Electric power thermal power generating unit operation method

Abstract

The invention relates to the field of electric power thermal power generating unit operation control, in particular to an electric power thermal power generating unit operation method. Screening basic data in the historical operation data according to the stability data, then selecting a plurality of groups of data with the maximum total relevancy according to the total relevancy among the groups of basic data to obtain standard data, and then performing working condition division on the standard data according to variable types and variable data to obtain a plurality of first working conditions; then, abnormal standard data are removed through a Gaussian mixture model to obtain a third working condition, so that an operation parameter library is established according to the third working condition, and operation parameters of all devices are configured according to the current working condition of the electric power thermal power generating unit; the problem that in the prior art, redundant data in the obtained operation data of all the devices are difficult to accurately screen out, so that when the operation parameter library is automatically generated and updated to control operation of all the devices in the electric power thermal power generating unit, preset requirements are difficult to meet is solved.

Description

Technical Field

[0001] The invention relates to the field of operation control of electric thermal power generation units, and in particular to an operation method of an electric thermal power generation unit. Background Art

[0002] At present, by obtaining data of various equipment in the power thermal power unit under different operating conditions, and then establishing corresponding objective functions, such as establishing objective functions with low carbon emissions and high efficiency, it is possible to obtain the operating parameter library of various equipment in the power thermal power unit under different operating conditions. Then, according to the actual operating conditions of the power thermal power unit during operation, the configuration parameters of each device are called according to the operating parameter library for configuration and operation.

[0003] However, due to the complex operation of thermal power units, the operation conditions of various devices, the working conditions and whether the operation is stable are different. According to statistics, the redundant data obtained that is not suitable for the establishment of an operation parameter library is approximately 10%. When the existing technology obtains such redundant data for comprehensive analysis, it is difficult to accurately screen out such unsuitable redundant data, which makes it difficult to meet the predetermined needs when automatically generating and updating the operation parameter library to control the operation of various devices in the thermal power units. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a method for operating a thermal power unit, which solves the problem that the prior art is difficult to accurately screen out redundant data in the acquired operating data of each device, resulting in difficulty in meeting predetermined needs when automatically generating and updating an operating parameter library to control the operation of each device in the thermal power unit.

[0005] To achieve the above object, the present invention provides the following technical solutions: A method for operating a thermal power unit, the method specifically comprising the following steps: S1. Obtain the historical operation data of each device in the thermal power unit; S2. Based on historical data, basic data containing key factors and within the corresponding stability threshold at each moment are obtained, and several groups of standard data are determined with the goal of maximizing the total correlation; S3, setting variable types and collecting variable data according to variables corresponding to the standard data operation of the thermal power unit, and dividing the standard data operation conditions according to the variable data to obtain a plurality of first conditions; S4, establishing a Gaussian mixture model with convergent log-likelihood function according to each first operating condition, and identifying and removing abnormal standard data in the first operating condition to obtain a third operating condition; S5. Establish an objective function according to the third operating condition and output the corresponding optimal parameters to establish an operating parameter library, obtain the current operating condition of the thermal power unit and control each module of the thermal power unit to operate according to the optimal parameter configuration according to the operating parameter library.

[0006] Preferably, in step S2, the following steps are specifically included: S21, arranging the historical operation data in chronological order to obtain a plurality of historical data strings; S22, establishing a filtering window centered on each moment on the historical data string, where the filtering window spans across several moments; S23, judging whether the stability data of key factors at each moment are all within the corresponding stability threshold, where the key factors include active power, main steam pressure, main steam temperature, reheat steam temperature and feed water flow; If so, extract the historical operation data in the time period included in the filter window to obtain several groups of basic data; If not, the historical operation data in the historical data string is not extracted; S24, calculating the average values ​​of any two groups of basic data in sequence, and calculating the first correlation between the two groups; S25, calculating the mean value and standard deviation of any two groups of basic data in sequence, and calculating the second correlation between the two; S26, calculating the difference between the data values ​​of corresponding data in any two sets of basic data, and calculating the inverse of the standard deviation of the difference between the data values, to obtain a third correlation between the any two sets of basic data; S27, calculating the total correlation of any two sets of basic data according to the first correlation, the second correlation and the third correlation; ; In the above formula, represents the total correlation between the k1th group of basic data and the k2th group of basic data, , and Represent the first correlation, the second correlation and the third correlation respectively, , and denote the first correlation coefficient, the second correlation coefficient and the third correlation coefficient respectively; S28. Set a correlation threshold, and select several groups of basic data whose total correlation is greater than the correlation threshold as standard data.

