New energy power station network source coordination performance detection method and system

Abstract

The invention discloses a new energy power station network source coordination performance detection method and system, and relates to the technical field of new energy power grids, and the method comprises the steps: building an interaction model between a new energy power station and a power grid based on operation data, and analyzing the interaction behavior of the energy power station and the power grid through the interaction model; identifying a coordination contradiction point between the random gene of the new energy power station and the deterministic demand of the power grid, and optimizing the coordination contradiction point into a coordination balance point by using a distributed balance control model; and simulating a real-time operation state when the new energy power station accesses the power grid, comparing the real-time operation state with the initial state, and identifying the abnormal coordination performance when the new energy power station accesses the power grid. Through the distributed balance control model driven by the homotopy algorithm, the abstract coordination contradiction point is converted into the specific optimal adjustment set value, a dynamic self-adaptive coordination mechanism between the new energy and the power grid is formed, and the power grid stability and scheduling efficiency after the new energy power station is connected are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of new energy power grid, in particular to a new energy power station grid-source coordination performance detection method and system. BACKGROUND

[0002] The new energy power station grid-source refers to the new energy power station (such as wind power, photovoltaic) as a power source, directly connected with the power grid and participating in the dynamic regulation and stability of the system. It is not only an independent power source, but also a part of the power grid. Through interaction with the power grid, it affects the power dispatching, frequency control, stability and other operating characteristics of the power grid.

[0003] New energy power station grid-source coordination performance detection refers to evaluating whether the power matching, frequency response, voltage stability, and power flow regulation of the new energy power station after being connected to the grid are coordinated with the overall operation state of the grid in dynamic operation, detecting whether it causes disturbance amplification, stability deterioration, or dispatch imbalance, etc. If the coordination performance is not detected, the following defects exist:

[0004] It cannot identify the potential risks caused by the output fluctuation of the new energy power station to the operation of the power grid, such as power reversal, frequency out of control, voltage collapse, etc. The control system lacks feedback basis, leading to invalid operation strategy. The abnormal state of the system is misjudged as normal, which may cause cascading failures and even cause local power outage or system instability.

[0005] In view of the problems in the related art, no effective solution has been proposed so far. SUMMARY

[0006] In view of the problems in the related art, the present application proposes a new energy power station grid-source coordination performance detection method and system to overcome the above technical problems existing in the prior art.

[0007] To this end, the specific technical solutions adopted by the present application are as follows:

[0008] According to one aspect of the present application, a new energy power station grid-source coordination performance detection method is provided, which comprises:

[0009] S1, collecting the operation data of the new energy power station and the power grid, constructing an interaction model between the new energy power station and the power grid based on the operation data, and analyzing the interaction behavior of the new energy power station and the power grid by using the interaction model;

[0010] S2, based on the interaction behavior of the new energy power station and the power grid, identifying the coordination contradiction points between the randomness genes of the new energy power station and the deterministic demand of the power grid, and using a distributed balance control model to optimize the coordination contradiction points to coordination balance points;

[0011] S3, taking the balance coordination point as an initial state, simulating the real-time operation state of the new energy power station when accessing the power grid, and comparing the real-time operation state with the initial state to identify the abnormal coordination performance of the new energy power station when accessing the power grid.

[0012] Preferably, based on the interaction behavior of the new energy power station and the power grid, the coordination contradiction point between the randomness gene of the new energy power station and the determinacy demand of the power grid is identified, and the coordination contradiction point is optimized to a coordination balance point by using a distributed balance control model, which comprises:

[0013] S21, extracting the randomness gene index of the new energy and the determinacy demand index of the power grid according to the interaction behavior between the new energy power station and the power grid, and performing a space-time correlation analysis on the randomness gene index and the determinacy demand index;

[0014] S22, based on the space-time correlation analysis result, establishing a topological structure between the randomness gene index and the determinacy demand index, and identifying the uncoordinated index deviating from the preset condition as the coordination contradiction point from the topological structure;

[0015] S23, according to the coordination contradiction point, the cause-effect relationship between the output characteristics of the new energy power station and the state characteristics of the power grid is deduced;

[0016] S24, according to the cause-effect relationship, a distributed balance control model is constructed, the optimal adjustment set value of the distributed balance control model is generated by combining the homotopy algorithm, and the optimal adjustment set value is taken as the coordination balance point.

[0017] Preferably, according to the coordination contradiction point, the cause-effect relationship between the output characteristics of the new energy power station and the state characteristics of the power grid is deduced, and a distributed balance control model is constructed according to the cause-effect relationship, which comprises:

[0018] S231, a power grid constraint surface model is constructed, the boundary form influence of the new energy power station output node on the power grid constraint surface model is analyzed through the power grid constraint surface model, and the singularity position leading to surface distortion under the new energy power station output mode is identified;

[0019] S232, taking the output node of the new energy power station corresponding to the singularity position as the center, and constructing a multi-element causal diagram between the output state of the new energy power station and the operation state of the power grid according to the output node;

[0020] S233, converting the multi-element causal diagram into a Boolean causal diagram, and constructing a conditional probability table for each pair of nodes in the Boolean causal diagram;

[0021] S234, according to the conditional probability table, the cause-effect relationship between the output characteristics of the new energy power station and the state characteristics of the power grid is deduced reversely, and a feasibility function is introduced to correct the cause-effect relationship.

[0022] Preferably, the distributed balance control model is constructed according to the causal relationship, the optimal regulation set value of the distributed balance control model is generated by combining the homotopy algorithm, and the optimal regulation set value is included as a coordinated balance point:

[0023] S241, encoding the causal relationship into a causal constraint condition and a target function to construct a distributed balance control model;

[0024] S242, using the homotopy algorithm to iteratively update a homotopy path gradually transitioning from a stable state to an actual target state based on a state of a power grid not connected to a new energy power station;

[0025] S243, dynamically adjusting control variables of the distributed balance control model along the homotopy path, and gradually approaching a global optimal solution satisfying the causal constraint condition in combination with a gradual advancement strategy;

[0026] S244, performing consistency verification on the global optimal solution by introducing a degree theory to generate an optimal regulation set value satisfying the causal constraint condition, and taking the optimal regulation set value as a coordinated balance point of the energy power station and the power grid.

