Quantitative evaluation method, system and equipment for frequency-voltage supporting capability of multi-scene gas power station

Abstract

The invention discloses a multi-scene gas power station frequency-voltage supporting capability quantitative evaluation method, system and device, and relates to the technical field of power system stability control and energy evaluation, and the method comprises the steps: collecting multi-source data of a plurality of scenes in real time and carrying out the preprocessing of the multi-source data in the grid-connected operation process of a gas power station; determining a scene feature vector and an index feature vector based on the preprocessed multi-source data; inputting the scene feature vector and the index feature vector into a power grid frequency-voltage supporting capability comprehensive evaluation model, and outputting an index score of each index in the index feature vector and a predicted value of a power grid frequency-voltage supporting capability comprehensive score of the current scene; and determining an evaluation grade of the power grid frequency-voltage supporting capability in the current scene based on a predicted value of the comprehensive score of the power grid frequency-voltage supporting capability in the current scene and the index score of each index in the index feature vector. According to the invention, the evaluation precision of the frequency-voltage supporting capability of the gas power station is improved.

Description

Technical Field

[0002] This application relates to the field of power system stability control and energy assessment technology, and in particular to a quantitative assessment method, system and equipment for the frequency-voltage support capability of gas-fired power plants in multiple scenarios. Background Technology

[0003] With the large-scale integration of new energy sources such as wind and solar power, the inertia of the power system is decreasing, and gas-fired power plants need to shoulder a heavier dynamic frequency-voltage support task. For example, measured data from a provincial power grid show that when the penetration rate of new energy exceeds 30%, the system frequency drop rate increases by 40%, and the voltage fluctuation amplitude increases by 25%. Gas-fired power plants operate under complex scenarios, including sudden load changes, fluctuations in new energy output, and line faults.

[0004] Traditional assessment methods are limited to single scenarios and struggle to quantify support capabilities under extreme weather conditions or equipment aging. Existing assessment systems emphasize steady-state indicators but often neglect transient processes. Single models calculate the short-circuit ratio of renewable energy power plants using equivalent impedance matrices, but this is only applicable to renewable energy grid-connected scenarios and cannot be extended to multi-scenario assessments of gas-fired power plants. Traditional AHP relies on expert scoring to determine indicator weights, which can lead to significant weight biases due to expert experience and a lack of objective data support. Static hierarchical structures fail to reflect scenario changes (e.g., voltage support weight needs to be increased to 40% in fault scenarios), resulting in assessment results that are out of sync with actual needs. Scenario classification in related methods relies on manual experience (e.g., only distinguishing between "high / low load" levels), lacking refined classification based on measured data. Steady-state and transient indicators are not integrated (e.g., in the assessment of a gas-fired power plant, the steady-state voltage pass rate is 98%, but the voltage recovery time after a fault exceeds the standard by 50%), leading to distorted comprehensive assessments.

[0005] Therefore, it is necessary to provide a quantitative evaluation method for the frequency-voltage support capability of gas-fired power plants in multiple scenarios to solve the above problems. Summary of the Invention

[0006] The purpose of this application is to provide a quantitative evaluation method, system, and equipment for the frequency-voltage support capability of gas-fired power plants in multiple scenarios, thereby improving the evaluation accuracy of the frequency-voltage support capability of gas-fired power plants.

[0007] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a quantitative evaluation method for the frequency-voltage support capability of gas-fired power plants in multiple scenarios, the quantitative evaluation method for the frequency-voltage support capability of gas-fired power plants in multiple scenarios includes: During the grid-connected operation of the gas-fired power plant, multi-source data from multiple scenarios are collected in real time. The multi-source data includes: equipment parameters of the gas turbine, grid operating status parameters, and environmental parameters. Preprocessing of multi-source data from multiple scenarios yields preprocessed multi-source data for multiple scenarios; Based on the preprocessed multi-source data of the current scenario, the scenario feature vector and indicator feature vector of the current scenario are determined. The scenario feature vector includes: the scenario type of the current scenario and the corresponding scenario feature indicator value. The indicator feature vector includes: frequency support capability indicator and voltage support capability indicator. The scene feature vector and index feature vector of the current scene are both input into the comprehensive evaluation model of power grid frequency-voltage support capability. The output is the index score of each index in the index feature vector and the predicted value of the comprehensive score of power grid frequency-voltage support capability of the current scene. The comprehensive evaluation model of power grid frequency-voltage support capability is obtained by training an attention mechanism neural network model using a training set. Based on the predicted value of the comprehensive score of the grid frequency-voltage support capability in the current scenario and the index scores of each index in the index feature vector, the evaluation level of the grid frequency-voltage support capability in the current scenario is determined, thereby determining the evaluation level of the grid frequency-voltage support capability of the gas-fired power plant in multiple scenarios.

[0008] In one embodiment, the gas turbine equipment parameters include: active power, power setpoint, and reactive power; Grid operating parameters include: high-frequency grid frequency at the grid connection point, voltage amplitude at the grid connection point, current waveform, total system load, total output of new energy sources, and grid fault signals; Environmental parameters include: ambient temperature, wind speed, and ambient humidity.

[0009] In one embodiment, multi-source data from multiple scenarios are preprocessed to obtain preprocessed multi-source data for multiple scenarios, specifically including: The random forest algorithm is used to remove outliers from multi-source data of multiple scenarios, thus obtaining effective multi-source data for multiple scenarios; Normalize the effective multi-source data from multiple scenarios to obtain preprocessed multi-source data from multiple scenarios.

[0010] In one implementation, the current scenario is classified as either a basic scenario or an extreme scenario. Basic scenarios include peak load scenarios, off-peak load scenarios, and scenarios involving large-scale renewable energy generation. Extreme scenarios include cold wave scenarios, typhoon season scenarios, and equipment failure scenarios. The scenario characteristic indicators include: total system load, new energy penetration rate, ambient temperature, and fault indicators.

