Intelligent optimization control method for supercritical thermal power unit

Abstract

The application provides a kind of supercritical thermal power unit intelligent optimization control method, it is related to thermal power generation technical field, comprising: obtaining the real-time operating parameters and historical operating parameters of each generator unit in power plant, and initial data are summarized and output;Combining with the selected instruction, a first model is selected in the model database, and the first model is trained and optimized using the initial data, and a preset analysis model is output;Based on the preset analysis model, and combining the preset evaluation index, the initial data is optimized and analyzed, and the optimization analysis result is output;The optimization analysis result is extracted, and the control strategy and method matched with the extracted features are selected in the strategy database by combining the preset feature strategy look-up table, and the operating parameters of each generator unit are adjusted based on the control strategy and control method.The application can reduce the intervention of operating personnel on the key system on the basis of ensuring the safe operation of the unit and equipment, while tapping the potential of the unit, improving the economy of the unit.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of thermal power generation, in particular to an intelligent optimization control method for supercritical thermal power generating units. BACKGROUND

[0002] The production process of supercritical thermal power generating units is complex. With the continuous long-time operation of the units, the equipment will gradually age, and the characteristics will change greatly, causing the performance of the control systems of important parameters such as main steam temperature, steam pressure, and environmental protection to decline, forcing the operation personnel to frequently intervene manually, greatly increasing the operation amount and monitoring pressure. Moreover, the operation personnel intervene the control system, mostly from the perspective of safe operation of the units, often ignoring the economy of the units, so that the units in operation for a long time generally have the problem of keeping the set values of key systems such as main steam temperature, steam pressure, and denitration at a low level to prevent the parameters from exceeding the standard. This phenomenon will cause the coal consumption of the units to increase, the ammonia escape and the amount of urea used to increase, and even cause the AGC and primary frequency modulation performance of the units to decline, resulting in the decline of the economy of the units.

[0003] At present, most of the control methods or systems of supercritical units are built by standard modules such as PID, and the parameters of which cannot be adjusted adaptively with the change of the characteristics of the units. After the long-time operation of the units, the control performance will deteriorate.

[0004] Therefore, the present application provides an intelligent optimization control method for supercritical thermal power generating units. SUMMARY

[0005] The present application provides an intelligent optimization control method for supercritical thermal power generating units, which can reduce the intervention of the operation personnel on the key systems and reduce the monitoring pressure on the basis of ensuring the safe operation of the units and equipment, and at the same time, can tap the potential of the units and improve the economy of the units.

[0006] The present application provides an intelligent optimization control method for supercritical thermal power generating units, which includes:

[0007] Step 1: Obtain the real-time operation parameters and historical operation parameters of each generating unit in the power plant, and output the initial data after summarizing;

[0008] Step 2: Select a first model in the model database in combination with a selection instruction, and train and optimize the first model using the initial data to output a preset analysis model;

[0009] Step 3: Based on the preset analysis model, and in combination with a preset evaluation index, the initial data is optimized and analyzed to output an optimization analysis result;

[0010] Step 4: feature extraction is performed on the optimization analysis result, and a control strategy and method matched with the extracted features are selected from a strategy database by combining a preset feature strategy reference table, and operation parameters of each generator set are adjusted based on the control strategy and control method.

[0011] Preferably, the real-time operation parameters of each generator set in the power plant are obtained, including:

[0012] The monitoring requirement information of the user is obtained, and an appropriate preset monitoring device is selected from a device database based on the monitoring requirement information.

[0013] Real-time operation parameter data of each generator set is obtained based on the preset monitoring device, and real-time operation parameter data is output in combination with a time stamp corresponding to the operation parameter data.

[0014] Preferably, in step 1, it further includes:

[0015] Feature extraction is performed on the real-time operation parameter data, a first feature set is constructed, and corresponding historical operation parameters are matched from a historical database based on the first feature set, and historical operation parameter data is output.

[0016] Preferably, before the initial data is output, it includes:

[0017] A matched preprocessing method is selected from a method database based on the monitoring requirement information and the first feature set.

[0018] The real-time operation parameter data and the historical operation parameter data are preprocessed based on the preprocessing method.

[0019] Preferably, the first model is selected from the model database in combination with the selection instruction, including:

[0020] A model selection instruction is output in combination with a priority corresponding to each instruction by obtaining a model manual selection instruction input by the user and a model automatic selection instruction generated by the system.

[0021] A matched first model is selected from the model database based on the model selection instruction.

[0022] Preferably, the first model is trained and optimized by using the initial data, including:

[0023] The initial data is divided into a training data set, a verification data set and a test data set by using a preset data division method.