[0027] Preferably, the dynamically adjusting control variables of the distributed balance control model along the homotopy path and gradually approaching a global optimal solution satisfying the causal constraint condition in combination with a gradual advancement strategy comprises:

[0028] S2431, initializing a homotopy algorithm parameter, and setting an initial value interval of the control variables of the distributed balance control model and a maximum number of iterations;

[0029] S2432, gradually increasing the homotopy algorithm parameter to promote changes in the control variables of the distributed balance control model, and in each homotopy path, taking the current changed control variables as initial values to solve the distributed balance control model;

[0030] S2433, judging a convergence state of the distributed balance control model, if successfully converging, obtaining a current optimal solution and performing step S2435, if failing to converge, performing step S2434;

[0031] S2434, judging whether a step length at the time of current iteration convergence is a minimum step length, if not, backtracking the homotopy algorithm parameter to a previous state and returning to step S2432, otherwise, taking the current optimal solution as an initial value for the next iteration;

[0032] S2435, judging whether the homotopy algorithm parameter has reached a preset value after each homotopy algorithm promotion, if reaching the preset value, indicating that the distributed balance control model successfully converges, and outputting a global optimal solution satisfying the causal constraint condition, otherwise, increasing the number of iterations and returning to step S2432 until the convergence condition is met.

[0033] Preferably, the balance coordination point is taken as the initial state, the real-time operation state when the new energy power station is connected to the power grid is simulated, and the real-time operation state is compared with the initial state to identify the abnormal coordination performance of the new energy power station connected to the power grid, which includes:

[0034] S31, based on the balance coordination point, a Hamiltonian non-diagonal matrix of the power grid is constructed, and a critical behavior in the operation of the power grid is analyzed based on the non-diagonal matrix to obtain an initial coordination state of the power grid;

[0035] S32, a lattice Boltzmann model is constructed, and a dynamic evolution process after the new energy power station is connected to the power grid is simulated by using the lattice Boltzmann model to obtain a disturbed operation state of the new energy power station;

[0036] S33, the initial coordination state and the disturbed operation state are respectively mapped into a complex plane unit disc, and the difference between the initial coordination state and the disturbed operation state is calculated by dynamic Mobius transformation;

[0037] S34, the difference calculation result is compared with the balance coordination point, and the comparison result is input into the improved random forest to identify the abnormal coordination performance.

[0038] Preferably, the lattice Boltzmann model is constructed, and a dynamic evolution process after the new energy power station is connected to the power grid is simulated by using the lattice Boltzmann model to obtain a disturbed operation state of the new energy power station, which includes:

[0039] S321, the nodes of the power grid are mapped as the nodes of the lattice grid, and the power transmission lines and connection devices in the power grid are taken as the edges between the lattice grids to construct the lattice Boltzmann model;

[0040] S322, the output mode of the new energy power station is taken as the input of the lattice Boltzmann model, and the dynamic evolution process of the output in the power grid is simulated by using the lattice Boltzmann to obtain the disturbed operation state of the new energy power station.

[0041] Preferably, the comparison result is input into the improved random forest to identify the abnormal coordination performance, which includes:

[0042] Based on the comparison result of the difference calculation result and the balance coordination point, an embedded difference sample is screened, a self-help statistical analysis strategy is used to randomly select a plurality of instances from the embedded difference sample to form a disturbed coordination sample subset for training;

[0043] A corresponding decision tree is respectively constructed for the embedded difference sample, the disturbed coordination sample subset is used to train each decision tree, and the discrimination ability between the difference calculation result and the balance coordination point is weighted evaluated, and a feature subspace is constructed according to the score proportion of the representative features;

[0044] The maximum depth growth of each decision tree is guided on the constructed feature subspace until all the decision trees constitute the improved random forest model;

[0045] The predefined to-be-tested index is input to the improved random forest model, and whether the to-be-tested index exists coordination deviation is judged by each tree in the improved random forest model independently, so that the detection result of the abnormal coordination performance of the new energy power station connected to the power grid is obtained.

[0046] Preferably, the expression of the lattice Boltzmann model is:

[0047]

[0048] In the formula, f i (x, t) represents the particle distribution function along the lattice direction i at position x and time t; e i represents a discrete velocity vector; Δt represents a time step; τ represents a relaxation time; f i eq represents an equilibrium state distribution function along the lattice direction i; F i (x, t) represents an external force term along the lattice direction i.

[0049] According to another aspect of the present application, a new energy power station grid source coordination performance detection system is also provided, which comprises:

[0050] An interactive behavior acquisition module is configured to acquire operation data of the new energy power station and the power grid, construct an interactive model between the new energy power station and the power grid based on the operation data, and analyze the interactive behavior of the new energy power station and the power grid by using the interactive model;

[0051] A coordination performance analysis module is configured to identify coordination contradiction points between the randomness genes of the new energy power station and the deterministic demand of the power grid based on the interactive behavior of the new energy power station and the power grid, and optimize the coordination contradiction points to coordination balance points by using a distributed balance control model;

[0052] A coordination performance detection module is configured to take the balance coordination points as initial states, simulate real-time operation states of the new energy power station connected to the power grid, and compare the real-time operation states with the initial states to identify abnormal coordination performance of the new energy power station connected to the power grid.

[0053] The present application has the following advantages:

[0054] 1. The application can accurately reveal how the uncertainty of new energy power station output affects the stability demand of the power grid, avoid local uncoordination causing system-level risks, realize the interpretable modeling of new energy disturbance behavior by extracting interaction indicators and establishing a space-time correlation model, effectively locate the key sensitive area of new energy access by combining the power grid constraint surface and singularity identification technology, and through the distributed balance control model driven by homotopy algorithm, convert the abstract coordination contradiction point into specific optimal adjustment set value, form a dynamic self-adaptive coordination mechanism between new energy and power grid, and improve the stability, flexibility and scheduling efficiency of the power grid after new energy power station access.

