Gas turbine combustion optimization control system based on self-adaptive control

Abstract

The invention relates to the technical field of gas turbines, and discloses a gas turbine combustion optimization control system based on adaptive control, and the system comprises an obtaining module which is used for constructing an efficiency evaluation model and an emission evaluation model according to historical efficiency related data and historical emission related data; the evaluation module is used for inputting the real-time efficiency related data and the real-time emission related data into an efficiency evaluation model and an emission evaluation model to obtain an efficiency evaluation value and an emission evaluation value; the judgment module is used for judging whether optimization control is needed or not, and if yes, obtaining to-be-optimized control data and generating a first control optimization strategy; the optimization module is used for acquiring actual change characteristics after the control optimization instruction and generating a prediction efficiency evaluation value and a prediction emission evaluation value; and the correction module is used for judging whether the first control optimization strategy is corrected or not, if yes, a second control optimization strategy is obtained, it is guaranteed that pollutant emission reaches the standard, the combustion efficiency is improved, and the combustion optimization control effect of the gas turbine is improved.

Description

Technical Field

[0001] The present application relates to the technical field of gas turbines, and particularly to a combustion optimization control system for a gas turbine based on adaptive control. Background Art

[0002] With the continuous growth of global energy demand and the increasing awareness of environmental protection, gas turbines, as a kind of efficient and clean energy conversion equipment, have been widely used in the fields of electric power, industry, etc. However, in the combustion control process of existing gas turbines, problems such as low combustion efficiency or excessive emissions often exist. Therefore, there is an urgent need for a combustion optimization control method to optimize the combustion process of gas turbines, optimize the combustion process of gas turbines on the premise of meeting pollutant emission standards, and improve combustion efficiency. Summary of the Invention

[0003] To solve the above technical problems, the present application provides a combustion optimization control system for a gas turbine based on adaptive control. By determining historical efficiency-related data and historical emission-related data, an efficiency evaluation model and an emission evaluation model are constructed to obtain an efficiency evaluation value and an emission evaluation value, and a preset analysis model is combined to judge whether optimization control is needed. If so, a first control optimization strategy is obtained. According to the predicted efficiency evaluation value and the predicted emission evaluation value after the first control optimization strategy, it is judged whether to correct the first control optimization strategy, timely discover and adjust the control optimization strategy, ensure that pollutant emissions meet the standards and improve combustion efficiency, and improve the combustion optimization control effect of the gas turbine.

[0004] In some embodiments of the present application, a combustion optimization control system for a gas turbine based on adaptive control is provided, including: An acquisition module, configured to acquire and analyze the historical operation data of the gas turbine, determine a number of historical efficiency-related data and historical emission-related data of the gas turbine according to the analysis results, and respectively construct an efficiency evaluation model and an emission evaluation model; An evaluation module, configured to acquire real-time efficiency-related data and real-time emission-related data and input them into the corresponding efficiency evaluation model and emission evaluation model to obtain an efficiency evaluation value and an emission evaluation value; A judgment module, configured to judge whether optimization control is needed according to a preset optimization analysis model, the efficiency evaluation value, and the emission evaluation value. If so, obtain the data to be optimized and generate a first control optimization strategy; An optimization module, configured to generate a control optimization instruction according to the first control optimization strategy, obtain the actual change characteristics of the real-time efficiency-related data and the real-time emission-related data after the control optimization instruction, and generate a predicted efficiency evaluation value and a predicted emission evaluation value for a future period; A correction module, configured to determine whether to correct the first control optimization strategy according to the predicted efficiency evaluation value and the predicted emission evaluation value. If so, obtain a second control optimization strategy and generate a control adjustment instruction.

[0005] In some embodiments of the present application, a number of historical efficiency-related data and historical emission-related data of the gas turbine are determined according to the analysis result, including: Preset characteristic efficiency data and characteristic emission data; Obtain the historical operation log of the gas turbine, extract the historical characteristic efficiency data and historical characteristic emission data in the historical operation log, and generate a historical initial efficiency evaluation value and a historical initial emission evaluation value in combination with a preset efficiency evaluation index and a preset emission evaluation index; Establish a time reference line based on the historical operation duration of each historical operation log, and set data acquisition nodes based on a preset time interval; Collect the historical operation data, historical initial efficiency evaluation value, and historical initial emission evaluation value in the corresponding historical operation log according to the data acquisition nodes, and map them to the corresponding time reference line to obtain a historical operation data analysis graph; Among them, the historical operation data analysis graph includes a number of historical operation data change curves, a historical initial efficiency evaluation value change curve, and a historical initial emission evaluation value change curve corresponding to the historical operation log; Obtain a first mutation feature of the historical initial efficiency evaluation value change curve, a second mutation feature of the historical initial emission evaluation value change curve, and a third mutation feature of the historical operation data change curve in the same historical operation data analysis graph; Among them, the first mutation feature includes a number of first mutation nodes and a first mutation variable value corresponding to each first mutation node, the second mutation feature includes a number of second mutation nodes and a second mutation variable value corresponding to each second mutation node, and the third mutation feature includes a number of third mutation nodes and a third mutation variable value corresponding to each third mutation node; Obtain a first time interval between each third mutation node of the same historical operation data and the corresponding first mutation node adjacent to the rear, and a second time interval between each third mutation node and the corresponding second mutation node adjacent to the rear; Generate a first time correlation degree between the historical operation data and the historical initial efficiency evaluation value according to multiple first time intervals in the same historical operation data analysis graph; Generate a second time correlation degree between the historical operation data and the historical initial emission evaluation value according to multiple second time intervals in the same historical operation data analysis graph; Filter out historical operation data with a relevance degree greater than a preset first relevance degree threshold at the first time, and set it as the first pending historical operation data. Construct a first third mutation variable value sequence according to the third mutation variable values corresponding to each third mutation node of the first pending historical operation data; Filter out historical operation data with a relevance degree greater than a preset second relevance degree threshold at the second time, and set it as the second pending historical operation data. Construct a second third mutation variable value sequence according to the third mutation variable values corresponding to each third mutation node of the second pending historical operation data; Construct a first mutation variable value sequence according to the first mutation variable values corresponding to the first mutation nodes adjacent to the third mutation nodes in the first third mutation variable value sequence of each first pending historical operation data; Construct a second mutation variable value sequence according to the second mutation variable values corresponding to the second mutation nodes adjacent to each third mutation node of each second pending historical operation data; Compare the third mutation variable values in the first third mutation variable value sequence with the corresponding first mutation variable values in the first mutation variable value sequence to obtain the first quantity change synchronization degree, and generate the first sequence relevance degree corresponding to the first pending historical operation data and the historical initial efficiency evaluation value; Compare the third mutation variable values in the second third mutation variable value sequence with the corresponding second mutation variable values in the second mutation variable value sequence to obtain the second quantity change synchronization degree, and generate the second sequence relevance degree corresponding to the second pending historical operation data and the historical initial emission evaluation value; Generate the first comprehensive relevance degree corresponding to the first pending historical operation data and the historical initial efficiency evaluation value according to the first time relevance degree and the first sequence relevance degree; Generate the second comprehensive relevance degree corresponding to the second pending historical operation data and the historical initial emission evaluation value according to the second time relevance degree and the second sequence relevance degree; Determine a number of historical efficiency-related data and historical emission-related data according to the first comprehensive relevance degree and the second comprehensive relevance degree.

