] D. Reducing environmental impact: Reducing dependence on fossil fuels reduces greenhouse gas emissions and other pollutant emissions, which has a positive effect on environmental protection.

[0063] E. Achieving clean production: As a clean energy technology, the MVR technology helps to achieve clean industrial production, meeting current environmental protection regulations and sustainable development goals.

[0064] F. Improving system economy: Efficient energy utilization makes the MVR system more economically attractive, especially in areas with fluctuating energy prices or high energy costs.

[0065] G. Reducing the demand for cooling water: Since the MVR system reduces the use of fresh steam, it correspondingly reduces the demand for cooling water, further reducing the energy consumption and cost of the system.

[0066] H. Optimizing the process flow: Recycling secondary steam can optimize the whole process flow, improving production efficiency and product quality.

[0067] I. Reducing system scale: Due to the high thermal efficiency of the MVR system, the required equipment scale may be smaller, reducing equipment investment and floor area.

[0068] J. Improving system reliability and stability: The closed-loop cycle reduces the impact of external factors on system stability and improves production reliability.

[0069] K. Strong adaptability: The MVR system can adapt to different material characteristics and processing requirements, with good flexibility and adaptability.

[0070] L. Technological advancement: The adoption of MVR technology reflects the enterprise's application of advanced technology, enhancing the enterprise's technological image and market competitiveness.

[0071] In summary, the recycling of the MVR system is considered from multiple aspects such as economy, environment, and technology. It represents an efficient, environmentally friendly, and sustainable industrial production method.

[0072] However, the difficulties or problems that the MVR system may encounter during the recycling process mainly include:

[0073] The MVR system involves multiple technical links, including evaporation, compression, heat exchange, etc. These links need to be precisely controlled to achieve the best effect and need to be integrated with existing factory equipment and process flows. This may involve complex engineering and coordination work. The operation of each link requires professional and accurate instructions. Improper instructions may lead to a reduction in system efficiency or equipment damage, reducing the stability of the MVR system during operation. The surging phenomenon of the compressor may occur. In addition, although the MVR system is designed for energy conservation, in actual operation, the energy consumption may increase due to various reasons. Therefore, the MVR system needs to adapt to different environmental conditions, such as temperature, humidity, etc., to maintain the best performance. Therefore, the recycling control of the MVR system needs to be designed.

[0074] The MVR system needs to be integrated with existing factory equipment and process flows, which may involve complex engineering and coordination work. Considering multiple aspects such as system design, equipment selection, operation training, maintenance strategy, and environmental adaptability, to ensure that the MVR system can effectively recycle and reuse secondary steam and achieve energy conservation and environmental protection goals.

[0075] The core components of the MVR system select centrifugal or roots compressors, which have a relatively high volumetric flow rate in the compression ratio range of 1:1.2 to 1:2. Different types of MVR evaporators, such as falling film evaporators, forced circulation evaporators, or crystallizers, are selected to adapt to different process requirements. The condensate water and concentrated liquid are preheated through a waste heat recovery device to improve energy efficiency.

[0076] When designing the system, the maximization of heat energy recovery needs to be considered to reduce energy loss. The MVR system is usually equipped with an automated control system to achieve real-time monitoring and adjustment of key parameters such as temperature, pressure, and flow rate, ensuring the stable operation of the system. Designing a complete recovery control scheme for the MVR (Mechanical Vapor Recompression) system requires comprehensive consideration of system stability, efficiency, energy consumption, and integration with other process flows. The following is a detailed design scheme:

[0077] In the present invention, the control objectives are set as follows:

[0078] 1. Stabilize system operation: Ensure that the system can operate stably under different working conditions, maintain the stable operation of the compressor and evaporator, and prevent compressor surge.

[0079] 2. Improve energy efficiency: Minimize energy consumption to the greatest extent and optimize the energy utilization efficiency of the system, including maximizing the recovery of heat from secondary steam and reducing the use of fresh steam.

[0080] 3. Adapt to process changes: Achieve automatic control and real-time adjustment of the system, reduce human intervention, flexibly respond to changes in raw material concentration, temperature, etc., and maintain the best performance of the system.

[0081] 4. Ensure safe operation: Monitor and prevent potential safety hazards such as overpressure and overheating in the system.