[0024] The first model is trained and optimized based on the training data set, the verification data set and the test data set.

[0025] Preferably, in step 3, comprising:

[0026] Feature extraction is performed on the monitoring demand information, and a second feature set is constructed based on the extracted features;

[0027] Based on the second feature set, and in combination with a preset feature-factor reference table, the corresponding index screening factors are obtained, and a screening factor set is constructed;

[0028] In combination with a preset factor-index mapping table, the first index matching each index screening factor in the screening factor set is selected in the index database, and an index candidate pool is constructed;

[0029] The real-time acquisition system generates index self-selection instructions using a preset algorithm, simultaneously captures artificial index selection instructions input by humans in real time through a preset port, and outputs an index selection instruction set in combination with the priority information corresponding to each instruction;

[0030] Based on the index selection instruction set, the matching preset evaluation index is selected in the index candidate pool, and an evaluation index set is constructed;

[0031] The evaluation index set is input into a preset analysis model, and the preset analysis model is configured, adjusted and updated in combination with a preset adaptive algorithm, and the model format requirement data after configuration and update is output;

[0032] Meanwhile, the corresponding data format information of all data contained in the initial data is obtained, and an initial data format reference table is output;

[0033] A preset format conversion method matching the model format requirement information and the initial data format reference table is selected in the method database;

[0034] Based on the model format requirement data and the initial data format reference table, the initial data is format-converted using the preset format conversion method, and the to-be-analyzed data is output;

[0035] In combination with the evaluation index set, the to-be-analyzed data is optimized and analyzed by the preset analysis model after configuration and update, and the optimization analysis result is output.

[0036] Preferably, in step 4, comprising:

[0037] A preset running threshold condition is input into the preset analysis model, and a preset boundary condition is generated;

[0038] The self-generated optimization index output by the preset analysis model under the preset boundary condition and the artificial optimization index selected by humans are obtained, and an optimization index set is output based on the self-generated optimization index and the artificial optimization index.

[0039] The initial data is input into the preset analysis model for predictive analysis to obtain the first prediction result under the variable structure predictive control method and the second prediction result under the generalized predictive control method.

[0040] Based on the optimization index set, the first prediction result and the second prediction result, and combined with the preset strategy-method formulation process design, an intelligent control strategy and an intelligent control method are formulated, and an index-strategy-method comparison table is constructed based on the mapping relationship between the optimization index set, the intelligent control strategy and the intelligent control method.

[0041] Feature extraction is performed on the optimization analysis results, and an optimization feature set is constructed based on the extracted features;

[0042] By combining the preset feature-index mapping table, the optimization index that matches each feature in the optimization feature set is obtained, and the set of optimization indexes to be executed is output.

[0043] Based on the indicator-strategy-method comparison table, obtain the intelligent control strategies and intelligent control methods that match each optimization indicator in the set of optimization indicators to be executed, and output the list of strategies-methods to be executed;

[0044] Based on the intelligent control strategies and intelligent control methods in the list of strategies and methods to be executed, the corresponding operating parameters in each generator set are controlled and adjusted, and process data is generated in real time and fed back to the preset data analysis model.

[0045] This invention provides an intelligent optimization control method for supercritical thermal power units. By acquiring real-time and historical operating parameters, combining model training and optimization, and employing feature extraction and control strategy selection, this invention achieves intelligent optimization control of the generator unit's operating parameters. Based on real-time and historical data, this invention can perform intelligent optimization analysis to select the optimal control strategy and method, thereby improving the efficiency of the power plant's generator units, reducing emissions, optimizing system operation, and ultimately achieving the goals of energy conservation and emission reduction, thus enhancing the economic efficiency and reliability of power plant operation. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0047] Figure 1 This is a flowchart illustrating an intelligent optimization control method for supercritical thermal power units provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0049] As shown in Figure 1 The supercritical thermal power unit intelligent optimization control method provided by the embodiment of the present application comprises:

[0050] Step 1: Obtain real-time operation parameters and historical operation parameters of each generator unit in the power plant, and output initial data by summarizing;

[0051] Step 2: Select a first model in the model database in combination with a selection instruction, and train and optimize the first model by using the initial data to output a preset analysis model;

[0052] Step 3: Based on the preset analysis model, and in combination with a preset evaluation index, the initial data is analyzed and optimized to output an optimization analysis result;

[0053] Step 4: Feature extraction is performed on the optimization analysis result, and a control strategy and method matched with the extracted features are selected in the strategy database in combination with a preset feature strategy table, and the operation parameters of each generator unit are adjusted based on the control strategy and the control method.