[0011] In one embodiment, the frequency support capability indicators include: regulation capacity ratio, response time, steady-state frequency deviation, and comprehensive regulation capability index; Voltage support capability indicators include: reactive power, voltage regulation sensitivity, reactive response time, and harmonic suppression capability.

[0012] In one implementation, based on preprocessed multi-source data of the current scene, the scene feature vector and indicator feature vector of the current scene are determined, specifically including: Determine whether the values ​​of multiple scene feature indicators in the preprocessed multi-source data of the current scene meet the corresponding preset range, determine the scene type to which the current scene belongs, and use the scene type to which the previous scene belongs and the corresponding scene feature indicator values ​​as the scene feature vector of the current scene. Based on the preprocessed multi-source data of the current scenario, the corresponding frequency support capability index and voltage support capability index are calculated to obtain the index feature vector corresponding to the current scenario.

[0013] In one embodiment, the training process of the integrated assessment model for grid frequency-voltage support capability specifically includes: Construct a training set; the training set includes: scene feature vectors and indicator feature vectors corresponding to multi-source data of historical scenarios, the true values ​​of the comprehensive score of power grid frequency-voltage support capability marked by manual annotation, and the indicator scores of each indicator in the indicator feature vector marked by manual annotation; Build a neural network model for the attention mechanism; Using the scene feature vector and indicator feature vector corresponding to multi-source data of historical scenarios as input, and the true value of the comprehensive score of power grid frequency-voltage support capability marked by manual annotation and the indicator scores of each indicator in the indicator feature vector marked by manual annotation as output, the attention mechanism neural network model is trained until the loss function reaches the minimum value or the training rounds reach the maximum value, and then training is stopped to obtain the comprehensive evaluation model of power grid frequency-voltage support capability.

[0014] In one implementation, based on the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario and the index scores of each index in the index feature vector, the assessment level of the power grid frequency-voltage support capability in the current scenario is determined, specifically including: Determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to the first preset threshold, and obtain the first judgment result; If the first judgment result is yes, then the evaluation level of the power grid frequency-voltage support capability in the current scenario is determined to be "excellent"; If the first judgment result is negative, then determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to the second preset threshold, and obtain the second judgment result. If the second judgment result is yes, then the assessment level of the power grid frequency-voltage support capability in the current scenario is determined to be "good"; If the second judgment result is negative, then determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to the third preset threshold, and obtain the third judgment result. If the third judgment result is yes, then the assessment level of the power grid frequency-voltage support capability in the current scenario is determined to be "average". If the third judgment result is negative, the assessment level of the power grid frequency-voltage support capability in the current scenario is judged to be "insufficient". Then, the fault tree analysis method is used to trace the root cause based on the minimum value of the index score of each index in the index feature vector, and generate a three-level optimization scheme corresponding to the current scenario.

[0015] Secondly, this application provides a quantitative evaluation system for the frequency-voltage support capability of multi-scenario gas-fired power plants. This system is used to implement the aforementioned quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants. The quantitative evaluation system for the frequency-voltage support capability of multi-scenario gas-fired power plants includes: The data acquisition unit is used to collect multi-source data from multiple scenarios in real time during the grid-connected operation of the gas-fired power plant. The multi-source data includes: equipment parameters of the gas turbine, grid operating status parameters, and environmental parameters. The preprocessing unit is used to preprocess multi-source data from multiple scenarios to obtain preprocessed multi-source data from multiple scenarios. The feature vector determination unit is used to determine the scene feature vector and indicator feature vector of the current scene based on the preprocessed multi-source data of the current scene. The scene feature vector includes: the scene type of the current scene and the corresponding scene feature indicator value; the indicator feature vector includes: frequency support capability indicator and voltage support capability indicator. The indicator score and comprehensive score determination unit is used to input both the scene feature vector and indicator feature vector of the current scene into the comprehensive evaluation model of power grid frequency-voltage support capability, and output the indicator score of each indicator in the indicator feature vector and the predicted value of the comprehensive evaluation model of power grid frequency-voltage support capability of the current scene; the comprehensive evaluation model of power grid frequency-voltage support capability is obtained by training an attention mechanism neural network model using a training set; The assessment level determination unit is used to determine the assessment level of the grid frequency-voltage support capability under the current scenario based on the predicted value of the comprehensive score of the grid frequency-voltage support capability under the current scenario and the index scores of each index in the index feature vector, thereby determining the assessment level of the grid frequency-voltage support capability of the gas-fired power plant under multiple scenarios.

[0016] According to the specific embodiments provided in this application, this application has the following technical effects: This application discloses a quantitative evaluation method, system, and device for the frequency-voltage support capability of gas-fired power plants in multiple scenarios. By constructing a multi-level evaluation model (i.e., a comprehensive evaluation model for power grid frequency-voltage support capability) consisting of a target layer, a standard layer, and an indicator layer, it covers multiple indicators of frequency support and voltage support. Combined with different scenarios, it achieves a comprehensive evaluation of the overall support capability of gas-fired power plants, significantly improving the accuracy of the evaluation and more accurately reflecting the actual operating status of the power grid. At the same time, the multi-level comprehensive evaluation model for power grid frequency-voltage support capability based on the attention mechanism can dynamically calculate scenario weights to dynamically adjust the importance of different features, replacing the judgment matrix weight calculation steps in traditional AHP that rely on expert experience. This achieves a completely data-driven objective evaluation, adapts to the evaluation needs under different scenarios, and improves the adaptability and flexibility of the model. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a quantitative evaluation method for the frequency-voltage support capability of a gas-fired power plant in multiple scenarios, provided in an embodiment of this application. Figure 2 This is another flowchart illustrating a quantitative evaluation method for the frequency-voltage support capability of a gas-fired power plant in multiple scenarios, provided in an embodiment of this application. Figure 3 This is a schematic diagram of the multi-level structure of a comprehensive evaluation model for power grid frequency-voltage support capability provided in an embodiment of this application; Figure 4 This is a schematic diagram of the analytic hierarchy process provided in an embodiment of this application; Figure 5 A schematic diagram of the module structure of a quantitative evaluation system for the frequency-voltage support capability of a gas-fired power plant in a scenario provided in an embodiment of this application; Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] This application constructs a multi-level evaluation model that includes frequency support capability indicators and voltage support capability indicators. Combined with dynamic weight allocation for different typical scenarios, it achieves quantitative scoring and optimization guidance for the comprehensive support capability of gas-fired power plants. This model is applicable to decision support in scenarios such as grid peak shaving, frequency regulation, renewable energy consumption, and fault recovery, enabling dynamic support capability analysis and optimization of gas-fired power plants during grid-connected operation. While retaining the hierarchical evaluation framework of AHP (Advanced Performance Management), it innovatively introduces an attention mechanism to construct a dynamic weight calculation model, completely eliminating the judgment matrix construction stage that relies on expert experience in traditional AHP, and achieving a fully data-driven objective evaluation.

