Turbine flow characteristic identification and optimization method based on big data mining

Abstract

The invention discloses a steam turbine flow characteristic identification and optimization method based on big data mining. The method comprises the following steps that S1, influence factors of all valves are analyzed through downstream valve flow characteristics, and DEH valve set characteristics are obtained; s2, converting the original data into standardized data suitable for analysis and modeling through data preprocessing; s3, the power and the flow serve as entry points, valve set characteristic analysis is combined, and an identification method of the flow characteristics under the unit test working condition is studied; s4, performing big data mining clustering calculation by using a data mining algorithm in combination with historical data and stable working condition identification elements, and extracting target data; and S5, fitting the actual traffic characteristics by adopting a deep learning method. Through data collection and preparation, data analysis and modeling, model verification and adjustment, flow characteristic optimization and real-time monitoring and maintenance, the optimal flow characteristic of the regulating valve can be obtained, and the operation efficiency and reliability of a steam turbine are improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent control technology for steam turbine control systems in thermal power plants, and relates to a method for identifying and optimizing steam turbine flow characteristics based on big data mining. Background Technology

[0002] Most thermal power unit turbines employ digital electro-hydraulic control systems (DEH), which operate in two modes: single-valve and sequential valve, and also feature valve switching functionality. In single-valve mode, all high-pressure control valves maintain the same opening degree, resulting in more uniform thermal stress on the turbine, but with greater throttling losses, especially at lower loads. In sequential valve mode, the control valves open continuously in a set sequence, reducing throttling losses and improving the unit's economic efficiency. With the development of large-scale power grid interconnection technology, grid-connected units are required to possess frequency regulation and peak shaving capabilities. Automatic generation control (AGC) and primary frequency regulation are gradually becoming key performance indicators for the power grid. Both AGC and primary frequency regulation require a good linear relationship between the turbine control valve opening and the actual generated power to ensure load control accuracy and response rate. In actual production, the flow characteristic function of the control valves set in the DEH (Department of Power) settings of most thermal power units is the theoretical characteristic function calculated at the factory. Due to installation, commissioning, or long-term operation of the unit, the actual flow characteristics of the control valves vary greatly, manifesting as slow changes in AGC (Automatic Gauge Control) rate, poor primary frequency regulation capability, large load fluctuations during steam distribution mode switching, and poor unit coordination response capability. Therefore, it is objectively necessary to identify and optimize the flow characteristics of turbine control valves.

[0003] To address this issue, a method for identifying and optimizing the flow characteristics of steam turbines based on big data mining is designed to overcome the aforementioned problems. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method for identifying and optimizing the flow characteristics of steam turbines based on big data mining. Through data collection and preparation, data analysis and modeling, model verification and adjustment, flow characteristic optimization, and real-time monitoring and maintenance, this invention can obtain the optimal flow characteristics of regulating valves, thereby improving the operating efficiency and reliability of steam turbines.

[0005] This invention is achieved through the following technical solution: a method for identifying and optimizing the flow characteristics of steam turbines based on big data mining, comprising the following steps:

[0006] Step S1: Analyze the flow characteristics of each valve, flow distribution, overlap, and other influencing factors to obtain the characteristics of the DEH valve group;

[0007] Step S2: The process of transforming raw data into standardized data suitable for analysis and modeling through data preprocessing, with the aim of improving data quality and usability;

[0008] Step S3: Taking power and flow rate as the starting points respectively, and combining valve group characteristic analysis, study the identification method of flow characteristics under the test conditions of the unit to obtain the flow characteristic curve of the steam turbine under stable operating conditions.

[0009] Step S4: Using data mining algorithms, combined with historical data and stable operating condition identification elements, perform big data mining and clustering calculations to extract target data that reflects the flow characteristics of the unit under varying operating conditions.

[0010] Step S5: Use deep learning methods to fit the actual traffic characteristics and tune the logic configuration parameters based on the algorithm results.

