A method for optimizing shutdown rule of hydroelectric generating unit considering maximum value of rotating speed rise and maximum value of water head fluctuation

Abstract

This invention provides an optimization method for the shutdown behavior of hydropower units considering the extreme values ​​of speed rise and head fluctuation, comprising the following steps: S1, acquiring turbine modeling data; S2, constructing a high-precision turbine model based on a BP neural network; S3, constructing a high-precision servo system model; S4, building a turbine regulation system simulation model using Octave; S5, constructing a comprehensive objective function considering the extreme values ​​of speed rise, head fluctuation, and standard ITAE index; S6, optimizing the guide vane closing inflection point and closing speed of the hydropower unit by combining the snowmelt optimization algorithm, the turbine regulation system simulation model, and the comprehensive objective function. The optimization method for the shutdown behavior of hydropower units designed in this invention, considering the extreme values ​​of speed rise and head fluctuation, can fully account for the influence of the shutdown behavior on the extreme values ​​of speed rise and head fluctuation, effectively ensuring the safe and stable operation of the hydropower unit.

Description

Technical Field

[0001] This invention relates to the field of safe and stable operation technology of hydropower units, and in particular to an optimization method for the shutdown law of hydropower units that considers the maximum and minimum values ​​of speed rise and head fluctuation. Background Technology

[0002] Hydropower units play a vital role in modern power systems, providing stable and renewable energy. However, the operating environment of hydropower units is complex and variable, exhibiting significant nonlinearity, time-varying characteristics, and strong coupling. Achieving safe, efficient, and stable operation during the operation and shutdown of hydropower units is a critical issue. Especially during shutdown, improper handling can cause drastic fluctuations in speed and head, leading to equipment damage and system instability. Therefore, optimizing the shutdown process is of great significance for ensuring the safe operation of hydropower units and extending their service life.

[0003] Currently, most hydropower unit shutdown processes employ traditional methods, which rely heavily on experience and simple control strategies, thus having certain limitations. Traditional shutdown methods often overlook the maximum speed increase during shutdown. During rapid shutdown, the turbine speed may surge instantaneously, exceeding safety limits and leading to equipment wear or even damage. Head fluctuation is a significant factor affecting the stability of hydropower units. Traditional shutdown methods fail to adequately consider the maximum and minimum head fluctuations, potentially causing water hammer effects and damaging pipelines and equipment. Furthermore, traditional shutdown methods are mostly static strategies, failing to dynamically adjust according to actual operating conditions and thus unable to achieve optimal shutdown results. In particular, traditional shutdown operations heavily rely on operator experience, lacking systematic optimization strategies, making it difficult to guarantee optimal shutdown results in complex and variable operating environments.

[0004] To overcome the limitations of traditional shutdown methods, modern control theory and optimization algorithms have been introduced into the study of hydropower unit shutdown processes. These methods, by considering the system's dynamic characteristics and various constraints, can achieve better shutdown behavior design. Commonly used optimization methods include: Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Grey Wolf Optimization (GWO).

[0005] The Snow Melt Optimization Algorithm (SMOA) is an emerging intelligent optimization algorithm inspired by the physical phenomena of snow melting. This algorithm simulates the position and state changes of snowflakes during the melting process, iteratively optimizing to gradually approach the optimal solution. SMOA possesses advantages such as strong global search capability, fast convergence speed, and ease of implementation, making it suitable for solving complex nonlinear optimization problems.

[0006] Given the limitations of traditional hydropower unit shutdown methods in failing to adequately consider the extreme values ​​of speed rise and head fluctuation, it is necessary to propose a new method for optimizing the guide vane closing inflection point and closing speed of hydropower units, comprehensively taking into account factors such as the extreme values ​​of speed rise and head fluctuation. This optimization method will significantly improve the operational safety and economy of hydropower units, and reduce equipment damage and system instability caused by speed and head fluctuations. Summary of the Invention

[0007] This invention addresses the limitations of traditional hydropower unit shutdown methods that do not adequately consider the maximum and minimum values ​​of speed rise and head fluctuation. It proposes an optimization method for hydropower unit shutdown patterns that takes into account these maximum and minimum values ​​of speed rise and head fluctuation.