[0054] In this embodiment, the real-time operation parameters are the operation parameters of each generator unit in the power plant at the current time, such as temperature, pressure, power, etc.

[0055] In this embodiment, the historical operation parameters are operation parameter data of each generator unit in the power plant in the past period of time, which are used for analysis and comparison, for example, load curve, temperature change, etc. of the generator unit in the past week.

[0056] In this embodiment, the initial data is data integrated by the obtained real-time operation parameters and historical operation parameters, which is used as a starting point for optimization analysis, for example, including real-time load, historical fault records, generator unit model, etc.

[0057] In this embodiment, the selection instruction is an instruction or rule for guiding the selection of a suitable model in the model database.

[0058] In this embodiment, the model database is a database for storing various models, which is used to select a suitable model according to the selection instruction.

[0059] In this embodiment, the first model: the initial data model selected from the model database according to the selection instruction, is used for training and optimization;

[0060] In this embodiment, the preset analysis model: the model trained and optimized, is used for analyzing and predicting the initial data, for example, the neural network model trained and optimized by data;

[0061] In this embodiment, the preset evaluation index: the index used to evaluate the optimization analysis result, helps to judge the performance of the system;

[0062] In this embodiment, the optimization analysis result: the result obtained after optimization by the preset analysis model and evaluation index, guides the subsequent control decision;

[0063] In this embodiment, the preset feature strategy correspondence table: the table containing the correspondence between features and strategies, is pre-set;

[0064] In this embodiment, the strategy database: the database storing various control strategies and methods;

[0065] In this embodiment, the control strategy and method: the control strategy and method selected according to the feature extraction result, is used to adjust the operating parameters of each generator set.

[0066] The implementation principle and beneficial effects of the embodiment are as follows: the present application realizes intelligent optimization control of the operating parameters of the generator set by acquiring real-time and historical operating parameters, combining model training and optimization, and feature extraction and control strategy selection. The present application can intelligently optimize and analyze the real-time and historical data, select the optimal control strategy and method, thereby improving the efficiency of the power plant generator set, reducing emissions, optimizing system operation, and achieving the goal of energy saving and emission reduction, and improving the economic efficiency and reliability of the power plant operation.

[0067] The embodiment of the present application provides an intelligent optimization control method for a supercritical thermal power generator set, which acquires real-time operating parameters of each generator set in a power plant, including:

[0068] Obtain monitoring demand information of the user, and select a preset monitoring device that is adapted based on the monitoring demand information in the equipment database;

[0069] Real-time operating parameter data of each generator set is acquired based on the preset monitoring device, and the real-time operating parameter data is output in combination with the time stamp corresponding to the operating parameter data.

[0070] In this embodiment, the monitoring demand information: the demand information of the user for the parameters and frequency of the monitoring device, for example, the user needs to acquire the load, temperature and vibration data of the generator set every minute;

[0071] In this embodiment, the device database: a database storing various monitoring device information, used for selecting suitable monitoring devices, for example, containing technical parameters and model information of various sensors and monitoring instruments;

[0072] In this embodiment, the preset monitoring device: a suitable monitoring device selected from the device database according to the monitoring requirement information, for example, a temperature sensor and a vibration monitor are selected as the preset monitoring device;

[0073] In this embodiment, the running parameter data: real-time running parameter data of each generator set, such as load, temperature, vibration, etc., for example, the load of generator set A is 100 MW, and the temperature is 300℃;

[0074] In this embodiment, the timestamp: used to mark the time point of data collection, to ensure the time sequence and accuracy of the data, for example, each data point will be attached with a timestamp indicating the time of data collection;

[0075] In this embodiment, the real-time running parameter data: real-time running parameter data of each generator set obtained based on the preset monitoring device, combined with the timestamp output real-time running parameter data, for example, the load, temperature and vibration data of generator set A are obtained every minute, with the corresponding timestamp information.

[0076] The implementation principle and beneficial effects of this embodiment: the present application selects suitable monitoring devices according to the user's monitoring requirement information, obtains real-time running parameter data of each generator set, and outputs real-time running parameter data combined with timestamp information, to realize real-time monitoring of the running state of the generator set in the power plant. By real-time acquisition and monitoring of the running parameter data of the generator set, the present application can timely discover abnormal conditions, give early warning of possible faults, help optimize control strategies, improve power plant operation efficiency, reduce downtime, improve equipment reliability and safety, and also help save energy, reduce costs, and realize intelligent and optimized operation of the power plant.

[0077] The supercritical thermal power unit intelligent optimization control method provided in the embodiment of the present application, step 1 further comprises:

[0078] The real-time running parameter data is feature extracted to construct a first feature set, and the corresponding historical running parameters are matched in the historical database based on the first feature set, and the historical running parameter data is output.

