Combined cycle unit control method based on data fusion

Abstract

The invention discloses a combined cycle unit control method based on data fusion, and the method comprises the steps: obtaining historical operation data, and determining a plurality of pieces of feature operation data according to the correlation degree of the historical operation data and a comprehensive evaluation index; obtaining historical related data of the feature operation data, determining a historical operation state according to the historical related data, and obtaining multiple types of feature operation data; analyzing the feature operation data of the same category and the comprehensive evaluation value, and determining optimal operation data; extracting feature control data of the optimal operation data of the same category, and analyzing the feature control data to obtain optimal control data; an optimal control data prediction model is constructed according to the multiple types of optimal control data, predicted optimal control data is determined, whether an adjustment instruction is generated or not is judged according to the predicted optimal control data, power plant personnel are helped to adjust a control strategy in time, the overall goals of saving energy, reducing consumption and improving performance efficiency are achieved, and the production benefits of a power plant are increased.

Description

Technical Field

[0001] The present application relates to the technical field of unit control, and in particular to a combined cycle unit control method based on data fusion. Background Art

[0002] The combined cycle unit adopts a "two-to-one" operation mode, that is, two gas turbines drive one steam turbine, and the exhaust gas of the gas turbine is introduced into the waste heat boiler. The high-temperature and high-pressure steam generated drives the steam turbine to drive the steam turbine to generate electricity. During the operation of the combined cycle unit, a large amount of data affects the energy efficiency. In order to improve the operating efficiency of the combined cycle unit and achieve the overall goal of energy saving and consumption reduction, a combined cycle unit control method based on data fusion is proposed to help power plant personnel adjust the control strategy in time and increase the production revenue of the power plant. Summary of the invention

[0003] To solve the above technical problems, the present application provides a combined cycle unit control method based on data fusion, which determines the characteristic operating data by the degree of correlation between the historical operating data and the comprehensive evaluation index, classifies the characteristic operating data according to the historical related data, obtains characteristic operating data groups of multiple categories, and determines the preferred operating data and optimal control data of the characteristic operating data group of each category, constructs an optimal control data prediction model based on the optimal control data of multiple categories, and determines the predicted optimal control data, determines whether to generate adjustment instructions based on the predicted optimal control data and real-time control data, and helps power plant personnel to adjust the control strategy in time, so as to achieve the overall goal of energy saving and consumption reduction and improving performance efficiency, and increase the production efficiency of the power plant.

[0004] In some embodiments of the present application, a combined cycle unit control method based on data fusion is provided, comprising:

[0005] Obtain historical operation data of the combined cycle unit in multiple historical periods, and determine multiple characteristic operation data according to the degree of correlation between the historical operation data and the comprehensive evaluation index;

[0006] Acquire historical related data of the characteristic operation data, determine the historical operation status of the characteristic operation data according to the historical related data, and classify the characteristic operation data according to the historical operation status to obtain characteristic operation data of multiple categories;

[0007] Analyze the characteristic operation data of the same category and the corresponding comprehensive evaluation values ​​to determine multiple preferred operation data of each category;

[0008] Extracting characteristic control data of preferred operation data of the same category based on a preset variation coefficient, analyzing the characteristic control data of the same category, and obtaining optimal control data of the corresponding category;

[0009] An optimal control data prediction model is constructed according to multiple categories of optimal control data, predicted optimal control data is determined based on the optimal control data prediction model, and whether to generate an adjustment instruction is determined according to the predicted optimal control data.

