A flow self-distribution based water supply system control method

Abstract

The application discloses a kind of water supply system control methods based on flow self-distribution, specifically related to water supply system technical field, including by using historical record to construct the probability density function of monitoring feature, and by sensor determines the actual situation of current monitoring feature, based on bayes formula determines water supply probability information, and by cross entropy adjusts the weight of monitoring feature, by monitoring the steam flow of each position in boiler pipeline, based on historical data determines the influence of each position steam flow variation on water supply system water supply decision, by logistic regression model, determines water supply deviation information, by the comprehensive analysis of water supply probability information and water supply deviation information, quantifies the influence degree of each time point on water supply system decision, by comprehensively considering steam flow, water supply flow and drum water level, construct neural network model, further optimize the regulation and control of water supply system, the application is helpful to accurately adjust water supply rate, improve boiler operating efficiency and safety.

Description

Technical Field

[0001] The present invention relates to the technical field of water supply systems, and more particularly, to a water supply system control method based on flow self-distribution. Background Art

[0002] With the widespread application of industrial boilers and steam generators, precise regulation of the water supply system is crucial to improving energy efficiency and ensuring safe operation of the boiler. The boiler's water supply system is an important part of boiler operation. Its main function is to provide the water required by the steam generator to ensure efficient and stable operation of the boiler. However, in actual operation, due to the influence of multiple factors such as load fluctuations, changes in environmental factors, and differences in equipment status, the boiler's steam demand and water supply fluctuate significantly. Traditional water supply control methods generally rely on fixed control strategies or simple PID control, but these methods are difficult to fully cope with the dynamic changes of the boiler system under complex working conditions, and it is not easy to make auxiliary adjustments to the water supply rate based on the historical data and actual usage of the water supply system.

[0003] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a water supply system control method based on flow self-distribution to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A water supply system control method based on flow self-distribution specifically comprises the following steps:

[0007] S1: Construct the probability density function of the monitoring feature by using historical records, determine the actual situation of the current monitoring feature through sensors, determine the water supply probability information based on the Bayesian formula, and adjust the weight of the monitoring feature through cross entropy;

[0008] S2: By monitoring the steam flow at various locations in the boiler pipeline, the impact of steam flow changes at various locations on the water supply decision of the water supply system is determined based on historical data, and the water supply deviation information is determined through a logistic regression model;

[0009] S3: Through comprehensive analysis of water supply probability information and water supply deviation information, a water supply assessment model is constructed to determine the water supply assessment coefficient at each time point and quantify the impact of each time point on water supply system decision-making;

[0010] S4: By comprehensively considering the steam flow, feed water flow and drum water level, and combining the feed water assessment coefficient and actual steam demand, a neural network model is constructed to further optimize the regulation of the feed water system.

[0011] In a preferred embodiment, the probability density function of the monitoring feature is constructed using historical records, including:

[0012] Based on the specific steam demand, determine the historical data of the monitoring characteristics, where the monitoring characteristics include steam flow, feed water flow, and drum water level, determine the steam demand deviation in the current unit time period, and mark the steam demand deviation in the current unit time period as: ,in, , XXQ is the actual steam demand per unit time period, and YXQ is the original steam demand per unit time period;

[0013] Based on the historical data of monitoring characteristics under steam demand deviation, the probability density function of each monitoring characteristic under water supply instruction is calculated by kernel density estimation, and the probability density function of each monitoring characteristic under water supply instruction is marked as: 、 、 , is the probability density function of the steam flow rate under the current steam demand deviation under the water supply instruction, is the probability density function of the feedwater flow rate under the current steam demand deviation under the feedwater instruction, is the probability density function of the current steam demand deviation of the drum water level under the feedwater instruction, where , , ,i is the index of each monitoring feature sample data in the unit time period, i=1, 2, 3, ..., N, N is a positive integer, 、 、 is the smoothing parameter, K is the kernel function, ZQ is the steam flow rate, GS is the feed water flow rate, SW is the drum water level, is the effective sample number of steam flow, is the effective sample number of water flow rate, is the number of valid samples of drum water level, is the i-th steam flow sample value in the unit time period, is the i-th water flow sample value in the unit time period, is the i-th drum water level sample value in the unit time period;