[0007] Preferably, in step S26, the following steps are specifically included: S261, calculating the difference between the data values ​​of corresponding data in any two sets of basic data; S262. Calculate the inverse of the standard deviation of the difference between the data values ​​to obtain a third correlation between any two sets of basic data.

[0008] Preferably, in step S3, the following steps are specifically included: S31, obtaining several variable types and variable data of corresponding types that affect standard data, and establishing a variable data set including the variable types and variable data; S32, marking the standard data as first standard data; S33, selecting any variable type and variable data of a corresponding type from the variable data set, and dividing the operating conditions of the first standard data by a K-means clustering algorithm to obtain a plurality of first operating conditions; S34, removing the selected variable type and variable data of the corresponding type from the variable data set; S35, determining whether the variable data set is an empty set; If yes, the first working condition is output and the process goes to step S4; If not, proceed to step S36; S36. Select any variable type and variable data of the corresponding type from the variable data set, and divide the first operating condition into several second operating conditions through the K-means clustering algorithm, then mark the second operating condition as the first operating condition, and then return to step S34.

[0009] Preferably, in step S4, the following steps are specifically included: S41, establishing a Gaussian mixture model according to a plurality of first working conditions; S42, establishing a log-likelihood function according to the Gaussian mixture model, and determining whether the log-likelihood function converges; If yes, proceed to step S43; If not, proceed to step S47; S43, calculating the probability density of any standard data from each first working condition according to the Gaussian mixture model; the calculation formula of the probability density of the first working condition is: ; In the above formula, Indicates the jth standard data The probability density from the kth first operating condition, is the Gaussian distribution density function of the kth first operating condition, is the sample mean of the standard data in the kth first working condition, is the covariance of the kth first operating condition, is the weight of the Gaussian distribution of the kth first working condition, and M is the total number of first working conditions; S44, calculating the density radius according to the probability density of the standard data in the first working condition; the calculation formula of the density radius is: ; In the above formula, represents the density radius of the kth first operating condition, Indicates the jth standard data in the first working condition The probability density from the kth first operating condition, represents the number of standard data in the kth first working condition, represents the radius coefficient; S45, calculating the density center of each first working condition; S46, sequentially calculating the numerical difference between each data in the first working condition and the density center, and determining whether the numerical difference is less than or equal to the density radius; If yes, the data is marked as the third operating condition, and then the process goes to step S5; If not, the data is marked as abnormal standard data; S47, optimize the first operating condition until the log-likelihood function converges, and then return to step S43.

[0010] Preferably, in step S45, the following steps are specifically included: S451, calculating the initial statistical radius of each first working condition; the calculation formula of the initial statistical radius is: ; In the above formula, represents the initial statistical radius of any first operating condition, and They represent the maximum and minimum values ​​of the data in the first working condition, respectively, and C is the reduction coefficient, 0<C<1; S452, setting the initial number of iterations to 1, the initial statistical radius, and the initial statistical area to the first working condition; S453, obtaining the current number of iterations, the current statistical radius and the current statistical area; S454, calculating the density of any standard data in the current statistical area according to the current number of iterations and the statistical radius; the calculation formula of the density is: ; in, ; In the above formula, Indicates the density of the i-th standard data in the current statistical area, Indicates the number of standard data in the current statistical area. is the symbolic function, and Respectively represent the i-th and j-th data in the current statistical area, represents the statistical radius at the dth iteration; S455, obtaining a plurality of standard data with the largest density, and determining whether the current plurality of standard data with the largest density and the statistical area are the same as those in the previous iteration; If yes, proceed to step S458; If not, the current number of iterations is increased by 1, and then the process goes to step S456; S456, marking the standard data whose data value difference with the several standard data with the largest density is less than the current statistical radius as the current statistical area; S457, calculate the current statistical radius according to the current number of iterations, and then return to step S453; the calculation formula of the current statistical radius is: ; In the above formula, Indicates the current statistical radius, C is the reduction coefficient, 0<C<1, Indicates the statistical radius at the last iteration; S458. Calculate the average value of several standard data with the largest density to obtain the density center of the corresponding first working condition.

