Variable-speed pumped storage power station dynamic characteristic numerical simulation method, equipment and medium
By using the feature line method and the improved Sut curve method in variable speed pumping storage units to establish the runner system model, combined with CNN and vector control, the problems of insufficient modeling accuracy and single control strategy in the existing technology are solved, and the problem of High-precision hydraulic-mechanical-electrical coupled dynamic simulation, supporting fast switching of power-first and speed-first modes.
Patent Information
- Application Number
- CN202410980379.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-15
AI Technical Summary
The existing technology has problems such as biased research objects, insufficient modeling accuracy and single control strategy in the modeling and simulation of variable speed pumping storage units. In particular, the nonlinear characteristics on the hydraulic side and the complexity of the water pump turbine are ignored, making it difficult to accurately Simulate dynamic characteristics of hydraulic-mechanical-electrical coupling.
The flow channel system model and water pump turbine model based on the feature line method and the improved Suth curve method are adopted, and the convolutional neural network (CNN) and vector control strategies are combined to establish a double-feed induction motor model to realize active/reactive decoupling control. And simulate the multi-physical quantum transition process through a collaborative controller.
Improve the simulation accuracy of the hydraulic mechanical and electrical side, and can effectively simulate the dynamic response characteristics of the hydraulic-mechanical-electrical coupling of variable-speed pumped storage power station, realize the rapid switching of power-first and speed-first modes, and have high simulation flexibility. .
Smart Images

Figure CN118965644A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of hydropower generation, and more specifically, relates to a method, device and medium for numerical simulation of dynamic characteristics of a variable speed pumped storage power station, which are used for refined numerical simulation of a variable speed pumped storage power station using a doubly-fed induction motor technology, and for simulating the multi-physical quantity transition process of the hydraulic, mechanical, electrical and other subsystems in the system. Background Art
[0002] Variable speed pumped storage units have advantages such as speed, flexibility, high efficiency, and reliability in operation and regulation. They have now become a new development direction for the global pumped storage industry. However, there are currently no variable speed pumped storage units officially put into operation in my country. A pumped storage power station including a pumped storage unit is a complex nonlinear power system that integrates hydraulic, mechanical, and electrical characteristics. The application of power electronic equipment such as AC devices in variable speed units can realize the decoupling control of active power and reactive power, but it also changes the coupling mode and dynamic response characteristics of hydraulic, mechanical, and electrical multi-physical quantities of conventional fixed speed pumped storage units. my country's research on variable speed pumped storage technology is still in its infancy. Therefore, modeling and simulation of variable speed pumped storage power stations are very necessary. For variable speed pumped storage units with a rated power greater than 100MW, the doubly fed induction motor technology is currently more commonly used. The present invention mainly uses this technology to perform mathematical modeling and numerical simulation calculations on variable speed pumped storage power stations.
[0003] Summarizing the existing background technology, the research on modeling and simulation of variable speed pumped storage units has the following main deficiencies: 1) Research object: At present, the mathematical modeling and numerical simulation of pumped storage power stations in China are mainly aimed at conventional fixed speed units, but less at variable speed units; 2) Modeling accuracy: Other current studies have made a lot of simplifications in the modeling and simulation of variable speed pumped storage power stations, especially the simplification of the hydraulic machinery side. Many studies only focus on the modeling of the electrical side and the controller, while the hydraulic side uses a simplified linear model, ignoring the compressibility and head loss of hydraulic transient flow in the flow channel system. In addition, many studies ignore the nonlinear characteristics of the pump turbine and use a linear expression. Therefore, the application scope and simulation accuracy of this model are difficult to guarantee; 3) Control strategy: At present, most of the modeling studies on variable speed units using doubly fed induction motors only use a single control mode, and lack a method to simultaneously model and compare the speed priority mode and the power priority mode. Summary of the invention
[0004] The present invention mainly solves and improves the defects existing in the prior art, and provides a method for numerical simulation of the dynamic characteristics of a variable-speed pumped-storage power station. The method constructs a refined mathematical model of a variable-speed pumped-storage unit using a doubly-fed induction motor, in which the pump-turbine model can simulate the full-operating characteristics of the unit, establishes a doubly-fed induction motor model that can realize active / reactive decoupling control, reasonably simplifies the mathematical models of the machine-side converter and the grid-side converter, and adopts a vector control strategy. The method can simultaneously simulate the transition process of multiple physical quantities on the hydraulic side and the electrical side, and can fully reflect the hydraulic-mechanical-electrical coupled dynamic characteristics of a variable-speed pumped-storage power station.
[0005] In order to achieve the above object, the present invention provides a method for numerical simulation of dynamic characteristics of a variable speed pumped storage power station, comprising the following steps: Step (1): Establish a flow channel system model based on the characteristic line method and a pump-turbine model based on the improved Souter curve method Step 1: Construct the flow channel system model.
[0006] Step 1.1: According to the pipeline layout, the flow channel system is divided into several pipelines, where the pipelines are connected by different nodes, including pump turbines, upstream and downstream reservoirs, upstream and downstream surge tanks, pipeline bifurcation points or series connection points, etc. Step 1.2: Assign parameters to a single pipeline. The characteristic parameters of the pipeline include the length L , pipeline head and end elevation H , pipe diameter D , pipe cross-sectional area A , water hammer wave velocity a , loss coefficient f , the number of pipeline segments n , the following is a brief introduction to the characteristic line method model of a single pipeline: The basic equation for water hammer in a pipe taking friction into account is:
[0007]
[0008] The characteristic line method is usually used to solve this set of basic equations. The basic idea of the characteristic line method is to transform the quasi-linear partial differential equations into a set of equivalent ordinary differential equations through linear transformation.