[0008] To achieve the above-mentioned technical features, the objective of this invention is as follows: An optimization method for the shutdown behavior of a hydroelectric generator unit considering the maximum and minimum values ​​of rotational speed rise and head fluctuation, comprising the following steps:

[0009] S1, Obtain turbine modeling data;

[0010] S2, Construct a high-precision turbine model based on a BP neural network;

[0011] S3, construct a high-precision servo system model;

[0012] S4, a simulation model of the turbine regulation system was built based on the Octave tool;

[0013] S5, construct a comprehensive objective function that considers the maximum and minimum values ​​of rotational speed rise, head fluctuation, and standard ITAE index;

[0014] S6 combines the snowmelt optimization algorithm, the turbine regulation system simulation model, and the comprehensive objective function to optimize the guide vane closing inflection point and closing speed of the hydro-generator unit.

[0015] Preferably, step S1 includes the following steps:

[0016] Building a high-precision turbine model requires detailed data support. This data comes from actual operating turbine systems. The process of obtaining turbine modeling data includes data collection, data preprocessing, and data verification steps.

[0017] S11, Data Collection:

[0018] Collect operational and experimental data of the water turbine system;

[0019] S12, Data Preprocessing:

[0020] The collected data is preprocessed to ensure its quality and consistency.

[0021] S13, Data Verification:

[0022] The preprocessed data is validated to confirm its accuracy and reliability.

[0023] Preferably, the operational data in S11 includes:

[0024] Input parameters: head, flow rate, guide vane opening;

[0025] Output parameters: speed, output power, torque;

[0026] Experimental data include:

[0027] Laboratory test data: Conduct performance tests on the water turbine to obtain accurate input-output relationship data;

[0028] Field test data: Conduct field tests in a real operating environment to obtain real operating data.

[0029] Preferably, the data preprocessing in S12 specifically includes:

[0030] S121, Data Cleaning:

[0031] Outlier handling: Detect and remove obvious outliers;

[0032] Missing value handling: Use data completion methods to fill in missing data;

[0033] S122, Data Conversion:

[0034] Normalization: Scaling the data to the same range to speed up model training and improve accuracy;

[0035] Feature extraction: Extract useful features from the raw data: unit rotational speed, guide vane opening, unit flow rate, and unit torque;

[0036] S123, Data Partitioning:

[0037] Training set: The main dataset used for model training, accounting for 70%-80% of the total data;

[0038] Validation set: Used for model tuning and selection, accounting for 10%-15% of the total data;

[0039] Test set: Used for final evaluation of model performance, accounting for 10%-15% of the total data.

[0040] Preferably, the data verification in S13 specifically includes:

[0041] S131, Statistical Analysis:

[0042] Descriptive statistics: Analyzing the basic statistical properties of data, including: variance, root mean square error, and standard deviation;

[0043] Correlation analysis: Analyze the correlation between each input parameter and the output parameter, and select parameters with high correlation for modeling;

[0044] S132, Visual Analysis:

[0045] Scatter plot: Draw a scatter plot between input and output parameters to visually display the data distribution and relationships;

[0046] Time series plot: Draw a time series plot to observe the patterns and periodicity of data changes over time;

[0047] S133, Data Consistency Check:

[0048] Check the consistency of data across different time periods and operating conditions to ensure data reliability.

[0049] Preferably, step S2 specifically comprises:

[0050] Based on the BP neural network, a neural network model for flow characteristics and a neural network model for torque characteristics are constructed, with unit rotational speed and guide vane opening as inputs and turbine flow rate and torque as outputs.

[0051] Preferably, step S3 specifically comprises:

[0052] Construct a high-precision servo system model. The servo system includes nonlinear elements with time delay, saturation, speed limiting, and dead zone. Its linear transfer function is:

[0053] ;

[0054] In the formula, K y This refers to the overall amplifier coefficient; T y1 and T y These are the response time constants of the intermediate relay and the main relay, respectively. For the nonlinear part of the servo system, the focus is on the guide vane opening inflection point and closing speed.