[0082] Data acquisition module: Used to collect the comprehensive status data of the MVR system, and perform fusion processing on the collected data to obtain a comprehensive data matrix;

[0083] The data categories of the comprehensive status data include evaporator data, compressor data, heat exchanger data, condenser data, and separator data.

[0084] The evaporator data includes the liquid level of the evaporator. The liquid level sensor is used to monitor the liquid level of the liquid in the evaporator, and the flow rate of the feed pump is adjusted to maintain the liquid level stable.

[0085] The compressor data includes the steam pressure, temperature, and flow rate of the compressor. Pressure sensors, temperature sensors, and flow sensors are used to monitor the inlet and outlet parameters of the compressor in real time.

[0086] The surge protection of the compressor is achieved by installing an anti-surge controller to adjust the operating point of the compressor in real time. For example, a bypass valve is added to reduce the compressor load and prevent surge.

[0087] The heat exchanger data includes the temperature of the heat exchanger. According to the process requirements, the flow valve on the steam side of the heat exchanger is adjusted to control the steam flow rate in the heat exchanger, and then the temperature of the liquid to be evaporated is regulated to keep the outlet temperature of the liquid to be evaporated within the set range, ensuring the stability of the evaporation process.

[0088] The condenser data includes the flow rate, pressure, and temperature of the condenser. The main control objective is to condense the remaining steam into liquid and recover heat. By adjusting the flow rate of the cooling water and through the temperature and pressure controllers, the flow rate of the cooling water and the pressure inside the condenser are adjusted respectively to maintain the pressure and temperature inside the condenser at the target values.

[0089] The separator data includes the liquid level of the separator. In the separator, the separation of steam and liquid is linked with the drain valve through the liquid level controller (PID) to maintain a stable liquid level. The change of the liquid level is monitored by the liquid level sensor, and the opening degree of the drain valve is adjusted.

[0090] The above temperature, liquid level, pressure, and flow rate are all detected by temperature sensors, liquid level sensors, pressure sensors, and flow sensors.

[0091] The methods for fusing the collected data include:

[0092] It is set that the comprehensive state data is provided by q sensors. At time point t, the measured values of w data categories are obtained, and labels are distributed to each sensor. The labels can be numerically encoded through one-hot. The labels are used to mark the measured values and retrieve the corresponding sensors of the measured values, which is convenient for corresponding the measured values with the measured objects to obtain the measurement vector.

[0093] The compressor characteristic coefficient and the condenser characteristic coefficient are obtained through the compressor data and the condenser data. The measured values corresponding to the steam pressure, temperature, and flow rate of the compressor in the measurement vector, as well as the measured values of the flow rate, pressure, and temperature of the condenser, are deleted. The compressor characteristic coefficient and the condenser characteristic coefficient are filled into the measurement vector, and labels are distributed to obtain the fused characteristic vector.

[0094] Taking time point t as the base point, the characteristic vectors at the previous p_a time points are continuously collected and arranged in chronological order to form a time series matrix, which is recorded as the comprehensive data matrix at time point t.

[0095] The methods for obtaining the compressor characteristic coefficient and the condenser characteristic coefficient through the compressor data and the condenser data include:

[0096] The measured values corresponding to the steam pressure, temperature, and flow rate of the compressor in the measurement vector, as well as the measured values of the flow rate, pressure, and temperature of the condenser, are extracted, and curve fitting is performed to obtain the compressor characteristic coefficient and the condenser characteristic coefficient. The curve fitting formula is set as Among them, W_Y o is used to represent the compressor characteristic coefficient or the condenser characteristic coefficient at time point o, and o = t or 1,..., e,..., r. t - e represents the previous e time points of time point o; Obtained by the least squares method, \(x_g\) y represents the correlation coefficient between the \(y\)-th measurement value and the fault, represents the measurement error variance of the \(y\)-th sensor at time point \(t - 1\), \(z_t\) y represents the measurement value of the \(y\)-th sensor, and \(Y\) represents the number of measurement values.

[0097] The compressor data and condenser data are represented by the compressor characteristic coefficient and the condenser characteristic coefficient, which not only simplifies the data volume, but also reflects the comprehensive characteristics of the data, improves the data processing speed, enhances the data mining depth, and realizes the high-quality input of subsequent data.