[0010] In some embodiments of the present application, a plurality of characteristic operation data are determined, including:

[0011] The comprehensive evaluation index includes an energy-saving index and a plurality of preset efficiency indexes;

[0012] Determine the energy-saving index and the first evaluation values ​​of the plurality of preset efficiency indexes based on the historical management report corresponding to the historical operation data for the historical period, and generate the comprehensive evaluation value of the comprehensive evaluation index according to the energy-saving index and the first evaluation values ​​of the plurality of preset efficiency indexes and the corresponding weight coefficients;

[0013] With the historical operation data in the same historical period as the horizontal axis and the comprehensive evaluation value of the comprehensive evaluation index as the vertical axis, the correlation curves in multiple historical periods are constructed;

[0014] Determine the correlation coefficient between each historical operation data and the comprehensive evaluation index according to multiple correlation relationship curves, and generate the correlation degree between the corresponding historical operation data and the comprehensive evaluation index according to the number of correlation coefficients between the same historical operation data and the comprehensive evaluation index in multiple historical periods that are greater than a preset correlation coefficient threshold and the correlation coefficient difference that is greater than the preset correlation coefficient threshold;

[0015] The historical operation data with a correlation degree greater than a preset correlation degree threshold is set as characteristic operation data.

[0016] In some embodiments of the present application, the calculation formula of the degree of association is:

[0017] ;

[0018] Among them, G is the degree of correlation, g0 is the correlation degree conversion coefficient, m is the number of historical operation data in multiple historical periods and the correlation coefficient between the comprehensive evaluation index is greater than the preset correlation coefficient threshold, n is the total number of historical periods, Xi is the correlation coefficient between the historical operation data in the i-th historical period and the comprehensive evaluation index, and X0 is the preset correlation coefficient threshold.

[0019] In some embodiments of the present application, multiple categories of characteristic operation data are obtained, including:

[0020] Acquire historical data related to characteristic operation data, wherein the historical data related to characteristic operation data includes historical environment data, historical load data and historical operating condition data;

[0021] Compare the historical environmental data, historical load data and historical operating condition data with the corresponding preset standard data intervals, and generate the historical operating status of the corresponding characteristic operating data according to the comparison results;

[0022] According to the relationship between the historical operating state and the preset standard operating state, the state deviation of the characteristic operating data is determined, and the characteristic operating data corresponding to the state deviation within the preset fluctuation range is set as the characteristic operating data of the same category.

[0023] In some embodiments of the present application, extracting characteristic control data of preferred operation data of the same category based on a preset variation coefficient includes:

[0024] Presetting a comprehensive evaluation value threshold;

[0025] The characteristic operation data whose comprehensive evaluation value is greater than the comprehensive evaluation value threshold is set as the preferred operation data of the corresponding category;

[0026] According to the change relationship between the preferred operation data and the historical control data in the corresponding historical period, determine whether there is a dependency relationship between the historical control data and the corresponding preferred operation data, and if so, calculate the dependency change coefficient between the historical control data and the corresponding preferred operation data;

[0027] The historical control data corresponding to the priority operation data of the same category in the preferred operation data matrix are screened according to the preset change coefficient, and the historical control data with a dependency change coefficient greater than the preset change coefficient is set as the characteristic control data.

[0028] In some embodiments of the present application, characteristic control data of the same category are analyzed to obtain optimal control data of the corresponding category, including:

[0029] According to the correlation degree between the characteristic control data of the same category and the energy-saving index and the multiple preset efficiency indexes, determining the strongly correlated characteristic control data of the energy-saving index and the multiple preset efficiency indexes;

[0030] Generate a ranking result of the first evaluation value of the characteristic operation data of the strongly associated characteristic control data of the same indicator in the corresponding historical time period, and set the first ranked strongly associated characteristic control data as the preferred control data of the corresponding indicator;

[0031] Determine whether the preferred control data of the energy-saving index and multiple preset efficiency indexes at the same historical time node are repeated. If repeated, fuse the repeated preferred control data to obtain fused control data and replace the corresponding preferred control data. Set the fused control data and the non-repeated preferred control data as the optimal control data of the corresponding category;

[0032] If there is no duplication, the preferred control data of the corresponding indicator is set as the optimal control data of the corresponding category.

