Small-interference stability prediction and joint scheduling method and system for wind power-containing power system

Abstract

The invention discloses a small-interference stability prediction and joint scheduling method and system for a wind power-containing power system, and relates to the field of stability prediction and optimal scheduling of a power system, and the method comprises the steps: obtaining historical state information, and carrying out the first operation to determine accurate state information; historical stability information is calculated based on the accurate state information, and a system stability prediction model is built and trained; perturbation is conducted on the power of the generator set, and initial state information before and after perturbation is obtained; and finally, performing joint scheduling on the system by combining the initial state information after perturbation, the trained stability prediction model and the power adjustment model. According to the method, the accuracy and screening efficiency of the historical state information are improved, and the effectiveness of model training data is ensured; the scheduling calculation process is optimized, and the power scheduling amount is effectively distributed according to the scheduling demand, so that the operation efficiency and stability of the wind power-containing electric power system are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of power system stability prediction and optimal dispatching, and in particular to a method and system for predicting small disturbance stability and joint dispatching of a wind power power system. Background Art

[0002] In order to ensure the normal and stable operation of the power system, it is generally necessary to monitor and predict the power dispatch and stability information (minimum damping ratio) of the power system in real time. In the past, this was achieved through real-time monitoring and adjustment by power operation and maintenance personnel. With the development of the times, manual monitoring and adjustment methods have gradually been replaced by advanced machine learning technology.

[0003] In terms of stability information prediction and power dispatch of existing wind power power systems, the following problems mainly exist: wind power power systems are highly volatile and easily affected by multiple factors such as meteorological conditions and load changes, making the prediction of their stability information quite difficult; traditional prediction methods often rely too much on historical data and statistical models, making it difficult to adapt to the rapid changes of wind power power systems; the dispatch strategy of the energy storage system is not sophisticated and intelligent enough to adapt to the real-time needs of wind power power systems. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: how to improve the accuracy of stability information prediction of a wind power system under small disturbances, and optimize the scheduling of an energy storage system in the wind power system.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a method for predicting and jointly dispatching small disturbance stability of a wind power system, including:

[0008] Acquire historical status information of the wind power system, perform a first operation, and determine accurate status information;

[0009] Based on accurate status information, calculate historical stability information, build and train a system stability prediction model;

[0010] Perturb the power of the generator set of the wind power system to obtain the initial state information before and after the perturbation;

[0011] Based on the initial state information before and after the perturbation, the trained system stability prediction model and the power adjustment model, the wind power power system is jointly dispatched.

[0012] As a preferred solution for the small disturbance stability prediction and joint dispatching method of the wind power system, the present invention provides:

[0013] The wind power system includes: a wind power generation system, a thermal power generation system and an energy storage system; the historical status information includes multiple groups of status information at different historical moments; the status information includes at least node voltage amplitude and phase angle;

[0014] Performing measurement mutation detection on historical status information and performing a first operation; determining accurate status information of the wind power system based on the historical status information after the first operation;

[0015] Based on accurate state information and a small disturbance stability model of the wind power system, calculate historical stability information; build a system stability prediction model based on machine learning technology; use historical state information as input and historical stability information as output to train the system stability prediction model;

[0016] The power of the generator sets of the wind power system is perturbed, and the initial state information of the wind power system before and after the perturbation is obtained; based on the initial state information of the wind power system before and after the perturbation, the trained system stability prediction model, and the power adjustment model of the wind power system, scheduling calculations are performed to obtain the power scheduling amount; the power scheduling amount is allocated to complete the joint scheduling of the wind power system.

[0017] As a preferred solution for the small disturbance stability prediction and joint dispatching method of the wind power system, the present invention provides:

[0018] The first operation of performing measurement mutation detection on the historical state information includes:

[0019] Establish the first objective function, expressed as:

[0020] J 1 (x 1 )=[z 1 -h(x 1 )] T R -1 [z 1 -h(x 1 )]

[0021] In the formula, J 1 (·) is the first objective function value, x 1 is the estimated state information, z 1 is each group of state information in the historical state information, h(·) is the measurement function, and R is the weight matrix;

[0022] Based on each group of state information in the historical state information, calculating the expected value and variance of the first objective function;

[0023] When the expected value or the variance is greater than a set threshold, deleting the current group state information;

[0024] Determining accurate status information of the wind power system based on historical status information after the first operation includes:

[0025] The second objective function is established and expressed as:

[0026] J 2 (x 2 )=[z 2 -h(x 2 )] T R -1 [z 2 -h(x 2 )]

[0027] In the formula, J 2 (·) is the second objective function value, x 2 For accurate status information, z 2 is each group of state information in the historical state information after mutation detection, h(·) is the measurement function, and R is the weight matrix;

[0028] Deriving a second objective function based on each group of state information in the historical state information after the first operation;

[0029] The derivative of the second objective function is converted into an equivalent equation and solved to obtain the optimal state information, thereby determining the accurate state information of the wind power system.