[0021] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, this application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] In one exemplary embodiment, such as Figure 1 and Figure 2 As shown, a quantitative evaluation method for the frequency-voltage support capability of gas-fired power plants in multiple scenarios is provided, including the following steps: Wherein: Step S1: During the grid-connected operation of the gas power plant, multi-source data from multiple scenarios are collected in real time; the multi-source data includes: equipment parameters of the gas turbine, grid operation status parameters, and environmental parameters.

[0023] As an optional implementation, in step S1, the gas turbine equipment parameters include: active power, power setpoint, and reactive power; the grid operation parameters include: high-frequency grid frequency at the grid connection point, voltage amplitude at the grid connection point, current waveform, total system load, total output of new energy sources, and grid fault signal; and the environmental parameters include: ambient temperature, wind speed, and ambient humidity.

[0024] Specifically, data acquisition begins with real-time monitoring of the gas turbine equipment status, grid operation status, and environmental parameters. This involves collecting active power, power setpoint, and high-frequency grid data at the grid connection point to calculate indicators such as regulation capacity, response time, and frequency deviation. Reactive power, grid connection point voltage amplitude, and current waveform are also collected to analyze reactive power regulation performance and harmonic characteristics. Simultaneously, temperature and humidity sensors and wind speed sensors are deployed to collect environmental parameters such as ambient temperature, wind speed, and humidity. Furthermore, to construct the scenario feature vector, external parameters such as total system load, total renewable energy output, and grid fault signals need to be integrated. All collected data is aggregated through the grid SCADA system, forming a multi-source data set including gas turbine equipment parameters, grid operation parameters, and environmental parameters. The sampling frequency is set to once per minute.

[0025] Step S2: Preprocess the multi-source data from multiple scenarios to obtain preprocessed multi-source data from multiple scenarios.

[0026] As an optional implementation method, step S2 specifically includes: Step S21: Use the random forest algorithm to remove outliers from the multi-source data of multiple scenarios to obtain valid multi-source data for multiple scenarios.

[0027] Step S22: Normalize the effective multi-source data from multiple scenarios to obtain preprocessed multi-source data from multiple scenarios.

[0028] Specifically, firstly, the random forest algorithm is used to remove outliers from the multi-source data of multiple scenarios, retaining the valid data; then, the valid data is normalized to map the data to the [0, 1] interval to eliminate the influence of units, resulting in preprocessed multi-source data for multiple scenarios. The normalization formula is as follows: (1) In the formula, x norrm These are preprocessed multi-source data values ​​from multiple scenarios; x This represents multi-source data values ​​from multiple scenarios. x min , x max These represent the minimum and maximum values ​​of multi-source data from multiple scenarios.

[0029] Step S3: Based on the preprocessed multi-source data of the current scenario, determine the scenario feature vector and indicator feature vector of the current scenario; the scenario feature vector includes: the scenario type of the current scenario and the corresponding scenario feature indicator value; the indicator feature vector includes: frequency support capability indicator and voltage support capability indicator.

[0030] As an optional implementation, step S3 specifically includes: Step S31: Determine whether the multiple scene feature index values ​​in the preprocessed multi-source data of the current scene meet the corresponding preset range, determine the scene type to which the current scene belongs, and use the scene type to which the previous scene belongs and the corresponding scene feature index value as the scene feature vector of the current scene.

[0031] Step S32: Based on the preprocessed multi-source data of the current scenario, calculate the corresponding frequency support capability index and voltage support capability index to obtain the index feature vector corresponding to the current scenario.

[0032] As an optional implementation, in step S3, the current scenario type is either a basic scenario or an extreme scenario; basic scenarios include: peak load scenario, low load scenario, and new energy power generation scenario; extreme scenarios include cold wave weather scenario, typhoon season scenario, and equipment failure scenario; scenario characteristic index values ​​include: total system load, new energy penetration rate, ambient temperature, and fault identifier.

[0033] As an optional implementation method, the frequency support capability indicators include: regulation capacity ratio, response time, steady-state frequency deviation, and comprehensive regulation capability index; the voltage support capability indicators include: reactive power, voltage regulation sensitivity, reactive response time, and harmonic suppression capability.

[0034] Specifically, taking a peak load scenario as an example, key attribute parameters (i.e., scenario feature indicators) of the peak load scenario are selected from the preprocessed multi-source data of the current scenario to form a scenario feature vector with dimension [1×4], denoted as . , =[Load rate, renewable energy penetration rate, ambient temperature, fault indicator]; based on preprocessed multi-source data for the current scenario. For the first i In each scenario, the quantitative values ​​of the frequency support capability index and the voltage support capability index are calculated using formulas (2)-(24), forming an index feature vector with dimension [1×8], denoted as... , For the first j The metrics ultimately output a scene feature vector. With indicator eigenvectors .

