A new energy scheduling intelligent monitoring analysis and control method

Abstract

The application belongs to the technical field of new energy dispatching, and specifically discloses a new energy dispatching intelligent monitoring analysis and regulation method, which comprises the following steps: realizing intelligent regulation of a new energy power station by constructing a power station equipment dispatching knowledge graph comprising a fault rule layer, an entity relationship layer and a dynamic weight layer, the core steps being: generating a fault rule by using a multi-parameter fusion technology, calculating a meteorological-output correlation weight by using a random forest algorithm, determining a dispatching priority coefficient based on an analytic hierarchy process to construct the knowledge graph, collecting electrical, meteorological and equipment state parameters by a sensor network, matching and judging the parameters with the fault rule to determine an abnormality, locating a fault source, generating a regulation instruction, and evaluating an effect after executing the regulation instruction; and the application realizes deep fusion and intelligent reasoning of multi-dimensional data by constructing a new energy dispatching knowledge graph comprising a fault rule layer, an entity relationship layer and a dynamic weight layer, thereby improving the diagnosis efficiency of complex faults.

Description

Technical Field

[0001] This invention belongs to the field of new energy dispatching technology and relates to a method for intelligent monitoring, analysis and control of new energy dispatching. Background Technology

[0002] New energy refers to non-fossil energy sources, represented by renewable energy such as solar, wind, and hydropower, and clean energy such as hydrogen and biomass energy. These sources offer advantages such as low carbon emissions, environmental friendliness, and resource sustainability. New energy dispatch refers to the real-time monitoring, optimized control, and dynamic adjustment of new energy power generation through intelligent technologies to address the grid stability challenges posed by their volatility and intermittency. The significant volatility and intermittency of new energy power generation place higher demands on the grid's real-time monitoring, dynamic adjustment, and fault response capabilities. Traditional dispatch methods often rely on economic benefit optimization or simple closed-loop control, which struggles to address the multi-dimensional parameter coupling problems under complex operating environments.

[0003] For example, Chinese invention patent CN117977712A discloses a scheduling method and scheduling device for new energy power plants. Its core is to generate scheduling strategies through predictive information (such as electricity price) and real-time data (such as power). However, this technology has the following limitations: (1) Functional limitation: It focuses on economic benefit optimization and grid stability, does not cover the complex fault diagnosis scenarios of new energy power plants, lacks the ability to dynamically adjust to abnormal operating conditions, and does not introduce multi-parameter fusion technology, so it cannot build a power plant equipment scheduling knowledge graph that supports intelligent reasoning, resulting in low fault location efficiency and delayed regulation.

[0004] For example, Chinese invention patent CN119093361A discloses a smart microgrid new energy centralized dispatching system, which realizes dispatching control through data acquisition, processing and economic optimization. However, its technical solution has the following defects: (2) Insufficient dynamic adaptability: It does not design a dynamic weighting mechanism, and cannot respond in real time to the impact of environmental parameter changes on power generation output. At the same time, the anomaly detection accuracy is low, it relies on traditional data processing methods, and does not build a fault rule layer and a threshold matching mechanism for real-time data, which cannot guarantee the accuracy of anomaly judgment, resulting in a high false alarm rate and limited effectiveness of the control strategy. Summary of the Invention

[0005] In view of this, in order to solve the problems mentioned in the background technology, a new energy dispatch intelligent monitoring, analysis and control method is proposed.

[0006] The objective of this invention can be achieved through the following technical solution: This invention provides a method for intelligent monitoring, analysis and control of new energy dispatch, including: S1, constructing a knowledge graph of power plant equipment dispatch based on historical fault data of new energy power plants, which includes a fault rule layer, an entity relationship layer and a dynamic weight layer.

[0007] S2. Real-time collection of electrical parameters, meteorological parameters and equipment status parameters of the new energy power station through sensor network technology, input of the parameters into the fault rule layer and matching with the preset fault rule condition threshold to generate anomaly judgment results.

[0008] S3. When the anomaly judgment result is an anomaly, the fault source is located based on the entity relationship layer and dynamic weight layer in the power plant equipment scheduling knowledge graph, and a control instruction containing the control strategy is generated.

[0009] S4. Execute control commands and collect operational response data after control through sensor networks to evaluate the control effect.

[0010] S5. If the control effect is satisfactory, the process ends; otherwise, S2 is executed and the control command is regenerated.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention constructs a power plant equipment scheduling knowledge graph that includes a fault rule layer, an entity relationship layer and a dynamic weight layer, thereby realizing the deep fusion and intelligent reasoning of multi-dimensional data and improving the diagnostic efficiency of complex faults.

[0012] (2) By introducing the meteorological-output correlation weight and scheduling priority coefficient, this invention quantifies the impact of the environment and equipment status on output in real time, ensuring the dynamic adaptability of the control strategy, and thus improving the control response speed of new energy power plants in a variable environment.

[0013] (3) After executing the control command, the present invention collects response data in real time and compares it with the qualified operating range. If the limit is exceeded, the strategy is automatically regenerated to form a "monitoring-control-evaluation" closed loop, realizing the synergistic optimization of anomaly detection accuracy and control efficiency, and reducing operation and maintenance costs. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram showing the connections between the steps of the method of the present invention.