[0011] Preferably, in step S47, the following steps are specifically included: S471, calculating and obtaining the probability density of any standard data from each first working condition according to step S43; S472, update the parameters in the Gaussian mixture model according to the calculated probability density, and then return to step S41; the update formula for updating the parameters in the Gaussian mixture model is as follows: ; In the above formula, Indicates the jth standard data The probability density of the first working condition of k, the total number of standard data is J, is the sample mean of the standard data in the kth first working condition, is the covariance of the kth first operating condition, is the weight of the Gaussian distribution of the kth first operating condition.

[0012] Preferably, in step S24, the calculation formula of the first correlation is: ; In the above formula, represents the first correlation between the k1th group of basic data and the k2th group of basic data, and Respectively represent the i-th data of the k1-th group and the k2-th group of basic data, and They respectively represent the average values ​​of the k1th group of basic data and the k2th group of basic data, and the number of data in the k1th group and the k2th group of basic data is n.

[0013] Preferably, in step S25, the calculation formula of the second correlation is: ; in, ; In the above formula, represents the second correlation between the k1th group of basic data and the k2th group of basic data, represents the unit correlation between the i-th data in the k1-th group of basic data and the k2-th group of basic data, and Respectively represent the i-th data of the k1-th group and the k2-th group of basic data, Differentiation coefficient .

[0014] Preferably, in step S262, the calculation formula of the third correlation is: ; in, ; In the above formula, represents the third correlation, Indicates the number of data in the basic data. It represents the difference between the data value of the i-th data in the k1-th group and the k2-th group of basic data. Represents the average of the differences between the data values.

[0015] Compared with the prior art, the present invention provides a method for operating a thermal power unit, which has the following beneficial effects: 1. The present invention screens out basic data in historical operating data of a power thermal power unit according to stability data, and then selects several groups of data with the largest total correlation as key data affecting the power thermal power unit according to the total correlation between each group of basic data to obtain standard data, and then divides the standard data into operating conditions according to variable types and variable data to obtain several first operating conditions, and then further eliminates abnormal standard data through a Gaussian mixture model to obtain a third operating condition, thereby establishing an operating parameter library according to the third operating condition, so that the operating parameters of each device can be configured according to the operating parameter library and the current operating condition of the power thermal power unit.

[0016] 2. The present invention determines whether the thermal power unit is in a stable state by determining whether the stability data of active power, main steam pressure, main steam temperature, reheat steam temperature and feed water flow rate are within the corresponding stability thresholds, and selects basic data from a large amount of historical operation data accordingly; 3. The present invention selects parameters that play a key role in the state of the power thermal power unit under different conditions by calculating the total correlation of each group of basic data, thereby further eliminating a large number of groups of basic data with weak correlation, and finally obtaining standard data.

[0017] 4. The present invention sets several variable types and divides the working conditions of the standard data according to different variable types to obtain several first working conditions, so as to facilitate the subsequent configuration of various equipment of the power thermal power unit by calling the corresponding parameters according to the real-time working conditions after establishing the operating parameter library.

[0018] 5. The present invention establishes a Gaussian mixture model and a corresponding log-likelihood function, and then uses the parameter-corrected Gaussian mixture model to achieve the final elimination of abnormal standard data in the first operating condition, thereby obtaining the third operating condition, thereby greatly improving the accuracy of the subsequently established operating parameter library to meet the purpose of predetermined needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 The present invention is a flow chart of the method for operating a thermal power generating unit. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific implementation methods, so that the implementation process of how the present application uses technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly.