[0098] The correlation coefficient \(x_g\) y is used to evaluate the influence degree of the data categories represented by each measurement value on the equipment fault. In the present invention, the data categories represented by each measurement value are respectively the steam pressure, temperature and flow rate of the compressor or the flow rate, pressure and temperature of the condenser, and the equipment fault is the compressor or condenser fault. The equipment state is set as \(Z_u\), and \(Z_u = 1\) represents a fault, \(Z_u = 0\) represents normal. The fault correlation model is set as where \(\beta_0\) is the intercept term, and \(\beta_y\) is the regression coefficient corresponding to the data category of the \(y\)-th measurement value \(z_t\) y The regression coefficient \(\beta_y\) is estimated and calculated by maximizing the likelihood function, \(x_g\) y \((Z_u = 1|X_y)\) represents that the regression coefficient corresponding to the data category of the \(y\) measurement values and the equipment fault is used as the correlation coefficient;

[0099] The correlation coefficient \(x_g\) at time point \(t\) y is obtained by calculating through the fault correlation model using the measurement values at time point \(t - 1\).

[0100] Segmentation module: Obtain the initial countermeasure set from the comprehensive data matrix through the anomaly recognition model, perform clustering analysis on the initial countermeasure set, and obtain \(k\) groups of sub-countermeasure sets;

[0101] Obtaining the initial countermeasure set from the comprehensive data matrix through the anomaly recognition model specifically includes:

[0102] Set a sample set, the sample set includes \(H_k\) groups of subsets, each subset includes a comprehensive data matrix, abnormal measurement values, and the initial countermeasure set corresponding to the abnormal measurement values. The comprehensive data matrix is marked as the input of the anomaly recognition model. The anomaly recognition model matches the abnormal measurement values with the fixed countermeasure set and outputs the initial countermeasure set. The abnormal measurement values and the initial countermeasure set are marked as the output of the anomaly recognition model. The sample set is divided into a training set and a validation set and input into the anomaly recognition model in turn;

[0103] The set of fixed countermeasures includes all the regulation strategies when abnormal measurement values are regulated to the normal state. Each regulation strategy includes the parameter settings used by the device during the adjustment of the MVP system. The abnormal recognition model uses optimization algorithms (such as simulated annealing and genetic algorithms) to search for z_c groups of optimal regulation strategies in the set of fixed countermeasures to form the initial countermeasure set. The set of fixed countermeasures can also be the set of regulation strategies used at all historical time points, or an integrated set of regulation strategies independently set by the staff based on experience or experimental data analysis, or a set of strategies automatically generated by some learning machines based on known data.

[0104] Set up an abnormal recognition model. The abnormal recognition model is based on RBF and LSTM. Optimize the model parameters by minimizing the loss function, use the validation set for cross-validation, adjust the model hyperparameters, and use accuracy, recall rate, and F1-score as performance evaluation indicators. Stop until the expected performance evaluation indicators are reached to obtain the trained abnormal recognition model, and export the trained abnormal recognition model in a deployable format.

[0105] Import the comprehensive data matrix at time point t into the trained abnormal recognition model to obtain the initial countermeasure set at time point t.

[0106] The abnormal recognition model learns the change trends of measurement values at multiple time points in the comprehensive data matrix based on RBF and LSTM, analyzes and predicts the numerical values to obtain measurement values with abnormal or dangerous states, which is convenient for subsequent analysis of abnormal measurement values and reduces the processing of data volume.

[0107] The methods for performing cluster analysis on the initial countermeasure set include:

[0108] Step S1: Set the number of cluster clusters k_zg, and randomly select k_zg regulation strategies as the initial centroids.

[0109] Step S2: Calculate the distance between each regulation strategy and each centroid, and assign each regulation strategy to the cluster represented by the centroid with the closest distance.

[0110] Step S3: Calculate the new centroid for all the regulation strategies in each cluster. The centroid is the mean of all the regulation strategies in the cluster.

[0111] Step S4: Repeat steps S2 to S3 until the centroids no longer change in two adjacent times and then stop.