[0033] In some embodiments of the present application, obtaining fused control data and replacing corresponding preferred control data includes:

[0034] If the preferred control data of the energy-saving index and the plurality of preset efficiency indexes are repeated, extracting the repeated control duration of the repeated preferred control data;

[0035] A time reference line is established based on the duration of repeated control, and analysis nodes are set based on preset time intervals. Preferred control data of different indicators that appear repeatedly are collected according to each analysis node on the time reference line, and mapped to the time reference line to generate a data change curve of the preferred control data of the corresponding indicator;

[0036] Obtaining the relative distance between the preferred control data of two different indicators at each analysis node and the relative distance threshold of the preferred control data of all different indicators, and eliminating the preferred control data whose relative distance is greater than the relative distance threshold;

[0037] The remaining optimal control data at each analysis node are fused according to the weight coefficient of the corresponding indicator to obtain the fused control data at each analysis node;

[0038] Calculate the fluctuation degree of the fusion control data, and if the fluctuation degree is greater than a preset fluctuation degree threshold, fine-tune the fusion control data;

[0039] The fine-tuned fusion control data replaces the preferred control data within the repeated control duration.

[0040] In some embodiments of the present application, an optimal control data prediction model is constructed according to multiple categories of optimal control data, including:

[0041] Generate a training data set and a test data set based on the optimal control data of each category and the corresponding historical related data;

[0042] According to the optimal control data of each category as training output data, the historical related data of the characteristic operation data of each category is averaged and used as training input data to perform neural network training to obtain the initial prediction model;

[0043] Obtain the credibility of the initial prediction model based on the test data set;

[0044] Pre-set credibility threshold;

[0045] If the credibility is greater than the credibility threshold, the initial prediction model is set as the optimal control data prediction model;

[0046] If the credibility is less than the credibility threshold, the training data set is iteratively trained and the initial prediction model is re-obtained until the credibility is greater than the credibility threshold.

[0047] In some embodiments of the present application, determining whether to generate an adjustment instruction according to the predicted optimal control data includes:

[0048] Acquire real-time relevant data, and input the real-time relevant data into an optimal control data prediction model to obtain predicted optimal control data;

[0049] Compare the predicted optimal control data with the corresponding real-time control data, and generate a control data deviation amount according to the comparison result;

[0050] If the control data deviation is greater than a preset deviation threshold, a corresponding adjustment coefficient is set and an adjustment instruction is generated.

[0051] Compared with the prior art, the combined cycle unit control method based on data fusion in the embodiment of the present application has the following beneficial effects:

[0052] Characteristic operation data is determined by the degree of correlation between historical operation data and comprehensive evaluation indicators. The characteristic operation data is classified according to historical related data to obtain characteristic operation data groups of multiple categories, and the preferred operation data and optimal control data of the characteristic operation data groups of each category are determined. An optimal control data prediction model is constructed based on the optimal control data of multiple categories, and the predicted optimal control data is determined. It is determined whether to generate adjustment instructions based on the predicted optimal control data and real-time control data, so as to help power plant personnel adjust control strategies in a timely manner, achieve the overall goal of energy saving and consumption reduction and improving performance efficiency, and increase the production benefits of power plants. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of a combined cycle unit control method based on data fusion in the preferred embodiment of the present application. DETAILED DESCRIPTION

[0054] The specific implementation methods of the present application are further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present application but are not intended to limit the scope of the present application.

[0055] In the description of the present application, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply 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 understood as a limitation on the present application.

[0056] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0057] In the description of this 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 can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0058] like Figure 1 As shown, a combined cycle unit control method based on data fusion in a preferred embodiment of the present application includes:

[0059] Step S101: acquiring historical operation data of a combined cycle unit in multiple historical periods, and determining multiple characteristic operation data according to the degree of correlation between the historical operation data and the comprehensive evaluation index;

[0060] Step S102: acquiring historical related data of the characteristic operation data, determining the historical operation status of the characteristic operation data according to the historical related data, and classifying the characteristic operation data according to the historical operation status to obtain characteristic operation data of multiple categories;

[0061] Step S103: analyzing the characteristic operation data of the same category and the corresponding comprehensive evaluation values ​​to determine a plurality of preferred operation data of each category;

[0062] Step S104: extracting characteristic control data of the preferred operation data of the same category based on the preset variation coefficient, analyzing the characteristic control data of the same category, and obtaining optimal control data of the corresponding category;

[0063] Step S105: construct an optimal control data prediction model according to multiple categories of optimal control data, determine predicted optimal control data based on the optimal control data prediction model, and determine whether to generate an adjustment instruction according to the predicted optimal control data.