[0021] Those skilled in the art can understand that all or part of the steps in the following embodiments can be completed by instructing the relevant hardware through a program, so the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0022] In the current operation of thermal power units, the methods for determining the optimal reference values ​​of energy efficiency parameters mainly include mechanism modeling and thermal test methods. In order to solve the problem that after obtaining the operation data of each device in the thermal power unit, it is difficult to accurately screen out the unsuitable redundant data in the prior art, which leads to the problem that it is difficult to meet the predetermined requirements when automatically generating and updating the operation parameter library to control the operation of each device in the thermal power unit, the present invention provides a method for operating a thermal power unit, by further screening the acquired data to obtain standard data under various working conditions, thereby improving the accuracy of the equipment operation parameter library. The method specifically includes the following steps: S1. Obtain the historical operation data of each device in the thermal power unit; the historical operation data obtained mainly include main steam pressure, main steam temperature, reheat steam temperature and feed water flow, as well as the heat exchanger cleaning factor of the boiler system, superheater desuperheating water volume, feed water temperature, reheater desuperheating water volume, superheater outlet pressure, heat exchanger wall temperature, superheater outlet temperature, reheater outlet temperature, reheater outlet pressure, fly ash carbon content, furnace pressure and coal consumption, as well as coal mill operating current and coal feeder coal supply data, as well as primary fan current, primary fan rotor blade opening and blower current in the air and smoke system. Each data can be obtained using the corresponding sensor; S2. Based on historical data, basic data containing key factors at each moment and within the corresponding stability threshold are obtained, and several groups of standard data are determined with the goal of maximizing the total correlation. By analyzing the power thermal power unit, it is found that active power, main steam pressure, main steam temperature, reheat steam temperature and feed water flow are several key factors for determining whether the power thermal power unit is stable as a whole. Therefore, the historical operation data in this state is the required basic data. At the same time, there is a strong or weak relationship in the obtained data. Therefore, the basic data with a higher total correlation is the key factor for the operation of the power thermal power unit. Therefore, as standard data, further subsequent analysis is performed. In step S2, the following steps are specifically included: S21, arranging the historical operation data in chronological order to obtain a plurality of historical data strings; each historical data string corresponds to historical data acquired by a sensor; S22, establishing a filtering window centered on each moment on the historical data string, where the filtering window spans across several moments; S23. Determine whether the stability data of key factors at each moment are within the corresponding stability thresholds. The key factors include active power, main steam pressure, main steam temperature, reheat steam temperature and feed water flow rate. The stability thresholds corresponding to active power, main steam pressure, main steam temperature, reheat steam temperature and feed water flow rate are all different. For example, they are set to 2 kW, 1 kPa, 10 °C, 5 °C and 50 tons / hour respectively, which are specifically set according to the equipment parameters of the thermal power unit.

[0023] If yes, extract the historical operation data in the time contained in the filter window to obtain several sets of basic data; this indicates that the fluctuation of the stability data is very small, so the operation of the thermal power unit is relatively stable, and the corresponding historical data at this time is the basic data; If not, the historical operation data in the historical data string is not extracted; S24, calculating the average values ​​of any two sets of basic data in sequence, and calculating the first correlation between the two; the calculation formula of the first correlation is: ; In the above formula, represents the first correlation between the k1th group of basic data and the k2th group of basic data, and Respectively represent the i-th data of the k1-th group and the k2-th group of basic data, and They represent the average values ​​of the k1th group of basic data and the k2th group of basic data respectively, and the number of data in the k1th group and the k2th group of basic data is n; when When , it indicates that the two sets of basic data k1 and k2 are positively correlated. When , it indicates that the two sets of basic data k1 and k2 are negatively correlated. When , it indicates that the two sets of basic data k1 and k2 are unrelated. The closer the value is to 0, the lower the correlation between the two sets of basic data k1 and k2.