[0112] Step S5: Output the regulation strategies in each cluster to obtain k groups of sub-countermeasure sets.

[0113] In the present invention, when performing calculations between regulation strategies, operations are carried out on the same-category numerical values in the regulation strategies.

[0114] The ways to obtain the number of clustering clusters \(k_{zg}\) include:

[0115] Pre-set a cluster number selection data set \(K = \{k_1,\ldots,k_n,\ldots,k_N\}\), where \(k_n\) represents the number of clusters represented by the \(n\)-th value in the cluster number selection data set, and \(n = 1,\ldots,N\);

[0116] Randomly select a \(k_n\) as the pre-set number of clusters, and calculate the average silhouette coefficient when the number of clusters is \(k_n\) where \(z_c\) represents the number of regulation strategies in the initial countermeasure set;

[0117] Traverse and calculate the average silhouette coefficients of all cluster numbers in the cluster number selection data set, select the cluster number with the largest silhouette coefficient as the optimal cluster number, and set the optimal cluster number as the clustering cluster number \(k_{zg}\);

[0118] For any regulation strategy, calculate its average distance from all other regulation strategies in the same cluster where \(|C_{sj}|\) represents the number of regulation strategies in cluster \(C_j\), and \(d(j,l)\) represents the distance (such as Euclidean distance) between the \(j\)-th regulation strategy and the \(l\)-th regulation strategy; and calculate its average distance from the points in all other clusters, and find the nearest cluster (i.e., the cluster with the smallest average distance) where cluster \(C_h\) is different from cluster \(C_j\), \(d(j,z)\) is the distance between the \(j\)-th regulation strategy and the \(z\)-th regulation strategy in cluster \(C_h\), and \(C_{sh}\) represents the number of regulation strategies in cluster \(C_h\).

[0119] By dividing the initial countermeasure set into multiple sub-countermeasure sets, similar regulation strategies are aggregated, reducing the complexity of the data, simplifying subsequent analysis and processing tasks, reducing the local stagnation caused by overly scattered data in the optimization process, reducing the computing power used for data retrieval, optimizing resource allocation and strategy formulation, facilitating precise optimization for different similar strategy groups, and achieving high-quality data processing.

[0120] Scheme output module: Use an optimization algorithm to optimize each of the \(k\) groups of sub-countermeasure sets to obtain a candidate set, and then use the optimization algorithm to optimize the candidate set again to obtain the optimal regulation scheme for the MVP system recovery control;

[0121] The ways to use an optimization algorithm to optimize each of the \(k\) groups of sub-countermeasure sets to obtain a candidate set include:

[0122] For each regulation strategy in any group of sub-countermeasure sets, it is regarded as a hunting dog individual. The hunting dogs cooperate with each other, continuously adjust their positions (i.e., the parameter settings in the regulation strategy), and approach the optimal prey (the optimal regulation strategy);

[0123] Set the position update formula for each hunting dog individual: Where, represents the position of the v-th hunting dog at the o-th iteration, wp v is the local best position of the v-th hunting dog, wp is the global best position, β1, β2 are learning factors, and ε1, ε2 are random factors between 0 and 1;

[0124] Set the fitness Where, and are the stability and energy consumption indicators of the regulation strategy after the update of the v-th hunting dog respectively, γ1, γ2 are weight coefficients. If the fitness exceeds the current global best fitness, then update the global best position, that is Stop when the change rate of the fitness between two adjacent iterations is less than the set threshold for several consecutive iterations;

[0125] Set the constraint condition as Where, SF represents the number of stability determination parameters, θ wx represents the weight coefficient of the wx-th parameter, Δh_l represents the absolute difference between the updated regulation strategy and the standard value, f_op, YU_D are preset thresholds, θ1, θ2 represent weight coefficients, Q RE represents the recovered heat, Q IN represents the input heat, Q FR represents the usage amount of fresh steam;

[0126] Output the current global best position wp, which is the optimal regulation strategy corresponding to the sub-game set.

[0127] Use the optimization algorithm to search for the optimal regulation plan for the MVP system recovery control again for the candidate set, including:

[0128] Collect the optimal regulation strategies of all sub-game sets to form a candidate set. Use the optimization algorithm to search for the regulation strategies in the candidate set again, output the optimal regulation strategy in the candidate set, denoted as the optimal regulation strategy at time point t, and the MVP system adjusts the parameters at the next time point according to the optimal regulation strategy at this time.