[0006] In some embodiments of the present application, the calculation formula of the first comprehensive relevance degree is: ; where G1 is the first comprehensive relevance degree, a1 is the weight coefficient of the time relevance degree, g1 is the first comprehensive conversion coefficient, t0 is the time conversion coefficient, n is the total number of third mutation nodes corresponding to the first pending historical operation data, is the first time interval between the i-th third mutation node and the adjacent first mutation node corresponding to the first pending historical operation data, Let a1 be the preset first-time correlation degree threshold, a2 be the weight coefficient of the sequence correlation degree, g2 be the second comprehensive correlation conversion coefficient, and b0 be the sequence correlation conversion coefficient. is the first quantization change synchronization degree. Preset the first sequence correlation degree threshold. When the first comprehensive correlation degree of the first pending historical operation data is greater than the preset first comprehensive correlation degree threshold, set the corresponding first pending historical operation data as historical efficiency-related data. The calculation formula for the second comprehensive correlation degree is: ; where G2 is the second comprehensive correlation degree, m is the total number of the third mutation nodes corresponding to the second pending historical operation data, is the second time interval between the s-th third mutation node corresponding to the second pending historical operation data and the adjacent second mutation node behind, is the preset second-time correlation degree threshold, is the second quantization change synchronization degree, Preset the first sequence correlation degree threshold. When the second comprehensive correlation degree of the second pending historical operation data is greater than the preset second comprehensive correlation degree threshold, set the corresponding second pending historical operation data as historical emission-related data.

[0007] In some embodiments of the present application, an efficiency evaluation model and an emission evaluation model are respectively constructed, including: Compare a number of historical efficiency-related data at the same data acquisition node with the corresponding standard efficiency-related data to obtain the difference in historical efficiency-related data, and generate the historical corrected efficiency evaluation value at the corresponding data acquisition node; Compare a number of historical emission-related data at the same data acquisition node with the corresponding standard emission-related data to obtain the difference in historical emission-related data, and generate the historical corrected emission evaluation value at the corresponding data acquisition node; Use the historical efficiency-related data and the corresponding historical corrected efficiency evaluation value at the same data acquisition node as an efficiency training group, where the historical efficiency-related data is used as the training input data, and the corresponding historical corrected efficiency evaluation value is used as the training output data; Perform neural network training based on the efficiency training groups at multiple data acquisition nodes to obtain the efficiency evaluation model of the gas turbine; Use the historical emission-related data and the corresponding historical corrected emission evaluation value at the same data acquisition node as an emission training group, where the historical emission-related data is used as the training input data, and the corresponding historical emission efficiency evaluation value is used as the training output data; Train a neural network based on the emission training groups at multiple data acquisition nodes to obtain an emission evaluation model for a gas turbine.

[0008] In some embodiments of the present application, the preset analysis model includes: The preset analysis model includes a first preset analysis model and a second preset analysis model; Obtain the historical optimization control logs of the gas turbine, and extract several historical corrected efficiency evaluation values and historical corrected emission evaluation values in each historical optimization control log; Compare several historical corrected efficiency evaluation values in the same historical optimization control log with a preset efficiency evaluation threshold to obtain a first initial historical node where the historical corrected efficiency evaluation value is less than the preset efficiency evaluation threshold, and a first end historical node where the historical corrected efficiency evaluation value after the first initial historical node is greater than the preset efficiency evaluation threshold; Compare the historical efficiency-related data at the first initial historical node with the corresponding standard efficiency-related data to obtain a difference in historical efficiency-related data. Set the historical efficiency-related data with a difference in historical efficiency-related data less than the preset difference threshold of efficiency-related data as the historical efficiency-related data to be optimized, and obtain the corresponding first historical difference to be optimized; Collect the historical control data between the first initial historical node and the first end historical node in the same historical optimization control log to generate a first historical control data set; Perform a change degree analysis on each historical control data in the first historical control data set, mark the historical control data with a change degree greater than the preset change degree, set the marked historical control data as the first historical optimization control data, and obtain the first historical control change amount value of the first historical optimization control data between the first initial historical node and the first end historical node; Take the historical efficiency-related data to be optimized, the first historical difference to be optimized, several first historical optimization control data between the first initial historical node and the first end historical node, and the corresponding first historical control change amount value at the first initial historical node of the same historical optimization control log as a set of first training data; Obtain multiple sets of first training data based on multiple historical optimization control logs and perform neural network training to obtain a first preset analysis model; Compare several historical corrected emission evaluation values in the same historical optimization control log with a preset emission evaluation threshold to obtain a second initial historical node where the historical corrected emission evaluation value is less than the preset emission evaluation threshold, and a second end historical node where the historical corrected emission evaluation value after the second initial historical node is greater than the preset emission evaluation threshold; Compare the historical emission-related data at the second initial historical node with the corresponding standard emission-related data to obtain the difference in historical emission-related data. Set the historical emission-related data with the difference in historical emission-related data less than the preset difference threshold of emission-related data as the historical emission-related data to be optimized, and obtain the corresponding second historical difference to be optimized; Collect the historical control data between the second initial historical node and the second terminal historical node in the same historical optimization control log to generate a second historical control data set; Analyze the degree of change for each historical control data in the second historical control data set, mark the historical control data with a degree of change greater than the preset degree of change, set the marked historical control data as the second historical optimized control data, and obtain the second historical control change amount value of the second historical optimized control data between the second initial historical node and the second terminal historical node; Take the second historical efficiency-related data to be optimized, the second historical difference to be optimized, the second historical optimized control data between the second initial historical node and the second terminal historical node, and the corresponding second historical control change amount value at the second initial historical node of the same historical optimization control log as a set of second training data; Obtain multiple sets of second training data based on multiple historical optimization control logs and perform neural network training to obtain a second preset analysis model.

