Energy-saving operation control optimization method and system for thermal power plant

Abstract

The invention relates to an energy-saving operation control optimization method and system for a thermal power plant, and relates to the field of thermal power plant energy-saving optimization. And performing operation energy-saving evaluation calculation by using the thermal fuzzy coefficient sequence, the power utilization fuzzy coefficient sequence, the heat supply fuzzy coefficient sequence and the power supply fuzzy coefficient sequence to obtain thermoelectric energy-saving fitness. And according to the thermoelectric energy-saving fitness, the operation parameters are adjusted and optimized, and the optimal operation parameters are obtained for energy-saving operation control. The technical problem that the energy utilization efficiency of the thermal power plant is low due to the fact that matched operation parameters cannot be adjusted and set in real time according to the actual heat and electricity supply and demand states in a traditional thermal power plant operation control method can be solved; the timeliness and accuracy of operation parameter setting can be improved, the optimal energy scheduling is realized, and the technical effects of high efficiency and energy conservation are achieved.

Description

Technical Field

[0001] The present invention relates to the field of energy-saving optimization of thermal power plants, and particularly to an energy-saving operation control optimization method and system for thermal power plants. Background Art

[0002] As a common energy production method, thermal power plants are responsible for providing both heat energy and electrical energy simultaneously, playing an important role in industrial and civil fields. In order to cope with energy shortages and environmental protection requirements, improving the operation efficiency and energy utilization rate of thermal power plants has become a key technical direction.

[0003] Currently, traditional thermal power plant operation control methods mostly rely on fixed scheduling plans or simple empirical rules, and fail to fully consider the differences between real-time changing heat supply, power supply and heat consumption, power consumption demands; therefore, when facing load fluctuations and demand changes, thermal power plants cannot flexibly adjust operation parameters, often resulting in excessive heat supply, insufficient power supply or energy waste, and being unable to effectively match the supply and demand status, thus causing inefficient utilization of energy resources. Summary of the Invention

[0004] Aiming at the technical problem in the traditional thermal power plant operation control method that due to the inability to adjust and set matching operation parameters in real time according to the actual heat and power supply and demand status, resulting in low energy utilization efficiency of thermal power plants, the present invention provides an energy-saving operation control optimization method and system for thermal power plants to solve this problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In the first aspect, the present invention provides an energy-saving operation control optimization method for thermal power plants, including: obtaining the heat consumption demand parameters and power consumption demand parameters of heat-consuming entities and power-consuming entities for combined heat and power supply of a thermal power plant within a past preset time range, performing integrated prediction of future heat and power consumption, obtaining a predicted heat parameter sequence and a predicted power parameter sequence, and performing predictive fuzzy analysis to obtain a heat fuzzy coefficient sequence and a power fuzzy coefficient sequence; obtaining the current real-time operation parameters of the thermal power plant, and randomly configuring adjustment operation parameters for control adjustment, performing integrated control adjustment prediction to obtain a predicted heat supply parameter sequence and a predicted power supply parameter sequence, and performing predictive fuzzy analysis to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; according to the predicted heat supply parameter sequence, predicted power supply parameter sequence, predicted heat parameter sequence and predicted power parameter sequence, as well as the heat fuzzy coefficient sequence, power fuzzy coefficient sequence, heat supply fuzzy coefficient sequence and power supply fuzzy coefficient sequence, performing operation energy-saving evaluation calculation to obtain a thermoelectric energy-saving fitness; according to the thermoelectric energy-saving fitness, optimizing the adjustment operation parameters to obtain optimal operation parameters, and performing energy-saving operation control on the thermal power plant.

[0007] Preferably, the energy-saving operation control optimization method for a thermal power plant further includes: obtaining the heat demand parameters and power demand parameters of the heat-consuming entities and power-consuming entities in the thermal power plant for combined heat and power supply and heat supply within a preset time range in the past, and recording them as the historical heat demand parameter sequence and the historical power demand parameter sequence; training a heat and power predictor, wherein the heat and power predictor includes a plurality of heat prediction paths and a plurality of power prediction paths; inputting the historical heat demand parameter sequence and the historical power demand parameter sequence into the plurality of heat prediction paths and the plurality of power prediction paths respectively, predicting the heat parameters and power parameters at multiple time nodes within a preset future time window, and outputting to obtain a plurality of heat parameter sequences and a plurality of power parameter sequences; performing prediction fuzzy analysis according to the plurality of heat parameter sequences and the plurality of power parameter sequences to obtain a heat fuzzy coefficient sequence and a power fuzzy coefficient sequence; calculating the average heat parameter and the average power parameter at each time node according to the plurality of heat parameter sequences and the plurality of power parameter sequences, and arranging them in time sequence to obtain a predicted heat parameter sequence and a predicted power parameter sequence.

[0008] Preferably, the energy-saving operation control optimization method for a thermal power plant further includes: according to the recorded data of the heat-consuming entities and power-consuming entities in the historical time, collecting a sample historical heat demand parameter sequence set and a sample historical power demand parameter sequence set, and collecting the heat parameter sequence or the power parameter sequence within a preset time window after each sample historical heat demand parameter sequence or sample historical power demand parameter sequence, and labeling them as a sample heat parameter sequence set and a sample power parameter sequence set; randomly dividing the sample historical heat demand parameter sequence set and the sample heat parameter sequence set multiple times to obtain multiple pieces of heat prediction training data; based on ensemble machine learning, using the multiple pieces of heat prediction training data to train and obtain a plurality of heat prediction paths; using the sample historical power demand parameter sequence set and the sample power parameter sequence set to train and obtain a plurality of power prediction paths; integrating the plurality of heat prediction paths and the plurality of power prediction paths to obtain a heat and power predictor.

[0009] Preferably, the energy-saving operation control optimization method for a thermal power plant further includes: within the multiple heat consumption parameter sequences and multiple power consumption parameter sequences, obtaining multiple first heat consumption parameters and multiple first power consumption parameters at a first time node within a future preset time window; respectively calculating the differences between the maximum value and the minimum value within the multiple first heat consumption parameters and multiple first power consumption parameters to obtain a first heat consumption parameter difference and a first power consumption parameter difference, and respectively calculating the means to obtain a first average heat consumption parameter and a first average power consumption parameter; respectively calculating the ratios of the first heat consumption parameter difference and the first power consumption parameter difference to the first average heat consumption parameter and the first average power consumption parameter to obtain a first heat consumption fuzzy coefficient and a first power consumption fuzzy coefficient; continuing to calculate the heat consumption fuzzy coefficients and power consumption fuzzy coefficients at multiple time nodes within the future preset time window to obtain a heat consumption fuzzy coefficient sequence and a power consumption fuzzy coefficient sequence.

[0010] Preferably, the energy-saving operation control optimization method for a thermal power plant further includes: obtaining the current real-time operation parameters of the thermal power plant, and obtaining the operation parameter range of the thermal power plant, and randomly configuring adjusted operation parameters for control adjustment within the operation parameter range; training a heat and power supply predictor, wherein the heat and power supply predictor includes multiple heat and power supply prediction paths; inputting the real-time operation parameters and the adjusted operation parameters into the heat and power supply predictor, and predicting and outputting multiple heat supply parameter sequences and multiple power supply parameter sequences at multiple time nodes within a future preset time window; performing predictive fuzzy analysis based on the multiple heat supply parameter sequences and multiple power supply parameter sequences to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; calculating the average heat supply parameter and the average power supply parameter at each time node based on the multiple heat supply parameter sequences and multiple power supply parameter sequences, and arranging them in time sequence to obtain a predicted heat supply parameter sequence and a predicted power supply parameter sequence.