[0129] In this embodiment, the present invention uses the compressor characteristic coefficient and the condenser characteristic coefficient to represent the compressor data and the condenser data, which not only simplifies the data volume, but also reflects the comprehensive characteristics of the data, improves the data processing speed, enhances the data mining depth, and realizes the high-quality input of subsequent data. The anomaly recognition model learns the change trend of the measured values at multiple time points in the comprehensive data matrix based on RBF and LSTM, analyzes and predicts the values, and obtains the measured values with abnormal or dangerous states, which is convenient for subsequent analysis of abnormal measured values, reduces the processing of data volume, and improves the optimized use of data. The optimization algorithm implemented by the present invention realizes multi-layer optimization of the control strategy and outputs the optimal control strategy. Through the segmentation of the multi-layer control strategy, it is convenient for centralized analysis and processing of data with the same characteristics. The layer-by-layer optimization not only reduces the data processing volume, but also the parallel calculation improves the operation rate; the adaptive control optimization mechanism continuously improves the overall efficiency of intelligent analysis, and the multi-modules cooperate with each other to realize the comprehensive acquisition, processing and analysis of data, and provide real-time and accurate analysis of detection data.

[0130] Embodiment 2

[0131] Refer to Figure 1 and Figure 2 , Embodiment 2 further describes a recovery control system based on the MVR system proposed by the present invention.

[0132] The recovery control method based on the MVR system includes the following steps:

[0133] Step S1: Collect the comprehensive state data of the MVR system at time point t, and perform fusion processing on the collected data to obtain a comprehensive data matrix;

[0134] Step S2: Construct an anomaly recognition model and train it, and import the comprehensive data matrix at time point t into the trained anomaly recognition model to obtain an initial countermeasure set;

[0135] Step S3: Set the number of clustering clusters k_zg, perform clustering analysis on the initial countermeasure set, and obtain k groups of sub-countermeasure sets;

[0136] Step S4: Use the optimization algorithm to optimize each of the k groups of sub-countermeasure sets respectively, gather the control strategies obtained by the optimization to form a candidate set, and then use the optimization algorithm to optimize the candidate set again to obtain the optimal control scheme, and apply the output optimal control scheme to the recovery control of the MVP system.

[0137] The recovery control method based on the MVR system is applied to a recovery control system based on the MVR system, including:

[0138] Data acquisition module: It is used to collect the comprehensive status data of the MVR system, and perform fusion processing on the collected data to obtain a comprehensive data matrix;

[0139] Segmentation module: Obtain an initial countermeasure set from the comprehensive data matrix through an anomaly recognition model, and perform clustering analysis on the initial countermeasure set to obtain k sets of sub-countermeasure sets;

[0140] Solution output module: Use an optimization algorithm to optimize each of the k sets of sub-countermeasure sets to obtain a set of candidates, and then use the optimization algorithm to optimize the set of candidates again to obtain the optimal control scheme for the recovery control of the MVP system;

[0141] Each module is connected by wired and / or wireless means.

[0142] In addition, according to the embodiments of the present application, the process described in the accompanying drawings of a recovery control system based on the MVR system can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be run by a processor to execute instructions corresponding to the method steps provided by the present application. Of course, the architecture shown in the accompanying drawings of a recovery control system based on the MVR system is only exemplary. When implementing different devices, adaptive selection or adjustment can be made according to actual needs.

[0143] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by software simulation of a large amount of collected data to get a formula closest to the real situation. The preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.

[0144] The above are only the preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the concept of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art in this technical field, several improvements and refinements made without departing from the principle of the present invention should also be regarded as within the protection scope of the present invention.