[0014] Based on the historical data of monitoring characteristics under steam demand deviation, the probability density function of each monitoring characteristic under no-water supply instruction is calculated by kernel density estimation, and the probability density function of each monitoring characteristic under no-water supply instruction is marked as: 、 、 ,in, is the probability density function of the steam flow rate at the current steam demand deviation under no water supply instruction, is the probability density function of the feedwater flow rate at the current steam demand deviation without feedwater instruction, is the probability density function of the drum water level deviation from the current steam demand under no feedwater instruction.

[0015] In a preferred embodiment, the water supply probability information is determined based on the Bayesian formula, and the weight of the monitoring feature is adjusted by cross entropy, including:

[0016] The water supply probability information is represented by the water supply probability adjustment coefficient;

[0017] Substitute the actual value of the monitoring feature of the current unit time period into the probability density function, determine the prior probability and posterior probability of the monitoring feature at each time point in the current time period, calculate the posterior probability of water supply, and mark the posterior probability of water supply as: ,in, 、 、 are the weights of steam flow, feed water flow, and drum water level, respectively. is the prior probability of water supply, is the posterior probability of water supply;

[0018] Calculate the posterior probability of no water supply, the calculation formula is: ;in, is the prior probability of no water supply, is the posterior probability of not giving water;

[0019] Among them, a weight is assigned to each monitoring feature, the prediction probability of the model is calculated through historical data, and the cross entropy loss is minimized using gradient descent to adjust the weight of each monitoring feature;

[0020] Calculate the water supply probability adjustment coefficient, the calculation formula is: ;in, is the water supply probability adjustment coefficient.

[0021] In a preferred embodiment, determining water supply deviation information includes:

[0022] The water supply deviation information is expressed by the water supply influence coefficient;

[0023] The logic for obtaining the feedwater influence coefficient is as follows: determining the steam flow rate at each location of the boiler pipeline, and determining the degree of influence of the steam flow rate at each location on the feedwater system based on historical data, and determining the proportion of the influence of the steam flow rate at each location on the feedwater system based on the historical data;

[0024] Calculate the water supply influence coefficient, the calculation formula is: ;in, is the water supply influence coefficient, 、 、……、 is the influence of steam flow at each location on the water supply of the water supply system, 、 、……、 is the proportion of the impact of changes in steam flow at various locations in the historical data on the water supply of the water supply system, 、 、……、 is the steam flow rate at each location at each time point in a unit time period, and e is a natural number.

[0025] In a preferred embodiment, constructing a water supply assessment model includes:

[0026] The water supply probability adjustment coefficient and water supply influence coefficient at each time point in the unit time period are weighted and calculated to construct a water supply assessment model and generate a water supply assessment coefficient. The calculation formula of the water supply assessment coefficient is: ;in, is the water supply assessment coefficient at each time point in the unit time period, 、 are the proportional coefficients of the water supply probability adjustment coefficient and the water supply influence coefficient, 、 Both are greater than 0.

[0027] In a preferred embodiment, constructing a neural network model includes:

[0028] The feedwater assessment coefficient at each time point, the steam flow rate at each time point, the feedwater flow rate at each time point, the drum water level at each time point, and the steam demand deviation in the current unit time period are used as the input of the neural network model, and the feedwater warning coefficient is used as the output of the neural network model;

[0029] The input layer inputs time series data with a dimension of 5*k, where k is the number of time points. The RNN layer captures the temporal dependencies in the time series data and maps the features extracted by the RNN layer to the output of the water supply warning coefficient. The output layer is the water supply warning coefficient.

[0030] Historical data is used for model training. Each input sample corresponds to a water supply warning coefficient. The output layer is the target value, and the mean square error is used as the loss function.