[0009]
[0010]
[0011] On the basis of dividing the grid into equal time periods, this set of ordinary differential equations is integrated along the characteristic lines, and the friction terms are numerically integrated to obtain the following algebraic equation:
[0012]
[0013] After sorting, it can be rewritten as
[0014]
[0015] Where:
[0016] Step 1.3: Assign parameters of upstream and downstream reservoirs: It is necessary to clarify the parameters of the pipeline connected to the reservoir and the flow and head of the pipeline; For upstream reservoirs:
[0017] Energy equation:
[0018] Where: is the pressure of the upstream reservoir; is the area of the pipe inlet. The above two equations can be transformed into a quadratic equation, which can be solved using analytical formulas.
[0019] For downstream reservoirs:
[0020] Energy equation:
[0021] Where: is the pressure of the downstream reservoir; is the area of the pipe inlet. The above two equations can be transformed into a quadratic equation, which can be solved using analytical formulas.
[0022] Step 1.4: To model and assign parameters to pipeline connection points (including series connection points, bifurcation points, intersection points, etc.), it is necessary to clarify the connection pipeline parameters of the node, and pass the pipeline parameters and the flow and head of the pipeline as input signals to the connection point module. The following takes the bifurcation point as an example to briefly introduce its mathematical model. Pipeline 1 is upstream of the bifurcation point, and pipelines 2 and 3 are downstream of the bifurcation point.
[0023]
[0024]
[0025]
[0026]
[0027]
[0028]
[0029] Step 1.5: Model and assign parameters to the upstream and downstream surge tanks. Ignore the water inertia and along-the-line losses in the surge tanks during the calculation process.
[0030] Energy equation for surge chamber:
[0031] In the formula Impedance loss factor.
[0032] Continuity equation:
[0033]
[0034]
[0035] Basic equation of surge chamber:
[0036] Step 1.6: Connect the components of the runner system to complete the modeling of the runner system.
[0037] Step 2: Build a pump-turbine model.
[0038] Step 2.1: Use convolutional neural network (CNN) to interpolate and encrypt the full characteristic curve of the pump turbine. The process is as follows: Step 2.1.1: Data collection and preparation: Collect the full characteristic curve data of the pump-turbine, including input variables (flow rate, head) and output variables (power, efficiency). Ensure that the data set covers the full operating range of the pump-turbine, covering different flow rates and head conditions.
[0039] Step 2.1.2: Data preprocessing: Preprocess the collected data, including removing outliers, filling missing values, and standardizing data. Data enhancement techniques, such as rotation, translation, and scaling, can also be used to expand the training data set and increase the diversity and generalization ability of the data.
[0040] Step 2.1.3: Build CNN model: Build a CNN model of the full characteristic curve of the pump-turbine in the deep learning framework, including convolutional layers, pooling layers, and fully connected layers. Design a suitable network structure, including the size of the convolution kernel, step size, padding method, and the selection of activation function.
[0041] Step 2.1.4: Model training: Divide the preprocessed dataset into a training set and a validation set. Use the data from the training set to input the CNN model and optimize the weights and biases of the model through back propagation. Select a suitable loss function to measure the difference between the predicted value and the true value through mean square error or mean absolute error. By iterative training and adjusting model parameters, gradually improve the performance and fitting ability of the model.
[0042] Step 2.1.4: Interpolation and encryption: Use the trained CNN model to interpolate and encrypt the original pump-turbine full characteristic curve data. Input the specific input variable data of the pump-turbine (e.g., flow and head), and obtain the corresponding output variable data (e.g., power and efficiency) through the forward propagation of the CNN model. Linear interpolation or quadratic B-spline interpolation can be used to fill in the missing data of the full characteristic curve to obtain a complete characteristic curve.
[0043] Step 2.2: Extract characteristic parameters, i.e. guide vane opening, based on the full characteristic curve of the pump turbine y , Unit speed n 11 , Unit moment M 11 and unit flow Q 11 .
[0044]
[0045] in D 1 is the inlet diameter of the pump turbine.
[0046] Step 2.3 Use the improved Souter curve processing method to solve the interpolation problem of the S area in the full characteristic curve of the pump-turbine.
[0047] Since the full characteristic curve of the pump-turbine may cross, overlap, or have multiple values in the S region (the region between the turbine region and the anti-pump region), it is difficult to use for interpolation calculation. In order to solve this problem, the present invention uses an improved logarithmic projection method to process the full characteristic curve of the pump-turbine, converting the torque characteristic curve of the pump-turbine into a WM curve and converting the flow characteristic curve of the pump-turbine into a WH curve. The conversion expression is as follows:
[0048] The converted horizontal axis x The expression is:
[0049] In the formula, .
[0050] Step 2.4 In the model, the volute pressure and tailwater pipe pressure of the computer group are calculated using the upstream and downstream pipes connected to the unit, so as to obtain physical quantities such as unit speed, and use them for the simulation of the pump-turbine.