[0009] In some embodiments of the present application, determine whether optimization control is required according to a preset optimization analysis model, an efficiency evaluation value, and an emission evaluation value. If so, obtain the control data to be optimized and generate a first control optimization strategy, including: When the efficiency evaluation value is greater than the preset efficiency evaluation threshold and the emission evaluation value is greater than the preset emission evaluation threshold, no control data to be optimized is generated; When the efficiency evaluation value is less than the preset efficiency evaluation threshold and the emission evaluation value is greater than the preset emission evaluation threshold, filter out the real-time efficiency-related data less than the corresponding standard efficiency-related data, set it as the real-time efficiency-related data to be optimized, and obtain the corresponding first real-time difference to be optimized; Analyze the real-time efficiency-related data to be optimized and the corresponding first real-time difference to be optimized based on the first preset optimization analysis model to obtain the first optimized control data and the corresponding first control change amount value; Set the first optimized control data as the control data to be optimized, and generate a first control optimization strategy according to multiple control data to be optimized and the corresponding first control change amount value; When the efficiency evaluation value is greater than the preset efficiency evaluation threshold and the emission evaluation value is less than the preset emission evaluation threshold, filter out the real-time emission-related data less than the corresponding standard emission-related data, set it as the real-time emission-related data to be optimized, and obtain the corresponding second real-time difference to be optimized; Analyze the real-time emission-related data to be optimized and the corresponding second real-time difference to be optimized based on the second preset optimization analysis model to obtain the second optimized control data and the corresponding second control change value; Set the second optimized control data as the control data to be optimized, and generate the first control optimization strategy according to multiple control data to be optimized and the corresponding second control change value; When the efficiency evaluation value is less than the preset efficiency evaluation threshold and the emission evaluation value is less than the preset emission evaluation threshold, set the first optimized control data and the second optimized control data as the control data to be optimized, and generate the first control optimization strategy by combining the corresponding first control change value and the second control change value.

[0010] In some embodiments of the present application, generating the predicted efficiency evaluation value and the predicted emission evaluation value after the inspection period includes: Preset the inspection period and several data monitoring nodes in the inspection period; Collect the real-time efficiency-related data and the real-time emission-related data according to the data monitoring nodes, and map them onto the inspection period to obtain the actual change curves and the corresponding actual change characteristics of each real-time efficiency-related data and real-time emission-related data in the inspection period; Wherein, the actual change characteristics include the actual change trend, the actual change rate between adjacent data monitoring nodes, and the actual change value; Perform curve trend extrapolation according to the actual change curve and the corresponding actual change characteristics to obtain the predicted efficiency-related data and the predicted emission-related data of each real-time efficiency-related data and real-time emission-related data in the future period; Input the predicted efficiency-related data and the predicted emission-related data in the future period into the corresponding efficiency evaluation model and emission evaluation model respectively to obtain the predicted efficiency evaluation value and the predicted emission evaluation value.

[0011] In some embodiments of the present application, judging whether to correct the first control optimization strategy according to the predicted efficiency evaluation value and the predicted emission evaluation value includes: Compare the predicted efficiency evaluation value with the preset efficiency evaluation threshold, and the predicted emission evaluation value with the preset emission evaluation threshold. If the predicted efficiency evaluation value is less than the preset efficiency evaluation threshold or the predicted emission evaluation value is less than the preset emission evaluation threshold, generate a correction instruction for the first control optimization strategy; Correct the first control optimization strategy according to the correction instruction to obtain the second control optimization strategy and generate the corresponding control adjustment instruction.

[0012] A gas turbine combustion optimization control system based on adaptive control according to an embodiment of the present application, compared with the prior art, its beneficial effects are: By determining the historical efficiency-related data and historical emission-related data, constructing an efficiency evaluation model and an emission evaluation model to obtain an efficiency evaluation value and an emission evaluation value, and combining a preset analysis model to judge whether optimization control is needed. If so, a first control optimization strategy is obtained, and it is judged whether to correct the first control optimization strategy according to the predicted efficiency evaluation value and the predicted emission evaluation value after the first control optimization strategy, so as to timely discover and adjust the control optimization strategy, ensure that the pollutant emissions meet the standards and improve the combustion efficiency, and improve the combustion optimization control effect of the gas turbine. Brief Description of the Drawings

[0013] Figure 1 It is a schematic diagram of a gas turbine combustion optimization control system based on adaptive control in an embodiment of the present application. Detailed Embodiments

[0014] The following combines the drawings and embodiments to further describe the detailed embodiments of the present application in detail. The following embodiments are used to illustrate the present application, but are not used to limit the scope of the present application.

[0015] In the description of the present application, it should be understood that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present application.

[0016] 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 such features. In the description of the present application, unless otherwise stated, the meaning of "plurality" is two or more.

[0017] In the description of the present application, it should be noted that unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected, or indirectly connected through an intermediate medium, and it may be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific situations.

[0018] As Figure 1 shown, a gas turbine combustion optimization control system based on adaptive control in an embodiment of the present application includes: An acquisition module, configured to acquire and analyze the historical operation data of a gas turbine, determine a number of historical efficiency-related data and historical emission-related data of the gas turbine according to the analysis results, and respectively construct an efficiency evaluation model and an emission evaluation model; An evaluation module, configured to acquire real-time efficiency-related data and real-time emission-related data and input them into the corresponding efficiency evaluation model and emission evaluation model to obtain an efficiency evaluation value and an emission evaluation value; A judgment module, configured to judge whether optimization control is required according to a preset optimization analysis model, an efficiency evaluation value, and an emission evaluation value. If so, obtain the data to be optimized and generate a first control optimization strategy; An optimization module, configured to generate a control optimization instruction according to the first control optimization strategy, obtain the actual change characteristics of the real-time efficiency-related data and the real-time emission-related data after obtaining the control optimization instruction, and generate a predicted efficiency evaluation value and a predicted emission evaluation value for a future period; A correction module, configured to judge whether to correct the first control optimization strategy according to the predicted efficiency evaluation value and the predicted emission evaluation value. If so, obtain a second control optimization strategy and generate a control adjustment instruction.