[0011] Preferably, step S1 includes the following steps:

[0012] Step S11: Valve characteristic calibration test;

[0013] Step S12: Analyze the influence of factors such as flow characteristics, flow distribution, and overlap of each valve through the flow characteristics analysis of the valve.

[0014] Preferably, step S12 includes the following steps:

[0015] Step S121: Determine the structural characteristics of the control valve (valve core shape, including linear, equal percentage, quick-opening, and parabolic characteristics): relative stroke & relative flow nonlinear characteristics, i.e., cam characteristics;

[0016] Step S122: Determine the flow characteristics of the regulating valve itself. These characteristics reveal the relationship between the flow rate through the valve and its lift and the pressure ratio before and after it. Specifically, this involves the relationship curves between the actual flow rate and the critical flow rate, and between the opening degree and the critical flow rate.

[0017] Step S123: Determine the flow characteristics of the steam turbine (opening degree-flow curve of the steam turbine regulating valve group). The regulation of the steam turbine flow is accomplished by the regulating valve and the subsequent nozzle group, that is, it is determined by the structural parameters and operating status parameters of the regulating valve and the nozzle group.

[0018] Preferably, step S2 includes the following steps:

[0019] Step S21: For conventional power plant turbine units, select the required parameters based on their principle thermodynamic system diagram;

[0020] Step S22: Encode the raw data parameters and perform data standardization operations such as data cleaning, data transformation, data integration, and data reduction;

[0021] Step S23: In a simple and comprehensive way, explore and observe the operating characteristics of the unit, and plot the changes in the total valve position value and main steam pressure value under the unit output load to obtain their changing patterns;

[0022] Step S24: Obtain a thermally stable operating condition dataset by filtering and processing stable data.

[0023] Preferably, step S3 includes the following steps:

[0024] Step S31: Calculate the main steam flow rate of the steam turbine using methods such as nozzle flow rate calculation, Freund's formula, and characteristic flow area;

[0025] Step S32: Taking power and flow rate as the starting point, different research objects are selected, and a flow characteristic identification method based on characteristic flow area is proposed.

[0026] Preferably, step S4 includes the following steps:

[0027] Step S41: Using this flow characteristic as the basis for linearity and optimization adjustment, analyze the basic requirements for optimizing the unit's flow characteristics;

[0028] Step S42: Use the piecewise linear optimization improved K-Medoids algorithm to perform mining calculations on the identification results.

[0029] Preferably, step S5 includes the following steps:

[0030] Step S51: Comprehensive valve position modeling and prediction based on BP neural network;

[0031] Step S52: Based on the neural network model estimate, tune the logic configuration correction parameters, and then perform a test on the corrected flow characteristics.

[0032] The beneficial effects of this invention are as follows:

[0033] The turbine flow characteristic identification and optimization method based on big data mining designed in this invention can obtain the optimal control valve flow characteristics and improve the operating efficiency and reliability of the turbine through data collection and preparation, data analysis and modeling, model verification and adjustment, flow characteristic optimization, and real-time monitoring and maintenance. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the overall process of the present invention. Detailed Implementation

[0035] To enable those skilled in the art to more clearly understand the purpose, technical solution, and advantages of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0036] In the description of this invention, it should be understood that the orientation or positional relationship indicated by terms such as "upper", "lower", "left", "right", "inner", "outer", "lateral", and "vertical" is based on the orientation or positional relationship shown in the accompanying drawings and is only for the convenience of describing this invention, and is not intended to indicate or imply that the device or component referred to must have a specific orientation, and therefore should not be construed as a limitation of this invention.

[0037] The invention will now be described in detail with reference to the accompanying drawings: Figure 1 As shown, a method for identifying and optimizing the flow characteristics of steam turbines based on big data mining includes the following steps:

[0038] Step S1: Analyze the flow characteristics of each valve, flow distribution, overlap, and other influencing factors to obtain the characteristics of the DEH valve group;

[0039] Step S2: The process of transforming raw data into standardized data suitable for analysis and modeling through data preprocessing, with the aim of improving data quality and usability;

[0040] Step S3: Taking power and flow rate as the starting points respectively, and combining valve group characteristic analysis, study the identification method of flow characteristics under the test conditions of the unit to obtain the flow characteristic curve of the steam turbine under stable operating conditions.