[0024] S25. Calculate the mean and standard deviation of any two sets of basic data in sequence, and calculate the second correlation between the two. The calculation formula of the second correlation is: ; in, ; In the above formula, represents the second correlation between the k1th group of basic data and the k2th group of basic data, represents the unit correlation between the i-th data in the k1-th group of basic data and the k2-th group of basic data, and Respectively represent the i-th data of the k1-th group and the k2-th group of basic data, Differentiation coefficient ; S26, calculating the difference between the data values ​​of the corresponding data in any two sets of basic data, and calculating the inverse of the standard deviation of the difference between the data values, to obtain the third correlation between the any two sets of basic data; in order to further illustrate the calculation method of the third correlation, in step S26, specifically including the following steps: S261, calculating the difference between the data values ​​of corresponding data in any two sets of basic data; S262, calculating the reciprocal of the standard deviation of the difference between the data values ​​to obtain the third correlation between any two sets of basic data; the calculation formula of the third correlation is: ; in, ; In the above formula, represents the third correlation, Indicates the number of data in the basic data. It represents the difference between the data value of the i-th data in the k1-th group and the k2-th group of basic data. Represents the average of the differences between the data values.

[0025] S27, calculating the total correlation of any two sets of basic data according to the first correlation, the second correlation and the third correlation; the calculation formula of the total correlation is: ; In the above formula, represents the total correlation between the k1th group of basic data and the k2th group of basic data, , and Represent the first correlation, the second correlation and the third correlation respectively, , and denote the first correlation coefficient, the second correlation coefficient and the third correlation coefficient respectively; S28. Set a correlation threshold, and select several groups of basic data whose total correlation is greater than the correlation threshold as standard data. The correlation threshold is generally 0.5.

[0026] S3, according to the variables corresponding to the standard data operation of the thermal power unit, the variable type is set and the variable data is collected, and the standard data operation conditions are divided according to the variable data to obtain a plurality of first operating conditions; the variable type includes ambient temperature, coal quality, ambient humidity and active power, etc., and the variable data is the specific value under the corresponding variable type, such as the ambient temperature is 25°C and 27°C, etc. According to different value ranges, the obtained standard data is divided into a plurality of operating conditions. In step S3, the following steps are specifically included: S31, obtaining several variable types and variable data of corresponding types that affect standard data, and establishing a variable data set including the variable types and variable data; S32, marking the standard data as first standard data; S33, selecting any variable type and variable data of a corresponding type from the variable data set, and dividing the operating conditions of the first standard data by a K-means clustering algorithm to obtain a plurality of first operating conditions; S34, removing the selected variable type and variable data of the corresponding type from the variable data set; S35, determining whether the variable data set is an empty set; If yes, the first working condition is output and the process goes to step S4; If not, proceed to step S36; S36. Select any variable type and variable data of the corresponding type from the variable data set, and divide the first operating condition into several second operating conditions through the K-means clustering algorithm, then mark the second operating condition as the first operating condition, and then return to step S34.