Claims

1. A recovery control system based on an MVR system, characterized in that, The recovery control system based on the MVR system includes: A data acquisition module: used to collect the comprehensive status data of the MVR system, and perform fusion processing on the collected data to obtain a comprehensive data matrix; The method of performing fusion processing on the collected data includes: It is assumed that the comprehensive status data is provided by q sensors. At time point t, the measured values of w data categories are obtained to get a measurement vector , where represents the measured value of the i-th sensor, and ; Obtain the compressor characteristic coefficient and the condenser characteristic coefficient through the compressor data and the condenser data, delete the measured values corresponding to the steam pressure, temperature, and flow rate of the compressor in the measurement vector, as well as the measured values of the flow rate, pressure, and temperature of the condenser, fill the compressor characteristic coefficient and the condenser characteristic coefficient into the measurement vector, and distribute labels to obtain the fused feature vector; Taking the time point t as the base point, continuously collect the feature vectors at the previous p_a time points and arrange them in chronological order to form a time series matrix, denoted as the comprehensive data matrix at the time point t, where p_a represents the number of time points; The method of obtaining the compressor characteristic coefficient and the condenser characteristic coefficient through the compressor data and the condenser data includes: extract the measured values corresponding to the steam pressure, temperature, and flow rate of the compressor in the measurement vector, as well as the measured values of the flow rate, pressure, and temperature of the condenser, and perform curve fitting to obtain the compressor characteristic coefficient and the condenser characteristic coefficient; A segmentation module: obtain the initial countermeasure set from the comprehensive data matrix through an anomaly recognition model, and perform clustering analysis on the initial countermeasure set to obtain k groups of sub-countermeasure sets; A solution output module: use an optimization algorithm to optimize each of the k groups of sub-countermeasure sets to obtain a candidate set, and use the optimization algorithm again to optimize the candidate set to obtain the optimal regulation solution for the recovery control of the MVP system; Each module is connected in a wired and / or wireless manner.

2. The recovery control system based on the MVR system according to claim 1, wherein The data categories of the comprehensive status data include evaporator data, compressor data, heat exchanger data, condenser data, and separator data. The evaporator data includes the liquid level of the evaporator, the compressor data includes the steam pressure, temperature, and flow rate of the compressor, the heat exchanger data includes the temperature of the heat exchanger, the condenser data includes the flow rate, pressure, and temperature of the condenser, and the separator data includes the liquid level of the separator.

3. The recovery control system based on the MVR system according to claim 2, wherein The method of obtaining the compressor characteristic coefficient and the condenser characteristic coefficient through the compressor data and the condenser data also includes: Set the curve fitting formula as , where is used to represent the compressor characteristic coefficient or condenser characteristic coefficient at time point o, and , represents the number of time points selected for curve fitting among p_a time points, represents the e time points before time point o; is obtained by the least squares method, represents the correlation coefficient between the y-th measurement value and the fault, represents the measurement error variance of the y-th sensor at time point t - 1, represents the measurement value of the y-th sensor, and Y represents the number of measurement values.

4. The recovery control system based on the MVR system according to claim 3, wherein, The correlation coefficient is used to evaluate the influence degree of the data category represented by each measurement value on the equipment failure. The equipment state is set to , and represents a failure, represents normal. The failure correlation model is set to , where is the intercept term, is the regression coefficient of the y-th measurement value corresponding to the data category, and the regression coefficient is estimated and calculated by maximizing the likelihood function, represents that the regression coefficient corresponding to the data category of the y measurement values and the equipment failure is used as the correlation coefficient; At the previous time point t-1 of the time point t, obtain the measurement values of each data category in the comprehensive status data, and input the measurement values into the correlation coefficient calculated by the fault correlation model , and use this correlation coefficient as the correlation coefficient at the time point t.

5. The recovery control system based on the MVR system according to claim 4, characterized in that The method of obtaining the initial countermeasure set from the comprehensive data matrix through the anomaly recognition model specifically includes: Set a sample set, which includes H_k groups of subsets. Each subset includes a comprehensive data matrix, abnormal measurement values, and the initial countermeasure set corresponding to the abnormal measurement values. The comprehensive data matrix is marked as the input of the anomaly recognition model. The anomaly recognition model matches the abnormal measurement values with the fixed countermeasure set and outputs the initial countermeasure set. The abnormal measurement values and the initial countermeasure set are marked as the output of the anomaly recognition model. Divide the sample set into a training set and a validation set and input them into the anomaly recognition model in sequence; The fixed countermeasure set includes all the regulation strategies when the abnormal measurement values are adjusted to the normal state. Each regulation strategy includes the parameter settings used by the equipment during the adjustment of the MVP system. The anomaly recognition model uses an optimization algorithm to search for z_c groups of optimal regulation strategies in the fixed countermeasure set to form the initial countermeasure set; Set up an anomaly recognition model. The anomaly recognition model is based on RBF and LSTM. Optimize the model parameters by minimizing the loss function, use the validation set for cross-validation, adjust the model hyperparameters, and use accuracy, recall, and F1-score as performance evaluation metrics. Stop until the expected performance evaluation metrics are achieved to obtain a trained anomaly recognition model, and export the trained anomaly recognition model into a deployable format; Import the comprehensive data matrix at time point t into the trained anomaly recognition model to obtain the initial countermeasure set at time point t.