[0031] Technical effects and advantages of the present invention:

[0032] The present invention constructs a probability density function and applies Bayesian inference, combines characteristics such as steam flow and feed water flow, quantifies the impact of each time point on the feed water decision, uses logistic regression to analyze the deviation impact of steam flow changes on the feed water system, and then optimizes the feed water system regulation. By applying the neural network model to the feed water decision of the boiler system, it can provide a more accurate and intelligent basis for the adjustment of the feed water rate based on the capture of complex temporal dependencies and multi-dimensional input characteristics. The present invention helps to accurately adjust the feed water rate and improve the efficiency and safety of boiler operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0034] Figure 1 The figure is a flow chart of a water supply system control method based on flow self-distribution according to the present invention. DETAILED DESCRIPTION

[0035] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0036] Example 1

[0037] like Figure 1 The flow chart of a water supply system control method based on flow self-distribution of the present invention specifically includes the following steps:

[0038] S1: Construct the probability density function of the monitoring feature by using historical records, determine the actual situation of the current monitoring feature through sensors, determine the water supply probability information based on the Bayesian formula, and adjust the weight of the monitoring feature through cross entropy;

[0039] S2: By monitoring the steam flow at various locations in the boiler pipeline, the impact of steam flow changes at various locations on the water supply decision of the water supply system is determined based on historical data, and the water supply deviation information is determined through a logistic regression model;

[0040] S3: Through comprehensive analysis of water supply probability information and water supply deviation information, a water supply assessment model is constructed to determine the water supply assessment coefficient at each time point and quantify the impact of each time point on water supply system decision-making;

[0041] S4: By comprehensively considering the steam flow, feed water flow and drum water level, and combining the feed water assessment coefficient and actual steam demand, a neural network model is constructed to further optimize the regulation of the feed water system.

[0042] In this embodiment, the feedwater system adopts a three-impulse control strategy, which comprehensively considers steam flow, feedwater flow, and drum water level to achieve precise control of feedwater. Through real-time monitoring and adjustment, the system can automatically allocate water flow according to changes in steam load, ensuring the safety and efficiency of boiler operation. Specifically, the following are included:

[0043] Steam flow measurement: Real-time monitoring of the boiler's steam output flow serves as a feedforward signal, providing a reference for feedwater control. When steam demand changes, such as load increases (increases steam consumption by equipment) or load decreases (decreases steam consumption by equipment), steam flow measurement can quickly capture these changes and provide real-time data to the control system.

[0044] Feedwater flow measurement: Monitor the feedwater flow entering the boiler to ensure that it matches the steam flow. The feedwater flow matches the steam output flow of the boiler. An increase in steam flow means an increase in the amount of evaporated water, so the feedwater flow needs to be increased accordingly. Conversely, if the steam demand decreases, the feedwater flow should also be reduced.

[0045] Drum water level measurement: Continuously monitor the water level in the drum to prevent it from being too high or too low, which could affect safe boiler operation. Accurate water level control helps maintain stable steam quality. A low water level can cause the boiler to produce overheated steam, while an excessively high water level can result in high-humidity steam, affecting subsequent steam efficiency and equipment operation.

[0046] By using historical records to construct the probability density function of the monitoring feature, and determining the actual situation of the current monitoring feature through sensors, the water supply probability information is determined by the Bayesian formula, and the weight of the monitoring feature is adjusted by cross entropy. The water supply probability information is represented by the water supply probability adjustment coefficient. The advantages of the water supply probability adjustment coefficient are:

[0047] By using historical records to construct the probability density function of monitoring features, the long-term operating characteristics of the boiler system can be captured, allowing the system to dynamically adjust the weights of monitoring features based on historical data. In particular, when making water supply decisions, the water supply probability adjustment coefficient can be adjusted in real time based on the relationship between current sensor data and historical data.