[0051] Step 3: Interpolation to obtain the pump-turbine head Step 3.1: Take the relay stroke output by the speed regulator as input and calculate the guide vane opening per unit value based on the base value , by looking up the table to obtain the actual opening of the guide vane ; Step 3.2: Get the horizontal coordinate of the improved Suter curve based on the flow rate and speed of the previous step x ; Step 3.3: Guide vane opening a (Step 3.1) and the horizontal coordinate of the improved Suter curve x (Step 3.2) Based on the improved Suter curve interpolation WH ; Step 3.4: According to WH Calculate actual head H ; Step 4: Interpolate to obtain the pump-turbine torque Step 4.1: Guide vane opening a (Step 3.1) and the horizontal coordinate of the improved Suter curve x (Step 3.2) Based on the improved Suter curve interpolation WM ; Step 4.2: Calculate the actual torque ; Step 5: Calculate the speed of the unit. Based on the electromagnetic torque command and the mechanical torque calculated by the model, the current speed of the unit can be calculated according to the motor balance equation:
[0052] Step (2): Establish the doubly-fed induction motor, converter, power grid and speed regulator system model Step 1: Build a doubly-fed induction motor model Step 1.1: The doubly-fed induction motor is decoupled from the power grid. The relationship between the angular frequency of the rotor current and voltage of the doubly-fed induction motor and the angular frequency of the stator current and voltage and the mechanical angular frequency of the rotor is:
[0053] Where: ω r is the rotor current voltage angular frequency, ω s is the stator current voltage angular frequency, ω m is the rotor angular frequency, ,Ω m is the mechanical rotation frequency of the rotor.
[0054] Step 1.2: For the model of the doubly-fed induction motor, the present invention adopts a state space matrix representation based on the dq axis, and the specific form is shown in the following figure:
[0055]
[0056]
[0057] Where: i ds , i qs , i dr , i qr They are stator d-axis current, stator q-axis current, rotor d-axis current, and rotor q-axis current respectively; L s , L r , L m They are stator inductance, rotor inductance and field inductance respectively; v ds , v qs , v dr , v qr They are stator d-axis voltage, stator q-axis voltage, rotor d-axis voltage, and rotor q-axis voltage respectively; R s and R r are the stator resistance and the rotor resistance respectively.
[0058] Step 1.3: The active and reactive powers on the rotor side and stator side are decoupled and controlled. The expressions of the active and reactive powers of the doubly-fed induction motor are as follows: P s , Q s , P r , Q r They are stator active power, stator reactive power, rotor active power and rotor reactive power respectively:
[0059] Step 1.4: The expression of the electromagnetic torque of the doubly-fed induction motor is as follows:
[0060] Step 1.5: The expression of the balance equation of the doubly-fed induction motor is as follows:
[0061] Where: T a is the unit inertia time constant, p is the number of motor pole pairs, T m is the mechanical torque, T em is the electromagnetic torque.
[0062] Step 2: Build the converter model Step 2.1: The present invention only considers the machine-side converter, which adopts a vector control strategy. It should be noted that the present invention does not consider the complex power electronic components and corresponding electromagnetic characteristics of the converter; Step 2.2: In the present invention, if the unit is in the power priority control mode, the converter is used to control the active power and reactive power. It should be noted that the active power here refers to the sum of the rotor active power and the stator active power. The output signal of the converter is the rotor voltage, and its corresponding expression is:
[0063]
[0064] Step 2.3: In the present invention, if the unit is in the speed priority control mode, the converter is used to control the speed and reactive power; Step 3: Construct a power grid model. The present invention simplifies the power grid model and only considers the electrical side physical quantities such as the active power command, reactive power command, and power grid frequency given by the power grid to the corresponding variable speed pumped storage unit.
[0065] Step 4: Build the governor model Step 4.1: Using the parallel PID regulation rule, after the signal is output, it is first amplified by the auxiliary relay, and then the hard feedback link performs negative feedback regulation on it, and then it is input into the main relay to control the guide vane opening. Assume that the time constant of the auxiliary relay with a smaller magnitude is ignored. ; Step 4.2: If the input value is the frequency setting and the feedback value is the unit frequency, the speed regulator model under the frequency control mode is obtained:
[0066] in is the permanent slip coefficient; is the main servomotor time constant; is the proportional gain; is the integral gain; is the differential gain; is the frequency deviation; Step 4.3: If the input value is the power given and the feedback value is the unit mechanical power, the governor model under the power control mode is obtained:
[0067] in is the power deviation; Step (3): Establish the cooperative controller model Step 1: Using the full characteristic curve of the pump-turbine, find the optimal speed for the unit to achieve the highest efficiency under different operating conditions under the corresponding parameter settings; Step 2: Given the initial net water head value of the unit; Step 3: Set different guide vane openings first y , initial speed N , the unit torque is obtained by interpolating the characteristic curve M 11 and unit flow Q 11 Then, the unit power and flow rate at this time are calculated based on the expressions of unit torque, unit flow rate, absolute torque and absolute flow rate; Step 4: The unit efficiency can be calculated by calculating the power, flow rate and net water head of the unit under a certain working condition. The expression of the unit efficiency is:
[0068] Step 5: Change the variable parameters of the unit (such as speed and guide vane opening) but keep the initial power of the unit unchanged, and obtain the unit operating efficiency at different parameter points. By comparing the different efficiencies of the unit under the same head and the same power when the speed and guide vane opening change, the highest efficiency and the corresponding optimal speed of the unit under the head and power are obtained; Step 6: Change the initial water head and power values and start the next solution process from Step 2; Step 7: Collect all the points corresponding to different water heads and different powers, and use the radial basis function interpolation method to interpolate and encrypt the interpolation table of the cooperative controller.