[0011] Preferably, the energy-saving operation control optimization method for a thermal power plant further includes: according to the operation data record of the thermal power plant, collecting a sample real-time operation parameter set and a sample adjusted operation parameter set, and obtaining the heat supply parameters and power supply parameters at multiple time nodes within a future preset time window after being controlled according to different sample real-time operation parameters and sample adjusted operation parameters, to obtain a sample heat supply parameter sequence set and a sample power supply parameter sequence set; performing multiple random partitions on the sample real-time operation parameter set, the sample adjusted operation parameter set, the sample heat supply parameter sequence set, and the sample power supply parameter sequence set to obtain multiple sets of heat and power supply prediction training data; respectively using the multiple sets of heat and power supply prediction training data, and training multiple heat and power supply prediction paths based on ensemble machine learning; integrating the multiple heat and power supply prediction paths to obtain a heat and power supply predictor.

[0012] Preferably, the energy-saving operation control optimization method for a thermal power plant further includes: according to the heat consumption fuzzy coefficient sequence and the heat supply fuzzy coefficient sequence, adding and calculating to obtain a heat fuzzy coefficient sequence, calculating the ratio of each heat fuzzy coefficient to the sum of the heat fuzzy coefficient sequence as the heat weight, and allocating to obtain a heat weight sequence; according to the electricity consumption fuzzy coefficient sequence and the power supply fuzzy coefficient sequence, adding and calculating to obtain an electricity fuzzy coefficient sequence, calculating the ratio of each electricity fuzzy coefficient to the sum of the electricity fuzzy coefficient sequence as the electricity weight, and allocating to obtain an electricity weight sequence; constructing an operation energy-saving evaluation function as follows: where, TPPES fit is the thermoelectric energy-saving fitness, N is the number of multiple time nodes within a preset future time window, is the heat weight at the i-th time node, is the predicted heat supply parameter at the i-th time node, is the predicted heat consumption parameter at the i-th time node, is the electricity weight at the i-th time node, is the predicted power supply parameter at the i-th time node,

[0013] is the predicted electricity consumption parameter at the i-th time node; according to the predicted heat supply parameter sequence, predicted power supply parameter sequence, predicted heat consumption parameter sequence and predicted electricity consumption parameter sequence, based on the operation energy-saving evaluation function, calculate to obtain the thermoelectric energy-saving fitness.

[0014] Preferably, the energy-saving operation control optimization method for a thermal power plant further includes: continuing to randomly configure new adjusted operation parameters, and processing and calculating to obtain a new thermoelectric energy-saving fitness; performing iterative optimization of the adjusted operation parameters until convergence, outputting the adjusted operation parameters with the maximum thermoelectric energy-saving fitness to obtain the optimal operation parameters, and performing energy-saving operation control on the thermal power plant.

[0015] In a second aspect, the present invention provides an energy-saving operation control optimization system for a thermal power plant, comprising: a heat and electricity integrated prediction module, configured to obtain the heat demand parameters and electricity demand parameters of heat-consuming entities and electricity-consuming entities in a thermal power plant for combined heat and power generation and supply within a preset time range in the past, perform integrated heat and electricity prediction for the future, obtain a predicted heat parameter sequence and a predicted electricity parameter sequence, and perform prediction fuzzy analysis to obtain a heat fuzzy coefficient sequence and an electricity fuzzy coefficient sequence; an integrated control adjustment prediction module, configured to obtain the current real-time operation parameters of the thermal power plant, randomly configure adjustment operation parameters for control adjustment, perform integrated control adjustment prediction, obtain a predicted heat supply parameter sequence and a predicted power supply parameter sequence, and perform prediction fuzzy analysis to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; an operation energy-saving evaluation module, configured to perform operation energy-saving evaluation calculation according to the predicted heat supply parameter sequence, the predicted power supply parameter sequence, the predicted heat parameter sequence, the predicted electricity parameter sequence, the heat fuzzy coefficient sequence, the electricity fuzzy coefficient sequence, the heat supply fuzzy coefficient sequence, and the power supply fuzzy coefficient sequence, to obtain a thermal power energy-saving fitness; and an adjustment operation parameter optimization module, configured to optimize the adjustment operation parameters according to the thermal power energy-saving fitness, obtain optimal operation parameters, and perform energy-saving operation control on the thermal power plant.

[0016] The beneficial effects of the present invention are as follows: By obtaining the heat demand parameters and electricity demand parameters of heat-consuming entities and electricity-consuming entities in a thermal power plant for combined heat and power generation and supply within a preset time range in the past, performing integrated heat and electricity prediction for the future, obtaining a predicted heat parameter sequence and a predicted electricity parameter sequence, and performing prediction fuzzy analysis to obtain a heat fuzzy coefficient sequence and an electricity fuzzy coefficient sequence; then obtaining the current real-time operation parameters of the thermal power plant, randomly configuring adjustment operation parameters for control adjustment, performing integrated control adjustment prediction, obtaining a predicted heat supply parameter sequence and a predicted power supply parameter sequence, and performing prediction fuzzy analysis to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; then performing operation energy-saving evaluation calculation according to the predicted heat supply parameter sequence, the predicted power supply parameter sequence, the predicted heat parameter sequence, the predicted electricity parameter sequence, the heat fuzzy coefficient sequence, the electricity fuzzy coefficient sequence, the heat supply fuzzy coefficient sequence, and the power supply fuzzy coefficient sequence, to obtain a thermal power energy-saving fitness; and finally optimizing the adjustment operation parameters according to the thermal power energy-saving fitness, obtaining optimal operation parameters, and performing energy-saving operation control on the thermal power plant. That is to say, through accurate demand prediction, fuzzy error analysis, and real-time adjustment of operation parameters, the timeliness, scientificity, and accuracy of operation parameter setting can be improved, and optimal energy scheduling can be achieved under dynamically changing load demands, thereby effectively reducing energy waste, lowering operation costs, and achieving the technical effect of high energy efficiency. Description of the Drawings

[0017] Figure 1Schematic flow chart of an energy-saving operation control optimization method for a thermal power plant provided by the present invention;

[0018] Figure 2 Schematic structural diagram of an energy-saving operation control optimization system for a thermal power plant provided by the present invention.

[0019] In the drawings, the components represented by the reference numerals are described as follows:

[0020] Thermal and electrical integrated prediction module 11, integrated control adjustment prediction module 12, operation energy-saving evaluation module 13, adjustment of operation parameter optimization module 14. Specific embodiments

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present invention.

[0022] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically defined.

[0023] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or more advantageous than other embodiments. In order to enable any person skilled in the art to implement and use the present invention, the following description is given. In the following description, details are set forth for purposes of explanation. It should be understood that those skilled in the art can recognize that the present invention can be implemented without using these specific details. In other instances, well-known structures and processes are not described in detail to avoid unnecessary details from obscuring the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope that conforms to the principles and features disclosed in the present invention.

[0024] Embodiment 1, as Figure 1 shown, the embodiment of the present invention provides an energy-saving operation control optimization method for a thermal power plant, which specifically includes the following steps:

[0025] S10: Obtain the heat demand parameters and electricity demand parameters of the heat-consuming entities and electricity-consuming entities in cogeneration heat and power supply of the thermal power plant within a preset time range in the past, conduct integrated prediction of future heat and electricity consumption, obtain a predicted heat parameter sequence and a predicted electricity parameter sequence, and conduct prediction fuzzy analysis to obtain a heat fuzzy coefficient sequence and an electricity fuzzy coefficient sequence.

[0026] Further, step S10 of the present invention further includes:

[0027] S11: Obtain the heat demand parameters and electricity demand parameters of the heat-consuming entities and electricity-consuming entities in cogeneration heat and power supply of the thermal power plant within a preset time range in the past, and record them as a historical heat demand parameter sequence and a historical electricity demand parameter sequence.

[0028] Specifically, cogeneration is an efficient energy production method that can generate heat energy and electrical energy simultaneously, and is usually used in multiple fields such as industry, residential, and commercial buildings. In a cogeneration system, the heat-consuming entity and the electricity-consuming entity are two key participating entities. The heat-consuming entity refers to the entity or user that needs heat energy. The cogeneration system meets the needs of these users by providing hot water, steam or other forms of heat energy. Common heat-consuming entities include industrial enterprises, commercial buildings, etc.; the electricity-consuming entity refers to the entity or user that needs electrical energy. The electricity demand of these users is the main target of the power supply part in the cogeneration system, such as industrial enterprises, residential houses, commercial buildings, etc. In a cogeneration system, the thermal power plant can simultaneously meet the needs of heat-consuming entities and electricity-consuming entities by jointly producing electrical energy and heat energy. The supply-demand relationship between the two is interrelated and affected by multiple factors such as time changes, climate factors, and production plans. Effectively coordinating these two demands can improve energy utilization efficiency and reduce energy waste.