[0048] The feedwater probability information is calculated using the Bayesian formula and combined with cross-entropy to adjust parameter weights. The feedwater probability adjustment coefficient can optimize the boiler feedwater strategy under different operating conditions. Cross-entropy can quantify the accuracy of the prediction and gradually optimize the boiler control logic by minimizing the difference between the prediction and actual data. The feedwater probability adjustment coefficient serves as a feedback mechanism, helping the system identify which control parameters have a greater impact on the feedwater decision and adjust their weights, thereby improving the control effect.

[0049] In boiler control systems, overfitting and underfitting are common problems. Overfitting may cause the control strategy to over-adapt to specific data and ignore the influence of other important variables, while underfitting means that the model cannot fully capture the potential relationships in the boiler system. Through Bayesian updating and cross-entropy optimization, the feedwater probability adjustment coefficient can find a balance between historical data and real-time data, avoiding overfitting and underfitting and improving the generalization ability of the model.

[0050] The logic for obtaining the feedwater probability adjustment coefficient is as follows: based on the specific steam demand, determine the historical data of monitoring characteristics, where the monitoring characteristics include steam flow, feedwater flow, and drum water level, determine the steam demand deviation in the current unit time period, and mark the steam demand deviation in the current unit time period as: ,in, , XXQ is the actual steam demand per unit time period, and YXQ is the original steam demand per unit time period;

[0051] It should be noted that the unit time period refers to a fixed time interval used to measure and record key parameters in the system. The actual steam demand per unit time period refers to the required steam consumption calculated based on system data within a specific unit time period. The original steam demand per unit time period refers to the steam consumption when the steam demand remains unchanged within the unit time period.

[0052] Based on the historical data of monitoring characteristics under steam demand deviation, the probability density function of each monitoring characteristic under water supply instruction is calculated by kernel density estimation, and the probability density function of each monitoring characteristic under water supply instruction is marked as: 、 、 , is the probability density function of the steam flow rate under the current steam demand deviation under the water supply instruction, is the probability density function of the feedwater flow rate under the current steam demand deviation under the feedwater instruction, is the probability density function of the deviation of the drum water level from the current steam demand under the feedwater instruction, where , , ,i is the index of each monitoring feature sample data in the unit time period, i=1, 2, 3, ..., N, N is a positive integer, 、 、 is the smoothing parameter, K is the kernel function;

[0053] Based on the historical data of monitoring characteristics under steam demand deviation, the probability density function of each monitoring characteristic under no-water supply instruction is calculated by kernel density estimation, and the probability density function of each monitoring characteristic under no-water supply instruction is marked as: 、 、 ,in, is the probability density function of the steam flow rate at the current steam demand deviation under no water supply instruction, is the probability density function of the feedwater flow rate at the current steam demand deviation without feedwater instruction, is the probability density function of the deviation of the drum water level from the current steam demand under no feedwater instruction;

[0054] Substitute the actual value of the monitoring feature of the current unit time period into the probability density function, determine the prior probability and posterior probability of the monitoring feature at each time point in the current time period, calculate the posterior probability of water supply, and mark the posterior probability of water supply as: ,in, 、 、 are the weights of steam flow, feed water flow, and drum water level, respectively. is the prior probability of water supply, is the posterior probability of water supply;

[0055] It should be noted that the weights determine the impact of each monitoring feature on the final decision. The sum of the weight values ​​is not necessarily required to be 1, but their relative sizes will affect the contribution of each feature to the posterior probability. If a feature is more important, give it a larger weight; if a feature is relatively unimportant, give it a smaller weight. The prior probability of water supply is determined by the historical data of the monitoring feature under steam demand deviation.

[0056] Calculate the posterior probability of no water supply, the calculation formula is: ;in, is the prior probability of no water supply, is the posterior probability of not giving water;

[0057] Among them, a weight is assigned to each monitoring feature, the prediction probability of the model is calculated through historical data, and the cross entropy loss is minimized using gradient descent to adjust the weight of each monitoring feature;

[0058] Calculate the water supply probability adjustment coefficient, the calculation formula is: ;in, is the water supply probability adjustment coefficient.