[0019] In some embodiments of the present application, determining a number of historical efficiency-related data and historical emission-related data of the gas turbine according to the analysis results includes: Presetting characteristic efficiency data and characteristic emission data in advance; Obtaining the historical operation log of the gas turbine, extracting the historical characteristic efficiency data and historical characteristic emission data in the historical operation log, and generating a historical initial efficiency evaluation value and a historical initial emission evaluation value in combination with a preset efficiency evaluation index and a preset emission evaluation index; Establishing a time reference line based on the historical operation duration of each historical operation log, and setting data acquisition nodes based on a preset time interval; Collecting the historical operation data, the historical initial efficiency evaluation value, and the historical initial emission evaluation value in the corresponding historical operation log according to the data acquisition nodes, and mapping them to the corresponding time reference line to obtain a historical operation data analysis graph; Wherein, the historical operation data analysis graph includes a number of historical operation data change curves, a historical initial efficiency evaluation value change curve, and a historical initial emission evaluation value change curve corresponding to the historical operation log; Obtaining a first mutation characteristic of the historical initial efficiency evaluation value change curve, a second mutation characteristic of the historical initial emission evaluation value change curve, and a third mutation characteristic of the historical operation data change curve in the same historical operation data analysis graph; Among them, the first mutation feature includes a number of first mutation nodes and the first mutation quantity values corresponding to each first mutation node, the second mutation feature includes a number of second mutation nodes and the second mutation quantity values corresponding to each second mutation node, and the third mutation feature includes a number of third mutation nodes and the third mutation quantity values corresponding to each third mutation node; Obtain the first time interval between each third mutation node of the same historical operation data and the corresponding adjacent first mutation node behind it, and the second time interval between each third mutation node and the corresponding adjacent second mutation node behind it; Generate the first time correlation degree between the historical operation data and the historical initial efficiency evaluation value according to multiple first time intervals in the same historical operation data analysis graph; Generate the second time correlation degree between the historical operation data and the historical initial emission evaluation value according to multiple second time intervals in the same historical operation data analysis graph; Screen out the historical operation data with the first time correlation degree greater than the preset first correlation degree threshold, and set it as the first pending historical operation data. Construct the first third mutation quantity value sequence according to the third mutation quantity values corresponding to each third mutation node of the first pending historical operation data; Screen out the historical operation data with the second time correlation degree greater than the preset second correlation degree threshold, and set it as the second pending historical operation data. Construct the second third mutation quantity value sequence according to the third mutation quantity values corresponding to each third mutation node of the second pending historical operation data; Construct the first mutation quantity value sequence according to the first mutation quantity values corresponding to the first mutation nodes adjacent to the third mutation nodes in the first third mutation quantity value sequence of each first pending historical operation data; Construct the second mutation quantity value sequence according to the second mutation quantity values corresponding to the second mutation nodes adjacent to the third mutation nodes of each second pending historical operation data; Compare the third mutation quantity values in the first third mutation quantity value sequence with the corresponding first mutation quantity values in the first mutation quantity value sequence to obtain the first quantity change synchronization degree, and generate the first sequence correlation degree corresponding to the first pending historical operation data and the historical initial efficiency evaluation value; Compare the third mutation quantity values in the second third mutation quantity value sequence with the corresponding second mutation quantity values in the second mutation quantity value sequence to obtain the second quantity change synchronization degree, and generate the second sequence correlation degree corresponding to the second pending historical operation data and the historical initial emission evaluation value; Generate the first comprehensive correlation degree corresponding to the first pending historical operation data and the historical initial efficiency evaluation value according to the first time correlation degree and the first sequence correlation degree; Generate a second comprehensive correlation degree corresponding to the second to-be-determined historical operation data and the historical initial emission evaluation value according to the second time correlation degree and the second sequence correlation degree; Determine a number of historical efficiency-related data and historical emission-related data according to the first comprehensive correlation degree and the second comprehensive correlation degree.

[0020] In this embodiment, the first time interval refers to the time interval between each third mutation node and the first adjacent first mutation node behind it, and the second time interval refers to the time interval between each third mutation node and the first adjacent second mutation node behind it. When all the first time intervals are small, it indicates that the corresponding historical operation data has an impact on the historical initial efficiency evaluation value and the historical initial emission evaluation value, and the smaller the impact delay, that is, the faster the impact, the greater the corresponding first time correlation degree and the second time correlation degree.

[0021] In this embodiment, the first magnitude change synchronization degree refers to the synchronization degree of the first mutation magnitude value in the first mutation magnitude value sequence changing with the corresponding third mutation magnitude value in the third mutation magnitude value sequence, and the second magnitude change synchronization degree refers to the synchronization degree of the second mutation magnitude value in the second mutation magnitude value sequence changing with the corresponding third mutation magnitude value in the third mutation magnitude value sequence. When the first magnitude change synchronization degree and the second magnitude change synchronization degree are larger, the corresponding first sequence correlation degree and the second sequence correlation degree are larger.

[0022] In this embodiment, by constructing a historical operation data analysis graph in each historical operation log, the time correlation degree and the sequence correlation degree between each historical operation data and the historical initial efficiency evaluation value and the historical emission evaluation value are obtained, so as to determine a number of historical efficiency-related data and historical emission-related data, laying a foundation for subsequent construction of an efficiency evaluation model and an emission evaluation model, improving the combustion optimization control efficiency of the gas turbine, and improving the combustion efficiency on the premise of meeting the emission requirements.

[0023] In some embodiments of the present application, the calculation formula for the first comprehensive correlation degree is: ; wherein, G1 is the first comprehensive correlation degree, a1 is the weight coefficient of the time correlation degree, g1 is the first comprehensive correlation conversion coefficient, t0 is the time correlation conversion coefficient, n is the total number of third mutation nodes corresponding to the first to-be-determined historical operation data, is the first time interval between the i-th third mutation node corresponding to the first to-be-determined historical operation data and the corresponding first mutation node adjacent behind it, is the preset first time correlation degree threshold, a2 is the weight coefficient of the sequence correlation degree, g2 is the second comprehensive correlation conversion coefficient, b0 is the sequence correlation conversion coefficient, is the first quantization change synchronization degree, preset the first sequence correlation degree threshold; When the first comprehensive correlation degree of the first to-be-determined historical operation data is greater than the preset first comprehensive correlation degree threshold, set the corresponding first to-be-determined historical operation data as historical efficiency-related data; The calculation formula for the second comprehensive correlation degree is: ; where G2 is the second comprehensive correlation degree, m is the total number of the third mutation nodes corresponding to the second to-be-determined historical operation data, is the second time interval between the s-th third mutation node corresponding to the second to-be-determined historical operation data and the adjacent second mutation node behind, is the preset second time correlation degree threshold, is the second quantization change synchronization degree, preset the first sequence correlation degree threshold; When the second comprehensive correlation degree of the second to-be-determined historical operation data is greater than the preset second comprehensive correlation degree threshold, set the corresponding second to-be-determined historical operation data as historical emission-related data.

[0024] In this embodiment, the first comprehensive correlation conversion coefficient refers to converting the time correlation degree into a value with the same dimension as the comprehensive correlation degree, and the second comprehensive correlation conversion coefficient refers to converting the sequence correlation degree into a value with the same dimension as the comprehensive correlation degree. When the time correlation degree is larger and the sequence correlation degree is larger, the corresponding comprehensive correlation degree is larger, and vice versa.

[0025] In this embodiment, the time correlation conversion coefficient refers to converting the time interval into a value with the same dimension as the time correlation degree, and the sequence correlation conversion coefficient refers to converting the magnitude change synchronization degree into a value with the same dimension as the sequence correlation degree. When the time interval is shorter, the corresponding time correlation degree is larger, and vice versa.