[0079] In this embodiment, the first feature set: a feature set extracted from the real-time running parameter data, used to describe the running state of the generator set, for example, including load size, temperature change rate, vibration frequency, etc.;

[0080] In this embodiment, the historical database: a database that stores historical operation parameter data, used to save the operation data of the generator set in the past period of time, for example, containing the load, temperature, vibration data of the generator set in the past week;

[0081] In this embodiment, the historical operation parameter data: the operation parameter data of the generator set recorded in the past period of time, used for comparison and analysis of the operation in different time periods, for example, the load of the generator set A last week is 120MW, and the temperature is 280℃.

[0082] The implementation principle and beneficial effects of this embodiment: the present application extracts features from real-time operation parameter data, constructs a first feature set, and then matches corresponding historical operation parameter data in the historical database to obtain past operation data for analysis and comparison. Through the analysis of historical operation parameter data, the present application can find the change trend, periodicity and other information of the generator set operation, help to predict possible problems in the future, optimize the control strategy, and improve the operation efficiency and reliability of the power plant. At the same time, by combining real-time and historical data for comprehensive analysis, it is helpful to develop more accurate control scheme, reduce fault risk, and improve the overall performance and economic benefit of the power plant.

[0083] The supercritical thermal power unit intelligent optimization control method provided by the embodiment of the present application comprises the following steps before the initial data is aggregated and output:

[0084] Based on the monitoring demand information and the first feature set, a matched preprocessing method is selected in the method database;

[0085] The real-time operation parameter data and the historical operation parameter data are preprocessed based on the preprocessing method.

[0086] In this embodiment, the method database: a database that stores various preprocessing methods, used to select the appropriate preprocessing method according to the monitoring demand information and the feature set, for example, containing data cleaning, data smoothing, data normalization and other preprocessing methods;

[0087] In this embodiment, the preprocessing method: a method for preprocessing data, aiming to extract useful information, reduce noise or adjust data distribution, for example, can include data smoothing, data dimensionality reduction, outlier processing and other methods;

[0088] In this embodiment, preprocessing: a process of cleaning, converting or adjusting data for subsequent analysis and processing, for example, removing outliers, normalizing real-time operation parameter data, and smoothing historical operation parameter data.

[0089] Principles and benefits of the embodiment: The present application selects appropriate preprocessing methods in the method database according to monitoring demand information and the first feature set, and then uses these methods to preprocess real-time operation parameter data and historical operation parameter data to ensure data quality and applicability. By preprocessing the data with the preprocessing method, the present application can reduce noise and outliers in the data, making the data more accurate and reliable. This helps to improve the accuracy of subsequent data analysis and modeling, optimize the development of control strategies, and further improve the operation efficiency and stability of the power plant. Through preprocessing, data is easier to understand and apply, providing stronger support for subsequent data processing and decision-making.

[0090] The embodiment of the present application provides an intelligent optimization control method for a supercritical thermal power unit, which selects a first model in a model database in combination with a selection instruction, including:

[0091] The model manual selection instruction input by the user and the model self-selection instruction generated by the system are obtained, and a model selection instruction is output in combination with the priority of each instruction;

[0092] Based on the model selection instruction, a matched first model is selected in the model database.

[0093] In the embodiment, the model manual selection instruction: an instruction manually input by the user, used to guide the system to select a model suitable for the current situation, for example, the user may specify to select a neural network-based model for predicting the load change of the generator unit;

[0094] In the embodiment, the model self-selection instruction: an instruction automatically generated by the system according to certain rules or algorithms, used to select a model according to data characteristics or requirements, for example, the system may automatically select an ARIMA model for load prediction according to the time series characteristics of the data;

[0095] In the embodiment, the model selection instruction: in combination with the model manual selection instruction and the model self-selection instruction, the final model selection instruction is output according to the priority of each instruction, for example, if the priority of the model manual selection instruction specified by the user is higher than that of the system self-selection instruction, the system will give priority to the model specified by the user.

[0096] Principles and benefits of the embodiment: The present application combines the model selection instructions input by the user and the model self-selection instructions generated by the system, and outputs the final model selection instruction based on the priority of each instruction. Then, the first model is selected in the model database based on this instruction. The present application can fully consider the user's needs and the system's intelligent selection by combining the artificial selection instruction and the system self-selection instruction, thereby ensuring that the most suitable model for the current situation is selected. This can improve the accuracy and applicability of the model, further optimize the control strategy, and improve the operation efficiency and performance of the power plant. At the same time, this combination of artificial and automatic selection method can improve the flexibility and intelligence of the system, and bring more convenience and benefits to the power plant management and operation.