[0029] First, collect the heat demand parameters and electricity demand parameters of the heat-consuming entities and electricity-consuming entities in cogeneration heat and power supply of the thermal power plant at multiple time points within a preset time range in the past (such as 2 hours). Among them, the preset time range can be set according to actual needs, such as 2 hours. The preset time range includes multiple time points, which can be set according to the data collection time interval. For example, data collection is carried out every 5 minutes. Then, after obtaining the data, organize the parameters of heat demand and electricity demand in the order of time points and record them as a historical heat demand parameter sequence and a historical electricity demand parameter sequence. These data sequences contain the demand quantities at different time nodes. Among them, each time point corresponds to the actual demand value, forming a continuous data sequence. These data provide basic data support for subsequent heat and electricity consumption prediction.

[0030] S12: Train a heat and electricity consumption predictor, where the heat and electricity consumption predictor includes multiple heat prediction paths and multiple electricity prediction paths.

[0031] Furthermore, step S12 of the present invention further includes:

[0032] S121: According to the recorded data of the heat-consuming entity and the electricity-consuming entity within the historical time, collect the sample historical heat demand parameter sequence set and the sample historical electricity demand parameter sequence set, and collect the heat parameter sequence or the electricity parameter sequence within the preset time window after each sample historical heat demand parameter sequence or sample historical electricity demand parameter sequence, and label them as the sample heat parameter sequence set and the sample electricity parameter sequence set; S122: Randomly divide the sample historical heat demand parameter sequence set and the sample heat parameter sequence set multiple times to obtain multiple copies of heat prediction training data; S123: Based on ensemble machine learning, use the multiple copies of heat prediction training data to train and obtain multiple heat prediction paths; S124: Use the sample historical electricity demand parameter sequence set and the sample electricity parameter sequence set to train and obtain multiple electricity prediction paths; S125: Integrate the multiple heat prediction paths and the multiple electricity prediction paths to obtain a heat and electricity predictor.

[0033] Specifically, first, according to the actual operation records of the thermal power plant, collect the heat demand parameters and electricity demand parameters of the heat-consuming entity and the electricity-consuming entity within the historical time range (such as within the most recent half year). The heat demand parameters represent the heat energy demand values at different time nodes, such as steam, hot water volume, etc.; the electricity demand parameters represent the electricity demand at different time nodes, such as kilowatt-hours, etc.; then sort the collected heat demand data according to the time nodes to obtain the sample historical heat demand parameter sequence. Similarly, organize the collected electricity demand data according to the time nodes into the electricity demand parameter sequence to obtain the sample historical heat demand parameter sequence set and the sample historical electricity demand parameter sequence set. Then, collect the heat parameter sequence within the preset time window (such as 2 hours) after each sample historical heat demand parameter sequence and label it as the sample heat parameter sequence; collect the electricity parameter sequence within the preset time window after each sample historical electricity demand parameter sequence and label it as the sample electricity parameter sequence to obtain the sample heat parameter sequence set and the sample electricity parameter sequence set.

[0034] Then, use the sample historical heat demand parameter sequence set and the sample heat parameter sequence set as the sample data set, and divide it equally into M parts, where M is a positive integer, and the value of M can be set according to requirements. For example, set M to 10 to obtain M data sets; further randomly select M times with replacement from the M data sets to obtain the first training set, and use the same method to iterate and select M times to obtain M training sets, that is, multiple copies of heat prediction training data.

[0035] Construct multiple heat consumption prediction paths based on the principles of machine learning. For example, construct multiple heat consumption prediction paths based on feedforward neural networks. A feedforward neural network is a basic artificial neural network structure. Its basic idea is to receive data through the input layer, then extract features through the hidden layer, and finally obtain the prediction result through the output layer. The heat consumption prediction path is a feedforward neural network model that can be iteratively optimized in machine learning and is used to predict the heat consumption demand of a thermal power plant. It includes an input layer, multiple hidden layers, and an output layer. Among them, the input data of the input layer is the historical heat consumption demand parameter sequence. The hidden layer extracts and processes the input data through an activation function (such as ReLU). The output data of the output layer is the predicted heat parameter sequence. Further, using the sample historical heat consumption demand parameter sequence as the input and the sample heat parameter sequence as the supervision, multiple heat consumption prediction training data are used to supervise and train multiple heat consumption prediction paths respectively. During the training process, first, after inputting the sample historical heat consumption demand parameter sequence, the data is calculated through each layer of the feedforward neural network. Each node is transformed through a weighted sum and an activation function (such as ReLU), and finally the prediction result of the output layer is obtained. Then, the prediction result output by the model is compared with the actual sample heat parameter sequence (supervision data), and the prediction error is calculated through a loss function. Then, according to the calculated loss, the weights of the neural network are updated using the backpropagation algorithm. The backpropagation algorithm calculates the gradient of the loss with respect to each weight and bias through the chain rule. Further, the weights are updated according to the gradient, and usually an optimization algorithm (such as gradient descent) is used to adjust the network weights. Each heat consumption prediction path (i.e., each independent neural network) is independently trained according to its own training data. Multiple heat consumption prediction paths receive different historical heat consumption demand data respectively and perform corresponding training. Each path is independently optimized to ensure that the model can make effective predictions under different conditions. The entire training process is iterative. Each iteration calculates the output through forward propagation, calculates the gradient through backpropagation, and updates the network weights. The goal of each training is to minimize the loss function so that the predicted value is as close as possible to the actual heat parameter sequence. Until the model converges, multiple trained heat consumption prediction paths are obtained.

[0036] On the other hand, the set of sample historical electricity demand parameter sequences and the set of sample electricity parameter sequences are used as electricity consumption training data, which are equally divided into multiple parts and iteratively selected multiple times with replacement to obtain multiple pieces of electricity consumption prediction training data; then, based on the same method of constructing the heat consumption prediction path, multiple electricity consumption prediction paths are constructed based on machine learning (feedforward neural network). The electricity consumption prediction path includes an input layer, multiple hidden layers, and an output layer. The input data of the input layer is the historical electricity demand parameter sequence, and the output data of the output layer is the predicted electricity parameter sequence; further, using the sample historical electricity demand parameter sequence as the input and the sample electricity parameter sequence as the supervision, multiple pieces of electricity consumption prediction training data are used to respectively supervise and train the multiple electricity consumption prediction paths. Among them, the training process can refer to the training process of the heat consumption prediction path until the model converges to obtain multiple trained electricity consumption prediction paths.

[0037] Finally, integrate the multiple heat consumption prediction paths and multiple electricity consumption prediction paths to obtain a heat and electricity consumption predictor. The core objective of the heat and electricity consumption predictor is to be able to simultaneously predict the heat consumption demand and electricity consumption demand of the thermal power plant in the future for a period of time and achieve accurate prediction under the dynamically changing load demand, so as to provide a scientific basis for the energy-saving operation control of the thermal power plant.

[0038] S13: Input the historical heat consumption demand parameter sequence and the historical electricity demand parameter sequence into the multiple heat consumption prediction paths and multiple electricity consumption prediction paths respectively, predict the heat parameters and electricity parameters at multiple time nodes within a future preset time window, and output to obtain multiple heat parameter sequences and multiple electricity parameter sequences.

[0039] Specifically, input the historical heat consumption demand parameter sequence into the multiple heat consumption prediction paths to predict the heat parameters at multiple time nodes within a future preset time window (such as within the next 2 hours). Each heat consumption prediction path will use the input historical heat consumption demand parameter sequence as its input, calculate through each layer of the feedforward neural network, and output a predicted heat parameter sequence to obtain multiple heat parameter sequences. Similarly, input the historical electricity demand parameter sequence into multiple electricity consumption prediction paths respectively to predict the electricity parameters at multiple time nodes within a future preset time window, and output multiple electricity parameter sequences.