[0026] In this embodiment, by calculating the time correlation degree and sequence correlation degree of each historical operation data with the historical initial efficiency evaluation value and historical initial emission evaluation value, the influence delay time and influence amount of each historical operation data on the historical initial efficiency evaluation value and historical initial emission evaluation value are accurately evaluated, so as to determine the corresponding comprehensive correlation degree, and the corresponding historical efficiency-related data and historical emission-related data are determined according to the comprehensive correlation degree, laying a foundation for subsequent construction of the efficiency evaluation model and emission evaluation model, and improving the evaluation accuracy of subsequent efficiency evaluation values and emission evaluation values.

[0027] In some embodiments of the present application, an efficiency evaluation model and an emission evaluation model are respectively constructed, including: Compare a number of historical efficiency-related data at the same data acquisition node with the corresponding standard efficiency-related data to obtain the difference in historical efficiency-related data, and generate a historical corrected efficiency evaluation value at the corresponding data acquisition node; Compare a number of historical emission-related data at the same data acquisition node with the corresponding standard emission-related data to obtain the difference in historical emission-related data, and generate a historical corrected emission evaluation value at the corresponding data acquisition node; Use the historical efficiency-related data and the corresponding historical corrected efficiency evaluation value at the same data acquisition node as an efficiency training group, where the historical efficiency-related data is used as the training input data, and the corresponding historical corrected efficiency evaluation value is used as the training output data; Perform neural network training based on the efficiency training groups at multiple data acquisition nodes to obtain an efficiency evaluation model for the gas turbine; Use the historical emission-related data and the corresponding historical corrected emission evaluation value at the same data acquisition node as an emission training group, where the historical emission-related data is used as the training input data, and the corresponding historical emission efficiency evaluation value is used as the training output data; Perform neural network training based on the emission training groups at multiple data acquisition nodes to obtain an emission evaluation model for the gas turbine.

[0028] In this embodiment, the historical corrected efficiency evaluation value and the historical corrected emission evaluation value refer to those recalculated based on a number of historical efficiency-related data and a number of historical emission-related data, and the accuracy rates of the historical corrected efficiency evaluation value and the historical corrected emission evaluation value are higher than the historical initial efficiency evaluation value and the historical initial emission evaluation value at the same data acquisition node.

[0029] In this embodiment, the larger the historical corrected efficiency evaluation value, the higher the combustion efficiency of the gas turbine, and vice versa. The larger the historical corrected emission evaluation value, the less pollutants are generated and the lower the emissions from the gas turbine combustion, and vice versa. Controlling the pollutant emissions during the combustion process of the gas turbine and improving the combustion efficiency are of great significance for environmental protection and energy conversion.

[0030] In some embodiments of the present application, the preset analysis model includes: The preset analysis model includes a first preset analysis model and a second preset analysis model; Obtain the historical optimization control logs of the gas turbine, and extract a number of historical corrected efficiency evaluation values and historical corrected emission evaluation values from each historical optimization control log; Compare several historical correction efficiency evaluation values in the same historical optimization control log with a preset efficiency evaluation threshold to obtain a first initial historical node where the historical correction efficiency evaluation value is less than the preset efficiency evaluation threshold, and a first terminal historical node where the historical correction efficiency evaluation value after the first initial historical node is greater than the preset efficiency evaluation threshold; Compare the historical efficiency-related data at the first initial historical node with the corresponding standard efficiency-related data to obtain a difference in historical efficiency-related data. Set the historical efficiency-related data with the difference in historical efficiency-related data less than the preset difference threshold of efficiency-related data as the historical efficiency-related data to be optimized, and obtain the corresponding first difference to be optimized for history; Collect the historical control data between the first initial historical node and the first terminal historical node in the same historical optimization control log to generate a first historical control data set; Perform a change degree analysis on each historical control data in the first historical control data set, mark the historical control data with a change degree greater than the preset change degree, set the marked historical control data as the first historical optimization control data, and obtain the first historical control change amount value of the first historical optimization control data between the first initial historical node and the first terminal historical node; Take the historical efficiency-related data to be optimized, the first difference to be optimized for history, several first historical optimization control data between the first initial historical node and the first terminal historical node, and the corresponding first historical control change amount value at the first initial historical node of the same historical optimization control log as a set of first training data; Obtain multiple sets of first training data based on multiple historical optimization control logs and perform neural network training to obtain a first preset analysis model; Compare several historical correction emission evaluation values in the same historical optimization control log with a preset emission evaluation threshold to obtain a second initial historical node where the historical correction emission evaluation value is less than the preset emission evaluation threshold, and a second terminal historical node where the historical correction emission evaluation value after the second initial historical node is greater than the preset emission evaluation threshold; Compare the historical emission-related data at the second initial historical node with the corresponding standard emission-related data to obtain a difference in historical emission-related data. Set the historical emission-related data with the difference in historical emission-related data less than the preset difference threshold of emission-related data as the historical emission-related data to be optimized, and obtain the corresponding second difference to be optimized for history; Collect the historical control data between the second initial historical node and the second terminal historical node in the same historical optimization control log to generate a second historical control data set; Analyze the degree of change for each piece of historical control data in the second historical control dataset, mark the historical control data with a degree of change greater than the preset degree of change, set the marked historical control data as the second historical optimized control data, and obtain the second historical control change amount value between the second initial historical node and the second terminal historical node of the second historical optimized control data; Take the second historical data related to the efficiency to be optimized, the second historical difference to be optimized, the second historical optimized control data between the second initial historical node and the second terminal historical node, and the corresponding second historical control change amount value at the second initial historical node of the same historical optimized control log as a set of second training data; Obtain multiple sets of second training data based on multiple historical optimized control logs and perform neural network training to obtain the second preset analysis model.

[0031] In this embodiment, the first initial historical node and the second initial historical node refer to the first time node in each historical optimized control log where the historical correction efficiency evaluation value is less than the preset efficiency evaluation threshold and the historical correction emission evaluation value is less than the preset emission evaluation threshold. The first terminal historical node and the second terminal historical node refer to the first time node in each historical optimized control log where the historical correction efficiency evaluation value after the first initial historical node is greater than the preset efficiency evaluation threshold and the historical correction emission evaluation value after the second initial historical node is greater than the preset emission evaluation threshold.

[0032] In this embodiment, by constructing the first preset analysis model and the second preset analysis model, it lays a foundation for formulating the first control optimization strategy in the follow-up and improves the combustion optimization control effect of the gas turbine.