[0097] The supercritical thermal power unit intelligent optimization control method provided by the embodiment of the present application trains and optimizes the first model using initial data, including:

[0098] The initial data is divided into a training data set, a validation data set, and a test data set using a preset data division method;

[0099] The first model is trained and optimized based on the training data set, the validation data set, and the test data set.

[0100] In this embodiment, the preset data division method refers to a data set division method set in advance, which is used to divide the initial data into a training data set, a validation data set, and a test data set. This helps to evaluate the performance of the model on different data sets, for example, a common 70%-15%-15% division ratio can be used to divide the data into a training set, a validation set, and a test set;

[0101] In this embodiment, the training data set is a data set used to train the model, and the model learns parameters through these data, for example, it contains historical operating parameter data and corresponding output results;

[0102] In this embodiment, the validation data set is a data set used to adjust the model hyperparameters and evaluate the model performance. Through the validation data set, the model can avoid overfitting on the training data, for example, it is used to adjust the learning rate of the neural network or select the best model complexity;

[0103] In this embodiment, the test data set is a data set used to evaluate the performance of the model. The model is tested on this data set to evaluate its generalization ability, for example, future real-time operating parameter data is simulated to verify the performance of the model in actual operation.

[0104] Principle and beneficial effects of the embodiment: The present application divides the initial data set into training data set, validation data set and test data set according to the preset data division method, and then trains and optimizes the first model based on these data sets. Training data is used for model parameter learning, validation data is used for adjusting model hyperparameters, and test data is used for final evaluation of model performance. The present application can improve the generalization ability and stability of the model by reasonably dividing the data set and training and optimizing the model using them. The training data set is used for model learning, the validation data set is used for adjusting the model to avoid overfitting, and the test data set is used for final evaluation of model performance. In this way, the model can ensure good prediction ability and generalization ability in actual application, further improving the operation efficiency and control precision of the power plant.

[0105] The supercritical thermal power unit intelligent optimization control method provided by the embodiment of the present application comprises the following steps:

[0106] Feature extraction is performed on the monitoring demand information, and a second feature set is constructed based on the extracted features;

[0107] Based on the second feature set, and in combination with a preset feature-factor reference table, corresponding index screening factors are obtained, and a screening factor set is constructed;

[0108] In combination with a preset factor-index mapping table, first indexes matching the index screening factors in the screening factor set are selected in the index database, and an index candidate pool is constructed;

[0109] The index self-selection instruction generated by the preset algorithm is obtained in real time, at the same time, the artificial index selection instruction input by the artificial input is captured in real time through the preset port, and the index selection instruction set is output in combination with the priority information corresponding to each instruction;

[0110] Based on the index selection instruction set, a preset evaluation index matching the index selection instruction set is selected in the index candidate pool, and an evaluation index set is constructed;

[0111] The evaluation index set is input into a preset analysis model, and the preset analysis model is configured and updated in combination with a preset adaptive algorithm, and at the same time, the model format requirement data after configuration and update is output;

[0112] At the same time, the corresponding data format information of all data contained in the initial data is obtained, and an initial data format reference table is output;

[0113] A preset format conversion method matching the model format requirement information and the initial data format reference table is selected in the method database;

[0114] Based on the model format requirement data and the initial data format reference table, the initial data is format-converted by using the preset format conversion method, and the to-be-analyzed data is output;

[0115] In combination with the evaluation index set, the updated preset analysis model is configured to perform optimized analysis on the to-be-analyzed data, and output an optimized analysis result.

[0116] In this embodiment, the second feature set: a feature set obtained by feature extraction on the monitoring demand information, is used as a basis for constructing the evaluation index and the optimized analysis, for example, the slope of the load curve and the frequency change are extracted from the real-time operation data of the thermal power generating unit;

[0117] In this embodiment, the preset feature-factor correspondence table: a correspondence table containing the mapping relationship between each feature in the second feature set and the index screening factor, is preset;

[0118] In this embodiment, the index screening factor: a feature factor obtained according to the preset feature-factor correspondence table, is used for subsequent index selection and construction of the evaluation index set;

[0119] In this embodiment, the screening factor set: a set composed of the index screening factor;

[0120] In this embodiment, the preset factor-index mapping table: a table containing the mapping relationship between the index screening factor and the evaluation index, is preset;

[0121] In this embodiment, the index database: a database containing various indexes in various power generation systems;

[0122] In this embodiment, the first index: an evaluation index selected from the index database;