[0055] Preferably, step S4 specifically comprises:

[0056] A simulation model of the turbine regulation system was built using the Octave tool. The simulation model includes a water intake system model, a turbine model, a generator and power grid model, and a governor and servo system model.

[0057] (1) Water diversion system model:

[0058] The water intake system adopts a rigid water hammer model, which is mathematically described as follows:

[0059] ;

[0060] In the formula: T w The inertial time constant of the water flow;

[0061] (2) Generator and power grid model:

[0062] Based on an infinite power grid, the transfer function of the generator and load is:

[0063] ;

[0064] In the formula: T a The inertial time constant of the unit. e g This is the generator load self-adjustment coefficient;

[0065] (3) Controller model:

[0066] The transfer function of the linear part of the PID controller is:

[0067] ;

[0068] In the formula: T 1v The time constant of the differential element; u For controller output; e For tracking error; K P , K I and K D These are the proportional, integral, and derivative gains, respectively; considering practical situations, the values ​​are taken when the unit is operating in speed or power mode. K D =0, taken in no-load operation mode K D These are the original parameters of the power plant.

[0069] Preferably, step S5 specifically comprises:

[0070] Construct a system that considers the maximum and minimum speed increases, the maximum and minimum head fluctuations, and the standard ITAE index. J ITAE The comprehensive objective function; where, J ITAE This is a commonly used objective function for PID parameter optimization, as shown in the following formula:

[0071] ;

[0072] In the formula, t This refers to the actual time. t s This represents the total simulation time. e ( t The integral of the turbine control signal error is denoted as .

[0073] A comprehensive objective function that includes the maximum and minimum speed rise, the maximum and minimum head fluctuation, and the standard ITAE index. J 综合 As shown in the following formula:

[0074] ;

[0075] In the formula, f 1 represents the maximum increase in engine speed under the current shutdown pattern; f 2 represents the maximum head fluctuation under the current shutdown pattern.

[0076] Preferably, step S6 specifically comprises:

[0077] The optimal combination of guide vane closing inflection point and closing speed during the shutdown process of a hydroelectric generator unit is obtained by combining the snow melting optimization algorithm, the simulation model of the turbine regulation system, and the comprehensive objective function.

[0078] The present invention has the following beneficial effects:

[0079] 1. The shutdown law optimization method of hydropower unit designed in this invention, which considers the maximum and minimum values ​​of speed rise and head fluctuation, can fully take into account the influence of shutdown law on the maximum and minimum values ​​of speed rise and head fluctuation, and effectively ensure the safe and stable operation of hydropower unit. Attached Figure Description

[0080] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0081] Figure 1 The method design steps of this invention are described below.

[0082] Figure 2 This is a schematic diagram of the high-precision servo system model of the present invention.

[0083] Figure 3 This is a schematic diagram of the turbine regulation system model of the present invention.

[0084] Figure 4 The calculation process of the snow melting optimization algorithm of this invention is as follows. Detailed Implementation

[0085] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0086] Example 1:

[0087] Reference Figure 1-4This embodiment provides an optimization method for the shutdown behavior of hydropower units considering the maximum and minimum values ​​of speed rise and head fluctuation. It comprises three parts: constructing a high-precision numerical simulation model of the turbine regulating system; constructing a comprehensive objective function based on the snowmelt optimization algorithm; and optimizing the guide vane closing inflection point and closing speed of the hydropower unit. The main implementation steps include: acquiring turbine modeling data; constructing a high-precision turbine model based on a BP neural network; constructing a high-precision servo system model; building a turbine regulating system simulation model using Octave; constructing a comprehensive objective function considering the maximum and minimum values ​​of speed rise, head fluctuation, and the standard ITAE index; and combining the snowmelt optimization algorithm, the turbine regulating system simulation model, and the comprehensive objective function to optimize the guide vane closing inflection point and closing speed of the hydropower unit. This optimization method for the shutdown behavior of hydropower units considering the maximum and minimum values ​​of speed rise and head fluctuation can fully account for the impact of the shutdown behavior on these values, effectively ensuring the safe and stable operation of the hydropower unit.