[0064] In this embodiment, the historical operation data includes data such as GIS generator power, natural gas pressure regulating station flow, natural gas chromatograph, unit main transformer power, circulating water flow, NCS main transformer power, etc. The historical related data includes historical environmental data, historical load data and historical operating condition data. The historical operation status of the characteristic operation data is determined based on the historical related data and the characteristic operation data is classified. The historical control data of each category of characteristic operation data is analyzed to determine the optimal control data.

[0065] In this embodiment, an optimal control data prediction model is constructed through multiple categories of optimal control data, and real-time related data is analyzed to obtain predicted optimal control data, which helps power plant personnel to adjust control strategies in a timely manner, achieve the overall goals of energy saving and consumption reduction and improving performance efficiency, and increase the production benefits of the power plant.

[0066] In some embodiments of the present application, a plurality of characteristic operation data are determined, including:

[0067] The comprehensive evaluation index includes an energy-saving index and a plurality of preset efficiency indexes;

[0068] Determine the energy-saving index and the first evaluation values ​​of the plurality of preset efficiency indexes based on the historical management report corresponding to the historical operation data for the historical period, and generate the comprehensive evaluation value of the comprehensive evaluation index according to the energy-saving index and the first evaluation values ​​of the plurality of preset efficiency indexes and the corresponding weight coefficients;

[0069] With the historical operation data in the same historical period as the horizontal axis and the comprehensive evaluation value of the comprehensive evaluation index as the vertical axis, the correlation curves in multiple historical periods are constructed;

[0070] Determine the correlation coefficient between each historical operation data and the comprehensive evaluation index according to multiple correlation relationship curves, and generate the correlation degree between the corresponding historical operation data and the comprehensive evaluation index according to the number of correlation coefficients between the same historical operation data and the comprehensive evaluation index in multiple historical periods that are greater than a preset correlation coefficient threshold and the correlation coefficient difference that is greater than the preset correlation coefficient threshold;

[0071] The historical operation data with a correlation degree greater than a preset correlation degree threshold is set as characteristic operation data.

[0072] In this embodiment, the preset efficiency indicators include a gas turbine efficiency index, a steam turbine efficiency index and an auxiliary machine efficiency index. The first evaluation value refers to the energy-saving effect evaluation value or the operating efficiency evaluation value of the energy-saving index or the preset efficiency index. The comprehensive evaluation value refers to the comprehensive evaluation value of the energy-saving index and multiple preset efficiency indicators.

[0073] In this embodiment, the correlation coefficient refers to the degree of synchronous change between the historical operating data and the comprehensive evaluation value. If the degree of synchronous change is greater than the preset synchronous change degree threshold, there is a correlation between the historical operating data and the comprehensive evaluation index, and the corresponding correlation coefficient is set according to the difference between the degree of synchronous change and the preset synchronous change degree threshold. The larger the difference, the larger the correlation coefficient, and the corresponding degree of correlation is generated.

[0074] In this embodiment, by screening out characteristic operation data that has a high degree of correlation with the comprehensive evaluation index, the amount of data processing and analysis is reduced, the analysis efficiency is improved, and a data foundation is laid for the subsequent determination of historical operation status and classification.