[0123] In this embodiment, the index candidate pool: a set containing candidate evaluation indexes matched according to the factors in the screening factor set, for example, the indexes of “load balance degree” and “frequency stability”;

[0124] In this embodiment, the preset algorithm: a preset algorithm, is used to generate the index self-selection instruction;

[0125] In this embodiment, the index self-selection instruction: an instruction generated by the preset algorithm, is used to automatically select the evaluation index, for example, an instruction generated by the system according to the real-time data features to automatically generate the evaluation index;

[0126] In this embodiment, the preset port: a previously set interface or channel, is used for data transmission and interaction between the system and the user, for example, a port in the system is specially set to receive user input instructions;

[0127] In this embodiment, the artificial index selection instruction: an instruction input by a human, is used to guide the system to select specific indexes or parameters, for example, the user may specify that the system should give priority to some important operation indexes;

[0128] In this embodiment, priority information refers to information indicating the importance or priority order of different instructions or operations, for example, high-priority instructions will be executed first.

[0129] In this embodiment, index selection instruction set refers to a set of instructions generated by the system, including manual index selection instructions and self-selection instructions, for selecting specific indicators or factors, for example, a set of instructions generated by the system according to user input and algorithm.

[0130] In this embodiment, evaluation index set refers to a set of indicators for evaluating system performance or results, for example, including indicators such as power plant efficiency and load balancing.

[0131] In this embodiment, preset adaptive algorithm refers to an algorithm set in advance, which is used to automatically adjust parameters or configurations according to system status or demand, for example, the system may automatically adjust the parameters of the control algorithm according to real-time data.

[0132] In this embodiment, configuration adjustment and update refers to the operation of adjusting and updating the system configuration according to the evaluation results or demand, for example, adjusting the control strategy according to the evaluation index results.

[0133] In this embodiment, model format requirement data refers to data in a specific format required by the model, for example, a neural network model may require input data with specific dimensions and ranges.

[0134] In this embodiment, data format information refers to information describing the structure and type of data, for example, whether the data is time series data or categorical data.

[0135] In this embodiment, initial data format correspondence table refers to a table recording the correspondence between the format information of the initial data and the format information required by the model.

[0136] In this embodiment, preset format conversion method refers to a method set in advance, which is used to convert the initial data into the format required by the model, for example, converting time series data into matrix format.

[0137] In this embodiment, data to be analyzed refers to data prepared for analysis after format conversion, for example, data processed by the preset format conversion method.

[0138] The implementation principle and beneficial effects of the embodiment are as follows: the application constructs an evaluation index set through steps such as feature extraction, factor selection, and index matching, and configures and updates the preset analysis model and the adaptive algorithm, so as to perform optimized analysis on the data to be analyzed. The application can better meet the user demand and automatically adapt to the system state by combining manual and automatic instruction selection. Meanwhile, the accuracy of the analysis result and the intelligent degree of the system can be improved through format conversion and data optimized analysis, the operation efficiency and control precision of the power plant are further optimized, the system can more effectively cope with complex control requirements and data processing, and the overall performance and efficiency of the system are improved.

[0139] The supercritical thermal power unit intelligent optimization control method provided by the embodiment of the application comprises the following steps:

[0140] The preset operation threshold condition is input into the preset analysis model to generate a preset boundary condition;

[0141] The self-generated optimization index output by the preset analysis model under the preset boundary condition and the artificial optimization index obtained by manual selection are acquired, and an optimization index set is output based on the self-generated optimization index and the artificial optimization index;

[0142] The initial data is input into the preset analysis model for prediction analysis, and a first prediction result under a variable structure prediction control method and a second prediction result under a generalized prediction control method are obtained;

[0143] Based on the optimization index set, the first prediction result, and the second prediction result, and in combination with a preset strategy-method development process, an intelligent control strategy and an intelligent control method are developed, and an index-strategy-method correspondence table is constructed based on the mapping relationship between the optimization index set, the intelligent control strategy, and the intelligent control method;

[0144] Features are extracted from the optimized analysis result, and an optimized feature set is constructed based on the extracted features;

[0145] In combination with a preset feature-index mapping table, the optimization index matched with each feature in the optimized feature set is acquired, and a to-be-executed optimization index set is output;

[0146] In combination with the index-strategy-method correspondence table, the intelligent control strategy and the intelligent control method matched with each optimization index in the to-be-executed optimization index set are acquired, and a to-be-executed strategy-method list is output;

[0147] Based on the intelligent control strategy and the intelligent control method in the to-be-executed strategy-method list, the corresponding operation parameters in each generator unit are controlled and adjusted, and process data is generated in real time and fed back to the preset data analysis model.