[0035] Among the frequency support capability indicators, the regulation capacity ratio is a key indicator for evaluating the power system's frequency support capability. Essentially, it's the ratio of regulation capacity to total generation capacity, reflecting the system's ability to maintain frequency stability by adjusting generation output during load fluctuations or disturbances. Response time is the time from the occurrence of a disturbance to the start of gas-fired power generation adjusting its output. A shorter response time indicates a more timely response to changes in grid frequency, contributing to rapid frequency recovery. Steady-state frequency deviation is the degree of deviation between the actual and target output of gas-fired power generation. A smaller deviation indicates more accurate power generation control and better frequency support. The comprehensive regulation capability index represents the ability of gas-fired power generation to maintain stable regulation over a long period; gas-fired power generation with strong continuous regulation capability can provide stable support when grid frequency fluctuates continuously.

[0036] In voltage support capability indicators, reactive power refers to the reactive power output that gas-fired power generation can provide under voltage deviation. The stronger the reactive power output capability, the stronger the gas-fired power generation's ability to support voltage. Voltage regulation sensitivity is the ratio of voltage change to the reactive power change that caused that change, reflecting the gas-fired power generation's sensitivity to voltage changes. The higher the regulation sensitivity, the faster the gas-fired power generation responds to voltage changes and provides reactive power support. Reactive power response time refers to the time from the occurrence of voltage deviation to the start of reactive power generation adjusting reactive power output. The shorter the reactive power response time, the more promptly the gas-fired power generation can respond to voltage changes, helping to quickly restore voltage stability. Harmonic suppression capability: Gas-fired power generation with strong harmonic suppression capability can reduce the impact of harmonics on the power grid and improve power quality. Harmonic suppression capability is mainly described by the following three parameters: Total Harmonic Distortion (THD) is the ratio of the root mean square value of the harmonic content in a periodic AC quantity to the root mean square value of its fundamental component; Harmonic Content (HRUn) is the ratio of the effective value of the nth harmonic component to the effective value of the fundamental component.

[0037] The calculation formulas for each frequency support capability index are as follows: Adjusting capacity ratio for: (2) The constraints are: (3) in, For the first i The regulating capacity of the power supply; This represents the total power generation capacity of the system. n This represents the total number of power supply types involved in frequency regulation.

[0038] Response time for: (4) The constraints are: (5) in, The moment when the power generation unit or energy storage system begins to adjust its output; This refers to the moment when the frequency deviation exceeds the corresponding threshold.

[0039] steady-state frequency deviation for: (6) The constraints are: (7) in, This represents the actual steady-state frequency of the system. This represents the system's target frequency value.

[0040] Based on indicators such as regulation time, regulation capacity, and regulation efficiency, a comprehensive regulation capability index (CRII) is set, and its calculation formula is as follows: (8) (9) (10) (11) in, To adjust the time; This is the moment when the frequency returns to stability; To adjust capacity; This is the maximum output during the adjustment process; This is the minimum output force during the adjustment process; Rated capacity; To adjust efficiency; This refers to the output power. Input power; , , These are weighting coefficients, reflecting the importance of different indicators.

[0041] The calculation formulas for each voltage support capability index are as follows: The reactive power Q is: (12) The constraints are: (13) Where I is current and U is voltage; It is the angle between current and voltage.

[0042] Voltage regulation sensitivity for: (14) The constraints are: (15) in, This refers to the change in voltage. This refers to the change in reactive power that causes a change in voltage.

[0043] Reactive response time for: (16) The constraints are: (17) in, This is the moment to begin adjusting the reactive power output; This is to detect the moment when voltage changes or reactive power demand changes.

[0044] Total Harmonic Distortion for: (18) The constraints are: (19) Harmonic content for: (20) The constraint conditions for odd-order harmonics are: (twenty one) The constraint conditions for even-order harmonics are: (twenty two) Harmonic attenuation rate for: (twenty three) The constraints are: (twenty four) Among them, U n U1 is the effective value of the nth harmonic voltage; U2 is the effective value of the fundamental voltage; I in I represents the effective value of the harmonic current at the input of the filter device. out This represents the effective value of the harmonic current at the output of the filter device. It is expressed as harmonic attenuation rate.

[0045] Step S4: Input the scene feature vector and index feature vector of the current scene into the comprehensive evaluation model of power grid frequency-voltage support capability, and output the index score of each index in the index feature vector and the predicted value of the comprehensive score of power grid frequency-voltage support capability of the current scene; the comprehensive evaluation model of power grid frequency-voltage support capability is obtained by training the attention mechanism neural network model with the training set.

[0046] As an optional implementation, step S4, the training process of the comprehensive evaluation model for grid frequency-voltage support capability, specifically includes: Step S41: Construct a training set. The training set includes: scene feature vectors and indicator feature vectors corresponding to multi-source data of historical scenarios, the true values ​​of the manually annotated comprehensive score of power grid frequency-voltage support capability, and the indicator scores of each indicator in the manually annotated indicator feature vectors.

[0047] Step S42: Construct the attention mechanism neural network model. The attention mechanism neural network model includes an input layer, an encoding layer, an attention layer, and an output layer.

[0048] Step S43: Using the scene feature vector and indicator feature vector corresponding to the multi-source data of historical scenes as input, and the true value of the manually annotated comprehensive score of power grid frequency-voltage support capability and the indicator scores of each indicator in the manually annotated indicator feature vector as output, the attention mechanism neural network model is trained until the loss function reaches the minimum value or the training rounds reach the maximum value, and the training is stopped to obtain the comprehensive evaluation model of power grid frequency-voltage support capability.

[0049] Specifically, to comprehensively assess the frequency-voltage support capability of gas-fired power to the power grid, a comprehensive assessment model for the power grid's frequency-voltage support capability was constructed, such as... Figure 3 As shown, the comprehensive evaluation model for the frequency-voltage support capability of the power grid is a multi-level evaluation model composed of a target layer, a standard layer, and an indicator layer (the target layer refers to the frequency-voltage support capability; the standard layer is divided into frequency support capability and voltage support capability; the indicator layer consists of several indicators corresponding to the two support capabilities). That is, while retaining the hierarchical evaluation framework of AHP (target layer, standard layer, indicator layer), an attention mechanism is introduced to realize dynamic weight calculation, completely abandoning the judgment matrix construction link that relies on expert experience in the traditional AHP, and realizing a completely data-driven objective evaluation.