[0030] As a preferred solution for the small disturbance stability prediction and joint dispatching method of the wind power system, the present invention provides:

[0031] The calculation of historical stability information based on accurate state information and a small disturbance stability model of a wind power system includes:

[0032] Select the dependent variables of all derivatives in the equations of the small disturbance stability model as the state vector and construct the state matrix;

[0033] Calculating a plurality of eigenvalues ​​of the state matrix based on accurate state information;

[0034] Calculate multiple damping ratios according to each eigenvalue and the damping ratio formula;

[0035] All damping ratios are compared and the minimum damping ratio is determined.

[0036] As a preferred solution for the small disturbance stability prediction and joint dispatching method of the wind power system, the present invention provides:

[0037] The system stability prediction model based on machine learning technology includes:

[0038] Perform dimensionality reduction on state information and stability information to extract key features; construct predictions for stable information and combine them with optimization algorithms to improve prediction accuracy.

[0039] As a preferred solution for the small disturbance stability prediction and joint dispatching method of the wind power system, the present invention provides:

[0040] The power adjustment model is expressed as:

[0041]

[0042] In the formula, ΔPG i is the power dispatching amount of the i-th generator set, n is the number of generator sets, are the lower and upper power limits of the i-th generator set, respectively. is the initial power of the i-th generator set, ξ 0 is the initial damping ratio of the power system containing wind power, Δξ is the change in damping ratio of the power system containing wind power, ξ 1 is the minimum damping ratio threshold of the power system containing wind power, C i is the sensitivity of the i-th minimum damping ratio to the power of the generator set;

[0043] The scheduling calculation to obtain the power scheduling amount includes:

[0044] Perform measurement mutation detection on the initial state information of the wind power system before and after the perturbation, and determine the accurate state information of the wind power system before and after the perturbation;

[0045] The accurate state information of the wind power system before and after the perturbation is input into the trained system stability prediction model to obtain the minimum damping ratio of the wind power system before and after the perturbation;

[0046] The power adjustment model is solved based on the minimum damping ratio of the wind power system before and after the perturbation to obtain the power dispatch amount.

[0047] As a preferred solution for the small disturbance stability prediction and joint dispatching method of the wind power system, the present invention provides:

[0048] The allocating of the power scheduling amount includes:

[0049] Determine the adjustment range of each synchronous machine unit according to the output constraint and climbing constraint of each synchronous machine unit;

[0050] Based on the regulation range of all synchronous machine units, determine whether the regulation margin of all synchronous machine units meets the dispatching requirements; if so, allocate the power dispatching amount according to the proportion of the regulation margin of each synchronous machine unit; if not, dispatch according to the limit value of the regulation margin of each synchronous machine unit, and the excess power dispatching amount shall be borne by the energy storage system;

[0051] The power dispatch capacity of each synchronous machine unit meets the following requirements:

[0052]

[0053] In the formula, ΔPT i is the power dispatch amount undertaken by the i-th synchronous machine unit, ΔP is the power dispatch amount, ΔPT i min , ΔPT i max are the lower and upper limits of the regulation range of the i-th synchronous machine unit, are the lower and upper limits of the adjustment range of the jth synchronous machine unit, respectively, and m is the number of synchronous machine units;

[0054] The power dispatch capacity undertaken by the energy storage system meets the following requirements:

[0055]

[0056] Where ΔPB is the power dispatch amount borne by the energy storage system.

[0057] In a second aspect, an embodiment of the present invention provides a small disturbance stability prediction and joint dispatching system for a wind power system, including:

[0058] A data acquisition and processing module, used to acquire historical status information of the wind power system, perform a first operation, and determine accurate status information;

[0059] Stability analysis and modeling module, which is used to calculate historical stability information based on accurate state information, and to build and train system stability prediction models;

[0060] The disturbance test and initial state acquisition module is used to perturb the power of the generator set of the wind power system and obtain the initial state information before and after the perturbation;

[0061] The dispatch calculation and execution module is used to jointly dispatch the wind power system based on the initial state information before and after the perturbation, the trained system stability prediction model and the power adjustment model.

[0062] In a third aspect, an embodiment of the present invention provides a computing device, including:

[0063] Memory and processor;

[0064] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the one or more programs are executed by the one or more processors, the one or more processors implement the small disturbance stability prediction and joint scheduling method of the wind power system as described in any embodiment of the present invention.

[0065] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, can implement the small-disturbance stability prediction and joint dispatching method for a wind power system.