[0041] Step S4: Using data mining algorithms, combined with historical data and stable operating condition identification elements, perform big data mining and clustering calculations to extract target data that reflects the flow characteristics of the unit under varying operating conditions.

[0042] Step S5: Use deep learning methods to fit the actual traffic characteristics and adjust the logic configuration parameters based on the algorithm results.

[0043] This invention obtains the flow characteristics of a valve group through sequential valve flow tests. Steam turbines typically have four or six high-pressure regulating valves, each controlled by an independent servo system. Valve regulation methods include single-valve regulation and sequential valve regulation. Single-valve regulation is suitable for throttling and full-circuit steam intake. Its characteristic is that all high-pressure regulating valves open and close simultaneously to control the unit's speed or load. The advantages of this method are that it ensures uniform temperature distribution in the first-stage steam chamber of the steam turbine, reduces the temperature difference between the turbine rotor and stator during load changes, and allows the unit to withstand larger load change rates. The disadvantage is poor economic efficiency, as the main steam experiences significant throttling losses when passing through the regulating valves, reducing the unit's efficiency. Sequential valve regulation is characterized by opening and closing the regulating valves sequentially. As the unit load changes, only one regulating valve is in a semi-open state, while the others are fully open or fully closed. The advantage of this method is higher steam turbine efficiency, but the disadvantage is that the unit can only withstand a smaller load change rate. Testing the flow characteristics of a single valve in a steam turbine is relatively simple. The flow characteristic curve can be obtained by testing the flow curve corresponding to the opening degree of each individual valve and then interchanging the horizontal and vertical axes. However, the flow characteristics of a forward valve are more complex, influenced by factors such as the flow characteristics of each valve, flow distribution, and overlap. Therefore, tests are conducted specifically on the flow characteristics under the forward valve configuration.

[0044] Step S1 includes the following steps:

[0045] Step S11: Valve characteristic calibration test;

[0046] Step S12: Analyze the influence of factors such as flow characteristics, flow distribution, and overlap of each valve through the flow characteristics analysis of the valve.

[0047] Step S12 includes the following steps:

[0048] Step S121: Determine the structural characteristics of the control valve (valve core shape, including linear, equal percentage, quick-opening, and parabolic characteristics): relative stroke & relative flow nonlinear characteristics, i.e., cam characteristics;

[0049] Step S122: Determine the flow characteristics of the regulating valve itself. These characteristics reveal the relationship between the flow rate through the valve and its lift and the pressure ratio before and after it. Specifically, this involves the relationship curves between the actual flow rate and the critical flow rate, and between the opening degree and the critical flow rate.

[0050] Step S123: Determine the flow characteristics of the steam turbine (opening degree-flow curve of the steam turbine regulating valve group). The regulation of the steam turbine flow is accomplished by the regulating valve and the subsequent nozzle group, that is, it is determined by the structural parameters and operating status parameters of the regulating valve and the nozzle group.