[0050] First, the acquired multi-source data is merged into a comprehensive dataset through data integration. Based on historical operating data, the actual contribution of each evaluation index to frequency-voltage stability under different scenarios is automatically learned, and dynamic attention weights are generated to replace the judgment matrix weight calculation that relies on expert experience in traditional AHP, thus achieving completely objective weight allocation.

[0051] For the assessment of "frequency-voltage support capability," a technical approach of "scene recognition - feature fusion - dynamic weight generation - comprehensive evaluation" is designed. An attention mechanism neural network is trained with the goal of minimizing the fitting error between the indicator weights and the actual contribution of the indicators to frequency-voltage stability. The input layer receives a concatenated vector of scene feature vectors and indicator feature vectors. Scene feature vectors include basic scenarios such as peak load, off-peak load, and large-scale renewable energy generation, as well as extreme scenarios such as cold waves, typhoon season, and equipment failure. Different scenarios are distinguished by the different ranges of scene feature indicator values ​​under different scenarios. Scene feature indicators include load factor, renewable energy penetration rate, ambient temperature, and fault identification. The indicator feature vector represents the frequency-voltage support capability index for the corresponding scenario.

[0052] The encoding layer extracts high-dimensional correlation features through a fully connected network, performs a non-linear mapping on the input scene feature vector and indicator feature vector, captures implicit correlations, and outputs a feature vector E=[e1, e2, ..., e 10 e 11 e 12 ].

[0053] The attention layer calculates dynamic attention weights; individual metric scores are determined using a piecewise linear function combined with the quantified metric value and a preset threshold. Taking the frequency support capability metric as an example, the response time t... response 100 points are awarded for a time limit of ≤2s, with 20 points deducted for every additional 0.5s. 100 points are awarded for a regulation capacity ratio γ ≥ 27%, 80 points for 20% ≤ γ < 27%, 60 points for 15% ≤ γ < 20%, and 40 points for γ < 15%. Similarly, for voltage support capacity, 100 points are awarded for a total harmonic distortion (THD) ≤ 5%, and 10 points are deducted for every 1% exceeding 5% THD. The specific calculation formulas involved are as follows: (25) (26) (27) in, For the first i In the scenario, the first j The matching score corresponding to each indicator measures the first... i In the scenario, the first j The correlation strength of each indicator with the "frequency-voltage stability" target; This represents a feedforward neural network; k is the k-th term in the 8-item feature vector; The input feature vector; This involves concatenating vectors. For the first i In the scenario, the first j Attention weights corresponding to each metric; For the first i In the scenario, the first j The score of each indicator (or the score of a single indicator); This refers to a feedforward neural network specifically designed for scoring.

[0054] The attention weights and scores of the calculated indicators are weighted and summed to obtain the comprehensive score for gas-fired power's ability to support the grid's frequency and voltage. i (That is, the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario), the calculation formula is as follows: (28) After the model training is complete, the current scene will be... , The input is fed into the comprehensive evaluation model of grid frequency-voltage support capability, and the output is the index score of each index in the index feature vector and the predicted value of the comprehensive score of grid frequency-voltage support capability for the current scenario. The flowchart of the corresponding analytic hierarchy process in this application is as follows. Figure 4 As shown.

[0055] The quantitative comprehensive score is output using a three-level calculation model of "indicator score - dimensional score - comprehensive score". The indicator score is the score of each individual indicator, calculated using a piecewise linear function; the dimensional score is the weighted sum of the scores of each indicator and the dynamic weights output by the attention mechanism (i.e., the scores of the frequency support capability indicator and the voltage support capability indicator); the comprehensive score is obtained by superimposing the frequency and voltage support capability scores and their scenario weights (i.e., the frequency-voltage support capability score).

[0056] Step S5: Based on the predicted value of the comprehensive score of the grid frequency-voltage support capability in the current scenario and the index scores of each index in the index feature vector, determine the evaluation level of the grid frequency-voltage support capability in the current scenario, thereby determining the evaluation level of the grid frequency-voltage support capability of the gas-fired power plant in multiple scenarios.

[0057] As an optional implementation, step S5 specifically includes: Step S51: Determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to a first preset threshold, and obtain a first judgment result. Preferably, the first preset threshold is 85.

[0058] Step S52: If the first judgment result is yes, then the evaluation level of the power grid frequency-voltage support capability in the current scenario is determined to be "excellent".

[0059] Step S53: If the first judgment result is negative, then determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to the second preset threshold, and obtain the second judgment result. Preferably, the second preset threshold is 70.

[0060] Step S54: If the second judgment result is yes, then the evaluation level of the power grid frequency-voltage support capability in the current scenario is determined to be "good".

[0061] Step S55: If the second judgment result is negative, then determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to a third preset threshold, and obtain the third judgment result. Preferably, the third preset threshold is 55.

[0062] Step S56: If the third judgment result is yes, then the assessment level of the power grid frequency-voltage support capability in the current scenario is determined to be "general".

[0063] Step S57: If the third judgment result is negative, the assessment level of the power grid frequency-voltage support capability under the current scenario is judged to be "insufficient". Then, the fault tree analysis method is used to trace the root cause based on the minimum value of the index score of each index in the index feature vector and generate a three-level optimization scheme corresponding to the current scenario.

[0064] Specifically, the evaluation levels are divided into: Score i A score of ≥85 is considered excellent; a score of 70≤Score i A score below 85 is considered good; a score of 55 or below is considered good. i A score below 70 indicates an average assessment level. i If the score is less than 55, the assessment level is insufficient.