[0066] The beneficial effects of the present invention are as follows: the present invention utilizes historical state information and historical stability information of a wind power system to train a system stability prediction model; on the one hand, the present invention uses accurate state information after measurement mutation detection and a small disturbance model of a wind power system to calculate historical stability information, which not only realizes the screening and precision of historical state information, but also fully considers the impact of small disturbances on the stability of the wind power system, thereby ensuring the validity of the model training data; on the other hand, the present invention first utilizes kernel principal component analysis (KPCA) to perform dimensionality reduction processing on state information and stability information, extracts key features, and then constructs a prediction of stability information through a support vector machine (SVM), and then combines the K-Nearest Neighbor (KNN) algorithm to optimize and improve its prediction accuracy, thereby finally realizing accurate prediction of stability information (minimum damping ratio) of a wind power system. In addition, the present invention also uses the trained system stability prediction model and the power adjustment model of the wind power system to perform scheduling calculations, and effectively allocates the power scheduling amount according to the scheduling requirements, determines the power scheduling amount borne by the synchronous machine unit and / or the energy storage system, and further improves the operating efficiency of the wind power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0068] Figure 1 It is an overall flow chart of the small disturbance stability prediction and joint dispatching method of the wind power system according to the first embodiment of the present invention;

[0069] Figure 2It is a prediction effect diagram of a system stability prediction model in a simulation example of a method for predicting stability and joint dispatching of a wind power system containing small disturbances according to the second embodiment of the present invention;

[0070] Figure 3 This is a fitting effect diagram of a system stability prediction model in a simulation example of a method for small disturbance stability prediction and joint dispatching of a wind power system according to a second embodiment of the present invention. DETAILED DESCRIPTION

[0071] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0072] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0073] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0074] Example 1

[0075] Reference Figure 1 , which is the first embodiment of the present invention, provides a method for predicting and jointly dispatching small disturbance stability of a wind power system, including:

[0076] S1: Obtain historical status information of the wind power system, perform a first operation, and determine accurate status information;

[0077] In an embodiment of the present application, the status information of the wind power system is collected using a supervisory control and data acquisition (SCADA) system, and the collected status information includes node voltage amplitude and phase angle, line current, line active power and reactive power, load status, etc.

[0078] In another possible implementation, the following system can also be used to collect historical status information of a wind power system:

[0079] Distributed Digital Protective Relays:

[0080] Equipment used to monitor and protect power systems can collect current, voltage, power and other information in real time and provide data logging functions.

[0081] Smart Metering Systems:

[0082] Monitor the power consumption of users and equipment, collect data such as power quality and load conditions, and provide detailed power usage analysis.

[0083] Monitoring Sensor Networks:

[0084] Various sensors deployed on wind turbines, substations and transmission lines (such as temperature sensors, humidity sensors, vibration sensors, etc.) can provide detailed data for equipment health monitoring and performance analysis.

[0085] Energy Management Systems (EMS):

[0086] A tool specifically designed to monitor and control power systems that can collect status information on generators, energy storage devices, and loads, providing a more comprehensive view of the power system status.

[0087] Real-time Monitoring and Data Logging Systems:

[0088] This type of system focuses on real-time collection and recording of the status of power equipment, and can provide real-time data such as voltage, power, and frequency.

[0089] Meteorological Monitoring Systems:

[0090] It is used to collect meteorological data such as wind speed, wind direction, temperature and air pressure to evaluate the availability and efficiency of wind power resources, which indirectly affects the operating status of wind power and power systems.

[0091] Performing a measurement mutation detection on the historical state information and performing a first operation;

[0092] In the embodiment of the present application, the historical state information is subjected to a measurement mutation detection, and the first operation includes:

[0093] Establish the first objective function, expressed as:

[0094] J 1 (x 1 )=[z 1 -h(x 1 )] T R -1 [z 1 -h(x 1 )]

[0095] In the formula, J 1 (·) is the first objective function value, x 1 is the estimated state information, z 1 is each group of state information in the historical state information, h(·) is the measurement function, and R is the weight matrix;

[0096] Based on each group of state information in the historical state information, calculating the expected value and variance of the first objective function;

[0097] Specifically, the expected value and variance of the first objective function are calculated as follows:

[0098]

[0099] In the formula, E[J 1 (x 1 )] is the expected value of the first objective function, var[J 1 (x 1 )] is the variance of the first objective function, K is the order, is the bad data value weighted by the state information, w w,ii is the influence coefficient.

[0100] When the expected value or the variance is greater than a set threshold, the current group state information is deleted.

[0101] Exemplarily, the threshold of the expected value is 0.05, and the threshold of the variance is 0.015.

[0102] determining accurate status information of the wind power system based on historical status information after performing the first operation;

[0103] In the embodiment of the present application, determining accurate status information of the wind power system according to historical status information after the first operation includes:

[0104] The second objective function is established and expressed as:

[0105] J 2 (x 2 )=[z 2 -h(x 2)] T R -1 [z 2 -h(x 2 )]

[0106] In the formula, J 2 (·) is the second objective function value, x 2 For accurate status information, z 2 is each group of state information in the historical state information after mutation detection, h(·) is the measurement function, and R is the weight matrix;

[0107] Deriving a second objective function based on each group of state information in the historical state information after the first operation;

[0108] Specifically, the second objective function is derived as follows:

[0109]

[0110] The derivative of the second objective function is converted into an equivalent equation for solution to obtain the optimal state information, thereby determining the accurate state information of the wind power system;

[0111] The equivalent equation is expressed as:

[0112] f(x 2 )=H T (x 2 )R -1 [z 2 -h(x 2 )]=0

[0113] Where f(·) is the equivalent equation value, and H(·) is the derivative of the measurement function.