[0059] The feedwater deviation information is determined by the feedwater influence coefficient. According to the actual steam demand, the steam flow rate at each location of the boiler pipeline is monitored to determine the deviation between the actual and demanded steam. The feedwater influence coefficient is determined through a logistic regression model. Specifically, it includes:

[0060] Collect historical data, obtain steam flow at various locations through sensors, and determine whether water supply was carried out in each time period or the specific value of water supply flow. Usually, this data is used to decide whether to start water supply based on boiler steam demand;

[0061] Calculate the change in steam flow at each location based on historical data. To determine the impact of steam flow change on feedwater decision-making, use a regression model (such as linear regression or logistic regression) to establish the relationship between steam flow change and feedwater decision-making.

[0062] Based on the results of regression analysis, the impact of steam flow changes at each location on water supply decisions is calculated.

[0063] It should be noted that the steam flow rates at various locations may be inconsistent, mainly because the flow rate in the steam pipe will be affected by the pipe layout, pressure loss, flow resistance and other environmental factors. Therefore, the steam flow rates at various locations in the water supply system may have different impacts on the water supply decision. The steam flow rate close to the boiler has a greater impact on the water supply system and can be given a higher weight. The pipes far away from the boiler may have increased steam flow resistance and a reduced flow rate, so their impact on the water supply decision may be smaller and can be given a lower weight.

[0064] The advantages of the water supply influence coefficient are:

[0065] Taking into account the spatial differences in steam flow, by assigning different weights to the steam flow at different pipeline locations, the impact of steam flow changes at each location on water supply decisions can be more accurately captured, avoiding the errors caused by ignoring spatial differences;

[0066] By calculating the contribution of each pipeline location to historical data, we can determine which location has the greatest impact on water supply decisions, thereby better understanding and explaining the model's decision-making process.

[0067] The weight value can be adjusted according to different working conditions, so that the model can be adaptively optimized according to actual steam flow demand, pipeline conditions and other factors;

[0068] By analyzing the impact of steam flow changes on feed water at each pipeline location, the model can more accurately allocate water flow, ensure efficient operation of the boiler, avoid excessive or insufficient feed water, and optimize resource utilization.

[0069] The logic for obtaining the feedwater influence coefficient is as follows: determining the steam flow rate at each location of the boiler pipeline, and determining the degree of influence of the steam flow rate at each location on the feedwater system based on historical data, and determining the proportion of the influence of the steam flow rate at each location on the feedwater system based on the historical data;

[0070] Calculate the water supply influence coefficient, the calculation formula is: ;in, is the water supply influence coefficient, 、 、……、 is the influence of steam flow at each location on the water supply of the water supply system, 、 、……、 is the proportion of the impact of changes in steam flow at various locations in the historical data on the water supply of the water supply system, 、 、……、 is the steam flow rate at each location at each time point in a unit time period, and e is a natural number.

[0071] It can be seen from the formula that the larger the water supply influence coefficient is, the more likely the water supply system needs to open the gate opening more quickly to increase the water supply rate.

[0072] The water supply probability adjustment coefficient and water supply influence coefficient at each time point in the unit time period are weighted and calculated to construct a water supply assessment model and generate a water supply assessment coefficient. The calculation formula of the water supply assessment coefficient is: ;in, is the water supply assessment coefficient at each time point in the unit time period, 、 are the proportional coefficients of the water supply probability adjustment coefficient and the water supply influence coefficient, 、 Both are greater than 0.

[0073] It can be seen from the formula that the larger the water supply assessment coefficient is, the larger the water supply probability adjustment coefficient and the water supply influence coefficient at each time point are, which means that it is more likely that the water supply needs to be adjusted through the water supply system at this time point. Conversely, the smaller the water supply assessment coefficient is, the smaller the water supply probability adjustment coefficient and the water supply influence coefficient at each time point are, which means that it is more likely that the water supply does not need to be adjusted through the water supply system at this time point.

[0074] The feedwater assessment coefficient at each time point, the steam flow at each time point, the feedwater flow at each time point, the drum water level at each time point and the steam demand deviation in the current unit time period are used as the input of the neural network model, and the feedwater warning coefficient is used as the output of the neural network model. The feedwater warning coefficient is used to control the quantitative basis of the feedwater rate.