[0051] This invention completes data preparation, cleaning, and feature extraction through data preprocessing. In modern scientific research and practical work, all industries need to process various types of collected data. How to discover deeper and more important information from this massive amount of data, enabling it to describe the overall characteristics of the data, predict development trends, and thus generate decisions, requires data mining. However, if the focus is only on data mining algorithms while neglecting data preprocessing, some important aspects of data mining are often lost. This is because data in real-world systems is generally incomplete, redundant, and ambiguous, rarely directly meeting the requirements of data mining algorithms. Furthermore, massive amounts of data contain many meaningless components, severely affecting the execution efficiency of data mining algorithms, and noise interference can also cause deviations in the mining results. Therefore, effective preprocessing of imperfect raw data has become a key issue in the implementation of data mining systems. The unit's DCS historical data is the result of sampling and recording various operating parameters of the unit during actual operation according to a set sampling period and arranged parameter measurement points, possessing authenticity and objectivity. As sampling time continues, the amount of historical data increases significantly, becoming increasingly reflective of the unit's operating characteristics. However, historical data not only records a large amount of data under thermally stable operating conditions but also stores much data under conditions in the adjustment and transition phases. Unsteady-state data during these transition phases may be coupled with numerous known and unknown factors, rendering them invalid for this study. Furthermore, historical data inevitably contains duplicate data under the same operating conditions, resulting in significant redundancy. Therefore, a data preprocessing method is proposed to obtain a concise dataset of thermally stable operating conditions.

[0052] Step S2 includes the following steps:

[0053] Step S21: For conventional power plant turbine units, select the required parameters based on their principle thermodynamic system diagram;

[0054] Step S22: Encode the raw data parameters and perform data standardization operations such as data cleaning, data transformation, data integration, and data reduction;

[0055] Step S23: In a simple and comprehensive way, explore and observe the operating characteristics of the unit, and plot the changes in the total valve position value and main steam pressure value under the unit output load to obtain their changing patterns;

[0056] Step S24: Obtain a thermally stable operating condition dataset through stable data filtering and processing. This invention takes power and flow rate as the starting point, selects different research objects, and successfully reproduces the turbine flow characteristics that reflect the actual operating characteristics of the unit through parameter identification.

[0057] Step S3 includes the following steps:

[0058] Step S31: Calculate the main steam flow rate of the steam turbine using methods such as nozzle flow rate calculation, Freund's formula, and characteristic flow area;

[0059] Step S32: Taking power and flow rate as the starting point, different research objects are selected, and a flow characteristic identification method based on characteristic flow area is proposed.

[0060] Step S4 includes the following steps:

[0061] Step S41: Using this flow characteristic as the basis for linearity and optimization adjustment, analyze the basic requirements for optimizing the unit's flow characteristics;

[0062] Step S42: Use the piecewise linear optimization improved K-Medoids algorithm to perform mining calculations on the identification results.

[0063] Preferably, step S5 includes the following steps:

[0064] Step S51: Comprehensive valve position modeling and prediction based on BP neural network;

[0065] Step S52: Based on the neural network model estimate, tune the logic configuration correction parameters, and then perform a test on the corrected flow characteristics.

[0066] Example

[0067] A method for identifying and optimizing the flow characteristics of steam turbines based on big data mining, the method comprising the following steps:

[0068] Step S1: Analyze the flow characteristics of each valve, including flow distribution, overlap, and other influencing factors, to obtain the characteristics of the DEH valve group. Step S1 includes the following steps:

[0069] When the unit is under rated load, adjust the main steam pressure to make the regulating valve fully open and maintain the main steam pressure of the unit stable.

[0070] Once the unit load, main steam pressure and other parameters stabilize, the unit exits the coordinated control system, and the main control of the boiler and turbine is switched to manual mode.

[0071] Manually and gradually reduce the turbine integrated valve position by 2% increments until the unit load reaches the boiler's minimum stable combustion load.

[0072] Manually and gradually increase the turbine integrated valve position in 2% increments until the unit is under rated load.

[0073] During unit load changes, operators must make timely adjustments to ensure stable main steam pressure and temperature. Test data must be recorded.

[0074] Based on the data, plot the relationships between actual flow rate and critical flow rate, valve opening and critical flow rate, and the valve opening-flow rate curve. This characteristic reveals the relationship between the flow rate through the valve and its lift and the pressure ratio before and after it.