[0027] S4, establishing a Gaussian mixture model with convergent log-likelihood function according to each first working condition, and identifying and removing abnormal standard data in the first working condition to obtain a third working condition; under the same working condition, the operating parameters of each device have the characteristics of normal distribution, therefore, the data is fitted and the abnormal standard data is removed to obtain the third working condition, in step S4, specifically including the following steps: S41, establishing a Gaussian mixture model according to a plurality of first working conditions; S42, establishing a log-likelihood function according to the Gaussian mixture model, and determining whether the log-likelihood function converges; the process of establishing the log-likelihood function is not described in detail here; If yes, proceed to step S43; If not, proceed to step S47; S43, calculating the probability density of any standard data from each first working condition according to the Gaussian mixture model; the calculation formula of the probability density of the first working condition is: ; In the above formula, Indicates the jth standard data The probability density from the kth first condition, is the Gaussian distribution density function of the kth first operating condition, is the sample mean of the standard data in the kth first working condition, is the covariance of the kth first operating condition, is the weight of the Gaussian distribution of the kth first working condition, and M is the total number of first working conditions; S44, calculating the density radius according to the probability density of the standard data in the first working condition; the calculation formula of the density radius is: ; In the above formula, represents the density radius of the kth first operating condition, Indicates the jth standard data in the first working condition The probability density from the kth first operating condition, represents the number of standard data in the kth first working condition, represents the radius coefficient; S45, calculating the density center of each first working condition; the density center of the first working condition is the standard data of the center of the range corresponding to the range with the most concentrated data in the first working condition. Now a method for obtaining the density center of the first working condition is provided. In step S45, the method specifically includes the following steps: S451, calculating the initial statistical radius of each first working condition; the calculation formula of the initial statistical radius is: ; In the above formula, represents the initial statistical radius of any first operating condition, and They represent the maximum and minimum values ​​of the data in the first working condition, respectively, and C is the reduction coefficient, 0<C<1; S452, setting the initial number of iterations to 1, the initial statistical radius, and the initial statistical area to the first working condition; S453, obtaining the current number of iterations, the current statistical radius and the current statistical area; S454, calculating the density of any standard data in the current statistical area according to the current number of iterations and the statistical radius; the calculation formula of the density is: ; in, ; In the above formula, Indicates the density of the i-th standard data in the current statistical area, Indicates the number of standard data in the current statistical area. is the symbolic function, and Respectively represent the i-th and j-th data in the current statistical area, represents the statistical radius at the dth iteration; S455, obtaining a plurality of standard data with the largest density, and determining whether the current plurality of standard data with the largest density and the statistical area are the same as those in the previous iteration; If yes, proceed to step S458; If not, the current number of iterations is increased by 1, and then the process goes to step S456; S456, marking the standard data whose data value difference with the several standard data with the largest density is less than the current statistical radius as the current statistical area; S457, calculate the current statistical radius according to the current number of iterations, and then return to step S453; the calculation formula of the current statistical radius is: ; In the above formula, Indicates the current statistical radius, C is the reduction coefficient, 0<C<1, Indicates the statistical radius at the last iteration; S458. Calculate the average value of several standard data with the largest density to obtain the density center of the corresponding first working condition.

[0028] S46, sequentially calculating the numerical difference between each data in the first working condition and the density center, and determining whether the numerical difference is less than or equal to the density radius; If yes, the data is marked as the third operating condition, and then the process goes to step S5; If not, the data is marked as abnormal standard data; S47, optimize the first working condition until the log-likelihood function converges, and then return to step S43; before using the Gaussian mixture model to eliminate abnormal standard data to obtain the third working condition, it is necessary to ensure the convergence of the Gaussian mixture model, so it is necessary to adjust the parameters in the established Gaussian mixture model, in step S47, specifically including the following steps: S471, calculating and obtaining the probability density of any standard data from each first working condition according to step S43; S472, update the parameters in the Gaussian mixture model according to the calculated probability density, and then return to step S41; the update formula for updating the parameters in the Gaussian mixture model is as follows: ; In the above formula, Indicates the jth standard data The probability density of the first working condition of k, the total number of standard data is J, is the sample mean of the standard data in the kth first working condition, is the covariance of the kth first operating condition, is the weight of the Gaussian distribution of the kth first operating condition.

[0029] S5. Establish an objective function according to the third operating condition and output the corresponding optimal parameters to establish an operating parameter library, obtain the current operating condition of the thermal power unit and control each module of the thermal power unit to operate according to the optimal parameter configuration according to the operating parameter library. The establishment of the objective function can select goals such as maximum efficiency or minimum carbon emissions.

[0030] The present invention screens out basic data in historical operating data of a power thermal power unit according to stability data, and then selects several groups of data with the largest total correlation as key data affecting the power thermal power unit according to the total correlation between each group of basic data, so as to obtain standard data, and then divides the standard data into working conditions according to variable types and variable data, so as to facilitate subsequent direct adjustment of corresponding equipment operating parameters according to corresponding working conditions. Before generating an operating parameter library, the abnormal standard data therein is further eliminated through a Gaussian mixture model to obtain a third working condition, so as to establish an operating parameter library according to the third working condition, so as to be able to configure the operating parameters of each device according to the operating parameter library and the current working condition of the power thermal power unit.