[0075] In some embodiments of the present application, the calculation formula of the degree of association is:

[0076]

[0077] Among them, G is the degree of correlation, g0 is the correlation degree conversion coefficient, m is the number of historical operation data in multiple historical periods and the correlation coefficient between the comprehensive evaluation index is greater than the preset correlation coefficient threshold, n is the total number of historical periods, Xi is the correlation coefficient between the historical operation data in the i-th historical period and the comprehensive evaluation index, and X0 is the preset correlation coefficient threshold.

[0078] In some embodiments of the present application, multiple categories of characteristic operation data are obtained, including:

[0079] Acquire historical data related to characteristic operation data, wherein the historical data related to characteristic operation data includes historical environment data, historical load data and historical operating condition data;

[0080] Compare the historical environmental data, historical load data and historical operating condition data with the corresponding preset standard data intervals, and generate the historical operating status of the corresponding characteristic operating data according to the comparison results;

[0081] According to the relationship between the historical operating state and the preset standard operating state, the state deviation of the characteristic operating data is determined, and the characteristic operating data corresponding to the state deviation within the preset fluctuation range is set as the characteristic operating data of the same category.

[0082] In this embodiment, the operating status of the combined cycle unit under different environments, different loads and different operating conditions is determined based on historical environmental data, historical load data and historical operating condition data, and the characteristic operating data is classified according to different operating statuses to obtain the optimal control data for each category, thereby helping power plant personnel to adjust the control strategy in a timely manner and increase the production efficiency of the power plant.

[0083] In some embodiments of the present application, extracting characteristic control data of preferred operation data of the same category based on a preset variation coefficient includes:

[0084] Presetting a comprehensive evaluation value threshold;

[0085] The characteristic operation data whose comprehensive evaluation value is greater than the comprehensive evaluation value threshold is set as the preferred operation data of the corresponding category;

[0086] According to the change relationship between the preferred operation data and the historical control data in the corresponding historical period, determine whether there is a dependency relationship between the historical control data and the corresponding preferred operation data, and if so, calculate the dependency change coefficient between the historical control data and the corresponding preferred operation data;

[0087] The historical control data corresponding to the priority operation data of the same category in the preferred operation data matrix are screened according to the preset change coefficient, and the historical control data with a dependency change coefficient greater than the preset change coefficient is set as the characteristic control data.

[0088] In this embodiment, the dependency change coefficient refers to the change of the corresponding preferred operating data within a preset time period after the historical control data changes. In this case, the preferred operating data depends on the corresponding historical control data, and the dependency change coefficient is set according to the difference between the degree of change and the preset degree of change threshold. The larger the difference, the larger the dependency change coefficient.

[0089] In this embodiment, by analyzing the historical control data of the preferred operating data of the same category, the historical control data that has a dependency relationship with the preferred operating data is screened out, the data processing and analysis volume is reduced, the analysis efficiency of the optimal control data and the control timeliness are improved, and the power plant personnel are helped to adjust the control strategy in time to increase the production efficiency of the power plant.

[0090] In some embodiments of the present application, characteristic control data of the same category are analyzed to obtain optimal control data of the corresponding category, including:

[0091] According to the correlation degree between the characteristic control data of the same category and the energy-saving index and the multiple preset efficiency indexes, determining the strongly correlated characteristic control data of the energy-saving index and the multiple preset efficiency indexes;

[0092] Generate a ranking result of the first evaluation value of the characteristic operation data of the strongly associated characteristic control data of the same indicator in the corresponding historical time period, and set the first ranked strongly associated characteristic control data as the preferred control data of the corresponding indicator;

[0093] Determine whether the preferred control data of the energy-saving index and multiple preset efficiency indexes at the same historical time node are repeated. If repeated, fuse the repeated preferred control data to obtain fused control data and replace the corresponding preferred control data. Set the fused control data and the non-repeated preferred control data as the optimal control data of the corresponding category;

[0094] If there is no duplication, the preferred control data of the corresponding indicator is set as the optimal control data of the corresponding category.