[0114] S2: Based on accurate status information, calculate historical stability information, build and train a system stability prediction model;

[0115] In the embodiment of the present application, the small disturbance stability model includes: a wind turbine model, a thermal power generator model, a coordinate transformation equation, a network equation and a load equation; the wind turbine model includes: a wind turbine motion equation, a wind turbine rotor equation and a wind turbine stator equation; the thermal power generator model includes: a thermal power generator motion equation, a thermal power generator rotor equation and a thermal power generator stator equation; the historical stability information includes the minimum damping ratio of the wind power system at different historical moments;

[0116] Specifically, the motion equation of the wind turbine generator set is:

[0117]

[0118] Where H is the inertia constant, s is the rotor slip; X m is the magnetizing reactance, X s is the stator reactance, U t is the terminal voltage, I qr is the q-axis component of the rotor current, T m is the mechanical torque of the motor.

[0119] The rotor equation of a wind turbine is:

[0120]

[0121]

[0122] In the formula, I dr is the d-axis component of the rotor current; ω s is the motor angular velocity; T r ' is the synchronous reactance, X r ' is the leakage reactance, is the d-axis component of the virtual rotor voltage, I dr_ref is the reference current of the d-axis component of the rotor current, K 1 , K 2 、T 1 、T 2 is the PI controller parameter, Q sref is the stator reference reactive power, is the q-axis component of the virtual rotor voltage, I qr_ref is the reference current of the q-axis component of the rotor current, P sref is the stator reference active power.

[0123] The stator equation of a wind turbine is:

[0124]

[0125] In the formula, I qs is the q-axis component of the stator current, I ds is the d-axis component of the stator current, I dr is the d-axis component of the rotor current.

[0126] The motion equation of the thermal power generator set is:

[0127]

[0128] In the formula, δ is the power angle of the generator set, ω, ω s are the speed and initial speed of the generator set respectively, P e , P m are the electromagnetic and mechanical power of the generator set, D, T j are the damping coefficient and inertia time constant of the generator set respectively.

[0129] The rotor equation of a thermal power generator is:

[0130]

[0131]

[0132] Where, E' q , E' d are the transient electromotive force of the quadrature axis and direct axis of the generator set, E″ q , E″ d are the sub-transient electromotive force of the quadrature axis and direct axis of the generator set, T′ d0 , T′ q0 are the transient open-circuit time constants of the direct-axis and quadrature-axis of the generator set, T″ d0 , T″ q0 are the subtransient open-circuit time constants of the direct and quadrature axes of the generator set, respectively, d 、x' d , x″ d are the synchronous reactance, transient reactance and subtransient reactance of the direct axis of the generator set, respectively, q 、x' q 、x' q ' are the synchronous reactance, transient reactance and subtransient reactance of the generator set quadrature axis, I d ,I q They are the direct-axis component and quadrature-axis component of the stator current of the generator set respectively.

[0133] The stator equation of a thermal power generator set is:

[0134]

[0135] Where V q 、V d are the quadrature-axis and direct-axis components of the stator voltage of the generator set, R a is the stator resistance of the generator set.

[0136] The coordinate transformation equation is:

[0137]

[0138] In the formula, δ 0 is the initial power angle of the generator set, V Gx 、V Gy are the x-axis and y-axis components of the generator set terminal current, I Gx ,I Gy They are the x-axis and y-axis components of the generator set current respectively.

[0139] The network equation is:

[0140]

[0141] In the formula, I i is the current at node i, V i is the voltage at node i, Y ij is the admittance between nodes i and j, and n is the number of nodes in the power system containing wind power.

[0142] The load equation is:

[0143]

[0144] Where Y L is the load admittance, is the load power conjugate, V L is the load node voltage.

[0145] In the embodiment of the present application, based on the accurate state information and the small disturbance stability model of the wind power system, calculating the historical stability information includes:

[0146] Select the dependent variables of all derivatives in the equations of the small disturbance stability model as the state vector and construct the state matrix;

[0147] For example, all the formulas in the small disturbance stability model are integrated into a set of differential algebraic equations, and the dependent variables of all derivatives in the equations are selected as the state vectors. For example, the equation contains derivatives Then select its dependent variable δ as the state vector, and write all the state vectors in this set of differential algebraic equations in matrix form.

[0148] Calculating a plurality of eigenvalues ​​of the state matrix based on accurate state information;

[0149] Specifically, the expression of the eigenvalue is:

[0150] λ=σ±jω;

[0151] Where λ is the eigenvalue, σ is the real part of the eigenvalue, and ω is the imaginary part of the eigenvalue.