[0069] Step 7.1: Select the Gaussian function as the basis for radial basis function interpolation and determine its standard deviation σ ; Step 7.2: Construct a system of linear equations. For each data point, use it as the center of the radial basis function, calculate the Euclidean distance between it and other data points, and calculate the value of the radial basis function based on the distance.
[0070] Step 7.3: Construct a linear equation system based on the constructed radial basis function and the corresponding optimal speed value. The form of the linear equation system is AX = B ,in A is the coefficient matrix of the radial basis function, X is the unknown coefficient vector, B is the known optimal speed vector; Step 7.4: Solve the linear equations: Use Gaussian elimination to solve the linear equations and get the unknown coefficient vector X ; Step 7.5: Prediction and interpolation: For the encrypted head and power data points, calculate the distance between them and the known data points according to the form of radial basis function. Based on the distance and the solved unknown coefficient vector X , calculate the predicted value of the optimal speed.
[0071] Step (4): Connect the above model modules to build the overall model and set the model parameters Step 1: The input of the speed regulator module is ω m , ω mref (in power priority mode) or and (In speed priority mode), output y Connect to the pump-turbine module; Step 2: Speed regulator output yAs the input of the pump-turbine module, the electromagnetic torque output by the doubly-fed induction motor T em It also serves as the input of the pump turbine and the output of the pump turbine ω m Connect to the speed regulator module; Step 3: For the flow system, the flow and water head of the pipeline will be transmitted as output signals to the node modules connected to both ends of the pipeline module (including pump turbines, upstream and downstream reservoirs, upstream and downstream surge tanks, connection points, etc.), and each node will also transmit the flow and water head of the node as input signals to the pipeline module; Step 4: For the doubly-fed induction motor, the input signals given by the power grid are the power grid frequency ω s , stator voltage V s , the input signal given by the pump turbine is the unit speed ω m , the input signal given by the converter is the rotor voltage V r The output signals of the doubly-fed induction motor module are: the electromagnetic torque output to the pump-turbine module T em , the rotor current input to the converter module i r , stator q-axis voltage V qs , rotor current voltage angular frequency ω r , Unit active power P , Unit reactive power Q .
[0072] Step 5: The input of the machine-side converter is the input signal stator electrical angle frequency given by the power grid ω s , Active power command P * , reactive power command Q * , Unit active power P , Unit reactive power Q and the rotor current i r , stator q-axis voltage V qs , rotor current voltage angular frequency ω r .
[0073] Step (5): Change the module connection logic according to the control strategy and set the control command for simulation Step 1: When the unit adopts power priority control, the unit power is controlled by the converter, the speed is controlled by the governor, and the power control command P * Input converter module, optimal speed Speed control command Input speed regulator module; Step 2: When speed priority control is adopted, the speed is controlled by the converter and the power is controlled by the speed regulator. The optimal speed Speed control command Input converter module, power control command Input speed regulator module; Step 3: Set the power control command or speed control command and perform simulation calculation; Step 4: Check the time domain simulation results of each physical quantity; Compared with the prior art, the present invention has the following advantages and effects: An electronic device comprises a computer-readable storage medium storing computer-executable instructions; and one or more processors, the one or more processors being coupled to the computer-readable storage medium and configured to execute the computer-executable instructions so that the device performs a method according to any one of claims 1-6.
[0074] A readable storage medium stores computer executable instructions, which, when executed by a processor, configure the processor to execute the method according to any one of claims 1 to 6.
[0075] Advantages: (1) The mathematical model and numerical simulation calculation method of the variable speed pumped storage unit using the doubly fed induction motor technology designed in the present invention adopts a pump-turbine model processed by the improved Souter curve method, a flow system model processed by the characteristic line method, and a relatively complete doubly fed induction motor and converter model. It has high simulation accuracy on the hydraulic mechanical side and the electrical side, and can effectively simulate the hydraulic-mechanical-electrical coupling dynamic response characteristics of the variable speed unit.
[0076] (2) The modeling method proposed in the present invention can realize rapid switching between the power priority mode and the speed priority control mode of the variable speed pumped storage unit using a doubly fed induction motor by changing the connection logic between subsystem modules, and has high simulation flexibility. BRIEF DESCRIPTION OF THE DRAWINGS
[0077] Figure 1 It is a connection diagram of the flow channel system of the variable speed pumped storage power station of the present invention; Figure 2 A model diagram of a pipeline element in the flow channel system of the present invention; Figure 3A model diagram of a reservoir in the flow channel system of the present invention; Figure 4 A model diagram of the series connection points in the flow channel system of the present invention; Figure 5 A model diagram of a water pump turbine in the flow channel system of the present invention; Figure 6 This is a vector control block diagram of the machine-side converter of the present invention; Figure 7 This is a control block diagram of the speed regulator of the present invention; Figure 8 It is a connection diagram of all modules of the present invention; Figure 9a -f is the calculation simulation result of the model in the present invention under a certain working condition.
[0078] Fig.10 It is a simulation flow chart of the present invention. DETAILED DESCRIPTION
[0079] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in each step of the present invention described below can be applied to simulate any other simulation object with a reasonable parameter combination.
[0080] First, let me introduce the principle of this method. The present invention provides a method for numerically simulating dynamic characteristics of a variable speed pumped storage power station, comprising the following steps: Step (1): Establish a flow channel system model based on the characteristic line method and a pump-turbine model based on the improved Souter curve method Step 1: Construct the flow channel system model.