6. The recovery control system based on the MVR system according to claim 5, characterized in that The method for performing clustering analysis on the initial countermeasure set includes: Step S1: Set the number of clustering clusters k_zg, and randomly select k_zg control strategies as the initial centroids; Step S2: Calculate the distance between each control strategy and each centroid, and assign each control strategy to the cluster represented by the centroid with the closest distance; Step S3: Calculate the new centroid for all control strategies in each cluster. The centroid is the mean of all control strategies in the cluster; Step S4: Repeat steps S2 to S3 until the centroids no longer change between two adjacent times; Step S5: Output the control strategies in each cluster to obtain k sets of sub-countermeasure sets.

7. The recovery control system based on the MVR system according to claim 6, characterized in that, The method for obtaining the number of clustering clusters k_zg includes: Pre-set a cluster number selection data set , where represents the number of clusters represented by the th value in the cluster number selection data set, and ; Randomly select one as the preset number of clusters, and calculate the average silhouette coefficient when the number of clusters is where denotes the number of regulatory strategies in the initial countermeasure set; ​ Traverse and calculate the average silhouette coefficient for all cluster numbers in the dataset of cluster numbers, select the cluster number with the largest silhouette coefficient as the optimal cluster number, and set the optimal cluster number as the number of clustering clusters k_zg; For any regulatory strategy, calculate its average distance from all other regulatory strategies in the same cluster , where represents the number of regulatory strategies in cluster , represents the distance between the th regulatory strategy and the th regulatory strategy; and calculate its average distance from the points in all other clusters to find the nearest cluster , where cluster is a cluster different from , is the distance between the th regulatory strategy and the th regulatory strategy in cluster , represents the number of regulatory strategies in cluster .

8. A recovery control system based on an MVR system according to claim 7, characterized in that The method for using an optimization algorithm to optimize each of the k sets of sub-countermeasure sets to obtain a candidate set includes: For each control strategy in any set of sub-countermeasure sets, it is regarded as a hunting dog individual. The hunting dogs cooperate with each other, continuously adjust their positions, and approach the optimal prey; Set the position update formula for each hunting dog individual: , where represents the position of the th hunting dog at the th iteration, is the local best position of the th hunting dog, is the global best position, , are learning factors, , are random factors between 0 and 1; Set fitness , where and are the stability and energy consumption indicators of the th hound's updated regulation strategy respectively. , are weight coefficients. If the fitness exceeds the current global best fitness, then update the global best position, that is . Stop when the change rate of adjacent two fitness values is less than the set threshold in several consecutive iterations. Set the constraint conditions as , , where represents the number of stability determination parameters, represents the weight coefficient of the th parameter, represents the absolute difference between the updated regulation strategy and the standard value, , are preset thresholds, , represent weight coefficients, represents the recovered heat, represents the input heat, represents the consumption of fresh steam; Output the current globally optimal position , which is the corresponding optimal control strategy in the sub-game set.

9. The recovery control system based on the MVR system according to claim 8, wherein, The method for using an optimization algorithm to optimize the candidate set again to obtain the optimal control scheme for the MVP system recovery control includes: Collect the optimal control strategies of all sub-countermeasure sets to form a candidate set, use an optimization algorithm to optimize the control strategies in the candidate set again, output the optimal control strategy in the candidate set, which is denoted as the optimal control strategy at time point t, and the MVP system adjusts the parameters at the next time point according to the optimal control strategy at this time.

Images