[0088] To address the limitations of traditional hydropower unit shutdown methods that fail to adequately consider the extreme values ​​of speed rise and head fluctuation, this invention proposes an optimization method for hydropower unit shutdown patterns that takes into account these factors comprehensively. This method will significantly improve the operational safety and economy of hydropower units, and reduce equipment damage and system instability caused by speed and head fluctuations.

[0089] Example 2:

[0090] This embodiment provides an optimization method for the shutdown pattern of a hydropower unit that considers the maximum and minimum values ​​of rotational speed rise and head fluctuation, including the following steps:

[0091] (1) Obtain water turbine modeling data.

[0092] Building a high-precision hydro turbine model requires detailed data support, which typically comes from actual operating hydro turbine systems. The process of acquiring hydro turbine modeling data includes multiple steps such as data collection, data preprocessing, and data validation. Detailed steps are explained below:

[0093] 1) Data collection:

[0094] Data collection is the foundation of modeling and mainly includes the following aspects of data:

[0095] 1.1) Running data:

[0096] Input parameters: head, flow rate, guide vane opening, etc.

[0097] Output parameters: speed, output power, torque, etc.

[0098] 1.2) Experimental data:

[0099] Laboratory test data: Conduct performance tests on the water turbine to obtain accurate input-output relationship data.

[0100] Field test data: Conduct field tests in a real operating environment to obtain real operating data.

[0101] 2) Data preprocessing:

[0102] Collected data often requires preprocessing to ensure its quality and consistency. Data preprocessing includes the following steps:

[0103] 2.1) Data cleaning:

[0104] Outlier handling: Detect and remove obvious outliers to avoid negative impacts on model training.

[0105] Missing value handling: Use interpolation, mean imputation, or other data completion methods to fill in missing data.

[0106] 2.2) Data Transformation:

[0107] Normalization: Scaling data to the same range (such as [0,1] or [-1,1]) helps to speed up model training and improve accuracy.

[0108] Feature extraction: Extract useful features from the raw data, such as unit rotational speed, guide vane opening, unit flow rate, and unit torque.

[0109] 2.3) Data partitioning:

[0110] Training set: The main dataset used for model training, accounting for 70%-80% of the total data.

[0111] Validation set: Used for model tuning and selection, accounting for 10%-15% of the total data.

[0112] Test set: Used for final evaluation of model performance, accounting for 10%-15% of the total data.

[0113] 3) Data validation:

[0114] Data validation is an important step in ensuring data quality, as it confirms the accuracy and reliability of the data.

[0115] 3.1) Statistical Analysis:

[0116] Descriptive statistics: Analyzing the basic statistical characteristics of data, such as variance, root mean square error, and standard deviation.

[0117] Correlation analysis: Analyze the correlation between each input parameter and the output parameter, and select parameters with high correlation for modeling.

[0118] 3.2) Visual Analysis:

[0119] Scatter plot: Draw a scatter plot between input and output parameters to visually display the data distribution and relationships.

[0120] Time series plot: Draw a time series plot to observe the patterns and periodicity of data changes over time.

[0121] 3.3) Data consistency check:

[0122] Check the consistency of data across different time periods and operating conditions to ensure data reliability.

[0123] (2) To ensure model accuracy, a flow characteristic neural network model (DCNN) was constructed based on a BP neural network, with unit rotational speed and guide vane opening as inputs and turbine flow rate and torque as outputs. Q 11 = Q 11 ( n 11 , Y ()) and torque characteristic neural network model (TCNN, M 11 = M 11 ( n 11 , Y )).

[0124] (3) Construct a high-precision servo system model.

[0125] The servo system includes nonlinear elements such as time delay, saturation, speed limiting, and dead zone. Its linear transfer function is:

[0126] (1)

[0127] in: K y This refers to the overall amplifier coefficient; T y1 and T y These are the response time constants of the intermediate relay and the main relay, respectively. For the nonlinear part of the servo system, the focus is on the guide vane opening inflection point and closing speed.