[0031] The above implementation methods have been described in detail. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for operating a thermal power unit, characterized in that: The method specifically comprises the following steps: S1. Obtain the historical operation data of each device in the thermal power unit; S2. Based on historical data, basic data containing key factors and within the corresponding stability threshold at each moment are obtained, and several groups of standard data are determined with the goal of maximizing the total correlation; S3, setting variable types and collecting variable data according to variables corresponding to the standard data operation of the thermal power unit, and dividing the standard data operation conditions according to the variable data to obtain a plurality of first conditions; S4, establishing a Gaussian mixture model with convergent log-likelihood function according to each first operating condition, and identifying and removing abnormal standard data in the first operating condition to obtain a third operating condition; S5. Establish an objective function according to the third operating condition and output the corresponding optimal parameters to establish an operating parameter library, obtain the current operating condition of the thermal power unit and control each module of the thermal power unit to operate according to the optimal parameter configuration according to the operating parameter library.

2. The operating method according to claim 1, characterized in that: In step S2, the following steps are specifically included: S21, arranging the historical operation data in chronological order to obtain a plurality of historical data strings; S22, establishing a filtering window centered on each moment on the historical data string, where the filtering window spans across several moments; S23, judging whether the stability data of key factors at each moment are all within the corresponding stability threshold, where the key factors include active power, main steam pressure, main steam temperature, reheat steam temperature and feed water flow; If so, extract the historical operation data in the time period included in the filter window to obtain several groups of basic data; If not, the historical operation data in the historical data string is not extracted; S24, calculating the average values ​​of any two groups of basic data in sequence, and calculating the first correlation between the two groups; S25, calculating the mean value and standard deviation of any two groups of basic data in sequence, and calculating the second correlation between the two; S26, calculating the difference between the data values ​​of corresponding data in any two sets of basic data, and calculating the inverse of the standard deviation of the difference between the data values, to obtain a third correlation between the any two sets of basic data; S27, calculating the total correlation of any two sets of basic data according to the first correlation, the second correlation and the third correlation; ； In the above formula, represents the total correlation between the k1th group of basic data and the k2th group of basic data, , and Represent the first correlation, the second correlation and the third correlation respectively, , and denote the first correlation coefficient, the second correlation coefficient and the third correlation coefficient respectively; S28. Set a correlation threshold, and select several groups of basic data whose total correlation is greater than the correlation threshold as standard data.

3. The operating method according to claim 2, characterized in that: In step S26, the following steps are specifically included: S261, calculating the difference between the data values ​​of corresponding data in any two sets of basic data; S262. Calculate the inverse of the standard deviation of the difference between the data values ​​to obtain a third correlation between any two sets of basic data.

4. The operating method according to claim 1, characterized in that: In step S3, the following steps are specifically included: S31, obtaining several variable types and variable data of corresponding types that affect standard data, and establishing a variable data set including the variable types and variable data; S32, marking the standard data as first standard data; S33, selecting any variable type and variable data of a corresponding type from the variable data set, and dividing the operating conditions of the first standard data by a K-means clustering algorithm to obtain a plurality of first operating conditions; S34, removing the selected variable type and variable data of the corresponding type from the variable data set; S35, determining whether the variable data set is an empty set; If yes, the first working condition is output and the process goes to step S4; If not, proceed to step S36; S36. Select any variable type and variable data of the corresponding type from the variable data set, and divide the first operating condition into several second operating conditions through the K-means clustering algorithm, then mark the second operating condition as the first operating condition, and then return to step S34.