[0065] Furthermore, in step S57, when the assessment level of the power grid frequency-voltage support capability in the current scenario is "insufficient," a fault tree analysis method is used to trace the root cause based on the minimum value of the index score of each index in the index feature vector, generating a three-level optimization scheme corresponding to the current scenario, as follows: (1) When the evaluation level is "insufficient" (i.e., the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is less than 55), based on the index scores of each index in the index feature vector, the lowest-scoring index is automatically located (such as insufficient continuous output stability in the peak load scenario), and the fault tree analysis (FTA) module is called to trace the root cause (such as the hysteresis of the compressor inlet guide vane adjustment) and generate a three-level optimization scheme: ① Level 1 optimization scheme (to be implemented immediately): Activate the guide vane servo valve gain compensation mode to increase the adjustment response speed by 15%.

[0066] ② Secondary optimization plan (to be implemented within 24 hours): Start the redundant fan of the air cooler to reduce the temperature fluctuation of the compressor inlet (controlled within ±2℃).

[0067] ③ Level 3 optimization scheme (to be implemented within 72 hours): calibrate the flow rate of the fuel nozzles to ensure that the output deviation of each burner is less than 3%.

[0068] (2) If Score i When the score is between 55 and 84, suggestions for potential improvement are given, such as optimizing the combustion chamber air distribution ratio during peak load scenarios to improve operational stability in the high-temperature section and extend the rated output maintenance time by 1.5 hours. If the score... i If the score is ≥85 (excellent), no optimization solution will be generated; only the evaluation result will be recorded.

[0069] In addition, the optimization plan needs to be implemented by connecting to a digital twin model to simulate the adjustment response under the same scenario after optimization, and ensure that the overall score is improved by ≥10 points before it can be put into the execution plan.

[0070] Overall, a three-tiered calculation model of "indicator score - dimensional score - comprehensive score" is adopted to output a quantitative comprehensive score. The dynamic attention weight is learned entirely from the actual contribution of indicators to frequency-voltage stability in historical evaluation data. Targeted optimization schemes are generated based on the comprehensive score, and the schemes must be verified by a digital twin system to ensure that the comprehensive score is improved by ≥10 points. Based on the predicted value of the comprehensive score of the grid frequency-voltage support capability in the current scenario, the support capability of gas-fired power to the grid is comprehensively evaluated, the scores of each indicator are analyzed, and targeted improvement suggestions are proposed to further enhance the support capability of gas-fired power to the grid.

[0071] Furthermore, taking the GE9HA.01 H-class gas turbine under peak load conditions as an example, 1) its corresponding frequency-voltage support capability constraints are as follows: (29) (30) (31) (32) (33) (34) Odd-order harmonics: (35) Even-order harmonics: (36) (37) 2) Real-time acquisition of multi-source data during peak load scenarios.

[0072] 3) Preprocess the multi-source data of the peak load scenario to obtain the preprocessed multi-source data of the peak load scenario.

[0073] 4) Using formulas (2)-(24) and the corresponding constraints of formulas (29)-(37), the scene feature vector and index feature vector of the load peak scenario are determined based on the preprocessed multi-source data of the load peak scenario.

[0074] 5) Input both the scenario feature vector and the indicator feature vector of the peak load scenario into the comprehensive evaluation model of the grid frequency-voltage support capability, and output the indicator score of each indicator in the indicator feature vector and the predicted value of the comprehensive score of the grid frequency-voltage support capability of the peak load scenario.

[0075] 6) Based on the predicted value of the comprehensive score of the grid frequency-voltage support capability under the peak load scenario and the index scores of each index in the index feature vector, determine the evaluation level of the grid frequency-voltage support capability under the peak load scenario.

[0076] Beneficial effects: 1) By constructing a multi-level evaluation model consisting of a target layer, a standard layer, and an indicator layer, covering multiple indicators of frequency support and voltage support, and combining basic scenarios such as peak load, off-peak load, and large-scale renewable energy generation with extreme scenarios such as cold waves, typhoons, and equipment failures, a comprehensive assessment of the integrated support capabilities of gas-fired power plants can be achieved.

[0077] 2) Introduce an attention mechanism to learn the actual contribution of each indicator in different scenarios through historical evaluation data, dynamically calculate scenario weights, replace the judgment matrix weight calculation steps that rely on expert experience in traditional AHP, and achieve a completely data-driven objective evaluation.

[0078] 3) Scene segmentation based on measured data: Collect real-time parameters and combine them with parameters such as total system load, new energy penetration rate, ambient temperature, and wind speed to achieve automatic identification of typical scenes and parameterized configuration of custom scenes, thereby improving the precision of scene segmentation.

[0079] 4) In the frequency support capability index, not only steady-state frequency deviation is considered, but also transient indicators such as response time and comprehensive regulation capability are introduced to achieve the integration of steady-state and transient indicators and ensure that the evaluation model can consider both steady-state and transient processes at the same time. In the voltage support capability index, reactive power, voltage regulation sensitivity, reactive response time and harmonic suppression capability are combined to comprehensively reflect the support capability of the gas-fired power plant and provide more comprehensive evaluation results.

[0080] 5) Adaptability to scene changes: The dynamic weight adjustment mechanism can automatically adjust the weights according to real-time data and scene changes to ensure the timeliness and accuracy of the evaluation results.

[0081] 6) Compared with traditional static evaluation methods, this application improves the evaluation accuracy to over 95%, effectively supporting frequency and voltage regulation decisions and optimization control of the power grid under complex operating conditions.

[0082] Based on the same inventive concept, this application also provides a system for quantitatively evaluating the frequency-voltage support capability of a multi-scenario gas-fired power plant, used to implement the aforementioned quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants. The solution provided by this system is similar to the implementation described in the above method. Therefore, the specific limitations in one or more embodiments of the quantitative evaluation system for the frequency-voltage support capability of multi-scenario gas-fired power plants provided below can be found in the limitations of the quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants described above, and will not be repeated here.