[0095] In this embodiment, based on the degree of correlation between the characteristic control data of the same category and the energy-saving index and multiple preset efficiency indicators, the strongly correlated characteristic control data corresponding to the energy-saving index and multiple preset efficiency indicators are determined. The strongly correlated characteristic control data refers to characteristic control data whose degree of correlation with the energy-saving index or multiple preset efficiency indicators is greater than the correlation degree threshold.

[0096] In this embodiment, the strongly correlated feature control data corresponding to the largest first evaluation value in each indicator is set as the preferred control data of the corresponding indicator, and whether repeated historical time nodes appear is determined according to the control duration of the preferred control data of each indicator. If repetition occurs, the repeated preferred control data are fused to obtain fused control data.

[0097] In this embodiment, based on the degree of correlation between characteristic control data of the same category and energy-saving indicators and multiple preset efficiency indicators, strongly correlated characteristic control data and preferred control data of each indicator are obtained. If the preferred control data are repeated, they are fused to obtain the optimal control data of each category, laying the foundation for the subsequent construction of the optimal control data prediction model, helping power plant personnel to adjust the control strategy in a timely manner, achieving the overall goal of energy saving and consumption reduction and improving performance efficiency, and increasing the production efficiency of the power plant.

[0098] In some embodiments of the present application, obtaining fused control data and replacing corresponding preferred control data includes:

[0099] If the preferred control data of the energy-saving index and the plurality of preset efficiency indexes are repeated, extracting the repeated control duration of the repeated preferred control data;

[0100] A time reference line is established based on the duration of repeated control, and analysis nodes are set based on preset time intervals. Preferred control data of different indicators that appear repeatedly are collected according to each analysis node on the time reference line, and mapped to the time reference line to generate a data change curve of the preferred control data of the corresponding indicator;

[0101] Obtaining the relative distance between the preferred control data of two different indicators at each analysis node and the relative distance threshold of the preferred control data of all different indicators, and eliminating the preferred control data whose relative distance is greater than the relative distance threshold;

[0102] The remaining optimal control data at each analysis node are fused according to the weight coefficient of the corresponding indicator to obtain the fused control data at each analysis node;

[0103] Calculate the fluctuation degree of the fusion control data, and if the fluctuation degree is greater than a preset fluctuation degree threshold, fine-tune the fusion control data;

[0104] The fine-tuned fusion control data replaces the preferred control data within the repeated control duration.

[0105] In this embodiment, data fusion refers to the center point of the remaining preferred control data at the same analysis node being determined according to the weight coefficient of the corresponding indicator, and the relative distance between the center point and the remaining preferred control data satisfies the weight coefficient of the corresponding indicator.

[0106] In this embodiment, the preferred control data at the same historical time node is adjusted through data fusion to obtain fused control data, thereby maximally meeting the energy-saving indicators and the operating goals of multiple preset efficiency indicators, determining the optimal control data, and laying the foundation for the subsequent construction of the optimal control data prediction model, helping power plant personnel to adjust the control strategy in a timely manner and increase the production efficiency of the power plant.

[0107] In some embodiments of the present application, an optimal control data prediction model is constructed according to multiple categories of optimal control data, including:

[0108] Generate a training data set and a test data set based on the optimal control data of each category and the corresponding historical related data;

[0109] According to the optimal control data of each category as training output data, the historical related data of the characteristic operation data of each category is averaged and used as training input data to perform neural network training to obtain the initial prediction model;

[0110] Obtain the credibility of the initial prediction model based on the test data set;

[0111] Pre-set credibility threshold;

[0112] If the credibility is greater than the credibility threshold, the initial prediction model is set as the optimal control data prediction model;

[0113] If the credibility is less than the credibility threshold, the training data set is iteratively trained and the initial prediction model is re-obtained until the credibility is greater than the credibility threshold.