[0081] Step 1.1: According to the pipeline layout, the flow channel system is divided into several pipelines, where the pipelines are connected by different nodes, including pump turbines, upstream and downstream reservoirs, upstream and downstream surge tanks, pipeline bifurcation points or series connection points, etc. Step 1.2: Assign parameters to a single pipeline. The characteristic parameters of the pipeline include the length L , pipeline head and end elevation H , pipe diameter D , pipe cross-sectional area A , water hammer wave velocity a , loss coefficient f , the number of pipeline segments n , the following is a brief introduction to the characteristic line method model of a single pipeline: The basic equation for water hammer in a pipe taking friction into account is:
[0082]
[0083] The characteristic line method is usually used to solve this set of basic equations. The basic idea of the characteristic line method is to transform the quasi-linear partial differential equations into a set of equivalent ordinary differential equations through linear transformation.
[0084]
[0085]
[0086] On the basis of dividing the grid into equal time periods, this set of ordinary differential equations is integrated along the characteristic lines, and the friction terms are numerically integrated to obtain the following algebraic equation:
[0087]
[0088] After sorting, it can be rewritten as
[0089]
[0090] Where:
[0091] Step 1.3: Assign parameters of upstream and downstream reservoirs: It is necessary to clarify the parameters of the pipeline connected to the reservoir and the flow and head of the pipeline; For upstream reservoirs:
[0092] Energy equation:
[0093] Where: is the pressure of the upstream reservoir; is the area of the pipe inlet. The above two equations can be transformed into a quadratic equation, which can be solved using analytical formulas.
[0094] For downstream reservoirs:
[0095] Energy equation:
[0096] Where: is the pressure of the downstream reservoir; is the area of the pipe inlet. The above two equations can be transformed into a quadratic equation, which can be solved using analytical formulas.
[0097] Step 1.4: To model and assign parameters to pipeline connection points (including series connection points, bifurcation points, intersection points, etc.), it is necessary to clarify the connection pipeline parameters of the node, and pass the pipeline parameters and the flow and head of the pipeline as input signals to the connection point module. The following takes the bifurcation point as an example to briefly introduce its mathematical model. Pipeline 1 is upstream of the bifurcation point, and pipelines 2 and 3 are downstream of the bifurcation point.
[0098]
[0099]
[0100]
[0101]
[0102]
[0103]
[0104] Step 1.5: Model and assign parameters to the upstream and downstream surge tanks. Ignore the water inertia and along-the-line losses in the surge tanks during the calculation process.
[0105] Energy equation for surge chamber:
[0106] In the formula Impedance loss factor.
[0107] Continuity equation:
[0108]
[0109]
[0110] Basic equation of surge chamber:
[0111] Step 1.6: Connect the components of the runner system to complete the modeling of the runner system.
[0112] Step 2: Build a pump-turbine model.
[0113] Step 2.1: Use convolutional neural network (CNN) to interpolate and encrypt the full characteristic curve of the pump turbine. The process is as follows: Step 2.1.1: Data collection and preparation: Collect the full characteristic curve data of the pump-turbine, including input variables (flow rate, head) and output variables (power, efficiency). Ensure that the data set covers the full operating range of the pump-turbine, covering different flow rates and head conditions.
[0114] Step 2.1.2: Data preprocessing: Preprocess the collected data, including removing outliers, filling missing values, and standardizing data. Data enhancement techniques, such as rotation, translation, and scaling, can also be used to expand the training data set and increase the diversity and generalization ability of the data.
[0115] Step 2.1.3: Build CNN model: Build a CNN model of the full characteristic curve of the pump-turbine in the deep learning framework, including convolutional layers, pooling layers, and fully connected layers. Design a suitable network structure, including the size of the convolution kernel, step size, padding method, and the selection of activation function.
[0116] Step 2.1.4: Model training: Divide the preprocessed dataset into a training set and a validation set. Use the data from the training set to input the CNN model and optimize the weights and biases of the model through back propagation. Select a suitable loss function to measure the difference between the predicted value and the true value through mean square error or mean absolute error. By iterative training and adjusting model parameters, gradually improve the performance and fitting ability of the model.
[0117] Step 2.1.4: Interpolation and encryption: Use the trained CNN model to interpolate and encrypt the original pump-turbine full characteristic curve data. Input the specific input variable data of the pump-turbine (e.g., flow and head), and obtain the corresponding output variable data (e.g., power and efficiency) through the forward propagation of the CNN model. Linear interpolation or quadratic B-spline interpolation can be used to fill in the missing data of the full characteristic curve to obtain a complete characteristic curve.
[0118] Step 2.2: Extract characteristic parameters, i.e. guide vane opening, based on the full characteristic curve of the pump turbine y , Unit speed n 11 , Unit moment M 11 and unit flow Q 11 .
[0119]
[0120] in D 1 is the inlet diameter of the pump turbine.
[0121] Step 2.3 Use the improved Souter curve processing method to solve the interpolation problem of the S area in the full characteristic curve of the pump-turbine.
[0122] Since the full characteristic curve of the pump-turbine may cross, overlap, or have multiple values in the S region (the region between the turbine region and the anti-pump region), it is difficult to use for interpolation calculation. In order to solve this problem, the present invention uses an improved logarithmic projection method to process the full characteristic curve of the pump-turbine, converting the torque characteristic curve of the pump-turbine into a WM curve and converting the flow characteristic curve of the pump-turbine into a WH curve. The conversion expression is as follows:
[0123] The converted horizontal axis x The expression is:
[0124] In the formula, .