[0075] Step S2: This step involves preprocessing the raw data to transform it into standardized data suitable for analysis and modeling, with the aim of improving data quality and usability. Step S2 includes the following steps:

[0076] The determination of the actual flow rate of the steam turbine is related to the main steam parameters, the parameters after the regulating valve, the parameters after the regulating stage, the extraction and exhaust parameters of each cylinder of the turbine, and the unit load. Generally speaking, for a conventional power plant steam turbine unit, the required parameters are selected according to its principle thermodynamic system diagram as follows: DEH flow rate or total valve position (%); unit load (MW); main steam pressure (MPa); regulating stage pressure (MPa); feedback from high-pressure regulating valve 1 (%); feedback from high-pressure regulating valve 2 (%); feedback from high-pressure regulating valve 3 (%); feedback from high-pressure regulating valve 4 (%); flow rate feedback (%); main steam temperature (°C); regulating stage temperature (°C); speed (r / min); pressure after CV1 valve (MPa); pressure after CV2 valve (MPa); pressure after CV3 valve (MPa); pressure after CV4 valve (MPa); high-pressure cylinder exhaust pressure (MPa); high-pressure cylinder exhaust temperature (°C); reheat pressure (MPa); reheat temperature (°C).

[0077] Determining whether a generating unit is in a state of relative thermal stability is a problem involving a combination of multiple conditions. Currently, there are no relevant standards or regulations available for reference in the power industry. Based on actual production and operation requirements, the following principles are established for data selection:

[0078] When the unit is running, all important parameters remain stable with small fluctuations (the range of numerical fluctuations is no greater than their respective accuracy settings), indicating that the unit is in a state of thermal stability.

[0079] To obtain the complete operating characteristics of the unit to the greatest extent, the operating range where the unit load increases / decreases significantly is selected as the key area for data screening.

[0080] Therefore, the specific method for selecting stable data is as follows:

[0081] If the unit's key parameters remain stable for at least 3-4 sampling cycles (i.e., a stabilization time of at least 10 seconds), the unit is considered to be in thermal stability. Considering the correlation between parameters, the key parameters of the unit are simplified to: output power value, total valve position value, and main steam pressure.

[0082] Based on the previous point, strictly following the chronological order of sampling time, the operating condition data where the deviations of all important operating parameters of the unit are less than their respective specified accuracy are divided into the sampling data of the same operating condition in that time period.

[0083] Grubbs analysis was performed on the sampled datasets under the same working condition to remove outliers, and the results were used as the stable data for that working condition during that period.

[0084] Step S3: Taking power and flow rate as starting points respectively, and combining valve group characteristic analysis, research the identification method of flow characteristics under the test conditions of the unit to obtain the flow characteristic curve of the steam turbine under stable operating conditions. Step S3 includes the following steps:

[0085] Applying the Flueger formula to thermodynamic calculations of steam turbines under varying operating conditions. For a stage group consisting of multiple flow stages, if the flow area remains constant, the relationship between the flow rate and thermodynamic parameters of the stage group can be expressed by the following formula according to the general form of the Flueger formula:

[0086]

[0087] In the formula, G, p, and T represent the flow rate through the stage, the pressure of the stage, and the temperature before the stage, respectively; the superscript "'" indicates a variable operating condition; and the subscripts "1" and "2" represent the front and rear of the stage, respectively.

[0088] A flow characteristic identification method based on characteristic flow area. It includes the following steps:

[0089] First, the actual steam inlet flow rate of the unit is calculated: For a high-pressure cylinder where the extraction steam flow rate is directly proportional to the main steam flow rate and the steam velocity is subsonic, the stage group consisting of the pressure stages from the first pressure stage after the regulating stage to the exhaust port of the high-pressure cylinder is taken as the research object. Using the proposed improved method, the actual steam inlet flow rate G of the unit under various actual operating conditions is calculated:

[0090]

[0091] In the formula, G, p, and υ represent the main steam flow rate (%), working fluid pressure (Pa), and specific volume (m³ / kg), respectively; the subscript 0 represents a known per-unit operating condition or standard operating condition, 1 represents before the stage group, 2 represents after the stage group, and d represents the cylinder exhaust point where the stage group is located. The same improvement method is used to refine its analytical expression:

[0092]

[0093] Secondly, the section from the high-pressure cylinder inlet to the regulating stage is selected as the stage segment. With the total valve position and valve openings remaining constant, the flow area of ​​the studied stage segment remains unchanged. Therefore, the characteristic flow area method can be applied to calculate the steam flow rate G that corrects the actual main steam pressure (condition B) to a specific main steam pressure (condition A). A :

[0094]

[0095] By identifying and calculating, the relationship between the unit's total valve position and the steam inlet flow rate under a certain main steam pressure is obtained, i.e., the main steam control valve flow characteristic curve, thereby achieving the purpose of identifying the flow characteristics of the control valve group.

[0096] Step S4: Using data mining algorithms, combined with historical data and stable operating condition identification elements, perform big data data mining and clustering calculations to extract target data reflecting the flow characteristics of the unit under varying operating conditions. Step S4 includes the following steps:

[0097] Based on the analysis of various classic data mining algorithms, the K-Means and K-Medoids algorithms are selected as suitable for traffic characteristic mining. Both of these algorithms belong to the partitioning methods in cluster analysis and have been widely used. However, the K-Means algorithm suffers from the cusp problem, while the iterative calculation of the K-Medoids algorithm is blind. Therefore, this invention makes reasonable improvements to the K-Medoids data mining algorithm.

[0098] This improved mining algorithm still uses Euclidean distance to calculate the distance between objects, takes the data object in the cluster that is closest to the cluster center as the representative of the cluster, uses the squared error criterion as the objective function, and measures the cost of replacing the cluster center according to the cost function, that is, to measure the average dissimilarity between the data object and the representative object.

[0099] The computation steps of the improved K-Medoids algorithm are as follows:

[0100] First, based on the total valve position values ​​of all actual operating conditions, select at least N (N>250) data rows (objects) as the coarse initial centers of the clusters, and assign all operating condition data rows to the clusters closest to them.

[0101] y i ={x k ||x FDEM,k -FDEM i |≤δ}

[0102]

[0103] In the formula x k FDEM i , δ, y i P i These represent a data row, total valve position value, total valve position value scale, cluster center, and cluster set for a specific operating condition.

[0104] Then, the data row with the working condition closest to the cluster mean is used as the initial center of the cluster, and all data rows are assigned to the cluster closest to it.

[0105]

[0106] Finally, to improve clustering quality, iterative calculations are performed to replace representative data rows with non-representative data rows, and a cost function is used to estimate the clustering quality.

[0107]

[0108] This algorithm uses a cost function to measure the impact of replacing cluster centers with non-cluster centers on clustering quality. The specific calculation method for the cost function is as follows: if the non-cluster center O... h Replace cluster center O i Then for each non-cluster center object O j The cost of this replacement is C. jih The cost function is as follows:

[0109]

[0110]

[0111] The total cost of this replacement (TC) ih It is all non-cluster-center objects O j Cost C jih The sum, if TC ih A value less than zero indicates that the sum of squared errors decreases after the replacement. h It is O i A good alternative. Then, O h Replace O i Become the representative center, and re-cluster all non-cluster center objects according to the new cluster center set:

[0112]

[0113] The improved K-Medoids algorithm is implemented as follows: i) Select at least k data rows as coarse initial centers of clusters based on the total valve position (or flow command) value of the data rows, and assign all data rows to the nearest cluster; ii) Specify the data row closest to the mean of objects within the cluster as the initial center of the cluster; iii) Repeat; iv) Assign all data rows to the nearest cluster; v) Select the fifty non-center points O closest to the cluster center; vi) Calculate the total cost TC of replacing Oj with O to form a new set. j Select the smallest TC n vii) If TC n <0, use the corresponding O n Replace O j A new set of k center points is formed; viiii) Until no further changes are made.