[0152] Calculate multiple damping ratios based on each eigenvalue and the damping ratio formula;

[0153] The damping ratio formula is expressed as:

[0154]

[0155] Where ξ is the damping ratio, σ is the real part of the eigenvalue, and ω is the imaginary part of the eigenvalue;

[0156] All damping ratios are compared and the minimum damping ratio is determined.

[0157] In another possible implementation, the operating state of the wind power system may be further determined according to different damping ratios:

[0158] When 0<ξ<1, the power system containing wind power is in an underdamped state, and the system will experience damped oscillations until it finally stabilizes;

[0159] When ξ = 0, the power system containing wind power is in an undamped state. Ideally, the system will oscillate with constant amplitude, that is, the amplitude of the oscillation will not decay over time;

[0160] When ξ<1, this would mean that the wind power system has negative damping or anti-damping effect, in which case the amplitude of the system oscillations increases over time because the system gains additional energy.

[0161] Build a system stability prediction model based on machine learning technology;

[0162] In the embodiment of the present application, kernel principal component analysis (KPCA) is first used to reduce the dimension of state information and stability information to extract key features, and then a prediction of stability information is constructed through a support vector machine (SVM), and then the K-Nearest Neighbor (KNN) algorithm is combined to optimize and improve its prediction accuracy.

[0163] Specifically, the construction of the KPCA model:

[0164] The function of the KPCA model is to extract the main features of the data set, reduce the correlation between the data, and reduce the dimensionality of the data. The input of the KPCA model is a data set containing state information and stability information, and the output is a data set containing reduced dimensionality state information and stability information.

[0165] The first step is to map the data set containing state information and stability information into a high-dimensional space:

[0166] x→φ(x);

[0167] Where x is a data set containing state information and stability information, and φ(x) is a data set of state information and stability information mapped to a high-dimensional space.

[0168] The second step is to generate the covariance matrix of the mapped data and obtain the eigenvector of the covariance matrix. The covariance matrix is:

[0169]

[0170] Where Cov is the covariance matrix in the high-dimensional space K, N is the total number of samples in the data set, and x i is the i-th sample data in the data set.

[0171] Characteristic equation:

[0172] λ·ω=Cov·ω;

[0173] Where λ is the eigenvalue of the covariance matrix in the high-dimensional space K, and ω is the eigenvector of the covariance matrix in the high-dimensional space K.

[0174] The third step is to project the data set containing state information and stability information mapped to the high-dimensional space into the low-dimensional space to achieve data dimensionality reduction:

[0175] Ker(x i ,x j )=[φ(x i ),φ(x j )];

[0176]

[0177] In the formula, Ker(x i ,x j ) is the kernel function, Proj i is a dataset projected into a low-dimensional space containing dimensionality reduction state information and stability information, v i is the ith eigenvector.

[0178] Construction of SVM model:

[0179] The function of the SVM model is to map the data set containing dimensionality reduction state information and stability information to the feature space through nonlinear changes, and find the optimal linear classification surface in this feature space for classification prediction. If the optimal linear classification surface found by the mapped part of the data is less than the set threshold, the KNN model is used for classification prediction.

[0180] The first step is to perform nonlinear mapping on the data set containing dimensionality reduction state information and stability information:

[0181] Project i →ψ(Proj i );

[0182] Where ψ(·) is the nonlinear mapping function.

[0183] The second step is to construct the objective function and constraints:

[0184]

[0185] Where Φ(x) is the objective function, ||w|| 2 is the square of the hyperplane width, C is the regularization parameter, ξ i is the ith slack variable of the nonlinear mapping data, n is the total number of samples of the nonlinear mapping data, x i is the eigenvector of the i-th nonlinear mapping data, y i is the label of the i-th nonlinear mapping data, and b is the bias term of the hyperplane.

[0186] The third step is to solve the nonlinear programming problem, determine the hyperplane in the feature space, classify the data set containing the dimensionality reduction state information and the stability information, and determine the classification structure that preserves the SVM model.

[0187] Construction of KNN model:

[0188] The role of the KNN model is to hand over the data set that cannot be classified and contains dimensionality reduction state information and stability information to the KNN model for classification prediction when the SVM model cannot find a hyperplane that meets the conditions.

[0189] The first step is to choose the value of k, which is the number of classes in each class.

[0190] The second step is to define the distance metric, mainly using the Euclidean distance metric:

[0191]

[0192] Where, d euclidean (x i ,x j ) represents the distance between each data in the data set that cannot be classified and contains dimensionality reduction state information and stable information.

[0193] The third step is to determine the distance between each data in the data set that cannot be classified and contains dimensionality reduction state information and stability information, and save the classification structure of the KNN model.

[0194] In another possible implementation, a dimensionality reduction method may be used instead of KPCA: Principal Component Analysis (PCA): is suitable for linear dimensionality reduction and may perform well in some scenarios.

[0195] t-Distributed Stochastic Neighbor Embedding (t-SNE): Suitable for visualization of high-dimensional data, especially helpful in discovering data structure.

[0196] Linear Discriminant Analysis (LDA): If the class labels are known, it can maximize the separation between classes while reducing the dimensionality.