[0128] (4) A simulation model of the turbine regulation system was built based on the Octave tool.

[0129] The simulation model includes a water diversion system model, a turbine model, a generator and power grid model, and a governor and servo system model.

[0130] 1) Water diversion system model:

[0131] For simplicity, the water intake system can be modeled using a rigid water hammer model, mathematically described as follows:

[0132] (2)

[0133] In the formula T w is the inertial time constant of the water flow.

[0134] 2) Generator and power grid model:

[0135] Considering an infinite power grid, the transfer functions of the generator and load are:

[0136] (3)

[0137] In the formula T a The inertial time constant of the unit. e g This is the generator load self-adjustment coefficient.

[0138] 3) Controller model:

[0139] The transfer function of the linear part of the PID controller is:

[0140] (4)

[0141] In the formula: T 1v The time constant of the differential element; u For controller output; e For tracking error; K P , K I and K D These are the proportional, integral, and derivative gains, respectively. Considering practical situations, the following values ​​are used when the unit is operating in either speed or power mode: K D =0, taken in no-load operation mode K D These are the original parameters of the power plant.

[0142] (5) Construct a system that considers the maximum and minimum values ​​of rotational speed rise, head fluctuation, and standard ITAE index ( J ITAE The comprehensive objective function of ).

[0143] J ITAE This is a commonly used objective function for PID parameter optimization, as shown in the following formula:

[0144] (5)

[0145] In the formula, t This refers to the actual time. t s This represents the total simulation time. e ( t ) represents the integral of the turbine control signal error.

[0146] A comprehensive objective function that includes the maximum and minimum speed rise, the maximum and minimum head fluctuation, and the standard ITAE index. J 综合 As shown in the following formula:

[0147] (6)

[0148] In the formula, f 1 represents the maximum increase in engine speed under the current shutdown pattern; f 2 represents the maximum head fluctuation under the current shutdown pattern.

[0149] (6) The optimal combination of the guide vane closing inflection point and closing speed during the shutdown process of the hydropower unit is obtained by combining the snow melting optimization algorithm, the turbine regulation system simulation model and the comprehensive objective function.

[0150] The Snow Ablation Optimizer (SAO) is a nature-inspired metaheuristic optimization algorithm that solves optimization problems by mimicking the natural melting process of snow. The algorithm mainly includes an initialization phase, an exploration phase, a mining phase, and a dual-population mechanism. First, it initializes a set of solutions as a snow pile. Then, it updates the solutions by simulating the natural melting process of snow. This process calculates the melting rate of each solution, adjusts the position of the solution based on the rate, and iteratively optimizes through a fitness function until the optimal solution is obtained.

[0151] The advantage of the SAO algorithm lies in its simulation of the snow melting process, which achieves a good balance between exploration and utilization, effectively avoiding premature convergence and improving solution accuracy. Furthermore, it demonstrates high efficiency and robustness when handling complex optimization problems.

[0152] (1) Initialization phase:

[0153] In SAO, the iterative process begins with a randomly generated population, as shown in Equation (7). The entire population is modeled as a matrix with N rows and D columns, where N represents the size of the population and D represents the dimension of the solution space.

[0154] = (7)

[0155] In the formula, Z For the initial population, and Let them represent the lower and upper bounds of the solution space, respectively. This represents a random number in the range [0, 1].

[0156] (2) Exploration stage:

[0157] When snow or liquid water derived from snow is converted into steam, the search subject exhibits a highly dispersed characteristic due to irregular movement. Brownian motion is used to simulate this situation. Brownian motion, as a stochastic process, is widely used to simulate the foraging behavior of animals. For standard Brownian motion, the step size is obtained from the probability density function based on a normal distribution with a mean of zero and a variance of 1, and the relevant mathematical representation is shown in equation (8).