[0033] In some embodiments of the present application, it is judged whether optimization control is needed according to the preset optimization analysis model, the efficiency evaluation value, and the emission evaluation value. If so, the data to be optimized for control is obtained and the first control optimization strategy is generated, including: When the efficiency evaluation value is greater than the preset efficiency evaluation threshold and the emission evaluation value is greater than the preset emission evaluation threshold, no data to be optimized for control is generated; When the efficiency evaluation value is less than the preset efficiency evaluation threshold and the emission evaluation value is greater than the preset emission evaluation threshold, filter out the real-time efficiency-related data that is less than the corresponding standard efficiency-related data, set it as the real-time data related to the efficiency to be optimized, and obtain the corresponding first real-time difference to be optimized; Analyze the real-time data related to the efficiency to be optimized and the corresponding first real-time difference to be optimized based on the first preset optimization analysis model to obtain the first optimized control data and the corresponding first control change amount value; Set the first optimized control data as the control data to be optimized, and generate a first control optimization strategy based on multiple control data to be optimized and the corresponding first control change amount values; When the efficiency evaluation value is greater than the preset efficiency evaluation threshold and the emission evaluation value is less than the preset emission evaluation threshold, filter out the real-time emission-related data that is less than the corresponding standard emission-related data, set it as the real-time emission-related data to be optimized, and obtain the corresponding second real-time difference to be optimized; Analyze the real-time emission-related data to be optimized and the corresponding second real-time difference to be optimized based on the second preset optimization analysis model to obtain the second optimized control data and the corresponding second control change amount values; Set the second optimized control data as the control data to be optimized, and generate a first control optimization strategy based on multiple control data to be optimized and the corresponding second control change amount values; When the efficiency evaluation value is less than the preset efficiency evaluation threshold and the emission evaluation value is less than the preset emission evaluation threshold, set the first optimized control data and the second optimized control data as the control data to be optimized, and generate a first control optimization strategy by combining the corresponding first control change amount values and the second control change amount values.

[0034] In this embodiment, the first real-time difference to be optimized refers to the difference between the real-time efficiency-related data to be optimized and the corresponding standard efficiency-related data, and the second real-time difference to be optimized refers to the difference between the real-time emission-related data to be optimized and the corresponding standard emission-related data.

[0035] In this embodiment, through different analysis results of the efficiency evaluation value and the emission evaluation value, and by combining the first preset analysis model and the second preset analysis model, the corresponding control data to be optimized and the corresponding control change amount values are obtained, and a first control optimization strategy is generated to realize the adaptive control of the combustion optimization of the gas turbine, improve the combustion optimization control effect, and improve the combustion efficiency on the premise of ensuring the combustion emission requirements.

[0036] In some embodiments of the present application, generating a predicted efficiency evaluation value and a predicted emission evaluation value after the inspection period includes: Preset the inspection period and several data monitoring nodes in the inspection period; Collect real-time efficiency-related data and real-time emission-related data according to the data monitoring nodes, and map them to the inspection period to obtain the actual change curves and corresponding actual change characteristics of each real-time efficiency-related data and real-time emission-related data in the inspection period; Wherein, the actual change characteristics include the actual change trend, the actual change rate between adjacent data monitoring nodes, and the actual change amount value; Based on the actual change curve and the corresponding actual change characteristics, perform curve trend extrapolation to obtain the predicted efficiency-related data and predicted emission-related data of each real-time efficiency-related data and real-time emission-related data in the future period; Input the predicted efficiency-related data and predicted emission-related data of the future period into the corresponding efficiency evaluation model and emission evaluation model respectively to obtain the predicted efficiency evaluation value and the predicted emission evaluation value.

[0037] In some embodiments of the present application, judging whether to correct the first control optimization strategy according to the predicted efficiency evaluation value and the predicted emission evaluation value includes: Compare the predicted efficiency evaluation value with the preset efficiency evaluation threshold, and compare the predicted emission evaluation value with the preset emission evaluation threshold. If the predicted efficiency evaluation value is less than the preset efficiency evaluation threshold or the predicted emission evaluation value is less than the preset emission evaluation threshold, generate a correction instruction for the first control optimization strategy; Correct the first control optimization strategy according to the correction instruction to obtain the second control optimization strategy and generate the corresponding control adjustment instruction.

[0038] In this embodiment, when the predicted efficiency evaluation value is less than the preset efficiency evaluation threshold or the predicted emission evaluation value is less than the preset emission evaluation threshold, screen out the predicted efficiency-related data less than the corresponding standard efficiency-related data or the predicted emission-related data less than the corresponding standard emission-related data, and set them as the predicted efficiency-related data to be optimized or the predicted emission-related data to be optimized, and obtain the corresponding first predicted difference to be optimized or the second predicted difference to be optimized. Input the predicted efficiency-related data to be optimized and the corresponding first predicted difference to be optimized into the first preset analysis model, or input the predicted emission-related data to be optimized and the corresponding second predicted difference to be optimized into the second preset analysis model to obtain the first predicted optimization control data and the corresponding first predicted control change value or the second predicted optimization control data and the corresponding second predicted control change value in the future period, and generate the corresponding predicted control optimization strategy in the future period. Correct or adjust the optimization strategy of the control data to be optimized in the corresponding period of the first control optimization strategy according to the predicted control optimization strategy to obtain the second control optimization strategy.

[0039] In this embodiment, by setting the inspection period and obtaining the actual change curve and actual change characteristics in the inspection period, the predicted efficiency evaluation value and the predicted emission evaluation value in the future period are obtained, and the deficiencies of the first control optimization strategy are timely discovered and adjusted to ensure the combustion emission requirements and improve the combustion efficiency, and improve the combustion optimization control effect of the gas turbine.

[0040] The above are only the preferred embodiments of the present application. It should be noted that for those of ordinary skill in the art, without departing from the technical principles of the present application, several improvements and substitutions can be made, and these improvements and substitutions should also be regarded as the protection scope of the present application.

Claims

1. A gas turbine combustion optimization control system based on adaptive control, characterized in that, Including: An acquisition module, configured to acquire and analyze historical operation data of a gas turbine, determine a number of historical efficiency-related data and historical emission-related data of the gas turbine according to the analysis results, and respectively construct an efficiency evaluation model and an emission evaluation model; An evaluation module, configured to acquire real-time efficiency-related data and real-time emission-related data and input them into the corresponding efficiency evaluation model and emission evaluation model to obtain an efficiency evaluation value and an emission evaluation value; A judgment module, configured to judge whether optimization control is required according to a preset optimization analysis model, the efficiency evaluation value, and the emission evaluation value. If so, obtain data to be optimized and controlled and generate a first control optimization strategy; An optimization module, configured to generate a control optimization instruction according to the first control optimization strategy, obtain the actual change characteristics of the real-time efficiency-related data and the real-time emission-related data after obtaining the control optimization instruction, and generate a predicted efficiency evaluation value and a predicted emission evaluation value for a future period; A correction module, configured to judge whether to correct the first control optimization strategy according to the predicted efficiency evaluation value and the predicted emission evaluation value. If so, obtain a second control optimization strategy and generate a control adjustment instruction.