[0125] Step 2.4 In the model, the volute pressure and tailwater pipe pressure of the computer group are calculated using the upstream and downstream pipes connected to the unit, so as to obtain physical quantities such as unit speed, and use them for the simulation of the pump-turbine.
[0126] Step 3: Interpolation to obtain the pump-turbine head Step 3.1: Take the relay stroke output by the speed regulator as input and calculate the guide vane opening per unit value based on the base value , by looking up the table to obtain the actual opening of the guide vane ; Step 3.2: Get the horizontal coordinate of the improved Suter curve based on the flow rate and speed of the previous step x ; Step 3.3: Guide vane opening a (Step 3.1) and the horizontal coordinate of the improved Suter curve x (Step 3.2) Based on the improved Suter curve interpolation WH ; Step 3.4: According to WH Calculate actual head H ; Step 4: Interpolate to obtain the pump-turbine torque Step 4.1: Guide vane opening a (Step 3.1) and the horizontal coordinate of the improved Suter curve x (Step 3.2) Based on the improved Suter curve interpolation WM ; Step 4.2: Calculate the actual torque ; Step 5: Calculate the speed of the unit. Based on the electromagnetic torque command and the mechanical torque calculated by the model, the current speed of the unit can be calculated according to the motor balance equation:
[0127] Step (2): Establish the doubly-fed induction motor, converter, power grid and speed regulator system model Step 1: Build a doubly-fed induction motor model Step 1.1: The doubly-fed induction motor is decoupled from the power grid. The relationship between the angular frequency of the rotor current and voltage of the doubly-fed induction motor and the angular frequency of the stator current and voltage and the mechanical angular frequency of the rotor is:
[0128] Where: ω r is the rotor current voltage angular frequency, ω s is the stator current voltage angular frequency, ω m is the rotor angular frequency, ,Ω m is the mechanical rotation frequency of the rotor.
[0129] Step 1.2: For the model of the doubly-fed induction motor, the present invention adopts a state space matrix representation based on the dq axis, and the specific form is shown in the following figure:
[0130]
[0131]
[0132] Where: i ds , i qs , i dr , i qr They are stator d-axis current, stator q-axis current, rotor d-axis current, and rotor q-axis current respectively; L s , L r , L m They are stator inductance, rotor inductance and field inductance respectively; v ds , v qs , vdr , v qr They are stator d-axis voltage, stator q-axis voltage, rotor d-axis voltage, and rotor q-axis voltage respectively; R s and R r are the stator resistance and the rotor resistance respectively.
[0133] Step 1.3: The active and reactive powers on the rotor side and stator side are decoupled and controlled. The expressions of the active and reactive powers of the doubly-fed induction motor are as follows: P s , Q s , P r , Q r They are stator active power, stator reactive power, rotor active power and rotor reactive power respectively:
[0134] Step 1.4: The expression of the electromagnetic torque of the doubly-fed induction motor is as follows:
[0135] Step 1.5: The expression of the balance equation of the doubly-fed induction motor is as follows:
[0136] Where: T a is the unit inertia time constant, p is the number of motor pole pairs, T m is the mechanical torque, T em is the electromagnetic torque.
[0137] Step 2: Build the converter model Step 2.1: The present invention only considers the machine-side converter, which adopts a vector control strategy. It should be noted that the present invention does not consider the complex power electronic components and corresponding electromagnetic characteristics of the converter; Step 2.2: In the present invention, if the unit is in the power priority control mode, the converter is used to control the active power and reactive power. It should be noted that the active power here refers to the sum of the rotor active power and the stator active power. The output signal of the converter is the rotor voltage, and its corresponding expression is:
[0138]
[0139] Step 2.3: In the present invention, if the unit is in the speed priority control mode, the converter is used to control the speed and reactive power; Step 3: Construct a power grid model. The present invention simplifies the power grid model and only considers the electrical side physical quantities such as the active power command, reactive power command, and power grid frequency given by the power grid to the corresponding variable speed pumped storage unit.
[0140] Step 4: Build the governor model Step 4.1: Using the parallel PID regulation rule, after the signal is output, it is first amplified by the auxiliary relay, and then the hard feedback link performs negative feedback regulation on it, and then it is input into the main relay to control the guide vane opening. Assume that the time constant of the auxiliary relay with a smaller magnitude is ignored. ; Step 4.2: If the input value is the frequency setting and the feedback value is the unit frequency, the speed regulator model under the frequency control mode is obtained:
[0141] in is the permanent slip coefficient; is the main servomotor time constant; is the proportional gain; is the integral gain; is the differential gain; is the frequency deviation; Step 4.3: If the input value is the power given and the feedback value is the unit mechanical power, the governor model under the power control mode is obtained:
[0142] in is the power deviation; Step (3): Establish the cooperative controller model Step 1: Using the full characteristic curve of the pump-turbine, find the optimal speed for the unit to achieve the highest efficiency under different operating conditions under the corresponding parameter settings; Step 2: Given the initial net water head value of the unit; Step 3: Set different guide vane openings first y , initial speed N , the unit torque is obtained by interpolating the characteristic curve M 11 and unit flow Q 11 Then, the unit power and flow rate at this time are calculated based on the expressions of unit torque, unit flow rate, absolute torque and absolute flow rate; Step 4: The unit efficiency can be calculated by calculating the power, flow rate and net water head of the unit under a certain working condition. The expression of the unit efficiency is:
[0143] Step 5: Change the variable parameters of the unit (such as speed and guide vane opening) but keep the initial power of the unit unchanged, and obtain the unit operating efficiency at different parameter points. By comparing the different efficiencies of the unit under the same head and the same power when the speed and guide vane opening change, the highest efficiency and the corresponding optimal speed of the unit under the head and power are obtained; Step 6: Change the initial water head and power values and start the next solution process from Step 2; Step 7: Collect all the points corresponding to different water heads and different powers, and use the radial basis function interpolation method to interpolate and encrypt the interpolation table of the cooperative controller.