[0055] 2. The application explicitly encodes the complex causal mechanism between new energy output and power grid response as an optimization objective and constraint condition, so that the control model has logical interpretability, and by introducing the stable state of the power grid without new energy access as the initial reference, the homotopy algorithm smoothly constructs a continuous mapping path from the ideal state to the actual state, effectively avoiding the solution failure caused by model nonlinearity or initial value sensitivity, dynamically adjusting the control variables along the homotopy path and combining the step-by-step advancement strategy to realize the approximation process of global optimization, ensuring that the optimal solution has stability guarantee in theory, and taking it as the adjustment reference point of new energy and power grid coordinated operation, significantly improving the system control precision and operation safety after new energy access.

[0056] 3. The application can realize dynamic evolution tracking from the theoretical stable state to the actual running state, identify potential unstable factors of the system by analyzing the power grid critical behavior with Hamilton non-diagonal matrix, simulate the propagation process of new energy output disturbance in the power grid with high fidelity through the lattice Boltzmann model, capture the dynamic changes of state variables such as voltage and frequency, and then map the initial coordination state and disturbance state to the complex plane unit disc and apply dynamic Mobius transformation to realize geometric difference measurement of multi-dimensional state, which improves the abnormal state recognition accuracy, speed and robustness in the process of new energy access, and provides more targeted and forward-looking early warning capability for power grid regulation. BRIEF DESCRIPTION OF DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0058] Figure 1 is a flow chart of a new energy power station grid-source coordination performance detection method according to an embodiment of the present application;

[0059] Figure 2This is a schematic diagram of a new energy power plant grid-source coordination performance testing system according to an embodiment of the present invention.

[0060] In the picture:

[0061] 1. Interactive behavior acquisition module; 2. Coordination performance analysis module; 3. Coordination performance detection module. Detailed Implementation

[0062] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention.

[0063] According to an embodiment of the present invention, a method and system for detecting the grid-source coordination performance of a new energy power plant are provided.

[0064] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, according to an embodiment of the present invention, a method for detecting the grid-source coordination performance of a new energy power plant includes:

[0065] S1. Collect operational data of new energy power plants and power grids, construct an interaction model between new energy power plants and power grids based on the operational data, and use the interaction model to analyze the interaction behavior between energy power plants and power grids.

[0066] It should be noted that the process of collecting operational data from new energy power plants and the power grid, constructing an interaction model between the two based on this data, and using this model to analyze their interactive behavior includes:

[0067] The system collects operational data from both the renewable energy power plant and the power grid, including time-series data such as voltage, current, frequency, and power. Using state-space modeling or graph neural networks, it constructs an interaction model between the renewable energy power plant and the power grid based on the time-series data, generating the energy coupling and dynamic response relationship between them. When the renewable energy power plant is not connected to the grid, the model analyzes its "islanding" behavior characteristics, such as output volatility, stability boundaries, and adaptive adjustment capabilities. Furthermore, it performs matching tests between the power plant's own dynamics and preset grid boundary conditions to assess its potential for interaction and coordination before grid connection. The interaction model can employ a state-space model, representing the system dynamics as differential / difference equations of state variables (such as voltage, frequency, and power), describing the interaction through inputs (control commands, grid disturbances) and outputs (measurement data).

[0068] S2, based on the interaction behavior of the new energy power station and the power grid, identify the coordination contradiction points between the randomness genes of the new energy power station and the deterministic demand of the power grid, and use the distributed balance control model to optimize the coordination contradiction points to the coordination balance points.

[0069] It should be noted that the interaction model reveals the interaction behavior between the two by collecting and analyzing the operation data of the new energy power station and the power grid, and further identifies the potential conflict between the randomness genes of the new energy power station and the deterministic demand of the power grid.

[0070] Specifically, the interaction model helps to establish the topology between the two by analyzing the interaction behavior of the energy power station and the power grid, and further backtracks the causal relationship by identifying the uncoordinated indicators (i.e. coordination contradiction points), and the distributed balance control model optimizes the interaction between the power station and the power grid based on these coordination contradiction points.

[0071] Among them, based on the interaction behavior of the new energy power station and the power grid, identify the coordination contradiction points between the randomness genes of the new energy power station and the deterministic demand of the power grid, and use the distributed balance control model to optimize the coordination contradiction points to the coordination balance points include:

[0072] S21, according to the interaction behavior between the new energy power station and the power grid, extract the randomness gene indicators of the new energy and the deterministic demand indicators of the power grid, and perform spatiotemporal correlation analysis on the randomness gene indicators and the deterministic demand indicators.

[0073] It should be noted that the randomness gene indicators of the new energy mainly include: photovoltaic or wind power fluctuation rate, power ramp change rate (power change gradient per unit time), output spectrum distribution characteristics (main frequency energy distribution based on wavelet transform or FFT), short-time prediction error variance, maximum instantaneous fluctuation amplitude, etc.

[0074] The deterministic demand indicators of the power grid include: frequency modulation and voltage regulation capacity margin (callable energy storage or conventional unit capacity), frequency offset tolerance threshold (maximum frequency deviation that the system can accept), real-time active / reactive demand, demand response sensitivity, load stability indicators, etc.

[0075] The spatiotemporal correlation analysis includes: time sequence alignment of the indicators, aligning the randomness indicators on the new energy side with the deterministic demand indicators on the power grid side according to a unified time window slice, and then using sliding window correlation analysis or Granger causality analysis to test whether some randomness fluctuations cause power grid demand fluctuations in time; in the spatial dimension, the correlation strength between different regions can be extracted by mapping the indicators to the geographical distribution map or the topology structure map, combined with spatial correlation analysis and graph attention mechanism, and finally a spatiotemporal influence mapping network of new energy randomness genes on power grid deterministic demand changes is established.

[0076] S22, based on the spatiotemporal correlation analysis result, establishing a topological structure between the random gene index and the deterministic demand index, and identifying a misalignment index deviating from a preset condition as a coordination contradiction point from the topological structure;

[0077] S23, according to the coordination contradiction point, deducing a causal relationship between the output characteristics of the new energy power station and the state characteristics of the power grid.