[0114] Step S5: Fit the actual traffic characteristics using deep learning methods, and tune the logic configuration parameters based on the algorithm results. Step S5 includes the following steps:

[0115] Comprehensive valve position modeling and prediction based on BP neural network.

[0116] Based on the estimates from the neural network model, the logic configuration correction parameters are tuned, and then the corrected flow characteristics are tested.

[0117] The specific embodiments described herein are merely illustrative of the principles and effects of the invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this invention should still be covered by the claims of this invention.

Claims

1. A method for identifying and optimizing the flow characteristics of steam turbines based on big data mining, characterized in that, Includes the following steps: Step S1: Analyze the flow characteristics, flow distribution, and overlap factors of each valve to obtain the characteristics of the DEH valve group; Step S2: The process of transforming raw data into standardized data suitable for analysis and modeling through data preprocessing, with the aim of improving data quality and usability; Step S3: Taking power and flow rate as the starting points respectively, and combining valve group characteristic analysis, study the identification method of flow characteristics under the test conditions of the unit to obtain the flow characteristic curve of the steam turbine under stable operating conditions. Step S4: Using data mining algorithms, combined with historical data and stable operating condition identification elements, perform big data mining and clustering calculations to extract target data that reflects the flow characteristics of the unit under varying operating conditions. Step S5: Use deep learning methods to fit the actual traffic characteristics and tune the logic configuration parameters based on the algorithm results.

2. The method for identifying and optimizing turbine flow characteristics based on big data mining according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Valve characteristic calibration test; Step S12: Analyze the flow characteristics, flow distribution, and overlap factors of each valve by following the valve flow characteristics.

3. The method for identifying and optimizing turbine flow characteristics based on big data mining according to claim 2, characterized in that, Step S12 includes the following steps: Step S121: Determine the structural characteristics of the control valve: nonlinear characteristics of relative stroke and relative flow, i.e., cam characteristics; Step S122: Determine the flow characteristics of the regulating valve itself. These characteristics reveal the relationship between the flow rate through the valve and its lift and the pressure ratio before and after it. Specifically, this involves the relationship curves between the actual flow rate and the critical flow rate, and between the opening degree and the critical flow rate. Step S123: Determine the flow characteristics of the steam turbine (opening degree of the steam turbine regulating valve group - flow curve). The regulation of the turbine flow rate is accomplished jointly by the regulating valve and the subsequent nozzle assembly, which is determined by the structural parameters and operating parameters of the regulating valve and the nozzle assembly.

4. The method for identifying and optimizing turbine flow characteristics based on big data mining according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: For conventional power plant turbine units, select the required parameters based on their principle thermodynamic system diagram; Step S22: Encode the raw data parameters and perform data cleaning, data transformation, data integration, and data reduction and normalization operations; Step S23: In a simple and comprehensive way, explore and observe the operating characteristics of the unit, and plot the changes in the total valve position value and main steam pressure value under the unit output load to obtain their changing patterns; Step S24: Obtain a thermally stable operating condition dataset by filtering and processing stable data.

5. The method for identifying and optimizing turbine flow characteristics based on big data mining according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Calculate the main steam flow rate of the steam turbine based on nozzle flow rate calculation, Freund's formula, and characteristic flow area path; Step S32: Taking power and flow rate as the starting point, different research objects are selected, and a flow characteristic identification method based on characteristic flow area is proposed.

6. The method for identifying and optimizing turbine flow characteristics based on big data mining according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Using this flow characteristic as the basis for linearity and optimization adjustment, analyze the basic requirements for optimizing the unit's flow characteristics; Step S42: Use the piecewise linear optimization improved K-Medoids algorithm to perform mining calculations on the identification results.

7. The method for identifying and optimizing turbine flow characteristics based on big data mining according to claim 1, characterized in that, Step S5 includes the following steps: Step S51: Comprehensive valve position modeling and prediction based on BP neural network; Step S52: Based on the neural network model estimate, tune the logic configuration correction parameters, and then perform a test on the corrected flow characteristics.

Images