[0144] Step 7.1: Select the Gaussian function as the basis for radial basis function interpolation and determine its standard deviation σ ; Step 7.2: Construct a system of linear equations. For each data point, use it as the center of the radial basis function, calculate the Euclidean distance between it and other data points, and calculate the value of the radial basis function based on the distance.
[0145] Step 7.3: Construct a linear equation system based on the constructed radial basis function and the corresponding optimal speed value. The form of the linear equation system is AX = B ,in A is the coefficient matrix of the radial basis function, X is the unknown coefficient vector, B is the known optimal speed vector; Step 7.4: Solve the linear equations: Use Gaussian elimination to solve the linear equations and get the unknown coefficient vector X ; Step 7.5: Prediction and interpolation: For the encrypted head and power data points, calculate the distance between them and the known data points according to the form of radial basis function. Based on the distance and the solved unknown coefficient vector X , calculate the predicted value of the optimal speed.
[0146] Step (4): Connect the above model modules to build the overall model and set the model parameters Step 1: The input of the speed regulator module is ω m , ω mref (in power priority mode) or and (In speed priority mode), output y Connect to the pump-turbine module; Step 2: Speed regulator output y As the input of the pump-turbine module, the electromagnetic torque output by the doubly-fed induction motor T em It also serves as the input of the pump turbine and the output of the pump turbine ω m Connect to the speed regulator module; Step 3: For the flow system, the flow and water head of the pipeline will be transmitted as output signals to the node modules connected to both ends of the pipeline module (including pump turbines, upstream and downstream reservoirs, upstream and downstream surge tanks, connection points, etc.), and each node will also transmit the flow and water head of the node as input signals to the pipeline module; Step 4: For the doubly-fed induction motor, the input signals given by the power grid are the power grid frequency ω s , stator voltage V s , the input signal given by the pump turbine is the unit speed ω m , the input signal given by the converter is the rotor voltage V r The output signals of the doubly-fed induction motor module are: the electromagnetic torque output to the pump-turbine module T em , the rotor current input to the converter module i r , stator q-axis voltage V qs , rotor current voltage angular frequency ω r , Unit active power P , Unit reactive power Q .
[0147] Step 5: The input of the machine-side converter is the input signal stator electrical angle frequency given by the power grid ω s , Active power command P * , reactive power command Q * , Unit active power P , Unit reactive power Q and the rotor current i r , stator q-axis voltage V qs , rotor current voltage angular frequency ωr .
[0148] Step (5): Change the module connection logic according to the control strategy and set the control command for simulation Step 1: When the unit adopts power priority control, the unit power is controlled by the converter, the speed is controlled by the governor, and the power control command P * Input converter module, optimal speed Speed control command Input speed regulator module; Step 2: When speed priority control is adopted, the speed is controlled by the converter and the power is controlled by the speed regulator. The optimal speed Speed control command Input converter module, power control command Input speed regulator module; Step 3: Set the power control command or speed control command and perform simulation calculation; Step 4: Check the time domain simulation results of each physical quantity; The following is an introduction with specific cases.
[0149] The system includes seven parts: station flow system, pump turbine, speed governor, double-fed induction motor, converter, cooperative controller and power grid. Figure 1 As shown. The speed regulator has a power control mode and a frequency control mode, and the converter can also control the active power of the motor or the speed of the unit. The purpose of the present invention is to propose a mathematical model and a numerical simulation calculation method for the system, so as to dynamically simulate the hydraulic-mechanical-electrical coupling transition process of the variable speed pumped storage unit using a double-fed induction motor, and simulate the dynamic response law of multiple physical quantities of the system.
[0150] To illustrate the effect of the present invention, a variable speed pumped storage unit using a double-fed induction motor in a power priority control mode under a turbine operating condition is used as an implementation object of the present invention to describe the method of the present invention in detail: Step (1): Establish a flow channel system model based on the characteristic line method and a pump-turbine model based on the improved Souter curve method Assume that the flow channel system of the pumped storage power station is designed as a one-hole-one-machine water diversion pipeline with upstream and downstream surge chambers. The pipeline can be divided into 6 sections. The schematic diagram of the flow channel system is shown in the figure below. Figure 2 As shown. From upstream to downstream, they are upstream reservoir, pipeline 1, upstream surge chamber, pipeline 2, series connection point 1, pipeline 3, pump turbine, pipeline 4, series connection point 2, pipeline 5, downstream surge chamber, pipeline 6. Among them, the model diagram of a single pipeline is as follows Figure 2 The model diagram of the reservoir is shown in Figure 3As shown, the model diagram of the series connection point is as follows Figure 4 The model diagram of the pump turbine is shown in Figure 5 shown.