[0040] S14: Perform prediction fuzzy analysis based on the multiple heat parameter sequences and multiple electricity parameter sequences to obtain a heat consumption fuzzy coefficient sequence and an electricity consumption fuzzy coefficient sequence.

[0041] Furthermore, step S14 of the present invention further includes:

[0042] S141: Within the multiple heat - using parameter sequences and multiple power - using parameter sequences, obtain multiple first heat - using parameters and multiple first power - using parameters at the first time node within a future preset time window; S142: Calculate the differences between the maximum and minimum values within the multiple first heat - using parameters and multiple first power - using parameters respectively to obtain a first heat - using parameter difference and a first power - using parameter difference, and calculate the means respectively to obtain a first average heat - using parameter and a first average power - using parameter; S143: Calculate the ratios of the first heat - using parameter difference and the first power - using parameter difference to the first average heat - using parameter and the first average power - using parameter respectively to obtain a first heat - using fuzzy coefficient and a first power - using fuzzy coefficient; S144: Continue to calculate the heat - using fuzzy coefficients and power - using fuzzy coefficients at multiple time nodes within the future preset time window to obtain a heat - using fuzzy coefficient sequence and a power - using fuzzy coefficient sequence.

[0043] Specifically, within the multiple heat - using parameter sequences and multiple power - using parameter sequences, obtain multiple first heat - using parameters and multiple first power - using parameters at the first time node (the first time node) within a future preset time window, that is, extract the predicted values of all heat - using prediction paths at the first time node and extract the predicted values of all power - using prediction paths at the first time node. Then, calculate the difference between the maximum and minimum values within the multiple first heat - using parameters, and set the heat - using parameter difference obtained by subtracting the minimum value from the maximum value as the first heat - using parameter difference; calculate the difference between the maximum and minimum values within the multiple first power - using parameters to obtain the first power - using parameter difference. On the other hand, calculate the mean of the multiple first heat - using parameters to obtain the first average heat - using parameter. The first average heat - using parameter represents the central tendency of multiple heat - using prediction paths and reflects the average heat - using demand at the future first time node; calculate the mean of the multiple first power - using parameters to obtain the first average power - using parameter. The first average power - using parameter represents the central tendency of multiple power - using prediction paths and reflects the average power - using demand at the future first time node.

[0044] Next, calculate the ratio of the first heat - using parameter difference to the first average heat - using parameter, and set it as the first heat - using fuzzy coefficient. The fuzzy coefficient is used to reflect the uncertainty of the prediction result. The larger the difference, the higher the fuzzy coefficient; the heat - using fuzzy coefficient characterizes the degree of fluctuation of the heat - using demand prediction relative to the average predicted value. On the other hand, set the ratio of the first power - using parameter difference to the first average power - using parameter as the first power - using fuzzy coefficient. The power - using fuzzy coefficient characterizes the degree of fluctuation of the power - using demand prediction relative to the average predicted value. These fuzzy coefficients provide important information for subsequent control and optimization, helping to judge which time nodes have higher prediction uncertainties, thus providing a basis for key optimization and adjustment.

[0045] Then, using the same method for calculating the first heat usage fuzzy coefficient and the first electricity usage fuzzy coefficient, continue to calculate the heat usage fuzzy coefficients and electricity usage fuzzy coefficients at multiple other time nodes within a future preset time window. By calculating the ratio of the difference between the predicted heat usage values at each time node to the mean value, obtain the heat usage fuzzy coefficient sequence within the entire time window, which reflects the uncertainty of the predicted heat demand at different time points. Similarly, by calculating the ratio of the difference between the predicted electricity usage values at each time node to the mean value, obtain the electricity usage fuzzy coefficient sequence within the entire time window, which reflects the uncertainty of the predicted electricity demand at different time points. These fuzzy coefficient sequences will provide valuable references for subsequent energy-saving control and scheduling optimization of the thermal power plant, helping to identify which time nodes have higher prediction uncertainties, thereby providing a basis for key optimization and adjustment.

[0046] S15: According to the multiple heat usage parameter sequences and multiple electricity usage parameter sequences, calculate the average heat usage parameter and the average electricity usage parameter at each time node, and obtain the predicted heat usage parameter sequence and the predicted electricity usage parameter sequence arranged in time series.

[0047] Specifically, according to the multiple heat usage parameter sequences and multiple electricity usage parameter sequences, calculate the average heat usage parameter at each time node respectively, obtain multiple average heat usage parameters, and obtain the predicted heat usage parameter sequence arranged in time series. On the other hand, calculate the average electricity usage parameter at each time node respectively, and obtain the predicted electricity usage parameter sequence arranged in time series. Among them, the predicted heat usage parameter sequence and the predicted electricity usage parameter sequence will serve as the basis for subsequent energy-saving operation control and scheduling of the thermal power plant, helping to more precisely adjust the operation parameters of the unit and achieve efficient energy utilization.

[0048] S20: Obtain the current real-time operation parameters of the thermal power plant, randomly configure the adjusted operation parameters for control adjustment, perform integrated control adjustment prediction, obtain the predicted heat supply parameter sequence and the predicted power supply parameter sequence, and perform predicted fuzzy analysis to obtain the heat supply fuzzy coefficient sequence and the power supply fuzzy coefficient sequence.

[0049] Furthermore, step S20 of the present invention further includes:

[0050] S21: Obtain the current real-time operation parameters of the thermal power plant, and obtain the operation parameter range of the thermal power plant. Randomly configure the adjusted operation parameters for control adjustment within the operation parameter range.

[0051] Specifically, first, monitor and obtain the current real-time operation parameters of the thermal power plant. The real-time operation parameters of the thermal power plant include, but are not limited to, boiler parameters (such as boiler pressure, temperature, steam flow rate, etc.), generator parameters (such as generator load, rotation speed, voltage, current, etc.), heating parameters (such as heating temperature, heating flow rate, etc.), etc. These parameters reflect the current states of each system during the operation of the thermal power plant. Next, obtain the operation parameter ranges of the thermal power plant. The operation parameter ranges refer to the ranges within which each operation parameter can be adjusted under normal operating conditions. For different operation parameters, their effective ranges will vary; these ranges represent the adjustable ranges of each operation parameter under the conditions of meeting safety and equipment limitations. The operation parameter ranges are generally determined by factors such as equipment performance, plant design, and process requirements.

[0052] Next, randomly configure the adjusted operation parameters for control adjustment within the operation parameter ranges, that is, for operation parameters (such as boiler temperature, steam flow rate, etc.), randomly adjust them within their corresponding ranges, and finally obtain a new set of adjusted operation parameters. By obtaining the real-time operation parameters of the thermal power plant and their operation parameter ranges, and randomly configuring the adjusted operation parameters based on this information, various possible control schemes can be explored. These schemes will help to further optimize the operation parameters of the thermal power plant, improve energy utilization efficiency, reduce energy waste, and thus achieve the energy-saving goal.

[0053] S22: Train the heating and power supply predictor, where the heating and power supply predictor includes multiple heating and power supply prediction paths.

[0054] Furthermore, step S22 of the present invention further includes:

[0055] S221: According to the operation data records of the thermal power plant, collect the sample real-time operation parameter set and the sample adjusted operation parameter set, and obtain the heating parameters and power supply parameters at multiple time nodes within a future preset time window after being controlled by different sample real-time operation parameters and sample adjusted operation parameters, to obtain the sample heating parameter sequence set and the sample power supply parameter sequence set; S222: Randomly divide the sample real-time operation parameter set, the sample adjusted operation parameter set, the sample heating parameter sequence set, and the sample power supply parameter sequence set multiple times to obtain multiple sets of heating and power supply prediction training data; S223: Respectively use the multiple sets of heating and power supply prediction training data, and based on ensemble machine learning, train to obtain multiple heating and power supply prediction paths; S224: Integrate the multiple heating and power supply prediction paths to obtain the heating and power supply predictor.