[0078] It should be noted that the time series data (such as power, frequency, voltage sequence) is used to test whether the power station output lag term significantly predicts the power grid state change (and vice versa). If the historical data of the power station output has statistical significance for the prediction of the power grid state, there is a causal direction from the power station to the power grid to verify the directionality of the causal relationship.

[0079] Among them, according to the coordination contradiction point, deducing a causal relationship between the output characteristics of the new energy power station and the state characteristics of the power grid, and constructing a distributed balance control model according to the causal relationship includes:

[0080] S231, constructing a power grid constraint surface model, analyzing the influence of the new energy power station output node on the boundary form of the power grid constraint surface model through the power grid constraint surface model, and identifying the singularity position leading to surface distortion under the new energy power station output mode.

[0081] It should be noted that the power grid constraint surface model is constructed, the influence of the new energy power station output node on the boundary form of the power grid constraint surface model is analyzed through the power grid constraint surface model, and the singularity position leading to surface distortion under the new energy power station output mode is identified, including:

[0082] Step 1, based on the power flow calculation result, obtain the power-voltage-frequency and other multi-dimensional state variables of the power grid node, and map the high-dimensional operating state to a low-dimensional surface representation in the multi-dimensional state space using kernel regression, principal component analysis or manifold learning method, forming a constraint surface model.

[0083] Step 2, connect the new energy power station output node to the system one by one, simulate its influence on the power grid state variable under different output modes (such as step, fluctuation, periodic disturbance, etc.), dynamically update the constraint surface and record its boundary changes, and focus on observing the smoothness and local curvature changes of the boundary.

[0084] Step 3, use surface differential geometry method to calculate high-order derivative, curvature tensor or Jacobian matrix, identify the singularity position leading to surface distortion, curvature mutation or continuity destruction under a certain new energy output mode, and further analyze its physical meaning and stability risk source combining with the power grid topology and node voltage sensitivity, for example:

[0085]

[0086] In the formula, f(x, y) represents a scalar function on the power grid constraint surface, such as the relationship of a certain node voltage V with active / reactive power output (P, Q); f x , f y represents the first-order partial derivative with respect to x=P and y=Q; f x , f y , f xy represents the corresponding second-order partial derivative; K represents the local Gaussian curvature of the corresponding surface at a certain output point, wherein when K→∞, or K→-∞, or the derivative term appears to be non-differentiable, the point can be identified as a candidate point of singularity of the surface distortion, and these singularities usually correspond to voltage collapse boundaries, power flow solution instability points or topology transition points caused by new energy disturbance modes.

[0087] S232, taking the output node of the new energy power station corresponding to the singularity position as the center, and constructing a multi-element causal graph between the output state of the new energy power station and the operation state of the power grid according to the output node;

[0088] S233, converting the multi-element causal graph into a Boolean causal graph, and constructing a conditional probability table for each pair of nodes in the Boolean causal graph;

[0089] S234, according to the conditional probability table, the causal relationship between the output characteristics of the new energy power station and the state characteristics of the power grid is reversely reasoned, and a feasibility function is introduced to correct the causal relationship.

[0090] It should be noted that by converting the multi-element causal graph into a Boolean causal graph, the complex multi-variable interaction relationship between the new energy power station and the power grid system can be effectively simplified, so that each node state is only represented as 0 or 1, thereby enhancing the clarity and calculability of the causal reasoning; after constructing the conditional probability table for each pair of nodes, the probability dependence relationship of the system state change under different output or state combinations can be quantitatively described, thereby providing a data basis for subsequent causal reasoning; on this basis, through reverse reasoning, the output behavior of the new energy power station can be reversely reasoned from the observed state characteristics of the power grid, and the potential causal driving mode thereof is identified, and in addition, the feasibility function can adopt a linear weighted form, a product form, a penalty function form, etc.

[0091] S24, constructing a distributed balance control model according to the causal relationship, generating optimal adjustment set values of the distributed balance control model by using the homotopy algorithm, and taking the optimal adjustment set values as the coordinated balance points.

[0092] Among them, constructing a distributed balance control model according to the causal relationship, generating optimal adjustment set values of the distributed balance control model by using the homotopy algorithm, and taking the optimal adjustment set values as the coordinated balance points include:

[0093] S241, encoding the causal relationship into causal constraint conditions and objective functions, and constructing a distributed balance control model.

[0094] It should be noted that encoding causal relationships into causal constraints and objective functions to construct a distributed equilibrium control model includes:

[0095] Step 1: Convert the causal relationships extracted from the interaction between the renewable energy power plant and the power grid into mathematical form, and set causal constraints between power output and grid status: P i (t)=f i (V i (t),Q i (t),θ i (t)), where P i (t) represents the output power of the i-th node, V i (t) represents the voltage at time t, Q i (t) represents the reactive power at time t, θ i (t) represents the phase angle of the power grid at time t, f i This function represents the causal relationship between new energy power plants and the power grid.

[0096] Step 2: Unify the causal relationships among multiple nodes in the system into constraints to ensure that the interaction between the power grid and the new energy power plant does not violate the system's physical constraints. For example, the voltage, frequency, and power in the power grid must all meet certain current flow constraints, A i P i (t)+B i Q i (t)+C i V i (t)+D i θ i (t) = 0, where A i B i C i D i All of these represent undetermined coefficients, calculated using the power flow equations of the power grid to ensure the stable operation of the power grid at all times.

[0097] Step 3: The objective function is usually set to minimize the total power loss of the power grid, maximize system stability, or maximize the matching degree between the output power of renewable energy power plants and the grid demand.

[0098]

[0099] In the formula, α i ,β i Both represent weighting coefficients, P target (t), Q target(t) represents the target power and reactive power demand of the power grid at time t, J represents the matching degree between the output power of the new energy power station and the demand of the power grid, and the objective function balances the output of the power grid and the new energy power station, so that the matching of the power is optimal, while ensuring the stability of the system.