5. The operating method according to claim 1, characterized in that: In step S4, the following steps are specifically included: S41, establishing a Gaussian mixture model according to a plurality of first working conditions; S42, establishing a log-likelihood function according to the Gaussian mixture model, and determining whether the log-likelihood function converges; If yes, proceed to step S43; If not, proceed to step S47; S43, calculating the probability density of any standard data from each first working condition according to the Gaussian mixture model; the calculation formula of the probability density of the first working condition is: ； In the above formula, Indicates the jth standard data The probability density from the kth first operating condition, is the Gaussian distribution density function of the kth first operating condition, is the sample mean of the standard data in the kth first working condition, is the covariance of the kth first operating condition, is the weight of the Gaussian distribution of the kth first working condition, and M is the total number of first working conditions; S44, calculating the density radius according to the probability density of the standard data in the first working condition; the calculation formula of the density radius is: ； In the above formula, represents the density radius of the kth first operating condition, Indicates the jth standard data in the first working condition The probability density from the kth first operating condition, represents the number of standard data in the kth first working condition, represents the radius coefficient; S45, calculating the density center of each first working condition; S46, sequentially calculating the numerical difference between each data in the first working condition and the density center, and determining whether the numerical difference is less than or equal to the density radius; If yes, the data is marked as the third operating condition, and then the process goes to step S5; If not, the data is marked as abnormal standard data; S47, optimize the first operating condition until the log-likelihood function converges, and then return to step S43.

6. The operating method according to claim 5, characterized in that: In step S45, the following steps are specifically included: S451, calculating the initial statistical radius of each first working condition; the calculation formula of the initial statistical radius is: ； In the above formula, represents the initial statistical radius of any first operating condition, and They represent the maximum and minimum values ​​of the data in the first working condition, respectively, and C is the reduction coefficient, 0<C<1; S452, setting the initial number of iterations to 1, the initial statistical radius, and the initial statistical area to the first working condition; S453, obtaining the current number of iterations, the current statistical radius and the current statistical area; S454, calculating the density of any standard data in the current statistical area according to the current number of iterations and the statistical radius; the calculation formula of the density is: ； in, ； In the above formula, Indicates the density of the i-th standard data in the current statistical area, Indicates the number of standard data in the current statistical area. is the symbolic function, and Respectively represent the i-th and j-th data in the current statistical area, represents the statistical radius at the dth iteration; S455, obtaining a plurality of standard data with the largest density, and determining whether the current plurality of standard data with the largest density and the statistical area are the same as those in the previous iteration; If yes, proceed to step S458; If not, the current number of iterations is increased by 1, and then the process goes to step S456; S456, marking the standard data whose data value difference with the several standard data with the largest density is less than the current statistical radius as the current statistical area; S457, calculate the current statistical radius according to the current number of iterations, and then return to step S453; the calculation formula of the current statistical radius is: ； In the above formula, Indicates the current statistical radius, C is the reduction coefficient, 0<C<1, Indicates the statistical radius at the last iteration; S458. Calculate the average value of several standard data with the largest density to obtain the density center of the corresponding first working condition.

7. The operating method according to claim 1, characterized in that: In step S47, the following steps are specifically included: S471, calculating and obtaining the probability density of any standard data from each first working condition according to step S43; S472, update the parameters in the Gaussian mixture model according to the calculated probability density, and then return to step S41; the update formula for updating the parameters in the Gaussian mixture model is as follows: ； In the above formula, Indicates the jth standard data The probability density of the first working condition of k, the total number of standard data is J, is the sample mean of the standard data in the kth first working condition, is the covariance of the kth first operating condition, is the weight of the Gaussian distribution of the kth first operating condition.

8. The operating method according to claim 2, characterized in that: In step S24, the calculation formula of the first correlation is: ； In the above formula, represents the first correlation between the k1th group of basic data and the k2th group of basic data, and Respectively represent the i-th data of the k1-th group and the k2-th group of basic data, and They respectively represent the average values ​​of the k1th group of basic data and the k2th group of basic data, and the number of data in the k1th group and the k2th group of basic data is n.

9. The operating method according to claim 2, characterized in that: In step S25, the calculation formula of the second correlation is: ； in, ； In the above formula, represents the second correlation between the k1th group of basic data and the k2th group of basic data, represents the unit correlation between the i-th data in the k1-th group of basic data and the k2-th group of basic data, and Respectively represent the i-th data of the k1-th group and the k2-th group of basic data, Differentiation coefficient .

10. The operating method according to claim 3, characterized in that: In step S262, the calculation formula of the third correlation is: ； in, ； In the above formula, represents the third correlation, Indicates the number of data in the basic data. It represents the difference between the data value of the i-th data in the k1-th group and the k2-th group of basic data. Represents the average of the differences between the data values.

Images