[0083] In one exemplary embodiment, a quantitative evaluation system for the frequency-voltage support capability of a multi-scenario gas-fired power plant is provided, comprising: The data acquisition unit is used to collect multi-source data from multiple scenarios in real time during the grid-connected operation of the gas power plant. The multi-source data includes: equipment parameters of the gas turbine, grid operating status parameters, and environmental parameters.

[0084] The preprocessing unit is used to preprocess multi-source data from multiple scenarios to obtain preprocessed multi-source data from multiple scenarios.

[0085] The feature vector determination unit is used to determine the scene feature vector and indicator feature vector of the current scene based on the preprocessed multi-source data of the current scene. The scene feature vector includes: the scene type of the current scene and the corresponding scene feature indicator value; the indicator feature vector includes: frequency support capability indicator and voltage support capability indicator.

[0086] The indicator score and comprehensive score determination unit is used to input the scene feature vector and indicator feature vector of the current scene into the comprehensive evaluation model of power grid frequency-voltage support capability, and output the indicator score of each indicator in the indicator feature vector and the predicted value of the comprehensive evaluation model of power grid frequency-voltage support capability of the current scene; the comprehensive evaluation model of power grid frequency-voltage support capability is obtained by training an attention mechanism neural network model using a training set.

[0087] The assessment level determination unit is used to determine the assessment level of the grid frequency-voltage support capability under the current scenario based on the predicted value of the comprehensive score of the grid frequency-voltage support capability under the current scenario and the index scores of each index in the index feature vector, thereby determining the assessment level of the grid frequency-voltage support capability of the gas-fired power plant under multiple scenarios.

[0088] Specifically, such as Figure 5As shown, the quantitative evaluation system for the frequency-voltage support capability of gas-fired power plants in multiple scenarios is based on the core logic of "scenario-driven, quantitative evaluation, and precise optimization," and constructs a four-layer architecture: "data layer, scenario layer, evaluation layer, and application layer." The data layer interacts in real time with the power grid SCADA system through IoT sensors, collecting support capability parameters and scenario parameters such as gas turbine inlet and outlet pressure, fuel flow, power grid frequency fluctuations, power grid voltage fluctuations, and new energy output curves. These parameters are then processed by edge computing nodes for noise filtering and normalization, forming a minute-level dynamic database. The scenario layer constructs a scenario generation engine based on the gas-electric coupling characteristics. Based on the different ranges of scenario indicator parameters under different scenarios, it achieves automated identification of typical scenarios and parameterized configuration of custom scenarios. The evaluation layer integrates the AHP quantitative model and constraint engine, outputting scenario adaptability scores. The application layer displays the evaluation results through a visualization platform and links with the PLC control system to generate executable optimization solutions.

[0089] Different gas-electric regulation scenarios are set up based on scenario indicators such as gas-electric regulation characteristics, power grid operation characteristics, and gas turbine operation characteristics. Based on the constructed quantitative evaluation model of frequency-voltage support capability, combined with the feature vectors of frequency-voltage support capability under different scenarios, a comprehensive score is obtained for each scenario. The evaluation process adopts a three-level calculation mode of "indicator score - dimensional score - comprehensive score". Based on the comprehensive score, a closed-loop optimization mechanism of "problem diagnosis - solution generation - effect prediction" is constructed. The final comprehensive score is divided into four levels from 0 to 100: "Excellent (≥85), Good (70-84), Average (55-69), and Insufficient (<55)". The comprehensive score determines the gas turbine's regulation capability in different scenarios, and a targeted optimization regulation scheme is given. The optimization scheme needs to be verified through a digital twin system to simulate the regulation effect under the same scenario and perform scenario reproduction verification to ensure that the comprehensive score improves by no less than 10 points after the implementation of the optimization scheme.

[0090] In one exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a quantitative evaluation method for the frequency-voltage support capability of a gas-fired power plant in multiple scenarios.

[0091] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 6As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a quantitative evaluation method for the frequency-voltage support capability of a multi-scenario gas-fired power plant.

[0092] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0093] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0094] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0095] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0096] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0097] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods, systems, and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A quantitative evaluation method for the frequency-voltage support capability of gas-fired power plants in multiple scenarios, characterized in that, The quantitative evaluation method for the frequency-voltage support capability of gas-fired power plants in various scenarios includes: During the grid-connected operation of the gas-fired power plant, multi-source data from multiple scenarios are collected in real time. The multi-source data includes: equipment parameters of the gas turbine, grid operation status parameters, and environmental parameters. Preprocessing of multi-source data from multiple scenarios yields preprocessed multi-source data for multiple scenarios; Based on the preprocessed multi-source data of the current scenario, the scenario feature vector and indicator feature vector of the current scenario are determined. The scenario feature vector includes: the scenario type of the current scenario and the corresponding scenario feature indicator value. The indicator feature vector includes: frequency support capability indicator and voltage support capability indicator. The scene feature vector and index feature vector of the current scene are both input into the comprehensive evaluation model of power grid frequency-voltage support capability. The output is the index score of each index in the index feature vector and the predicted value of the comprehensive score of power grid frequency-voltage support capability of the current scene. The comprehensive evaluation model of power grid frequency-voltage support capability is obtained by training an attention mechanism neural network model using a training set. Based on the predicted value of the comprehensive score of the grid frequency-voltage support capability in the current scenario and the index scores of each index in the index feature vector, the evaluation level of the grid frequency-voltage support capability in the current scenario is determined, thereby determining the evaluation level of the grid frequency-voltage support capability of the gas-fired power plant in multiple scenarios.

2. The quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants according to claim 1, characterized in that, Gas turbine equipment parameters include: active power, power setpoint, and reactive power; Grid operating parameters include: high-frequency grid frequency at the grid connection point, voltage amplitude at the grid connection point, current waveform, total system load, total output of new energy sources, and grid fault signals; Environmental parameters include: ambient temperature, wind speed, and ambient humidity.