2. The gas turbine combustion optimization control system based on adaptive control according to claim 1, wherein Determining a number of historical efficiency-related data and historical emission-related data of the gas turbine according to the analysis results, including: Presetting characteristic efficiency data and characteristic emission data; Obtaining the historical operation log of the gas turbine, extracting the historical characteristic efficiency data and historical characteristic emission data in the historical operation log, and generating a historical initial efficiency evaluation value and a historical initial emission evaluation value in combination with a preset efficiency evaluation index and a preset emission evaluation index; Establishing a time reference line based on the historical operation duration of each historical operation log, and setting data acquisition nodes based on a preset time interval; Collecting the historical operation data, the historical initial efficiency evaluation value, and the historical initial emission evaluation value in the corresponding historical operation log according to the data acquisition nodes, and mapping them to the corresponding time reference line to obtain a historical operation data analysis graph; Wherein, the historical operation data analysis graph includes a number of historical operation data change curves, a historical initial efficiency evaluation value change curve, and a historical initial emission evaluation value change curve corresponding to the historical operation log; Obtaining a first mutation characteristic of the historical initial efficiency evaluation value change curve, a second mutation characteristic of the historical initial emission evaluation value change curve, and a third mutation characteristic of the historical operation data change curve in the same historical operation data analysis graph; Wherein, the first mutation characteristic includes a number of first mutation nodes and a first mutation variable value corresponding to each first mutation node, the second mutation characteristic includes a number of second mutation nodes and a second mutation variable value corresponding to each second mutation node, and the third mutation characteristic includes a number of third mutation nodes and a third mutation variable value corresponding to each third mutation node; Obtaining a first time interval between each third mutation node and the corresponding first mutation node adjacent thereto in the same historical operation data, and a second time interval between each third mutation node and the corresponding second mutation node adjacent thereto; Generating a first time correlation degree between the historical operation data and the historical initial efficiency evaluation value according to a plurality of first time intervals in the same historical operation data analysis graph; Generate the second-time correlation degree between historical operation data and historical initial emission evaluation values based on multiple second-time intervals in the same historical operation data analysis graph; Screen out the historical operation data with the first-time correlation degree greater than the preset first correlation degree threshold, and set it as the first pending historical operation data. Construct the first third mutation variable value sequence according to the third mutation variable values corresponding to each third mutation node of the first pending historical operation data; Screen out the historical operation data with the second-time correlation degree greater than the preset second correlation degree threshold, and set it as the second pending historical operation data. Construct the second third mutation variable value sequence according to the third mutation variable values corresponding to each third mutation node of the second pending historical operation data; Construct the first mutation variable value sequence according to the first mutation variable values corresponding to the first mutation nodes adjacent to the third mutation nodes in the first third mutation variable value sequence of each first pending historical operation data; Construct the second mutation variable value sequence according to the second mutation variable values corresponding to the second mutation nodes adjacent to each third mutation node of each second pending historical operation data; Compare the third mutation variable values in the first third mutation variable value sequence with the corresponding first mutation variable values in the first mutation variable value sequence to obtain the first quantity change synchronization degree, and generate the first sequence correlation degree corresponding to the first pending historical operation data and the historical initial efficiency evaluation value; Compare the third mutation variable values in the second third mutation variable value sequence with the corresponding second mutation variable values in the second mutation variable value sequence to obtain the second quantity change synchronization degree, and generate the second sequence correlation degree corresponding to the second pending historical operation data and the historical initial emission evaluation value; Generate the first comprehensive correlation degree corresponding to the first pending historical operation data and the historical initial efficiency evaluation value according to the first-time correlation degree and the first sequence correlation degree; Generate the second comprehensive correlation degree corresponding to the second pending historical operation data and the historical initial emission evaluation value according to the second-time correlation degree and the second sequence correlation degree; Determine a number of historical efficiency-related data and historical emission-related data according to the first comprehensive correlation degree and the second comprehensive correlation degree.

3. The gas turbine combustion optimization control system based on adaptive control according to claim 2, wherein The calculation formula for the first comprehensive correlation degree is: ； Among them, G1 is the first comprehensive correlation degree, a1 is the weight coefficient of the time correlation degree, g1 is the first comprehensive correlation conversion coefficient, t0 is the time correlation conversion coefficient, and n is the total number of the third mutation nodes corresponding to the first undetermined historical operation data. is the first time interval between the i-th third mutation node corresponding to the first undetermined historical operation data and the adjacent first mutation node behind. is the preset first time correlation degree threshold, a2 is the weight coefficient of the sequence correlation degree, g2 is the second comprehensive correlation conversion coefficient, and b0 is the sequence correlation conversion coefficient. is the first quantization change synchronization degree. Preset the first sequence correlation degree threshold. When the first comprehensive correlation degree of the first pending historical operation data is greater than the preset first comprehensive correlation degree threshold, set the corresponding first pending historical operation data as historical efficiency-related data; The calculation formula for the second comprehensive correlation degree is: ； Among them, G2 is the second comprehensive correlation degree, m is the total number of third mutation nodes corresponding to the second undetermined historical operation data, is the second time interval between the s-th third mutation node corresponding to the second undetermined historical operation data and the adjacent second mutation node behind, is the preset second time correlation degree threshold, is the second quantization change synchronization degree, preset first sequence correlation degree threshold; When the second comprehensive correlation degree of the second pending historical operation data is greater than the preset second comprehensive correlation degree threshold, set the corresponding second pending historical operation data as historical emission-related data.

4. The gas turbine combustion optimization control system based on adaptive control according to claim 3, wherein Construct an efficiency evaluation model and an emission evaluation model respectively, including: Compare a number of historical efficiency-related data at the same data acquisition node with the corresponding standard efficiency-related data to obtain the difference in historical efficiency-related data, and generate the historical corrected efficiency evaluation value at the corresponding data acquisition node; Compare several historical emission-related data at the same data acquisition node with the corresponding standard emission-related data to obtain the difference in historical emission-related data, and generate the historical corrected emission evaluation value at the corresponding data acquisition node; Use the historical efficiency-related data and the corresponding historical corrected efficiency evaluation value at the same data acquisition node as an efficiency training group. Among them, the historical efficiency-related data is used as the training input data, and the corresponding historical corrected efficiency evaluation value is used as the training output data; Perform neural network training based on the efficiency training groups at multiple data acquisition nodes to obtain the efficiency evaluation model of the gas turbine; Use the historical emission-related data and the corresponding historical corrected emission evaluation value at the same data acquisition node as an emission training group. Among them, the historical emission-related data is used as the training input data, and the corresponding historical emission efficiency evaluation value is used as the training output data; Perform neural network training based on the emission training groups at multiple data acquisition nodes to obtain the emission evaluation model of the gas turbine.