[0114] In some embodiments of the present application, determining whether to generate an adjustment instruction according to the predicted optimal control data includes:

[0115] Acquire real-time relevant data, and input the real-time relevant data into an optimal control data prediction model to obtain predicted optimal control data;

[0116] Compare the predicted optimal control data with the corresponding real-time control data, and generate a control data deviation amount according to the comparison result;

[0117] If the control data deviation is greater than a preset deviation threshold, a corresponding adjustment coefficient is set and an adjustment instruction is generated.

[0118] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present application. These improvements and substitutions should also be regarded as the scope of protection of the present application.

Claims

1. A combined cycle unit control method based on data fusion, characterized in that: include: Obtain historical operation data of the combined cycle unit in multiple historical periods, and determine multiple characteristic operation data according to the degree of correlation between the historical operation data and the comprehensive evaluation index; Acquire historical related data of the characteristic operation data, determine the historical operation status of the characteristic operation data according to the historical related data, and classify the characteristic operation data according to the historical operation status to obtain characteristic operation data of multiple categories; Analyze the characteristic operation data of the same category and the corresponding comprehensive evaluation values ​​to determine multiple preferred operation data of each category; Extracting characteristic control data of preferred operation data of the same category based on a preset variation coefficient, analyzing the characteristic control data of the same category, and obtaining optimal control data of the corresponding category; An optimal control data prediction model is constructed according to multiple categories of optimal control data, predicted optimal control data is determined based on the optimal control data prediction model, and whether to generate an adjustment instruction is determined according to the predicted optimal control data.

2. The combined cycle unit control method based on data fusion according to claim 1, characterized in that: Determine multiple characteristic operating data, including: The comprehensive evaluation index includes an energy-saving index and a plurality of preset efficiency indexes; Determine the energy-saving index and the first evaluation values ​​of the plurality of preset efficiency indexes based on the historical management report corresponding to the historical operation data for the historical period, and generate the comprehensive evaluation value of the comprehensive evaluation index according to the energy-saving index and the first evaluation values ​​of the plurality of preset efficiency indexes and the corresponding weight coefficients; With the historical operation data in the same historical period as the horizontal axis and the comprehensive evaluation value of the comprehensive evaluation index as the vertical axis, the correlation curves in multiple historical periods are constructed; Determine the correlation coefficient between each historical operation data and the comprehensive evaluation index according to multiple correlation relationship curves, and generate the correlation degree between the corresponding historical operation data and the comprehensive evaluation index according to the number of correlation coefficients between the same historical operation data and the comprehensive evaluation index in multiple historical periods that are greater than a preset correlation coefficient threshold and the correlation coefficient difference that is greater than the preset correlation coefficient threshold; The historical operation data with a correlation degree greater than a preset correlation degree threshold is set as characteristic operation data.

3. The combined cycle unit control method based on data fusion according to claim 2, characterized in that: The calculation formula of the degree of association is: Among them, G is the degree of correlation, g0 is the correlation degree conversion coefficient, m is the number of historical operation data in multiple historical periods and the correlation coefficient between the comprehensive evaluation index is greater than the preset correlation coefficient threshold, n is the total number of historical periods, Xi is the correlation coefficient between the historical operation data in the i-th historical period and the comprehensive evaluation index, and X0 is the preset correlation coefficient threshold.

4. The combined cycle unit control method based on data fusion according to claim 3, characterized in that: Get multiple categories of feature operation data, including: Acquire historical data related to characteristic operation data, wherein the historical data related to characteristic operation data includes historical environment data, historical load data and historical operating condition data; Compare the historical environmental data, historical load data and historical operating condition data with the corresponding preset standard data intervals, and generate the historical operating status of the corresponding characteristic operating data according to the comparison results; According to the relationship between the historical operating state and the preset standard operating state, the state deviation of the characteristic operating data is determined, and the characteristic operating data corresponding to the state deviation within the preset fluctuation range is set as the characteristic operating data of the same category.