[0056] Specifically, first, according to the operation data records of the thermal power plant, a set of sample real-time operation parameters and a set of sample adjusted operation parameters are collected. The real-time operation parameters represent the actual operation state of the thermal power plant at a certain moment, including boiler pressure, generator load, heating temperature, power supply flow rate, etc.; the adjusted operation parameters are the new parameters obtained by adjusting the real-time operation parameters based on the control strategy, including the adjusted boiler pressure, steam flow rate, generator load, etc. Then, the heating parameters and power supply parameters at multiple time nodes within a future preset time window are obtained according to different sample real-time operation parameters and sample adjusted operation parameters, and a set of sample heating parameter sequences and a set of sample power supply parameter sequences are obtained.

[0057] Next, the set of sample real-time operation parameters, the set of sample adjusted operation parameters, the set of sample heating parameter sequences, and the set of sample power supply parameter sequences are used as a sample data set and divided into multiple equal parts to construct a first training set. Then, the same method is iteratively selected multiple times to obtain multiple sets of heating and power supply prediction training data. Further, based on machine learning (constructing a feedforward neural network), multiple heating and power supply prediction paths are constructed. The heating and power supply prediction path is a feedforward neural network model that can be iteratively optimized in machine learning, including an input layer, multiple hidden layers, and an output layer. Among them, the input data of the input layer is the real-time operation parameters and the adjusted operation parameters, and the output data of the output layer is the predicted heating parameter sequence and the predicted power supply parameter sequence. Further, the multiple sets of heating and power supply prediction training data are respectively used to perform supervised training on the multiple heating and power supply prediction paths. During the training process, first, the input data (real-time operation parameters and adjusted operation parameters) in the training data set are input into the input layer of the neural network and calculated through the hidden layers in the neural network to obtain the predicted output data (heating and power supply parameters); then, the error between the predicted output and the actual output (i.e., the supervised data) is calculated, and usually the mean square error is used to measure it; then the backpropagation algorithm is used to update the weights and biases of the neural network according to the calculated error, and the optimization algorithm (such as gradient descent) adjusts the parameters of each layer according to the gradient of the error; through multiple iterative optimizations, the network parameters are continuously updated until the loss function converges. After each iteration, the network weights and biases are gradually optimized, and the prediction ability is improved. When the error drops to a certain extent or reaches the predetermined maximum number of iterations, the training process is stopped, and the trained heating and power supply prediction paths are obtained, and multiple heating and power supply prediction paths are obtained.

[0058] Finally, the multiple heating and power supply prediction paths are integrated to obtain a heating and power supply predictor; by constructing a heating and power supply predictor based on machine learning, the energy demand of the thermal power plant can be efficiently predicted, and accurate data support can be provided for subsequent energy-saving operation control.

[0059] S23: Input the real-time operation parameters and adjusted operation parameters into the heating and power supply predictor to predict and output multiple heating parameter sequences and multiple power supply parameter sequences at multiple time nodes within a future preset time window; S24: Conduct predictive fuzzy analysis based on the multiple heating parameter sequences and multiple power supply parameter sequences to obtain a heating fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; S25: Calculate the average heating parameter and average power supply parameter at each time node based on the multiple heating parameter sequences and multiple power supply parameter sequences, and obtain a predicted heating parameter sequence and a predicted power supply parameter sequence by arranging them in time series.

[0060] Specifically, first, input the real-time operation parameters and adjusted operation parameters into the heating and power supply predictor, and based on the patterns learned during the training process, output the prediction results of heating and power supply parameters at multiple time nodes within a future time window, obtaining multiple heating parameter sequences and multiple power supply parameter sequences. Then, based on the above predictive fuzzy analysis method, conduct predictive fuzzy analysis based on the multiple heating parameter sequences, that is, conduct error analysis on the heating parameters at each time node, calculate the difference between the maximum value and the minimum value, and calculate the ratio of the difference to the mean value as the fuzzy coefficient; similarly, conduct predictive fuzzy analysis based on the multiple power supply parameter sequences to obtain a power supply fuzzy coefficient sequence.

[0061] On the other hand, calculate the average heating parameter at each time node based on the multiple heating parameter sequences, and obtain a predicted heating parameter sequence by arranging them in time series; calculate the average power supply parameter at each time node based on the multiple power supply parameter sequences, and obtain a predicted power supply parameter sequence by arranging them in time series. The predicted heating parameter sequence and the predicted power supply parameter sequence reflect the heating and power supply demands within a future time window and provide data support for subsequent adjustment and optimization.

[0062] S30: Conduct an operation energy-saving evaluation calculation based on the predicted heating parameter sequence, predicted power supply parameter sequence, predicted heat consumption parameter sequence, predicted power consumption parameter sequence, as well as the heat consumption fuzzy coefficient sequence, power consumption fuzzy coefficient sequence, heating fuzzy coefficient sequence, and power supply fuzzy coefficient sequence to obtain a thermoelectric energy-saving fitness.

[0063] Furthermore, step S30 of the present invention further includes:

[0064] S31: Based on the thermal fuzzy coefficient sequence and the heat supply fuzzy coefficient sequence, perform summation calculation to obtain the thermal fuzzy coefficient sequence, calculate the ratio of each thermal fuzzy coefficient to the total of the thermal fuzzy coefficient sequence as the thermal weight, and allocate to obtain the thermal weight sequence; S32: Based on the electricity consumption fuzzy coefficient sequence and the power supply fuzzy coefficient sequence, perform summation calculation to obtain the electricity fuzzy coefficient sequence, calculate the ratio of each electricity fuzzy coefficient to the total of the electricity fuzzy coefficient sequence as the electricity weight, and allocate to obtain the electricity weight sequence; S33: Construct an operation energy-saving evaluation function as follows: where, TPPES fit is the thermoelectric energy-saving fitness, N is the number of multiple time nodes within a preset future time window, is the thermal weight of the i-th time node, is the predicted heat supply parameter of the i-th time node, is the predicted heat consumption parameter of the i-th time node, is the electricity weight of the i-th time node, is the predicted power supply parameter of the i-th time node, is the predicted electricity consumption parameter of the i-th time node; S34: Based on the predicted heat supply parameter sequence, the predicted power supply parameter sequence, the predicted heat consumption parameter sequence, and the predicted electricity consumption parameter sequence, calculate and obtain the thermoelectric energy-saving fitness based on the operation energy-saving evaluation function.

[0065] Specifically, based on the heat consumption fuzzy coefficient sequence and the heat supply fuzzy coefficient sequence, add and sum the heat consumption fuzzy coefficient and the heat supply fuzzy coefficient at the same time point, and set it as the thermal fuzzy coefficient at this time point to obtain the thermal fuzzy coefficient sequence; further calculate the ratio of each thermal fuzzy coefficient to the total of the thermal fuzzy coefficient sequence, that is, for each time node, calculate the ratio of each thermal fuzzy coefficient to the total of the entire thermal fuzzy coefficient sequence as the thermal weight of each time point, which represents the relative importance of this time node to the overall heat energy demand. The greater the weight, the greater the impact of this node on the thermal energy scheduling of the system, and obtain the thermal weight sequence.

[0066] On the other hand, based on the electricity consumption fuzzy coefficient sequence and the power supply fuzzy coefficient sequence, add and sum the electricity consumption fuzzy coefficient and the power supply fuzzy coefficient at the same time point, and set it as the electricity fuzzy coefficient at this time point to obtain the electricity fuzzy coefficient sequence; further calculate the ratio of each electricity fuzzy coefficient to the total of the electricity fuzzy coefficient sequence as the electricity weight, and obtain the electricity weight sequence. The electricity weight of each time node reflects the impact of this node on the power scheduling of the system. The greater the weight, the greater the impact of this node on the power scheduling of the system.