[0148] In this embodiment, the preset operating threshold condition is a condition set in advance to guide the system's operating threshold in a specific situation, for example, setting the generator load rate not to exceed 90%;

[0149] In this embodiment, the preset boundary condition is a condition generated according to the preset operating threshold condition to define the boundary of the system's operation, for example, a boundary condition set according to the load rate;

[0150] In this embodiment, the self-generated optimization index is an automatically generated index for optimization output by the preset analysis model, for example, an efficiency index generated by the system according to data;

[0151] In this embodiment, the artificial optimization index is an index selected by humans for optimization, for example, a key performance indicator selected by an operator based on experience;

[0152] In this embodiment, the optimization index set includes both self-generated optimization indexes and artificial optimization indexes, for example, indexes such as efficiency and load balancing;

[0153] In this embodiment, the variable structure predictive control method is a model-based control method that uses the system model to predict the system's behavior for a future period of time within each control cycle and optimizes the control input to achieve optimization of system performance, for example, dynamically adjusting the control strategy according to real-time data;

[0154] In this embodiment, the first prediction result is the first prediction result obtained under the variable structure predictive control method, for example, predicting the load change in the next hour;

[0155] In this embodiment, the generalized predictive control method is a prediction-based control method that uses the system model to predict the system's response for a future period of time and adjusts the control input based on these prediction results to achieve optimal control of the system, for example, using a recursive algorithm for real-time calculation and prediction;

[0156] In this embodiment, the second prediction result is the second prediction result obtained under the generalized predictive control method, for example, predicting the power grid load demand for the next day;

[0157] In this embodiment, the preset strategy-method formulation process specifies the process of formulating intelligent control strategies and methods, for example, including steps such as determining optimization objectives, selecting control strategies, and formulating control methods;

[0158] In this embodiment, the intelligent control strategy and the intelligent control method are strategies and methods used to optimize the performance of the control system, with the strategy specifying the overall direction of control and the method specifying specific control means and algorithms;

[0159] In this embodiment, the index-strategy-method mapping table: a table that records the mapping relationship between the optimization index, the intelligent control strategy, and the intelligent control method, for example, specifying the use of a specific control strategy and control method under certain optimization indexes;

[0160] In this embodiment, the optimization feature set: a set of features extracted from the optimization analysis results, used to guide the selection of optimization indexes, which can reflect the performance, state, or other important information of the system;

[0161] In this embodiment, the preset feature-index mapping table: a table containing the mapping relationship between each feature in the optimization feature set and the optimization index, used to select the matching optimization index according to the optimization feature set;

[0162] In this embodiment, the set of optimization indexes to be executed: a set of optimization indexes determined according to the optimization feature set and the mapping table, which will be used to determine the intelligent control strategy and the intelligent control method;

[0163] In this embodiment, the list of strategies and methods to be executed: a list of intelligent control strategies and methods to be executed determined according to the index-strategy-method mapping table, which will be applied to the system for real-time control;

[0164] In this embodiment, the process data: real-time data generated during the control process, used to feed back to the preset data analysis model for further analysis and adjustment.

[0165] The implementation principle and beneficial effects of this embodiment: The present application generates a preset boundary condition according to a preset operating threshold condition, and then generates a set of optimization indexes according to automatically generated and manually selected optimization indexes. Then, the results obtained by prediction analysis and the set of optimization indexes are used to develop intelligent control strategies and methods, and an index-strategy-method mapping table is constructed. Finally, the features are extracted from the optimization analysis results to generate an optimization feature set, select the corresponding optimization index, and execute the intelligent control strategy and method. By combining automatically generated and manually selected indexes, the system can more comprehensively evaluate the system performance and develop intelligent control strategies and methods based on the prediction results. This method can help the system more accurately adjust the operating parameters, improve the stability and efficiency of the system, and further optimize the operation and control effect of the power plant. At the same time, by establishing the index-strategy-method mapping table, the system can be made more intelligent and adaptive, improving the overall performance and resilience of the system.