[0100] In step four, in the solution of the distributed balance control model, the distributed control can coordinate the behaviors of each node through local information interaction. In order to optimize the entire system, the Lagrange multiplier method, the KKT condition, the ADMM (alternating direction method) and other distributed optimization algorithms are used to gradually realize the balance control of the power grid and the effective access of the new energy power station through the joint optimization of the constraint conditions and the objective function.

[0101] S242, based on the state of the power grid without accessing the new energy power station as a stable state, a homotopy algorithm is used to iteratively update a homotopy path gradually transitioning from the stable state to an actual target state;

[0102] S243, the control variables of the distributed balance control model are dynamically adjusted along the homotopy path, and a step-by-step advancement strategy is combined to gradually approach a global optimal solution that meets the causal constraint conditions.

[0103] The dynamic adjustment of the control variables of the distributed balance control model along the homotopy path, and the step-by-step approximation of the global optimal solution that meets the causal constraint conditions in combination with the step-by-step advancement strategy include:

[0104] S2431, the homotopy algorithm parameters are initialized, and the initial value range of the control variables of the distributed balance control model and the maximum number of iterations are set.

[0105] It should be noted that the homotopy algorithm parameters include homotopy parameters, convergence threshold, control variable initial value range, maximum number of iterations and neighborhood communication radius, and the adjustment principle is: initialize according to typical values, then fine-tune based on offline simulation, if the shock increases, then reduce the step size or increase the penalty term; if the convergence is slow, then relax the threshold or increase the number of iterations.

[0106] S2432, gradually increase the homotopy algorithm parameters to drive the changes of the control variables of the distributed balance control model, and in each homotopy path, the current changed control variables are taken as the initial values to solve the distributed balance control model;

[0107] S2433, the convergence state of the distributed balance control model is judged, if successful convergence is obtained, the current optimal solution is executed and step S2435 is executed, if the convergence fails, step S2434 is executed;

[0108] S2434, it is judged whether the step length at the time of current iteration convergence is the minimum step length, if not, the homotopy algorithm parameters are backtracked to the previous state and step S2432 is returned, otherwise, the current optimal solution is taken as the initial value of the next iteration.

[0109] S2435、in each homotopy algorithm push judge homotopy algorithm parameters have reached the preset value, if reach the preset value, then indicate that the distributed balance control model converges successfully, and output the global optimal solution that meets the causal constraint condition, otherwise, increase the iteration number and return to step S2432, until the convergence condition is met.

[0110] It should be noted that the homotopy algorithm parameters include homotopy path parameters, iteration step, tolerance threshold and maximum iteration number, the preset value of the homotopy path parameter is linearly increased from 0 to 1, the specific iteration step is 0.01 or smaller, to ensure smooth path tracking, ensure that the adjustment of each iteration is not too large, and ensure convergence and stability; The tolerance threshold is set to a small value, for example, 0.001 or less, to determine whether the algorithm is close enough to the optimal solution, to avoid excessive calculation; The maximum iteration number is set to 1000 or higher, to prevent the algorithm from terminating too early due to convergence problems, and the specific settings of each parameter should be adjusted according to the complexity and accuracy requirements of the actual model operation.

[0111] It should be noted that the homotopy algorithm (Homotopy Method) is a numerical optimization technique that gradually deforms a simple problem (such as a stable grid state without new energy access) to a target problem (a complex state after actual new energy access) by constructing a continuously changing parameterized path, thereby avoiding the difficulty of directly solving high-dimensional nonlinear problems. When combined with the distributed balance control model, the homotopy algorithm takes the initial stable state of the grid as the starting point, adjusts the control variables (such as output, voltage, frequency, etc.) through iteration, and gradually approaches the global optimal solution that meets the causal constraints along the homotopy path. Its effect is reflected in three aspects:

[0112] (1) Enhance convergence, avoid local optimization or divergence caused by new energy randomness through path tracking;

[0113] (2) Dynamic adaptability, re-solve the distributed model after each homotopy parameter update, allowing the control strategy to adaptively adjust with new energy output changes;

[0114] (3) Physical feasibility guarantee, combined with the step-by-step advancement strategy and backtracking mechanism, to ensure that each solution meets the grid physical constraints, and ultimately achieve stable global optimization under high-proportion new energy access.

[0115] In addition, the grid state variable (such as voltage, power) dimension grows with the number of nodes, but the distributed control can reduce the single node calculation amount through local neighborhood interaction, avoid global high-dimensional matrix operation, and the path design is reasonable (such as physical-inspired parameterization) to avoid singular points, so the iteration number is controllable, and the quasi-steady-state characteristics of the grid help to linearly approach the path segment.

[0116] S244, the degree theory is used to verify the consistency of the global optimal solution, generate optimal regulation setting values that meet the causal constraint conditions, and use the optimal regulation setting values as the coordinated balance point of the energy power station and the power grid.

[0117] It should be noted that the degree theory is a tool in topology, which is used to judge the existence and number of solutions of nonlinear equations. The basic idea is to calculate the topological degree of the mapping on the boundary to infer whether there is a solution in a certain region and the properties of the solution.

[0118] When the degree theory is used to verify the consistency of the global optimal solution, it is mathematically verified whether the obtained distributed balance control solution exists uniquely within the given constraint space and whether it meets the causal constraint conditions, thereby excluding the possibility of false solutions or local abnormal solutions, ensuring that the final generated regulation setting values have global convergence, consistency and feasibility, and can be used as a real stable coordinated balance point between the new energy power station and the power grid.

[0119] It should be noted that the coordinated balance point is the optimal regulation setting value obtained on the basis of identifying the coordination contradiction point between the randomness of the new energy power station and the determinacy of the demand of the power grid. It represents the stable balance state reached by the new energy power station and the power grid under certain operating conditions. In this state, the output characteristics of the new energy power station match the response capability of the power grid, so the balance coordination point is used as the initial state for simulation, providing a comparison benchmark for identifying abnormal coordination performance during the process of connecting the new energy power station to the power grid.