[0197] The following algorithms can be used instead of SVM:

[0198] Random Forest: An ensemble learning method with strong robustness and the ability to handle high-dimensional data.

[0199] Gradient Boosting Trees: Improve the performance of weak learners through step-by-step learning, suitable for modeling nonlinear relationships.

[0200] Neural Networks: In particular, deep learning models can capture complex nonlinear patterns and are suitable for large-scale data.

[0201] The following algorithms can be used instead of KNN:

[0202] Decision Trees: Easy to understand and interpret, suitable for processing nonlinear and mixed types of data.

[0203] Locally Weighted Regression (LWR): performs regression modeling in a specific area and can flexibly adapt to data patterns.

[0204] Support Vector Regression (SVR): It is a regression version of SVM and is suitable for modeling regression problems, especially high-dimensional data.

[0205] The system stability prediction model is trained with historical state information as input and historical stability information as output;

[0206] In the embodiment of the present application, the acquired data set containing historical state information and historical stability information also needs to be preprocessed, that is, the data is standardized:

[0207]

[0208] Where μ is the mean of the historical state information and the historical stable information, σ is the standard deviation of the historical state information and the historical stable information, and x is the historical state information and the historical stable information.

[0209] S3: perturb the power of the generator set including the wind power system to obtain the initial state information before and after the perturbation;

[0210] S4: Based on the initial state information before and after the perturbation, the trained system stability prediction model and the power adjustment model, the wind power power system is jointly dispatched.

[0211] In the embodiment of the present application, the power adjustment model is expressed as:

[0212]

[0213] In the formula, ΔPG i is the power dispatching amount of the i-th generator set, n is the number of generator sets, are the lower and upper power limits of the i-th generator set, respectively. is the initial power of the i-th generator set, ξ 0 is the initial damping ratio of the power system containing wind power, Δξ is the change in damping ratio of the power system containing wind power, ξ 1 is the minimum damping ratio threshold of the power system containing wind power, C i is the sensitivity of the i-th minimum damping ratio to the power of the generator set.

[0214] Perform scheduling calculations to obtain the power scheduling amount, including:

[0215] Perform measurement mutation detection on the initial state information of the wind power system before and after the perturbation, and determine the accurate state information of the wind power system before and after the perturbation;

[0216] It should be noted that the detailed process is the same as that described in the above step S1, and will not be described again. In addition, the obtained accurate status information needs to be standardized.

[0217] The accurate state information of the wind power system before and after the perturbation is input into the trained system stability prediction model to obtain the minimum damping ratio of the wind power system before and after the perturbation;

[0218] The power adjustment model is solved based on the minimum damping ratio of the wind power system before and after the perturbation to obtain the power dispatch amount.

[0219] It should be noted that after calculating the minimum damping ratio of the wind power system before and after the perturbation, we can also determine parameters such as the sensitivity of the minimum damping ratio to the power of the generator set and the change in the damping ratio of the wind power system. Based on these parameters, the above-mentioned power adjustment model can be solved to obtain the power dispatch amount.

[0220] Allocate the power dispatching capacity and complete the joint dispatching of wind power power systems.

[0221] In the embodiment of the present application, allocating the power scheduling amount includes:

[0222] Determine the adjustment range of each synchronous machine unit according to the output constraint and climbing constraint of each synchronous machine unit;

[0223] Specifically, the adjustment range of each synchronous machine unit is determined according to the output constraint and climbing constraint of each synchronous machine unit:

[0224]

[0225] In the formula, ΔPT i min , ΔPT i max are the lower and upper limits of the regulation range of the i-th synchronous machine unit, PT i min PT i max are the lower and upper limits of the output of the i-th synchronous machine unit, PT i is the initial output of the i-th synchronous machine unit, are the ramp-down rate and ramp-up rate of the i-th synchronous machine unit respectively.

[0226] Based on the regulation range of all synchronous machine units, determine whether the regulation margin of all synchronous machine units meets the dispatching requirements; if so, allocate the power dispatching amount according to the proportion of the regulation margin of each synchronous machine unit; if not, dispatch according to the limit value of the regulation margin of each synchronous machine unit, and the excess power dispatching amount shall be borne by the energy storage system;

[0227] The power dispatch capacity of each synchronous machine unit meets the following requirements:

[0228]

[0229] In the formula, ΔPT i is the power dispatch amount undertaken by the i-th synchronous machine unit, ΔP is the power dispatch amount, ΔPT i min , ΔPT i max are the lower and upper limits of the regulation range of the i-th synchronous machine unit, are the lower and upper limits of the adjustment range of the jth synchronous machine unit, respectively, and m is the number of synchronous machine units;

[0230] The power dispatch capacity undertaken by the energy storage system meets the following requirements:

[0231]

[0232] Where ΔPB is the power dispatch amount borne by the energy storage system.