[0166] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; and although the present application has been described in detail with reference to the foregoing embodiments, it should be appreciated by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A supercritical thermal power unit intelligent optimization control method, characterized in that, The method comprises the following steps: Step 1: obtaining real-time operation parameters and historical operation parameters of each generator unit in the power plant, and outputting initial data; The real-time operation parameters of each generator unit in the power plant are obtained, including: obtaining the monitoring demand information of the user, and selecting the appropriate preset monitoring device in the device database based on the monitoring demand information; based on the preset monitoring device, the operation parameter data of each generator unit is obtained in real time, and the real-time operation parameter data is output in combination with the corresponding time stamp of the operation parameter data; Step 2: selecting a first model in the model database in combination with the selection instruction, and training and optimizing the first model by using the initial data to output a preset analysis model; Step 3: based on the preset analysis model, and in combination with the preset evaluation index, the initial data is analyzed and optimized to output an optimization analysis result; The step 3 further comprises: The monitoring demand information is feature extracted, and a second feature set is constructed based on the extracted features; Based on the second feature set, and in combination with a preset feature-factor reference table, the corresponding index screening factor is obtained, and a screening factor set is constructed; In combination with a preset factor-index mapping table, a first index matching each index screening factor in the screening factor set is selected in the index database, and an index candidate pool is constructed; The real-time acquisition system uses a preset algorithm to generate an index self-selection instruction, at the same time, an artificial index selection instruction is input by a preset port in real time, and an index selection instruction set is output in combination with the priority information corresponding to each instruction; Based on the index selection instruction set, a matching preset evaluation index is selected in the index candidate pool, and an evaluation index set is constructed; The evaluation index set is input into a preset analysis model, and the preset analysis model is configured and updated in combination with a preset adaptive algorithm, and at the same time, the model format requirement data after configuration and update is output; At the same time, the corresponding data format information of all data contained in the initial data is obtained, and an initial data format reference table is output; A preset format conversion method matching the model format requirement information and the initial data format reference table is selected in the method database; Based on the model format requirement data and the initial data format reference table, and by using the preset format conversion method, the initial data is format converted, and the to-be-analyzed data is output; In combination with the evaluation index set, the to-be-analyzed data is analyzed and optimized by the preset analysis model after configuration and update, and an optimization analysis result is output; Step 4: feature extraction is performed on the optimization analysis result, and a control strategy and method matching the extracted features are selected in a strategy database in combination with a preset feature strategy reference table, and the operation parameters of each generator unit are adjusted based on the control strategy and control method.

2. The intelligent optimization control method for supercritical thermal power generating units according to claim 1, characterized in that, In step 1, it further comprises: The real-time operation parameter data is feature extracted, a first feature set is constructed, and corresponding historical operation parameters are matched in a historical database based on the first feature set, and historical operation parameter data is output.

3. The intelligent optimization control method for supercritical thermal power generating units according to claim 2, characterized in that, Before the initial data is output, it comprises: Based on the monitoring requirement information and the first feature set, a matched preprocessing method is selected from a method database; Based on the preprocessing method, the real-time operation parameter data and the historical operation parameter data are preprocessed.

4. The intelligent optimization control method for supercritical thermal power generating units according to claim 1, characterized in that, The combination of the selection instruction selects a first model from a model database, including: Obtaining the model artificial selection instruction input by the user and the model self-selection instruction generated by the system, and outputting the model selection instruction in combination with the priority corresponding to each instruction; Based on the model selection instruction, a matched first model is selected from a model database.

5. The intelligent optimization control method for supercritical thermal power generating units according to claim 1, characterized in that, The training and optimization of the first model using the initial data include: Using a preset data division method to divide the initial data into a training data set, a validation data set, and a test data set; Based on the training data set, the validation data set, and the test data set, the first model is trained and optimized.

6. The intelligent optimization control method for supercritical thermal power generating units according to claim 1, characterized in that, Step 4 includes: Input the preset operation threshold condition into the preset analysis model to generate a preset boundary condition; Obtain the self-generated optimization index output by the preset analysis model under the preset boundary condition and the artificial optimization index obtained by artificial selection, and output an optimization index set based on the self-generated optimization index and the artificial optimization index; Input the initial data into the preset analysis model for prediction analysis to obtain a first prediction result under a variable structure prediction control method and a second prediction result under a generalized prediction control method; Based on the optimization index set, the first prediction result, and the second prediction result, and in combination with a preset strategy-method development process, an intelligent control strategy and an intelligent control method are designed and developed, and an index-strategy-method correspondence table is constructed based on the mapping relationship between the optimization index set, the intelligent control strategy, and the intelligent control method; Feature extraction is performed on the optimization analysis result, and an optimization feature set is constructed based on the extracted features; In combination with a preset feature-index mapping table, the optimization index matching each feature in the optimization feature set is obtained, and a to-be-executed optimization index set is output; In combination with the index-strategy-method correspondence table, the intelligent control strategy and the intelligent control method matching each optimization index in the to-be-executed optimization index set are obtained, and a to-be-executed strategy-method list is output; Based on the intelligent control strategy and the intelligent control method in the to-be-executed strategy-method list, the corresponding operation parameters in each generator set are controlled and adjusted, and process data is generated in real time and fed back to the preset data analysis model. ​

Images