[0151] Step (2): Establish the doubly-fed induction motor, converter, power grid and speed regulator system model Using the state space matrix based on the dq axis, the doubly fed induction motor model is obtained. Using the vector control strategy, the converter model is obtained, such as Figure 6 The power grid model is simplified to several sets of given signal inputs. The speed regulator model uses PID control. Figure 7 When the unit is in power priority control mode, the speed governor is in frequency control mode, and the control signal switch is turned to the unit speed control circuit ( ); When the unit is in speed priority control mode, the speed governor is in power control mode, and the control signal switch is turned to the power control circuit ( ). In this embodiment, the control signal switch is turned to the power control loop.
[0152] Step (3): Establish the cooperative controller model Input power command , assuming that the water levels of the upstream and downstream reservoirs remain unchanged during the simulation period, when the unit switches to different powers, the optimal speed is output through interpolation ; When in power priority control mode, Output to the speed regulator, which is the simulation working condition of this embodiment; when in the speed priority control mode, Output to the converter.
[0153] Step (5): Connect the above model modules to build the overall model and set the model parameters The above modules are connected according to the data flow transmission logic to form the overall model. The connection sequence between modules is as follows: Figure 8 shown.
[0154] Step (6): Change the module connection logic according to the control strategy and set the control command for simulation Since this embodiment adopts the power priority control mode, its module connection logic has been described in the above steps. The specific module connection sequence is as follows: Figure 8 To switch to the speed priority control mode, just follow Figure 7 Just switch the speed regulator mode.
[0155] Finally, a 40% load reduction disturbance condition is set for simulation. The specific simulation results are as follows: Figure 9a -f, Figure 9a For power simulation (including power command, electromagnetic power, stator active power and rotor active power), Figure 9bis the simulation result of torque and guide vane opening, Fig.9c is the simulation result of electromagnetic angular frequency of the unit rotor, Figure 9d is the simulation result of the pump turbine flow rate, Fig.9e is the simulation result of unit current / voltage, Figure 9f This is the pressure simulation at the volute and tailwater pipe of the pump-turbine. It can be seen that by using the converter to preferentially control the active power of the unit, the active power of the unit can quickly track the active power command and smoothly change the output; the unit speed is controlled by the speed governor; compared with conventional units, the variable speed pumped storage unit using the doubly fed induction motor realizes variable speed operation, reflecting a higher flexible adjustment performance.
[0156] It is easy for those skilled in the art to understand that the above-described working conditions are only preferred embodiments of the present invention and are not intended to limit the present invention. Any example analysis, equivalent substitution, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for numerical simulation of dynamic characteristics of a variable speed pumped storage power station, characterized in that: The following steps are involved: Use the MATLAB Function module in MATLAB / Simulink to establish a flow channel system model based on the characteristic line method and a pump-turbine full-characteristic model based on the improved Suter curve; Establish the doubly-fed induction motor, converter, power grid and speed regulator system model; Establish a cooperative controller model; Connect the above model modules to build the overall model and set the model parameters; Change the module connection logic according to different calculation conditions and set the control commands for simulation.
2. A method for numerical simulation of dynamic characteristics of a variable speed pumped storage power station according to claim 1, characterized in that: The flow channel system model includes upstream and downstream reservoirs, surge chambers, connection points, and pipelines. The pump-turbine adopts a nonlinear model under all working conditions, and the improved Souter curve method is used to process the full characteristic curve of the pump-turbine.
3. A method for numerical simulation of dynamic characteristics of a variable speed pumped storage power station according to claim 1, characterized in that: The doubly-fed induction power system has established a state space model based on the dq axis, the converter has adopted a simplified vector control strategy, the power grid has adopted a simplified model, given active and reactive power, voltage and other instructions, the speed regulator has adopted a PID control structure.
4. A method for numerical simulation of dynamic characteristics of a variable speed pumped storage power station according to claim 1, characterized in that: The cooperative controller adopts the optimal efficiency curve calculated based on the full characteristic curve of the pump turbine, and obtains the speed at which the unit can currently achieve the highest efficiency, that is, the optimal speed, under the condition of given head and power.
5. A method for numerical simulation of dynamic characteristics of a variable speed pumped storage power station according to claim 1, characterized in that: The overall model includes seven parts: flow channel system, pump turbine, doubly fed induction motor, converter, cooperative controller, speed regulator and power grid. The model parameters include hydraulic side parameters, mechanical side parameters and electrical side parameters.
6. A method for numerical simulation of dynamic characteristics of a variable speed pumped storage power station according to claim 1, characterized in that: The control strategies include power priority control strategy and speed priority control strategy, and the control commands will be transmitted from the cooperative controller to the speed regulator and the converter according to the control strategies.
7. An electronic device, characterized in that: A computer-readable storage medium storing computer-executable instructions; and one or more processors, the one or more processors being coupled to the computer-readable storage medium and configured to execute the computer-executable instructions so that the device performs the method according to any one of claims 1-6.
8. A readable storage medium, characterized in that: Computer executable instructions are stored, which, when executed by a processor, configure the processor to perform the method according to any one of claims 1-6.
Citation Information
Cited By
Dynamic management system for water turbine
CN119356117A
A dynamic management system for water turbines
CN119356117B
Doubly-fed variable-speed pumped storage unit control method and device based on model prediction
CN119393282A
Transition process simulation method for water-electricity-electricity long-time and short-time step length coupling of variable-speed pumped storage power station
CN119538677A
A transient simulation method for coupling long and short time steps of hydro-mechanical and electrical systems in variable-speed pumped storage power stations
CN119538677B