3. The quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants according to claim 2, characterized in that, Preprocessing of multi-source data from multiple scenarios yields preprocessed multi-source data for multiple scenarios, specifically including: The random forest algorithm is used to remove outliers from multi-source data of multiple scenarios, thus obtaining effective multi-source data for multiple scenarios; Normalize the effective multi-source data from multiple scenarios to obtain preprocessed multi-source data from multiple scenarios.

4. The quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants according to claim 3, characterized in that, The current scenario is classified as either a basic scenario or an extreme scenario. Basic scenarios include peak load scenarios, off-peak load scenarios, and scenarios involving large-scale renewable energy generation. Extreme scenarios include cold wave scenarios, typhoon season scenarios, and equipment failure scenarios. The scenario characteristic indicators include: total system load, new energy penetration rate, ambient temperature, and fault indicators.

5. The quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants according to claim 4, characterized in that, Frequency support capability indicators include: regulation capacity ratio, response time, steady-state frequency deviation, and comprehensive regulation capability index; Voltage support capability indicators include: reactive power, voltage regulation sensitivity, reactive response time, and harmonic suppression capability.

6. The quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants according to claim 5, characterized in that, Based on the preprocessed multi-source data of the current scene, determine the scene feature vector and indicator feature vector of the current scene, specifically including: Determine whether the values ​​of multiple scene feature indicators in the preprocessed multi-source data of the current scene meet the corresponding preset range, determine the scene type to which the current scene belongs, and use the scene type to which the previous scene belongs and the corresponding scene feature indicator values ​​as the scene feature vector of the current scene. Based on the preprocessed multi-source data of the current scenario, the corresponding frequency support capability index and voltage support capability index are calculated to obtain the index feature vector corresponding to the current scenario.

7. The quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants according to claim 6, characterized in that, The training process of the comprehensive evaluation model for power grid frequency-voltage support capability specifically includes: Construct a training set; the training set includes: scene feature vectors and indicator feature vectors corresponding to multi-source data of historical scenarios, the true values ​​of the comprehensive score of power grid frequency-voltage support capability marked by manual annotation, and the indicator scores of each indicator in the indicator feature vector marked by manual annotation; Build a neural network model for the attention mechanism; Using the scene feature vector and indicator feature vector corresponding to multi-source data of historical scenarios as input, and the true value of the comprehensive score of power grid frequency-voltage support capability marked by manual annotation and the indicator scores of each indicator in the indicator feature vector marked by manual annotation as output, the attention mechanism neural network model is trained until the loss function reaches the minimum value or the training rounds reach the maximum value, and then training is stopped to obtain the comprehensive evaluation model of power grid frequency-voltage support capability.

8. The quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants according to claim 7, characterized in that, Based on the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario and the index scores of each index in the index feature vector, the assessment level of the power grid frequency-voltage support capability in the current scenario is determined, specifically including: Determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to the first preset threshold, and obtain the first judgment result; If the first judgment result is yes, then the evaluation level of the power grid frequency-voltage support capability in the current scenario is determined to be "excellent"; If the first judgment result is negative, then determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to the second preset threshold, and obtain the second judgment result. If the second judgment result is yes, then the assessment level of the power grid frequency-voltage support capability in the current scenario is determined to be "good"; If the second judgment result is negative, then determine whether the predicted value of the comprehensive score of the power grid frequency-voltage support capability in the current scenario is greater than or equal to the third preset threshold, and obtain the third judgment result. If the third judgment result is yes, then the assessment level of the power grid frequency-voltage support capability in the current scenario is determined to be "average". If the third judgment result is negative, the assessment level of the power grid frequency-voltage support capability in the current scenario is judged to be "insufficient". Then, the fault tree analysis method is used to trace the root cause based on the minimum value of the index score of each index in the index feature vector, and generate a three-level optimization scheme corresponding to the current scenario.

9. A quantitative evaluation system for the frequency-voltage support capability of gas-fired power plants in multiple scenarios, characterized in that, The quantitative evaluation system for the frequency-voltage support capability of multi-scenario gas-fired power plants is used to implement the quantitative evaluation method for the frequency-voltage support capability of multi-scenario gas-fired power plants according to any one of claims 1-8. The quantitative evaluation system for the frequency-voltage support capability of multi-scenario gas-fired power plants includes: The data acquisition unit is used to collect multi-source data from multiple scenarios in real time during the grid-connected operation of the gas-fired power plant. The multi-source data includes: equipment parameters of the gas turbine, grid operating status parameters, and environmental parameters. The preprocessing unit is used to preprocess multi-source data from multiple scenarios to obtain preprocessed multi-source data from multiple scenarios. The feature vector determination unit is used to determine the scene feature vector and indicator feature vector of the current scene based on the preprocessed multi-source data of the current scene. The scene feature vector includes: the scene type of the current scene and the corresponding scene feature indicator value; the indicator feature vector includes: frequency support capability indicator and voltage support capability indicator. The indicator score and comprehensive score determination unit is used to input both the scene feature vector and indicator feature vector of the current scene into the comprehensive evaluation model of power grid frequency-voltage support capability, and output the indicator score of each indicator in the indicator feature vector and the predicted value of the comprehensive evaluation model of power grid frequency-voltage support capability of the current scene; the comprehensive evaluation model of power grid frequency-voltage support capability is obtained by training an attention mechanism neural network model using a training set; The assessment level determination unit is used to determine the assessment level of the grid frequency-voltage support capability under the current scenario based on the predicted value of the comprehensive score of the grid frequency-voltage support capability under the current scenario and the index scores of each index in the index feature vector, thereby determining the assessment level of the grid frequency-voltage support capability of the gas-fired power plant under multiple scenarios.

10. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the quantitative evaluation method for the frequency-voltage support capability of a multi-scenario gas-fired power plant as described in any one of claims 1-8.

Images