[0075] It should be noted that the input of the input layer in the neural network structure is time series data, and the dimension of the input data is 5*k, where k is the number of time points. The temporal dependency in the time series data is captured through the RNN layer, and the features extracted by the RNN layer are mapped to the output of the water supply warning coefficient. The output layer is the water supply warning coefficient.

[0076] Historical data is used for model training. Each input sample corresponds to a water supply warning coefficient. The output layer is the target value. The mean square error is used as the loss function. The water supply warning coefficient indicates the urgency of water supply adjustment in the current state.

[0077] The advantages of water supply early warning coefficient are:

[0078] The RNN layer can effectively process time series data. Through its internal loop structure, the RNN can "remember" the state of previous time points and pass this information to the current moment. This enables the neural network to better capture the temporal dependencies between data, identify potential temporal patterns in the water supply system, and ensure that the system can predict future water supply demand based on past conditions.

[0079] Through the training of the neural network model, the network can automatically learn the mapping relationship between input features and water supply warning coefficients. During the training process, by gradually minimizing the mean square error (MSE) loss function, the model can provide high-precision water supply warning coefficients.

[0080] The neural network can adjust the water supply warning coefficient under different input conditions and can dynamically respond to various real-time factors such as actual steam demand deviation and flow rate changes, without relying on fixed rules or linear models;

[0081] The water supply warning coefficient obtained through the neural network can be combined with the PID controller or other control strategies to further optimize the regulation of the water supply system and fine-tune the water supply rate to adapt to the dynamic changes of the boiler system.

[0082] The present invention constructs a probability density function and applies Bayesian inference, combines characteristics such as steam flow and feed water flow, quantifies the impact of each time point on the feed water decision, uses logistic regression to analyze the deviation impact of steam flow changes on the feed water system, and then optimizes the feed water system regulation. By applying the neural network model to the feed water decision of the boiler system, it can provide a more accurate and intelligent basis for the adjustment of the feed water rate based on the capture of complex temporal dependencies and multi-dimensional input characteristics. The present invention helps to accurately adjust the feed water rate and improve the efficiency and safety of boiler operation.

[0083] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0084] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0085] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0086] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0087] Those skilled in the art will 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.