[0120] S3, the balance coordination point is used as the initial state to simulate the real-time operating state of the new energy power station when it is connected to the power grid, and the real-time operating state is compared with the initial state to identify the abnormal coordination performance of the new energy power station when it is connected to the power grid.

[0121] Among them, the balance coordination point is used as the initial state to simulate the real-time operating state of the new energy power station when it is connected to the power grid, and the real-time operating state is compared with the initial state to identify the abnormal coordination performance of the new energy power station when it is connected to the power grid, including:

[0122] S31, based on the balance coordination point, a non-diagonal matrix of the Hamiltonian of the power grid is constructed, and based on the non-diagonal matrix, the critical behavior in the operation of the power grid is analyzed to obtain the initial coordination state of the power grid.

[0123] It should be noted that the state variables of this point (such as node power, voltage amplitude, phase angle difference, etc.) are used as the initial state of the system to input the Hamiltonian dynamics model, and a non-diagonal matrix form of the Hamiltonian H of the power grid is constructed. The matrix is based on the topological structure of the power grid and the coupling relationship between nodes, and the elements can represent the node energy exchange term or interconnection power gradient. The specific form is as follows: ij = -K ij cos(θi -θ j ), wherein K ij represents the node coupling strength, θ i represents the phase angle; by solving the eigenvalues and eigenvectors of the non-diagonal Hamiltonian matrix, the critical behavior of the power grid operation is analyzed, such as the critical instability boundary of the system when the eigenvalue approaches zero, or the energy catastrophe point; finally, according to the spectral distribution of the matrix, the characteristic structure of the initial coordination state of the system is identified, and the voltage-power stability distribution and the minimum disturbance energy state under the state are extracted.

[0124] S32, construct a lattice Boltzmann model, simulate the dynamic evolution process of the new energy power station after connecting to the power grid by using the lattice Boltzmann model, and obtain the disturbance operation state of the new energy power station.

[0125] Among them, the lattice Boltzmann model is constructed, the dynamic evolution process of the new energy power station after connecting to the power grid is simulated by using the lattice Boltzmann model, and the disturbance operation state of the new energy power station is obtained, which includes:

[0126] S321, map the nodes of the power grid to the nodes of the lattice grid, and the transmission lines and connecting devices in the power grid as the edges between the lattice grids, and construct a lattice Boltzmann model;

[0127] S322, take the output mode of the new energy power station as the input of the lattice Boltzmann model, simulate the dynamic evolution process of the output in the power grid by using the lattice Boltzmann, and obtain the disturbance operation state of the new energy power station.

[0128] It should be noted that the lattice Boltzmann model (LBM) is a calculation model based on local interaction and fluid dynamics, which can be used to simulate the evolution process of complex systems. In the application of the dynamic evolution process of new energy power stations and power grids, the lattice Boltzmann model can describe the state evolution of the power grid and the disturbance influence after the connection of new energy by mapping the power grid and the output mode of the new energy power station to the lattice system, which specifically includes:

[0129] Step one, take the output mode of the new energy power station (such as the volatility of wind power, photovoltaic power generation, etc.) as the input of the model, which is dynamically simulated by the state updating rule of the local node in the lattice model, reflecting the disturbance transmission and influence of the new energy power station after connecting to the power grid.

[0130] Step two, through the lattice Boltzmann simulation, the dynamic evolution process of each node in the power grid is obtained, including power flow, voltage fluctuation, etc. The disturbance caused by the output mode of the new energy power station propagates in the lattice network, and the change of the state of the power grid is obtained through time iteration, finally simulating the disturbance operation state of the new energy power station after connecting to the power grid.

[0131] Step three, compare the state changes before and after the disturbance, identify abnormal behaviors of the power grid after the access of new energy, such as frequency fluctuation, overload risk, voltage collapse, etc., which can reflect whether the coordination performance of the power grid is affected by the output fluctuation of new energy.

[0132] wherein the expression of the lattice Boltzmann model is:

[0133]

[0134] wherein f i (x,t) represents the particle distribution function along the lattice direction i at position x and time t, that is, the probability distribution of the state (such as power, voltage, etc.) of the power grid node; e i represents the discrete velocity vector (lattice velocity in direction i), corresponding to the propagation direction of energy or information in the power grid (such as the connection direction of the power transmission line); Δt represents the time step; τ represents the relaxation time, which controls the speed of the system tending to the equilibrium state, and reflects the inertia or damping characteristics of the power grid; represents the equilibrium state distribution function along the lattice direction i, that is, the steady-state distribution of the power grid without disturbance; F i (x,t) represents the external force term along the lattice direction i, that is, the disturbance input of the new energy power station.

[0135] S33, respectively map the initial coordination state and the disturbance operation state into the complex plane unit disc, and calculate the difference between the initial coordination state and the disturbance operation state through dynamic Mobius transformation.

[0136] It should be noted that respectively mapping the initial coordination state and the disturbance operation state into the complex plane unit disc, and calculating the difference between the initial coordination state and the disturbance operation state through dynamic Mobius transformation includes:

[0137] The state variables (such as voltage amplitude and phase angle) of each node of the power grid are converted into complex form and normalized into the unit disc (the unit disc is a region composed of all complex numbers in the complex plane that satisfy |z|≤1); complex point sets are generated for the initial coordination state and the disturbance operation state, respectively, and then a dynamic Mobius transformation is constructed, the least squares method is used to fit to minimize the difference, and finally the residual norm after transformation is used to quantify the difference between the two states, identify the phase shift and amplitude distortion region caused by new energy disturbance, and reveal the weakness of the power grid through sensitivity analysis of the transformation parameters.

[0138] S34, compare the difference calculation result with the equilibrium coordination point, and input the comparison result into the improved random forest to identify abnormal coordination performance.