5. The gas turbine combustion optimization control system based on adaptive control according to claim 4, characterized in that The preset analysis model includes: The preset analysis model includes a first preset analysis model and a second preset analysis model; Obtain the historical optimization control log of the gas turbine, and extract several historical corrected efficiency evaluation values and historical corrected emission evaluation values in each historical optimization control log; Compare several historical corrected efficiency evaluation values in the same historical optimization control log with the preset efficiency evaluation threshold to obtain the first initial historical node where the historical corrected efficiency evaluation value is less than the preset efficiency evaluation threshold, and the first terminal historical node where the historical corrected efficiency evaluation value after the first initial historical node is greater than the preset efficiency evaluation threshold; Compare the historical efficiency-related data at the first initial historical node with the corresponding standard efficiency-related data to obtain the difference in historical efficiency-related data. Set the historical efficiency-related data with the difference in historical efficiency-related data less than the preset difference threshold of efficiency-related data as the historical efficiency-related data to be optimized, and obtain the corresponding first difference to be optimized; Collect the historical control data between the first initial historical node and the first terminal historical node in the same historical optimization control log to generate the first historical control dataset; Perform a change degree analysis on each historical control data in the first historical control dataset, mark the historical control data with a change degree greater than the preset change degree, set the marked historical control data as the first historical optimized control data, and obtain the first historical control change amount value of the first historical optimized control data between the first initial historical node and the first terminal historical node; Use the historical efficiency-related data to be optimized, the first difference to be optimized, several first historical optimized control data between the first initial historical node and the first terminal historical node, and the corresponding first historical control change amount value at the first initial historical node of the same historical optimization control log as a group of first training data; Obtain multiple groups of first training data based on multiple historical optimization control logs and perform neural network training to obtain the first preset analysis model; Compare several historical corrected emission evaluation values in the same historical optimization control log with a preset emission evaluation threshold to obtain a second initial historical node where the historical corrected emission evaluation value is less than the preset emission evaluation threshold, and a second end historical node where the historical corrected emission evaluation value after the second initial historical node is greater than the preset emission evaluation threshold; Compare the historical emission-related data at the second initial historical node with the corresponding standard emission-related data to obtain a difference in historical emission-related data. Set the historical emission-related data with a difference in historical emission-related data less than the preset difference threshold of emission-related data as the historical emission-related data to be optimized, and obtain the corresponding second historical difference to be optimized; Collect the historical control data between the second initial historical node and the second end historical node in the same historical optimization control log to generate a second historical control data set; Analyze the degree of change for each historical control data in the second historical control data set, mark the historical control data with a degree of change greater than the preset degree of change, set the marked historical control data as the second historical optimized control data, and obtain the second historical control change amount value of the second historical optimized control data between the second initial historical node and the second end historical node; Take the second historical efficiency-related data to be optimized, the second historical difference to be optimized, the second historical optimized control data between the second initial historical node and the second end historical node, and the corresponding second historical control change amount value at the second initial historical node of the same historical optimization control log as a set of second training data; Obtain multiple sets of second training data according to multiple historical optimization control logs and perform neural network training to obtain a second preset analysis model; 6. The gas turbine combustion optimization control system based on adaptive control according to claim 5, wherein Judge whether optimization control is required according to the preset optimization analysis model, the efficiency evaluation value, and the emission evaluation value. If so, obtain the control data to be optimized and generate a first control optimization strategy, including: When the efficiency evaluation value is greater than the preset efficiency evaluation threshold and the emission evaluation value is greater than the preset emission evaluation threshold, no control data to be optimized is generated; When the efficiency evaluation value is less than the preset efficiency evaluation threshold and the emission evaluation value is greater than the preset emission evaluation threshold, filter out the real-time efficiency-related data less than the corresponding standard efficiency-related data, set it as the real-time efficiency-related data to be optimized, and obtain the corresponding first real-time difference to be optimized; Analyze the real-time efficiency-related data to be optimized and the corresponding first real-time difference to be optimized based on the first preset optimization analysis model to obtain the first optimized control data and the corresponding first control change amount value; Set the first optimized control data as the control data to be optimized, and generate a first control optimization strategy according to multiple control data to be optimized and the corresponding first control change amount value; When the efficiency evaluation value is greater than the preset efficiency evaluation threshold and the emission evaluation value is less than the preset emission evaluation threshold, filter out the real-time emission-related data less than the corresponding standard emission-related data, set it as the real-time emission-related data to be optimized, and obtain the corresponding second real-time difference to be optimized; Analyze the real-time data to be optimized for emissions and the corresponding second real-time difference to be optimized based on the second preset optimization analysis model to obtain the second optimized control data and the corresponding second control change value; Set the second optimized control data as the control data to be optimized, and generate the first control optimization strategy based on the multiple control data to be optimized and the corresponding second control change value; If the efficiency evaluation value is less than the preset efficiency evaluation threshold and the emission evaluation value is less than the preset emission evaluation threshold, set the first optimized control data and the second optimized control data as the control data to be optimized, and generate the first control optimization strategy by combining the corresponding first control change value and the second control change value.

7. The gas turbine combustion optimization control system based on adaptive control according to claim 6, wherein, Generate the predicted efficiency evaluation value and the predicted emission evaluation value after the inspection period, including: Preset the inspection period and several data monitoring nodes in the inspection period; Collect the real-time efficiency-related data and the real-time emission-related data according to the data monitoring nodes and map them to the inspection period to obtain the actual change curve and the corresponding actual change characteristics of each real-time efficiency-related data and real-time emission-related data in the inspection period; Among them, the actual change characteristics include the actual change trend, the actual change rate between adjacent data monitoring nodes, and the actual change value; Perform curve trend extrapolation based on the actual change curve and the corresponding actual change characteristics to obtain the predicted efficiency-related data and the predicted emission-related data of each real-time efficiency-related data and real-time emission-related data in the future period; Input the predicted efficiency-related data and the predicted emission-related data in the future period into the corresponding efficiency evaluation model and emission evaluation model respectively to obtain the predicted efficiency evaluation value and the predicted emission evaluation value.

8. The gas turbine combustion optimization control system based on adaptive control according to claim 7, characterized in that, Judge whether to correct the first control optimization strategy according to the predicted efficiency evaluation value and the predicted emission evaluation value, including: Compare the predicted efficiency evaluation value with the preset efficiency evaluation threshold, and the predicted emission evaluation value with the preset emission evaluation threshold. If the predicted efficiency evaluation value is less than the preset efficiency evaluation threshold or the predicted emission evaluation value is less than the preset emission evaluation threshold, generate a correction instruction for the first control optimization strategy; Correct the first control optimization strategy according to the correction instruction to obtain the second control optimization strategy and generate the corresponding control adjustment instruction.

Images