[0088] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0089] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0090] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A water supply system control method based on flow self-distribution, characterized in that: The specific steps include: S1: Construct the probability density function of the monitoring feature by using historical records, determine the actual situation of the current monitoring feature through sensors, determine the water supply probability information based on the Bayesian formula, and adjust the weight of the monitoring feature through cross entropy; S2: By monitoring the steam flow at various locations in the boiler pipeline, the impact of steam flow changes at various locations on the water supply decision of the water supply system is determined based on historical data, and the water supply deviation information is determined through a logistic regression model; S3: Through comprehensive analysis of water supply probability information and water supply deviation information, a water supply assessment model is constructed to determine the water supply assessment coefficient at each time point and quantify the impact of each time point on water supply system decision-making; S4: By comprehensively considering the steam flow rate, feedwater flow rate and drum water level, and combining the feedwater assessment coefficient and actual steam demand, a neural network model is constructed to further optimize the control of the feedwater system; Use historical records to construct probability density functions of monitoring features, including: Based on the specific steam demand, determine the historical data of the monitoring characteristics, where the monitoring characteristics include steam flow, feed water flow, and drum water level, determine the steam demand deviation in the current unit time period, and mark the steam demand deviation in the current unit time period as: ,in, , XXQ is the actual steam demand per unit time period, and YXQ is the original steam demand per unit time period; Based on the historical data of monitoring characteristics under steam demand deviation, the probability density function of each monitoring characteristic under water supply instruction is calculated by kernel density estimation, and the probability density function of each monitoring characteristic under water supply instruction is marked as: 、 、 , is the probability density function of the steam flow rate under the current steam demand deviation under the water supply instruction, is the probability density function of the feedwater flow rate under the current steam demand deviation under the feedwater instruction, is the probability density function of the current steam demand deviation of the drum water level under the feedwater instruction, where , , ,i is the index of each monitoring feature sample data in the unit time period, i=1, 2, 3, ..., N, N is a positive integer, 、 、 is the smoothing parameter, K is the kernel function, ZQ is the steam flow rate, GS is the feed water flow rate, SW is the drum water level, is the effective sample number of steam flow, is the effective sample number of water flow rate, is the number of valid samples of drum water level, is the i-th steam flow sample value in the unit time period, is the i-th water flow sample value in the unit time period, is the i-th drum water level sample value in the unit time period; Based on the historical data of monitoring characteristics under steam demand deviation, the probability density function of each monitoring characteristic under no-water supply instruction is calculated by kernel density estimation, and the probability density function of each monitoring characteristic under no-water supply instruction is marked as: 、 、 ,in, is the probability density function of the steam flow rate at the current steam demand deviation under no water supply instruction, is the probability density function of the feedwater flow rate at the current steam demand deviation without feedwater instruction, is the probability density function of the deviation of the drum water level from the current steam demand under no feedwater instruction; The water supply probability information is represented by the water supply probability adjustment coefficient; Substitute the actual value of the monitoring feature of the current unit time period into the probability density function, determine the prior probability and posterior probability of the monitoring feature at each time point in the current time period, calculate the posterior probability of water supply, and mark the posterior probability of water supply as: ,in, 、 、 are the weights of steam flow, feed water flow, and drum water level, respectively. is the prior probability of water supply, is the posterior probability of water supply; Calculate the posterior probability of no water supply, the calculation formula is: ;in, is the prior probability of no water supply, is the posterior probability of not giving water; Among them, a weight is assigned to each monitoring feature, the prediction probability of the model is calculated through historical data, and the cross entropy loss is minimized using gradient descent to adjust the weight of each monitoring feature; Calculate the water supply probability adjustment coefficient, the calculation formula is: ;in, is the water supply probability adjustment coefficient; The water supply deviation information is expressed by the water supply influence coefficient; The logic for obtaining the feedwater influence coefficient is as follows: determining the steam flow rate at each location of the boiler pipeline, and determining the degree of influence of the steam flow rate at each location on the feedwater system based on historical data, and determining the proportion of the influence of the steam flow rate at each location on the feedwater system based on the historical data; Calculate the water supply influence coefficient, the calculation formula is: ;in, is the water supply influence coefficient, 、 、……、 is the influence of steam flow at each location on the water supply of the water supply system, 、 、……、 is the proportion of the impact of changes in steam flow at various locations in the historical data on the water supply of the water supply system, 、 、……、 is the steam flow rate at each location at each time point in a unit time period, and e is a natural number; The water supply probability adjustment coefficient and water supply influence coefficient at each time point in the unit time period are weighted and calculated to construct a water supply assessment model and generate a water supply assessment coefficient. The calculation formula of the water supply assessment coefficient is: ;in, is the water supply assessment coefficient at each time point in the unit time period, 、 are the proportional coefficients of the water supply probability adjustment coefficient and the water supply influence coefficient, 、 are both greater than 0; Among them, building a neural network model includes: The feedwater assessment coefficient at each time point, the steam flow at each time point, the feedwater flow at each time point, the drum water level at each time point and the steam demand deviation in the current unit time period are used as the input of the neural network model, and the feedwater warning coefficient is used as the output of the neural network model.

2. A water supply system control method based on flow self-distribution according to claim 1, characterized in that: Neural network models, including: The input layer inputs time series data with a dimension of 5*k, where k is the number of time points. The RNN layer captures the temporal dependencies in the time series data and maps the features extracted by the RNN layer to the output of the water supply warning coefficient. The output layer is the water supply warning coefficient. Historical data is used for model training. Each input sample corresponds to a water supply warning coefficient. The output layer is the target value, and the mean square error is used as the loss function.

Images