[0139] wherein inputting the comparison result into the improved random forest to identify abnormal coordination performance includes:

[0140] Screen the embedded difference sample based on the comparison result of the difference calculation result and the balanced coordination point, use a self-help statistical analysis strategy to randomly extract a plurality of instances from the embedded difference sample to form a disturbance coordination sample subset for training;

[0141] Respectively construct a corresponding decision tree for the embedded difference sample, train each decision tree using the disturbance coordination sample subset, and perform a weighted evaluation on the discrimination ability between the difference calculation result and the balanced coordination point, select representative features according to the score ratio to construct a feature subspace;

[0142] On the constructed feature subspace, guide each decision tree to grow to the maximum depth, until all the decision trees constitute an improved random forest model;

[0143] Input the predefined to-be-measured index into the improved random forest model, independently judge whether the to-be-measured index exists coordination deviation through each tree in the improved random forest model, and obtain the detection result of the abnormal coordination performance of the new energy power station connected to the power grid.

[0144] It should be noted that by inputting the difference calculation result and the balanced coordination point into the improved random forest model, the abnormal coordination performance of the new energy power station connected to the power grid can be efficiently identified. The improved random forest selects a disturbance coordination sample subset through a self-help statistical analysis strategy, enhances the capture ability of small probability abnormal patterns, reduces redundant feature interference, and improves the sensitivity of the model to key state variables (such as voltage and frequency deviation) of the power grid. The depth growth strategy of the decision tree maximizes the information gain on the feature subspace, so that the model can more accurately distinguish between normal and abnormal coordination states. Finally, through the independent voting mechanism of multiple trees, the false positive rate is reduced, and a robust abnormal detection result is output. Compared with the traditional random forest, the improvement points include:

[0145] (1) Introduce difference sample screening and weighted evaluation, and specifically strengthen the identification of disturbance patterns;

[0146] (2) Feature subspace optimization instead of random feature selection, avoiding irrelevant feature dilution of discrimination ability;

[0147] (3) Depth growth constraint combined with voting mechanism, balancing the risk of overfitting and detection accuracy, significantly improving the abnormal diagnosis ability of high-dimensional nonlinear power grid dynamics.

[0148] According to another embodiment of the present application, as Figure 2 shown, a new energy power station grid-source coordination performance detection system is also provided, which comprises:

[0149] An interactive behavior acquisition module 1 is used to acquire the operation data of the new energy power station and the power grid, construct an interactive model between the new energy power station and the power grid based on the operation data, and analyze the interactive behavior of the new energy power station and the power grid by using the interactive model.

[0150] The coordination performance analysis module 2 is used for identifying the coordination contradiction point between the randomness gene of the new energy power station and the certainty demand of the power grid based on the interaction behavior of the new energy power station and the power grid, and optimizing the coordination contradiction point into a coordination balance point by using a distributed balance control model;

[0151] The coordination performance detection module 3 is used for taking the balance coordination point as an initial state, simulating the real-time running state of the new energy power station when the new energy power station is connected to the power grid, and comparing the real-time running state with the initial state to identify the abnormal coordination performance of the new energy power station when the new energy power station is connected to the power grid.

[0152] The above merely describes the preferred embodiments of the present application and is not used to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for detecting the grid-source coordination performance of a new energy power plant, characterized in that, The method includes: S1. Collect operational data of new energy power plants and power grids, construct an interaction model between new energy power plants and power grids based on the operational data, and use the interaction model to analyze the interaction behavior between energy power plants and power grids. S2. Based on the interaction behavior between new energy power plants and the power grid, identify the coordination contradictions between the randomness of new energy power plants and the deterministic demand of the power grid, and use a distributed balance control model to optimize the coordination contradictions into coordination balances. S3. Using the equilibrium coordination point as the initial state, simulate the real-time operating state of the new energy power station when it is connected to the grid, and compare the real-time operating state with the initial state to identify abnormal coordination performance when the new energy power station is connected to the grid.

2. The method for detecting the grid-source coordination performance of a new energy power station according to claim 1, characterized in that, The process of identifying coordination contradictions between the stochastic nature of new energy power plants and the deterministic demands of the power grid based on the interaction behavior between the power plants and the grid, and optimizing these contradictions into equilibrium points using a distributed balance control model, includes: S21. Extract the randomness gene indicators of new energy and the deterministic demand indicators of the power grid based on the interaction behavior between new energy power plants and the power grid, and conduct spatiotemporal correlation analysis on the randomness gene indicators and the deterministic demand indicators. S22. Based on the spatiotemporal correlation analysis results, establish the topological structure between random gene indicators and deterministic demand indicators, and identify incompatible indicators that deviate from the preset conditions from the topological structure as points of coordination contradictions. S23. Based on the points of coordination contradictions, infer the causal relationship between the output characteristics of new energy power plants and the state characteristics of the power grid; S24. Construct a distributed equilibrium control model based on causal relationships, combine the homotopy algorithm to generate the optimal adjustment setpoint of the distributed equilibrium control model, and use the optimal adjustment setpoint as the coordination equilibrium point.

3. The method for detecting the grid-source coordination performance of a new energy power station according to claim 2, characterized in that, The process of inferring the causal relationship between the output characteristics of new energy power plants and the state characteristics of the power grid based on the points of conflict, and constructing a distributed balance control model based on the causal relationship, includes: S231. Construct a power grid constrained surface model, analyze the influence of the power output nodes of new energy power plants on the boundary morphology of the power grid constrained surface model through the power grid constrained surface model, and identify the singularity locations that cause surface distortion under the power output mode of new energy power plants. S232. Taking the output node of the new energy power station corresponding to the singularity position as the center, and constructing a multivariate causal graph between the output state of the new energy power station and the operation state of the power grid based on the output node. S233. Convert the multivariate causal graph into a Boolean causal graph, and construct a conditional probability table for each pair of nodes in the Boolean causal graph. S234. Based on the conditional probability table, infer the causal relationship between the output characteristics of the new energy power station and the state characteristics of the power grid, and introduce a feasibility function to correct the causal relationship.