5. The combined cycle unit control method based on data fusion according to claim 4, characterized in that: Extracting characteristic control data of preferred operation data of the same category based on a preset variation coefficient includes: Presetting a comprehensive evaluation value threshold; The characteristic operation data whose comprehensive evaluation value is greater than the comprehensive evaluation value threshold is set as the preferred operation data of the corresponding category; According to the change relationship between the preferred operation data and the historical control data in the corresponding historical period, determine whether there is a dependency relationship between the historical control data and the corresponding preferred operation data, and if so, calculate the dependency change coefficient between the historical control data and the corresponding preferred operation data; The historical control data corresponding to the priority operation data of the same category in the preferred operation data matrix are screened according to the preset change coefficient, and the historical control data with a dependency change coefficient greater than the preset change coefficient is set as the characteristic control data.

6. The combined cycle unit control method based on data fusion according to claim 5, characterized in that: Analyze the characteristic control data of the same category to obtain the optimal control data of the corresponding category, including: According to the correlation degree between the characteristic control data of the same category and the energy-saving index and the multiple preset efficiency indexes, determining the strongly correlated characteristic control data of the energy-saving index and the multiple preset efficiency indexes; Generate a ranking result of the first evaluation value of the characteristic operation data of the strongly associated characteristic control data of the same indicator in the corresponding historical time period, and set the first ranked strongly associated characteristic control data as the preferred control data of the corresponding indicator; Determine whether the preferred control data of the energy-saving index and multiple preset efficiency indexes at the same historical time node are repeated. If repeated, fuse the repeated preferred control data to obtain fused control data and replace the corresponding preferred control data. Set the fused control data and the non-repeated preferred control data as the optimal control data of the corresponding category; If there is no duplication, the preferred control data of the corresponding indicator is set as the optimal control data of the corresponding category.

7. The combined cycle unit control method based on data fusion according to claim 6, characterized in that: Obtaining fused control data and replacing corresponding preferred control data includes: If the preferred control data of the energy-saving index and the plurality of preset efficiency indexes are repeated, extracting the repeated control duration of the repeated preferred control data; A time reference line is established based on the duration of repeated control, and analysis nodes are set based on preset time intervals. Preferred control data of different indicators that appear repeatedly are collected according to each analysis node on the time reference line, and mapped to the time reference line to generate a data change curve of the preferred control data of the corresponding indicator; Obtaining the relative distance between the preferred control data of two different indicators at each analysis node and the relative distance threshold of the preferred control data of all different indicators, and eliminating the preferred control data whose relative distance is greater than the relative distance threshold; The remaining optimal control data at each analysis node are fused according to the weight coefficient of the corresponding indicator to obtain the fused control data at each analysis node; Calculate the fluctuation degree of the fusion control data, and if the fluctuation degree is greater than a preset fluctuation degree threshold, fine-tune the fusion control data; The fine-tuned fusion control data replaces the preferred control data within the repeated control duration.

8. The combined cycle unit control method based on data fusion according to claim 7, characterized in that: The optimal control data prediction model is constructed based on multiple categories of optimal control data, including: Generate a training data set and a test data set based on the optimal control data of each category and the corresponding historical related data; According to the optimal control data of each category as training output data, the historical related data of the characteristic operation data of each category is averaged and used as training input data to perform neural network training to obtain the initial prediction model; Obtain the credibility of the initial prediction model based on the test data set; Pre-set credibility threshold; If the credibility is greater than the credibility threshold, the initial prediction model is set as the optimal control data prediction model; If the credibility is less than the credibility threshold, the training data set is iteratively trained and the initial prediction model is re-obtained until the credibility is greater than the credibility threshold.

9. The combined cycle unit control method based on data fusion according to claim 8, characterized in that: Determining whether to generate an adjustment instruction based on the predicted optimal control data includes: Acquire real-time relevant data, and input the real-time relevant data into an optimal control data prediction model to obtain predicted optimal control data; Compare the predicted optimal control data with the corresponding real-time control data, and generate a control data deviation amount according to the comparison result; If the control data deviation is greater than a preset deviation threshold, a corresponding adjustment coefficient is set and an adjustment instruction is generated.

Images