[0067] Then, construct an operation energy-saving evaluation function. In the operation energy-saving evaluation function, TPPES fitis the thermoelectric energy-saving fitness. The larger the fitness, the better the energy-saving effect is represented. N is the number of multiple time nodes within a preset future time window. is the thermal weight of the i-th time node (any one of the N time nodes). is the predicted heat supply parameter of the i-th time node. is the predicted heat consumption parameter of the i-th time node. is the electrical weight of the i-th time node. is the predicted power supply parameter of the i-th time node. is the predicted power consumption parameter of the i-th time node. Further, based on the operation energy-saving evaluation function, according to the predicted heat supply parameter sequence, predicted power supply parameter sequence, predicted heat consumption parameter sequence, and predicted power consumption parameter sequence, the thermoelectric energy-saving fitness is calculated and obtained.

[0068] S40: According to the thermoelectric energy-saving fitness, optimize the adjusted operation parameters to obtain the optimal operation parameters, and perform energy-saving operation control on the thermal power plant.

[0069] Further, step S40 of the present invention further includes:

[0070] S41: Continuously randomly configure new adjusted operation parameters, and process and calculate to obtain the new thermoelectric energy-saving fitness; S42: Perform iterative optimization of the adjusted operation parameters until convergence, output the adjusted operation parameters with the maximum thermoelectric energy-saving fitness, obtain the optimal operation parameters, and perform energy-saving operation control on the thermal power plant.

[0071] Specifically, continuously randomly configure new adjusted operation parameters. In each iteration, according to the operation data of the thermal power plant and the predetermined operation parameter range, randomly generate new adjusted operation parameters. These adjusted operation parameters are used to simulate the control of the energy dispatch of the thermal power plant, try different parameter combinations to evaluate their impact on the energy-saving effect. Then, according to the newly configured adjusted operation parameters, process and calculate to obtain the new thermoelectric energy-saving fitness. Perform iterative optimization of the adjusted operation parameters until reaching the predetermined number of adjustments (such as 100 times), output the adjusted operation parameters with the maximum thermoelectric energy-saving fitness during the entire optimization process as the optimal operation parameters, and perform energy-saving operation control on the thermal power plant according to the optimal operation parameters, that is, apply the optimal operation parameters to the control system of the thermal power plant to actually control its heat supply and power supply systems, ensuring that the thermal power plant can perform energy dispatch according to the optimal strategy during actual operation and achieve the energy-saving control goal.

[0072] By randomly configuring the adjusted operation parameters, calculating the thermoelectric energy-saving fitness, and performing iterative optimization, the optimal operation parameters can be finally obtained, enabling the thermal power plant to precisely adjust the energy supply and consumption under changing load and demand conditions, optimize the operation efficiency of the cogeneration system, and thus achieve a significant energy-saving effect.

[0073] An energy-saving operation control optimization method for a thermal power plant provided by an embodiment of the present invention has at least the following technical effects:

[0074] By obtaining the heat demand parameters and power demand parameters of the heat-consuming main body and the power-consuming main body for cogeneration of heat and power supply in a thermal power plant within a preset time range in the past, future heat and power integration prediction is carried out to obtain a predicted heat parameter sequence and a predicted power parameter sequence, and prediction fuzzy analysis is carried out to obtain a heat consumption fuzzy coefficient sequence and a power consumption fuzzy coefficient sequence; then the current real-time operation parameters of the thermal power plant are obtained, and the adjusted operation parameters for control adjustment are randomly configured, and integrated control adjustment prediction is carried out to obtain a predicted heat supply parameter sequence and a predicted power supply parameter sequence, and prediction fuzzy analysis is carried out to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; then, according to the predicted heat supply parameter sequence, the predicted power supply parameter sequence, the predicted heat parameter sequence and the predicted power parameter sequence, as well as the heat consumption fuzzy coefficient sequence, the power consumption fuzzy coefficient sequence, the heat supply fuzzy coefficient sequence and the power supply fuzzy coefficient sequence, operation energy-saving evaluation calculation is carried out to obtain the thermal power energy-saving fitness; finally, according to the thermal power energy-saving fitness, the optimization of the adjusted operation parameters is carried out to obtain the optimal operation parameters, and energy-saving operation control is carried out on the thermal power plant. That is to say, through accurate demand prediction, fuzzy error analysis and real-time adjustment of operation parameters, the timeliness, scientificity and accuracy of operation parameter setting can be improved, and the optimal energy scheduling can be achieved under dynamic load demand, so as to effectively reduce energy waste, reduce operation costs and achieve the technical effect of high-efficiency energy saving.

[0075] Example 2, as Figure 2As shown, based on the same inventive concept as the energy-saving operation control optimization method for a thermal power plant provided in Embodiment 1, an energy-saving operation control optimization system for a thermal power plant provided in an embodiment of the present invention includes: a heat and electricity integrated prediction module 11, configured to obtain the heat demand parameters and electricity demand parameters of the heat-consuming entities and electricity-consuming entities for cogeneration of heat and power in the thermal power plant within a past preset time range, perform future heat and electricity integrated prediction, obtain a predicted heat parameter sequence and a predicted electricity parameter sequence, and perform predicted fuzzy analysis to obtain a heat fuzzy coefficient sequence and an electricity fuzzy coefficient sequence; an integrated control adjustment prediction module 12, configured to obtain the current real-time operation parameters of the thermal power plant, randomly configure adjustment operation parameters for control adjustment, perform integrated control adjustment prediction, obtain a predicted heat supply parameter sequence and a predicted power supply parameter sequence, and perform predicted fuzzy analysis to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; an operation energy-saving evaluation module 13, configured to perform operation energy-saving evaluation calculation according to the predicted heat supply parameter sequence, the predicted power supply parameter sequence, the predicted heat parameter sequence, the predicted electricity parameter sequence, as well as the heat fuzzy coefficient sequence, the electricity fuzzy coefficient sequence, the heat supply fuzzy coefficient sequence, and the power supply fuzzy coefficient sequence, to obtain a thermal power energy-saving fitness; an adjustment operation parameter optimization module 14, configured to optimize the adjustment operation parameters according to the thermal power energy-saving fitness, obtain the optimal operation parameters, and perform energy-saving operation control on the thermal power plant.

[0076] Further, the energy-saving operation control optimization system for a thermal power plant is further configured to: obtain the heat demand parameters and electricity demand parameters of the heat-consuming entities and electricity-consuming entities for cogeneration of heat and power in the thermal power plant within a past preset time range, and record them as a historical heat demand parameter sequence and a historical electricity demand parameter sequence; train a heat and electricity predictor, where the heat and electricity predictor includes a plurality of heat prediction paths and a plurality of electricity prediction paths; input the historical heat demand parameter sequence and the historical electricity demand parameter sequence into the plurality of heat prediction paths and the plurality of electricity prediction paths respectively, predict the heat parameters and electricity parameters at multiple time nodes within a future preset time window, and output to obtain a plurality of heat parameter sequences and a plurality of electricity parameter sequences; perform predicted fuzzy analysis according to the plurality of heat parameter sequences and the plurality of electricity parameter sequences to obtain a heat fuzzy coefficient sequence and an electricity fuzzy coefficient sequence; calculate the average heat parameter and the average electricity parameter at each time node according to the plurality of heat parameter sequences and the plurality of electricity parameter sequences, and arrange them in time sequence to obtain a predicted heat parameter sequence and a predicted electricity parameter sequence.

[0077] Further, the energy-saving operation control optimization system for a thermal power plant is also used for: according to the recorded data of heat-consuming entities and power-consuming entities within a historical time, collecting a set of sample historical heat demand parameter sequences and a set of sample historical power demand parameter sequences, and collecting heat parameter sequences or power parameter sequences within a preset time window after each sample historical heat demand parameter sequence or sample historical power demand parameter sequence, and labeling them as a set of sample heat parameter sequences and a set of sample power parameter sequences; randomly dividing the set of sample historical heat demand parameter sequences and the set of sample heat parameter sequences multiple times to obtain multiple copies of heat prediction training data; based on ensemble machine learning, using the multiple copies of heat prediction training data to train and obtain multiple heat prediction paths; using the set of sample historical power demand parameter sequences and the set of sample power parameter sequences to train and obtain multiple power prediction paths; integrating the multiple heat prediction paths and the multiple power prediction paths to obtain a heat and power predictor.