[0233] The above is a schematic scheme of the small disturbance stability prediction and joint dispatch method of the wind power system in this embodiment. It should be noted that the technical scheme of the small disturbance stability prediction and joint dispatch system of the wind power system and the technical scheme of the small disturbance stability prediction and joint dispatch method of the wind power system mentioned above belong to the same concept. For the details not described in detail in the technical scheme of the small disturbance stability prediction and joint dispatch system of the wind power system in this embodiment, please refer to the description of the technical scheme of the small disturbance stability prediction and joint dispatch method of the wind power system mentioned above.

[0234] The small disturbance stability prediction and joint dispatching system of the wind power system in this embodiment includes:

[0235] A data acquisition and processing module, used to acquire historical status information of the wind power system, perform a first operation, and determine accurate status information;

[0236] Stability analysis and modeling module, which is used to calculate historical stability information based on accurate state information, and to build and train system stability prediction models;

[0237] The disturbance test and initial state acquisition module is used to perturb the power of the generator set of the wind power system and obtain the initial state information before and after the perturbation;

[0238] The dispatch calculation and execution module is used to jointly dispatch the wind power system based on the initial state information before and after the perturbation, the trained system stability prediction model and the power adjustment model.

[0239] This embodiment further provides a computing device, which is applicable to the case of a small disturbance stability prediction and joint dispatching method for a wind power system, including:

[0240] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the small disturbance stability prediction and joint scheduling method of the wind power system as proposed in the above embodiment.

[0241] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for predicting small disturbance stability and joint scheduling of a wind power system as proposed in the above embodiment is implemented.

[0242] The storage medium proposed in this embodiment and the small disturbance stability prediction and joint dispatching method of the wind power system proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0243] Example 2

[0244] Reference Figure 2-Figure 3 Tables 1 and 2 are an embodiment of the present invention, which provide a small disturbance stability prediction and joint dispatching method for a wind power system. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0245] In view of the small disturbance stability prediction and joint dispatch method of the wind power system provided in this application, and the system stability prediction model, this embodiment provides Figure 2 , Figure 3 The model prediction effect diagram and the data shown in Table 1 and Table 2 are used to verify its effectiveness; Figure 2 This is the prediction effect diagram of the system stability prediction model; Figure 3 This is the fitting effect diagram of the system stability prediction model.

[0246] Table 1 Prediction results of system stability prediction model

[0247]

[0248]

[0249] Table 2 Generator set adjustment table

[0250] Generator Set Adjustment amount Unit 1 33.1535969180495 Unit 2 37.0806689754761 Unit 3 29.7119619319553 Unit 4 35.2009890161429 Unit 5 37.4515093856529 Unit 6 42.0349079325850 Unit 7 43.5133712982194 Unit 8 33.4589114111537

[0251] It can be seen that the above-mentioned small disturbance stability prediction and joint dispatching method of the wind power power system can realize the accurate prediction of the stability information of the wind power power system by deeply mining the relationship between the system operation data and the small disturbance stability, and optimize the dispatching strategy of the energy storage system, thereby improving the operation efficiency and reliability of the wind power power system.

[0252] In summary, the above process can provide accurate stability information prediction for operators of wind power power systems, help them adjust the operating status of the energy storage system in a timely manner, and improve the overall operating efficiency and economy of wind power power systems. At the same time, the above process can also reduce dependence on traditional monitoring equipment, reduce operating and maintenance costs, and provide technical support for the sustainable development of wind power power systems. In short, this application has important application value in the stability prediction and optimal scheduling of wind power power systems, and is expected to promote the intelligent and efficient development of wind power power systems.

[0253] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for predicting small disturbance stability and joint dispatching of a wind power system, characterized in that: include: Acquire historical status information of the wind power system, perform a first operation, and determine accurate status information; Based on accurate status information, calculate historical stability information, build and train a system stability prediction model; Perturb the power of the generator set of the wind power system to obtain the initial state information before and after the perturbation; Based on the initial state information before and after the perturbation, the trained system stability prediction model and the power adjustment model, the wind power power system is jointly dispatched.

2. The method for predicting small disturbance stability and joint dispatching of a wind power system according to claim 1, characterized in that: The wind power system includes: a wind power generation system, a thermal power generation system and an energy storage system; the historical status information includes multiple groups of status information at different historical moments; the status information includes at least node voltage amplitude and phase angle; Performing measurement mutation detection on historical status information and performing a first operation; determining accurate status information of the wind power system based on the historical status information after the first operation; Based on accurate state information and a small disturbance stability model of the wind power system, calculate historical stability information; build a system stability prediction model based on machine learning technology; use historical state information as input and historical stability information as output to train the system stability prediction model; The power of the generator sets of the wind power system is perturbed, and the initial state information of the wind power system before and after the perturbation is obtained; based on the initial state information of the wind power system before and after the perturbation, the trained system stability prediction model, and the power adjustment model of the wind power system, scheduling calculations are performed to obtain the power scheduling amount; the power scheduling amount is allocated to complete the joint scheduling of the wind power system.