4. The method for detecting the grid-source coordination performance of a new energy power station according to claim 3, characterized in that, The step of constructing a distributed equilibrium control model based on causal relationships, generating the optimal adjustment setpoint of the distributed equilibrium control model using a homotopy algorithm, and using the optimal adjustment setpoint as the coordination equilibrium point includes: S241. Encode causal relationships into causal constraints and objective functions, and construct a distributed equilibrium control model; S242. Based on the grid state without access to new energy power stations as the stable state, the homotopy algorithm is used to iteratively update the homotopy path that gradually transitions from the stable state to the actual target state. S243. Dynamically adjust the control variables of the distributed equilibrium control model along the homotopy path, and gradually approach the global optimal solution that satisfies the causal constraints by combining the stepwise advancement strategy. S244. The degree theory is introduced to verify the consistency of the global optimal solution, generate the optimal adjustment setpoint that satisfies the causal constraint, and use the optimal adjustment setpoint as the coordination balance point between the power plant and the power grid.

5. The method for detecting the grid-source coordination performance of a new energy power station according to claim 4, characterized in that, The dynamic adjustment of control variables along the homotopy path in the distributed equilibrium control model, combined with a stepwise advancement strategy to gradually approach the global optimal solution satisfying causal constraints, includes: S2431. Initialize the homotopy algorithm parameters, and set the initial value range and maximum number of iterations for the control variables of the distributed equilibrium control model. S2432. Gradually increase the parameters of the homotopy algorithm to drive the change of the control variables in the distributed equilibrium control model, and in each homotopy path, use the currently changed control variables as the initial values ​​to solve the distributed equilibrium control model; S2433. Determine the convergence state of the distributed equilibrium control model. If convergence is successful, obtain the current optimal solution and execute step S2435. If convergence fails, execute step S2434. S2434. Determine whether the step size at the convergence of the current iteration is the minimum step size. If not, backtrack the homotopy algorithm parameters to the previous state and return to step S2432. Otherwise, use the current optimal solution as the initial value for the next iteration. S2435. After each iteration of the homotopy algorithm, determine whether the homotopy algorithm parameters have reached the preset value. If the preset value is reached, it means that the distributed equilibrium control model has successfully converged, and output the global optimal solution that satisfies the causal constraint. Otherwise, increase the number of iterations and return to step S2432 until the convergence condition is met.

6. The method for detecting the grid-source coordination performance of a new energy power station according to claim 1, characterized in that, The step of using the equilibrium coordination point as the initial state to simulate the real-time operating state of a new energy power station when it is connected to the grid, and comparing the real-time operating state with the initial state to identify abnormal coordination performance when the new energy power station is connected to the grid includes: S31. Based on the equilibrium coordination point, construct the off-diagonal matrix of the Hamiltonian of the power grid, and analyze the critical behavior of the power grid operation based on the off-diagonal matrix to obtain the initial coordination state of the power grid. S32. Construct a lattice Boltzmann model and use the lattice Boltzmann model to simulate the dynamic evolution process of new energy power plants after they are connected to the grid, and obtain the disturbance operation state of the new energy power plants. S33. Map the initial coordinated state and the disturbed running state onto the complex plane unit disk respectively, and calculate the difference between the initial coordinated state and the disturbed running state through dynamic Möbius transformation; S34. Compare the difference calculation results with the equilibrium coordination point, and input the comparison results into the improved random forest to identify abnormal coordination performance.

7. The method for detecting grid-source coordination performance of a new energy power station according to claim 6, characterized in that, The construction of the lattice Boltzmann model is used to simulate the dynamic evolution process of new energy power plants after they are connected to the grid, and the disturbance operation state of the new energy power plants includes: S321. Map the power grid nodes as nodes of a lattice grid, and use the transmission lines and connecting equipment in the power grid as edges between the lattice grids to construct a lattice Boltzmann model. S322. The output mode of the new energy power station is used as the input of the lattice Boltzmann model. The dynamic evolution process of the output in the power grid is simulated by the lattice Boltzmann model to obtain the disturbance operation state of the new energy power station.

8. The method for detecting the grid-source coordination performance of a new energy power station according to claim 7, characterized in that, The process of inputting the comparison results into the improved random forest to identify anomalies and coordinate performance includes: Based on the comparison between the difference calculation results and the equilibrium coordination points, embedded difference samples are selected, and a number of instances are randomly extracted from the embedded difference samples using a self-help statistical analysis strategy to form a subset of perturbation coordination samples for training. Decision trees are constructed for each embedded difference sample. Each decision tree is trained using a subset of perturbation-coordinated samples. The discriminative ability between the difference calculation results and the equilibrium coordination point is evaluated by weighting. Representative features are selected according to the score ratio to construct a feature subspace. In the constructed feature subspace, each decision tree is guided to grow to the maximum depth until all decision trees constitute an improved random forest model. Input the predefined test index into the improved random forest model. Each tree in the improved random forest model independently judges whether there is a coordination offset in the test index, so as to obtain the detection results of abnormal coordination performance of new energy power plants connected to the grid.

9. The method for detecting the grid-source coordination performance of a new energy power station according to claim 6, characterized in that, The expression for the lattice Boltzmann model is: In the formula, f i (x,t) represents the particle distribution function along the lattice direction i at position x and time t; e i The vector represents the discrete velocity vector; Δt represents the time step; τ represents the relaxation time; f i eq F represents the equilibrium distribution function along lattice direction i; i (x,t) represents the external force term along the lattice direction i.

10. A grid-source coordination performance testing system for new energy power plants, used to implement the grid-source coordination performance testing method for new energy power plants according to any one of claims 1-9, characterized in that, The system includes: The interaction behavior acquisition module is used to collect operational data between new energy power plants and the power grid, construct an interaction model between the new energy power plants and the power grid based on the operational data, and analyze the interaction behavior between the power plants and the power grid using the interaction model. The coordination performance analysis module is used to identify coordination contradictions between the randomness of new energy power plants and the deterministic demands of the power grid based on the interaction behavior between new energy power plants and the power grid, and to optimize the coordination contradictions into coordination equilibrium points using a distributed balance control model. The coordination performance detection module is used to simulate the real-time operating state of a new energy power station when it is connected to the grid, using the equilibrium coordination point as the initial state. It then compares the real-time operating state with the initial state to identify abnormal coordination performance when the new energy power station is connected to the grid.

Images