[0158] (8)

[0159] The exploration phase utilizes Gaussian Brownian motion to simulate the highly dispersed characteristics that occur when snow or water from melted snow transforms into vapor. Based on this process, the exploration location is updated, expanding the search area. The location update formula is as follows:

[0160] (9)

[0161] In the formula, for t In the nth iteration i The position of each particle. For random individuals in an elite collection, A random number between 0 and 1. Random numbers generated by a Gaussian distribution representing Brownian motion. Let be the center of mass of the entire particle. The corresponding formula is as follows:

[0162] (10)

[0163] (11)

[0164] In the formula, and These represent the second-best and third-best individuals in the current population, respectively. This indicates the centroid position of individuals whose fitness values ​​rank in the top 50%.

[0165] (12)

[0166] in The number of leaders is equal to half the size of the entire group. Therefore, in each iteration, Elite(t) is randomly selected from the set consisting of the current best solution, the second best individual, the third best individual, and the centroid position of the leaders.

[0167] (3) Mining stage:

[0168] The mining phase utilizes a snowmelt model (day-to-day method) to develop around the current optimal solution, thereby obtaining a better solution, rather than further expanding its highly dispersed functionality. The position update formula for this phase is as follows:

[0169] (13)

[0170] In the formula, A random number in the range [-1, 1] M The snow melting rate is calculated using the following formula:

[0171] (14)

[0172] In the formula: This represents the current iteration number; This represents the maximum number of iterations.

[0173] (4) Two-population mechanism

[0174] The SAO algorithm employs a dual-population mechanism to balance the needs of the exploration and mining phases. The entire particle population is randomly divided into two equal-sized subpopulations, each responsible for both exploration and mining. As the iteration progresses, the size of the subpopulation responsible for exploration gradually decreases, while the size of the subpopulation responsible for mining increases accordingly, ensuring that the algorithm maintains an effective balance between exploring new solution spaces and refining known solutions.

[0175] In summary, the complete position update equation of the SAO algorithm is as follows:

[0176] (15)

[0177] The present invention provides an optimization method for the shutdown behavior of hydropower units that considers the maximum and minimum values ​​of rotational speed rise and head fluctuation. This method can fully account for the impact of the shutdown behavior on the maximum and minimum values ​​of rotational speed rise and head fluctuation, effectively ensuring the safe and stable operation of hydropower units.

[0178] The specific embodiments of the present invention have been described in detail above, but these are merely one example, and the present invention is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications and substitutions to the present invention are also within the scope of the present invention. Therefore, all equivalent changes and modifications made without departing from the spirit and scope of the present invention should be covered within the scope of the present invention.

Claims

1. A method for optimizing shutdown rules of a hydroelectric generating unit considering maximum values of speed rise and water head fluctuation, characterized in that, The method comprises the following steps: S1, obtaining water turbine modeling data; S2, constructing a high-precision water turbine model based on a BP neural network; S3, constructing a high-precision servo system model; S4, building a water turbine regulating system simulation model based on an Octave tool; S5, constructing a comprehensive objective function considering the maximum value of speed rise, the maximum value of water head fluctuation and the standard ITAE index; S6, optimizing the guide vane closing inflection point and closing speed of the hydroelectric generating set by combining the snow melt optimization algorithm, the water turbine regulating system simulation model and the comprehensive objective function; The step S3 is specifically: The high-precision servo system model contains nonlinear links of time delay, saturation, speed limit and dead zone, and the linear part transfer function is: ； In the formula, K y is the comprehensive amplifier coefficient; T y1 and T y respectively, the intermediate force amplifier and the main force amplifier reaction time constant, for the nonlinear part of the follow-up system, focusing on the closing inflection point and closing speed of the guide vane opening; s is the Laplace operator; y is the guide vane opening control signal; u is the PID controller output signal; The step S4 is specifically: The water turbine regulating system simulation model is built based on the Octave tool, and the simulation model comprises a water diversion system model, a water turbine model, a generator and power grid model, and a governor and servo system model; (1) Water diversion system model: The rigid water hammer model is adopted for the water diversion system, and the mathematical description is: ； wherein: T w is the water flow inertia time constant; is the water system transfer function; (2) Generator and power grid model: Based on an infinite power grid, the transfer function of the generator and load is: ； In the formula: T a The inertial time constant of the unit. e g This is the generator load self-adjustment coefficient; (3) Controller model: The linear part transfer function of the PID controller is: ； where: T 1v is the differential element time constant; u is the controller output; e is the tracking error; K P , K I and K D are the proportional, integral and differential gains respectively; considering practical situations, take K D = 0 when the unit is running in speed or power mode, and take K D are the original parameters of the power plant; The step S5 is specifically: A comprehensive objective function considering the maximum value of speed rise, the maximum value of water head fluctuation and the standard ITAE index is constructed J ITAE A comprehensive objective function considering the maximum value of speed rise, the maximum value of water head fluctuation and the standard ITAE index is constructed wherein J ITAE is a commonly used objective function for PID parameter optimization, as shown in the following equation: ； wherein t is the actual time; t s is the total simulation time; e t is the water turbine control signal error integral;​ A comprehensive objective function including the maximum value of the rotating speed rise, the maximum value of the water head fluctuation, and a standard ITAE index J 综合 as shown in the following formula: ； In the formula, f 1 is the maximum value of the rotating speed under the current shutdown rule; f 2 is the maximum value of the water head fluctuation under the current shutdown rule; and is a weight coefficient.