3. The method for predicting small disturbance stability and joint dispatching of a wind power system according to claim 2, characterized in that: The first operation of performing measurement mutation detection on the historical state information includes: Establish the first objective function, expressed as: J1(x1)=[z1-h(x1)] T R -1 [z1-h(x1)] Where J1(·) is the value of the first objective function, x1 is the estimated state information, z1 is each group of state information in the historical state information, h(·) is the measurement function, and R is the weight matrix; Based on each group of state information in the historical state information, calculating the expected value and variance of the first objective function; When the expected value or the variance is greater than a set threshold, deleting the current group state information; Determining accurate status information of the wind power system based on historical status information after the first operation includes: The second objective function is established and expressed as: J2(x2)=[z2-h(x2)] T R -1 [z2-h(x2)] Where J2(·) is the value of the second objective function, x2 is the accurate state information, z2 is each group of state information in the historical state information after mutation detection, h(·) is the measurement function, and R is the weight matrix; Deriving a second objective function based on each group of state information in the historical state information after the first operation; The derivative of the second objective function is converted into an equivalent equation and solved to obtain the optimal state information, thereby determining the accurate state information of the wind power system.

4. The method for predicting small disturbance stability and joint dispatching of a wind power system according to claim 3, characterized in that: The calculation of historical stability information based on accurate state information and a small disturbance stability model of a wind power system includes: Select the dependent variables of all derivatives in the equations of the small disturbance stability model as the state vector and construct the state matrix; Calculating a plurality of eigenvalues ​​of the state matrix based on accurate state information; Calculate multiple damping ratios according to each eigenvalue and the damping ratio formula; All damping ratios are compared and the minimum damping ratio is determined.

5. The method for predicting small disturbance stability and joint dispatching of a wind power system according to claim 4, characterized in that: The system stability prediction model based on machine learning technology includes: Perform dimensionality reduction on state information and stability information to extract key features; construct predictions for stable information and combine them with optimization algorithms to improve prediction accuracy.

6. The method for predicting small disturbance stability and joint dispatching of a wind power system according to claim 5, characterized in that: The power adjustment model is expressed as: Where ΔPG i is the power dispatching amount of the i-th generator set, n is the number of generator sets, are the lower and upper power limits of the i-th generator set, respectively. is the initial power of the i-th generator set, ξ0 is the initial damping ratio of the power system containing wind power, Δξ is the change of the damping ratio of the power system containing wind power, ξ1 is the minimum damping ratio threshold of the power system containing wind power, C i is the sensitivity of the i-th minimum damping ratio to the power of the generator set; The scheduling calculation to obtain the power scheduling amount includes: Perform measurement mutation detection on the initial state information of the wind power system before and after the perturbation, and determine the accurate state information of the wind power system before and after the perturbation; The accurate state information of the wind power system before and after the perturbation is input into the trained system stability prediction model to obtain the minimum damping ratio of the wind power system before and after the perturbation; The power adjustment model is solved based on the minimum damping ratio of the wind power system before and after the perturbation to obtain the power dispatch amount.

7. The method for predicting small disturbance stability and joint dispatching of a wind power system according to claim 6, characterized in that: The allocating of the power scheduling amount includes: Determine the adjustment range of each synchronous machine unit according to the output constraint and climbing constraint of each synchronous machine unit; Based on the regulation range of all synchronous machine units, determine whether the regulation margin of all synchronous machine units meets the dispatching requirements; if so, allocate the power dispatching amount according to the proportion of the regulation margin of each synchronous machine unit; if not, dispatch according to the limit value of the regulation margin of each synchronous machine unit, and the excess power dispatching amount shall be borne by the energy storage system; The power dispatch capacity of each synchronous machine unit meets the following requirements: In the formula, ΔPT i is the power dispatch amount undertaken by the i-th synchronous machine unit, ΔP is the power dispatch amount, are the lower and upper limits of the regulation range of the i-th synchronous machine unit, are the lower and upper limits of the adjustment range of the jth synchronous machine unit, respectively, and m is the number of synchronous machine units; The power dispatch capacity undertaken by the energy storage system meets the following requirements: Where ΔPB is the power dispatch amount borne by the energy storage system.

8. A system using the method for predicting small disturbance stability and joint dispatching of a wind power system as claimed in any one of claims 1 to 7, characterized in that: include: A data acquisition and processing module, used to acquire historical status information of the wind power system, perform a first operation, and determine accurate status information; Stability analysis and modeling module, which is used to calculate historical stability information based on accurate state information, and to build and train system stability prediction models; The disturbance test and initial state acquisition module is used to perturb the power of the generator set of the wind power system and obtain the initial state information before and after the perturbation; The dispatch calculation and execution module is used to jointly dispatch the wind power system based on the initial state information before and after the perturbation, the trained system stability prediction model and the power adjustment model.

9. A computing device comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the small disturbance stability prediction and joint dispatching method for a wind power system described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for small-disturbance stability prediction and joint dispatch of a wind power system as claimed in any one of claims 1 to 7.

Images