[0078] Further, the energy-saving operation control optimization system for a thermal power plant is also used for: within the multiple heat parameter sequences and the multiple power parameter sequences, obtaining multiple first heat parameters and multiple first power parameters at the first time node within a future preset time window; respectively calculating the differences between the maximum value and the minimum value within the multiple first heat parameters and the multiple first power parameters to obtain a first heat parameter difference and a first power parameter difference, and respectively calculating the mean values to obtain a first average heat parameter and a first average power parameter; respectively calculating the ratios of the first heat parameter difference and the first power parameter difference to the first average heat parameter and the first average power parameter to obtain a first heat fuzzy coefficient and a first power fuzzy coefficient; continuing to calculate the heat fuzzy coefficients and the power fuzzy coefficients at multiple time nodes within a future preset time window to obtain a heat fuzzy coefficient sequence and a power fuzzy coefficient sequence.

[0079] Further, the energy-saving operation control optimization system for a thermal power plant is also used for: obtaining the current real-time operation parameters of the thermal power plant, and obtaining the operation parameter range of the thermal power plant, and randomly configuring adjustment operation parameters for control adjustment within the operation parameter range; training a heat and power predictor, where the heat and power predictor includes multiple heat and power prediction paths; inputting the real-time operation parameters and the adjustment operation parameters into the heat and power predictor to predict and output multiple heat parameter sequences and multiple power parameter sequences at multiple time nodes within a future preset time window; performing predictive fuzzy analysis according to the multiple heat parameter sequences and the multiple power parameter sequences to obtain a heat fuzzy coefficient sequence and a power fuzzy coefficient sequence; according to the multiple heat parameter sequences and the multiple power parameter sequences, calculating the average heat parameter and the average power parameter at each time node, and arranging them in time sequence to obtain a predicted heat parameter sequence and a predicted power parameter sequence.

[0080] Furthermore, the energy-saving operation control optimization system for a thermal power plant is also used for: according to the operation data records of the thermal power plant, collecting a set of sample real-time operation parameters and a set of sample adjusted operation parameters, and obtaining the heating parameters and power supply parameters at multiple time nodes within a future preset time window controlled by different sample real-time operation parameters and sample adjusted operation parameters, so as to obtain a set of sample heating parameter sequences and a set of sample power supply parameter sequences; performing multiple random partitions on the set of sample real-time operation parameters, the set of sample adjusted operation parameters, the set of sample heating parameter sequences, and the set of sample power supply parameter sequences to obtain multiple sets of heating and power supply prediction training data; respectively using the multiple sets of heating and power supply prediction training data, and training multiple heating and power supply prediction paths based on integrated machine learning; integrating the multiple heating and power supply prediction paths to obtain a heating and power supply predictor.

[0081] Furthermore, the energy-saving operation control optimization system for a thermal power plant is also used for: according to the sequence of heat usage fuzzy coefficients and the sequence of heating fuzzy coefficients, performing addition calculation to obtain a sequence of heat fuzzy coefficients, calculating the ratio of each heat fuzzy coefficient to the sum of the sequence of heat fuzzy coefficients as the heat weight, and allocating to obtain a sequence of heat weights; according to the sequence of electricity usage fuzzy coefficients and the sequence of power supply fuzzy coefficients, performing addition calculation to obtain a sequence of electricity fuzzy coefficients, calculating the ratio of each electricity fuzzy coefficient to the sum of the sequence of electricity fuzzy coefficients as the electricity weight, and allocating to obtain a sequence of electricity weights; constructing an operation energy-saving evaluation function as follows: where TPPES fit is the thermal power energy-saving fitness, N is the number of multiple time nodes within the future preset time window, is the heat weight at the i-th time node, is the predicted heating parameter at the i-th time node, is the predicted heat usage parameter at the i-th time node, is the electricity weight at the i-th time node, is the predicted power supply parameter at the i-th time node, is the predicted electricity usage parameter at the i-th time node; according to the predicted heating parameter sequence, the predicted power supply parameter sequence, the predicted heat usage parameter sequence, and the predicted electricity usage parameter sequence, calculating the thermal power energy-saving fitness based on the operation energy-saving evaluation function.

[0082] Furthermore, the energy-saving operation control optimization system for a thermal power plant is also used for: continuing to randomly configure new adjusted operation parameters, and processing and calculating to obtain a new thermal power energy-saving fitness; performing iterative optimization of the adjusted operation parameters until convergence, outputting the adjusted operation parameters with the maximum thermal power energy-saving fitness to obtain the optimal operation parameters, and performing energy-saving operation control on the thermal power plant.

[0083] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made to these embodiments by those skilled in the art once they learn of the basic inventive concept.

[0084] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalent technologies, the present invention also intends to include these modifications and variations.

Claims

1. An energy-saving operation control optimization method for a thermal power plant, characterized in that, The method includes: Obtaining the heat demand parameters and power demand parameters of the heat-consuming entities and power-consuming entities in cogeneration of heat and power in a thermal power plant within a past preset time range, performing integrated prediction of future heat and power consumption, obtaining a predicted heat parameter sequence and a predicted power parameter sequence, and performing predictive fuzzy analysis to obtain a heat fuzzy coefficient sequence and a power fuzzy coefficient sequence; Obtaining the current real-time operation parameters of the thermal power plant, randomly configuring the adjusted operation parameters for control adjustment, performing integrated control adjustment prediction, obtaining a predicted heat supply parameter sequence and a predicted power supply parameter sequence, and performing predictive fuzzy analysis to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; Performing an operation energy-saving evaluation calculation based on the predicted heat supply parameter sequence, the predicted power supply parameter sequence, the predicted heat parameter sequence, the predicted power parameter sequence, as well as the heat fuzzy coefficient sequence, the power fuzzy coefficient sequence, the heat supply fuzzy coefficient sequence, and the power supply fuzzy coefficient sequence to obtain a thermoelectric energy-saving fitness; Optimizing the adjusted operation parameters according to the thermoelectric energy-saving fitness to obtain the optimal operation parameters, and performing energy-saving operation control on the thermal power plant.

2. The energy-saving operation control optimization method for a thermal power plant according to claim 1, wherein Obtaining the heat demand parameters and power demand parameters of the heat-consuming entities and power-consuming entities in cogeneration of heat and power in a thermal power plant within a past preset time range, performing integrated prediction of future heat and power consumption, including: Obtaining the heat demand parameters and power demand parameters of the heat-consuming entities and power-consuming entities in cogeneration of heat and power in a thermal power plant within a past preset time range, and recording them as a historical heat demand parameter sequence and a historical power demand parameter sequence; Training a heat and power consumption predictor, where the heat and power consumption predictor includes multiple heat prediction paths and multiple power prediction paths; Inputting the historical heat demand parameter sequence and the historical power demand parameter sequence into the multiple heat prediction paths and the multiple power prediction paths respectively, predicting the heat parameters and power parameters at multiple time nodes within a future preset time window, and outputting to obtain multiple heat parameter sequences and multiple power parameter sequences; Performing predictive fuzzy analysis based on the multiple heat parameter sequences and the multiple power parameter sequences to obtain a heat fuzzy coefficient sequence and a power fuzzy coefficient sequence; Calculating the average heat parameter and the average power parameter at each time node according to the multiple heat parameter sequences and the multiple power parameter sequences, and arranging them in time sequence to obtain a predicted heat parameter sequence and a predicted power parameter sequence.