2. The method of claim 1, wherein the method further comprises: The step S1 comprises the following steps: The construction of the high-precision water turbine model needs detailed data support, and the data is obtained from the actual running water turbine system, and the process of obtaining the water turbine modeling data comprises data collection, data preprocessing and data verification steps; S11, data collection: Collecting the running data and experimental data of the water turbine system; S12, data preprocessing: The collected data is preprocessed to ensure the quality and consistency of the data; S13, data verification: The data after preprocessing is verified to confirm the accuracy and reliability of the data.

3. The method of claim 2, wherein the method further comprises: The running data in S11 comprises: Input parameters: water head, flow, guide vane opening; Output parameters: speed, output power, torque; The experimental data comprises: Laboratory test data: performance test of the water turbine is carried out to obtain accurate input-output relationship data; Field test data: field test is carried out in the actual running environment to obtain real running data.

4. The method of claim 3, wherein the method further comprises: The data preprocessing in S12 specifically comprises: S121, data cleaning: Outlier processing: detect and remove obvious outliers; Missing value processing: use data completion method to fill in missing data; S122, data conversion: Normalization processing: scale the data to the same range to speed up the model training speed and improve the accuracy; Feature extraction: extract useful features from the original data: unit speed, guide vane opening, unit flow and unit torque; S123, data segmentation: Training set: the main data set for model training, accounting for 70%-80% of the total data; Verification set: used for model optimization and selection, accounting for 10%-15% of the total data; Test set: used for final evaluation of model performance, accounting for 10%-15% of the total data.

5. The method of claim 4, wherein the method further comprises: The data verification in S13 specifically comprises: S131, statistical analysis: Descriptive statistics: analyze the statistical characteristics of the data, including: variance, root mean square error, standard deviation; Correlation analysis: analyze the correlation between input and output parameters, select parameters with high correlation for modeling; S132, visual analysis: Scatter plot: draw scatter plots between input and output parameters to visually display data distribution and relationships; Time series plot: draw time series plots to observe the time variation and periodicity of the data; S133, data consistency check: Check the consistency of data at different time periods and under different working conditions to ensure the reliability of the data.

6. The method of claim 1, wherein the method further comprises: The step S2 is specifically: Based on the BP neural network, a flow characteristic neural network model and a torque characteristic neural network model are constructed, with unit speed and guide vane opening as input, and water turbine flow and torque as output, respectively.

7. The method of claim 1, wherein the method further comprises: The step S6 is specifically: The optimal combination of guide vane closing inflection point and closing speed in the shutdown process of the hydroelectric generating set is obtained by combining the snow melt optimization algorithm, the water turbine regulating system simulation model and the comprehensive objective function.

Images