3. The energy-saving operation control optimization method for a thermal power plant according to claim 2, characterized in that Training a heat and power consumption predictor, including: According to the recorded data of the heat-consuming entities and power-consuming entities within a historical time, collecting a set of sample historical heat demand parameter sequences and a set of sample historical power demand parameter sequences, and collecting the heat parameter sequences or power parameter sequences within a preset time window after each sample historical heat demand parameter sequence or sample historical power demand parameter sequence, and labeling them as a set of sample heat parameter sequences and a set of sample power parameter sequences; Performing multiple random partitions on the set of sample historical heat demand parameter sequences and the set of sample heat parameter sequences to obtain multiple sets of heat prediction training data; Based on integrated machine learning, using the multiple pieces of heat consumption prediction training data, multiple heat consumption prediction paths are trained and obtained; Using the sample historical electricity demand parameter sequences set and the sample electricity parameter sequences set, multiple electricity consumption prediction paths are trained and obtained; Integrate the multiple heat consumption prediction paths and the multiple electricity consumption prediction paths to obtain a heat and electricity consumption predictor.

4. The energy-saving operation control optimization method for a thermal power plant according to claim 2, wherein According to the multiple heat parameter sequences and the multiple electricity parameter sequences, perform predictive fuzzy analysis to obtain a heat consumption fuzzy coefficient sequence and an electricity consumption fuzzy coefficient sequence, including: Within the multiple heat parameter sequences and the multiple electricity parameter sequences, obtain multiple first heat parameters and multiple first electricity parameters at the first time node within a future preset time window; Calculate the differences between the maximum value and the minimum value within the multiple first heat parameters and the multiple first electricity parameters respectively to obtain a first heat parameter difference and a first electricity parameter difference, and calculate the means respectively to obtain a first average heat parameter and a first average electricity parameter; Calculate the ratios of the first heat parameter difference and the first electricity parameter difference to the first average heat parameter and the first average electricity parameter respectively to obtain a first heat consumption fuzzy coefficient and a first electricity consumption fuzzy coefficient; Continue to calculate the heat consumption fuzzy coefficients and the electricity consumption fuzzy coefficients at multiple time nodes within the future preset time window to obtain a heat consumption fuzzy coefficient sequence and an electricity consumption fuzzy coefficient sequence.

5. The energy-saving operation control optimization method for a thermal power plant according to claim 1, characterized in that Obtain the current real-time operating parameters of the thermal power plant, and randomly configure the adjusted operating parameters for control adjustment, and perform integrated control adjustment prediction to obtain a predicted heat supply parameter sequence and a predicted power supply parameter sequence, including: Obtain the current real-time operating parameters of the thermal power plant, and obtain the operating parameter range of the thermal power plant, and randomly configure the adjusted operating parameters for control adjustment within the operating parameter range; Train a heat and power supply predictor, where multiple heat and power supply prediction paths are included in the heat and power supply predictor; Input the real-time operating parameters and the adjusted operating parameters into the heat and power supply predictor, and predict and output multiple heat supply parameter sequences and multiple power supply parameter sequences at multiple time nodes within a future preset time window; According to the multiple heat supply parameter sequences and the multiple power supply parameter sequences, perform predictive fuzzy analysis to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence; According to the multiple heat supply parameter sequences and the multiple power supply parameter sequences, calculate the average heat supply parameter and the average power supply parameter at each time node, and arrange them in time sequence to obtain a predicted heat supply parameter sequence and a predicted power supply parameter sequence.

6. The energy-saving operation control optimization method for a thermal power plant according to claim 5, characterized in that, Train a heat and power supply predictor, including: According to the operation data records of the thermal power plant, collect a sample real-time operating parameter set and a sample adjusted operating parameter set, and obtain the heat supply parameters and the power supply parameters at multiple time nodes within a future preset time window after being controlled according to different sample real-time operating parameters and sample adjusted operating parameters, to obtain a sample heat supply parameter sequence set and a sample power supply parameter sequence set; Perform multiple random partitions on the sample real-time operating parameter set, the sample adjusted operating parameter set, the sample heat supply parameter sequence set and the sample power supply parameter sequence set to obtain multiple pieces of heat and power supply prediction training data; Respectively adopt the multiple heat and power supply prediction training data, and based on integrated machine learning, train to obtain multiple heat and power supply prediction paths. Integrate the multiple heat and power supply prediction paths to obtain a heat and power supply predictor.

7. The energy-saving operation control optimization method for a thermal power plant according to claim 1, characterized in that According to the predicted heat supply parameter sequence, predicted power supply parameter sequence, predicted heat consumption parameter sequence, and predicted power consumption parameter sequence, as well as the heat consumption fuzzy coefficient sequence, power consumption fuzzy coefficient sequence, heat supply fuzzy coefficient sequence, and power supply fuzzy coefficient sequence, perform operation energy-saving evaluation calculations to obtain the thermoelectric energy-saving fitness, including: According to the heat consumption fuzzy coefficient sequence and the heat supply fuzzy coefficient sequence, perform summation calculation to obtain a heat fuzzy coefficient sequence, calculate the ratio of each heat fuzzy coefficient to the total sum of the heat fuzzy coefficient sequence as the heat weight, and allocate to obtain a heat weight sequence. According to the power consumption fuzzy coefficient sequence and the power supply fuzzy coefficient sequence, perform summation calculation to obtain an electricity fuzzy coefficient sequence, calculate the ratio of each electricity fuzzy coefficient to the total sum of the electricity fuzzy coefficient sequence as the electricity weight, and allocate to obtain an electricity weight sequence. Construct an operation energy-saving evaluation function as follows: Among them, TPPES fit is the thermoelectric energy-saving fitness, N is the number of multiple time nodes within a preset future time window, is the heat weight of the i-th time node, is the predicted heating parameter of the i-th time node, is the predicted heat consumption parameter of the i-th time node, is the electricity weight of the i-th time node, is the predicted power supply parameter of the i-th time node, is the predicted electricity consumption parameter of the i-th time node; According to the predicted heat supply parameter sequence, predicted power supply parameter sequence, predicted heat consumption parameter sequence, and predicted power consumption parameter sequence, based on the operation energy-saving evaluation function, calculate to obtain the thermoelectric energy-saving fitness.

8. The energy-saving operation control optimization method for a thermal power plant according to claim 1, characterized in that According to the thermoelectric energy-saving fitness, perform optimization of adjusting operation parameters to obtain the optimal operation parameters, and control the thermal power plant, including: Continue to randomly configure new adjusting operation parameters and process to calculate the new thermoelectric energy-saving fitness. Perform iterative optimization of adjusting operation parameters until convergence, output the adjusting operation parameters with the maximum thermoelectric energy-saving fitness to obtain the optimal operation parameters, and perform energy-saving operation control on the thermal power plant.

9. An energy-saving operation control optimization system for a thermal power plant, characterized in that, Steps for implementing the energy-saving operation control optimization method for a thermal power plant according to any one of claims 1 to 8, including: A heat consumption and power consumption integrated prediction module, used to obtain the heat consumption demand parameters and power consumption demand parameters of the heat consumption main body and power consumption main body for cogeneration heat and power supply of the thermal power plant within a past preset time range, perform future heat consumption and power consumption integrated prediction to obtain a predicted heat consumption parameter sequence and a predicted power consumption parameter sequence, and perform prediction fuzzy analysis to obtain a heat consumption fuzzy coefficient sequence and a power consumption fuzzy coefficient sequence. An integrated control adjustment prediction module, used to obtain the current real-time operation parameters of the thermal power plant, randomly configure the adjusting operation parameters for control adjustment, perform integrated control adjustment prediction to obtain a predicted heat supply parameter sequence and a predicted power supply parameter sequence, and perform prediction fuzzy analysis to obtain a heat supply fuzzy coefficient sequence and a power supply fuzzy coefficient sequence. An operation energy-saving evaluation module, used to perform operation energy-saving evaluation calculations according to the predicted heat supply parameter sequence, predicted power supply parameter sequence, predicted heat consumption parameter sequence, and predicted power consumption parameter sequence, as well as the heat consumption fuzzy coefficient sequence, power consumption fuzzy coefficient sequence, heat supply fuzzy coefficient sequence, and power supply fuzzy coefficient sequence, to obtain the thermoelectric energy-saving fitness. An adjusting operation parameter optimization module, used to perform optimization of adjusting operation parameters according to the thermoelectric energy-saving fitness to obtain the optimal operation parameters, and perform energy-saving operation control on the thermal power plant.

Images