[0016] Figure 2 This is a schematic diagram showing the connection steps of constructing the power plant equipment scheduling knowledge graph of the present invention.

[0017] Figure 3 This is a schematic diagram showing the connection steps for generating the anomaly judgment result of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 As shown, the present invention provides a new energy dispatch intelligent monitoring, analysis and control method, which includes: S1, constructing a power plant equipment dispatch knowledge graph based on historical fault data of new energy power plants, which includes a fault rule layer, an entity relationship layer and a dynamic weight layer.

[0020] Please see Figure 2 As shown, for example, the construction of the power plant equipment scheduling knowledge graph includes: S1.1, extracting environmental parameters, equipment status, performance indicators and time series of each fault of the new energy power plant from the historical fault data of the new energy power plant, as well as fault handling methods, and then generating the fault rule layer of the new energy power plant through multi-parameter fusion technology.

[0021] It should be added that environmental parameters include, but are not limited to: temperature, relative humidity, irradiance, wind speed, wind direction, air pressure, precipitation intensity, and dust concentration.

[0022] Equipment status includes, but is not limited to: power generation equipment: surface temperature of photovoltaic modules, open-circuit voltage, short-circuit current, impeller speed of wind turbine generators, gearbox oil temperature, vibration acceleration; battery packs of energy storage equipment: cell voltage, remaining capacity, charge / discharge rate, supercapacitor, equivalent series resistance, self-discharge rate; inverters of power electronic equipment: internal temperature, switching frequency, harmonic distortion rate; converters: DC bus voltage, power factor.

[0023] Performance indicators include, but are not limited to, power generation performance: actual power generation, power generation efficiency, and curtailment rate; equipment health: component degradation rate, bearing remaining life, and insulation resistance; and grid adaptability: voltage deviation, frequency deviation, and reactive power regulation capability.

[0024] Time series parameters include, but are not limited to, absolute time: the time of failure, the equipment running time; relative time: season (spring, summer, autumn, winter), time period (peak, peak, flat period, trough); meteorological cycle: daily radiation cycle (sunrise and sunset time), wind speed fluctuation cycle (such as the nighttime wind speed change of offshore wind power stations); event time: the time of the last maintenance, the fault interval.

[0025] The fault handling methods include, but are not limited to: 1. Electrical overload fault handling, (1) Equipment restart: technical means, restarting power electronic equipment such as inverters and circuit breakers through remote control, implementation parameters: restart interval time ≥ 30 seconds (to prevent frequent start and stop); (2) Parameter adjustment: voltage adjustment, reducing the inverter output voltage by 5% to 10% (e.g., from 315V to 290V), power derating: gradually reducing active power according to the preset gradient (e.g., derating by 10% each time); (3) Redundancy switching, equipment switching: automatically switching to the backup inverter or energy storage system, line switching: switching to the redundant transmission line through the intelligent switch.

[0026] 2. Handling mechanical wear-related faults: (1) Lubrication system startup: Technical parameters: Inject lubricating oil into the gearbox (pressure ≥ 0.5 MPa, flow rate 2 L / min); (2) Vibration suppression: Active damping control: Adjust the pitch angle of the wind turbine (e.g., increase the damping angle by 3°), passive vibration isolation: Trigger the hydraulic vibration isolation system (response time ≤ 50 ms); (3) Planned maintenance: Work order generation: Automatically generate bearing replacement work orders (remaining life < 1000 hours), spare parts scheduling: Trigger spare parts inventory warning through the Internet of Things system (e.g., bearing inventory < 3 pieces).

[0027] 3. Handling of environmental anomalies, (1) Component cleaning: Cleaning method: Start the water jet cleaning system (water pressure 50MPa, flow rate 15L / min), intelligent scheduling: according to the dust concentration (>500μg / m 3 (1) Automatically trigger cleaning; (2) Shadow avoidance: bracket adjustment: adjust the angle of the photovoltaic tracking bracket (accuracy ±0.5°), path planning: optimize the inspection path of the drone to identify obstructions; (3) Temperature control: heat dissipation system start-up: turn on the redundant fan (speed increased to 2000rpm), phase change material application: trigger the heat dissipation of the phase change material of the energy storage system (melting point 45℃), the above processing means correspond to the control instructions containing the control strategy.

[0028] Furthermore, the generation of the fault rule layer of the new energy power station through multi-parameter fusion technology includes: S1.1.1, based on the equipment status of the new energy power station at each fault, the fault type is divided into three categories: electrical overload, mechanical wear and environmental abnormality.

[0029] S1.1.2 Calculate the correlation of each parameter using the Pearson correlation coefficient, and select parameters whose correlation is greater than or equal to the set reference correlation as the set of fault-related parameters.

[0030] It should be added that the Pearson correlation coefficient, usually denoted by r, is a statistic used to measure the degree of linear correlation between two variables. For two variables X and Y, assuming they have b sets of observations (x1, y1), (x2, y2), ..., (x... a ,ya The calculation formula is as follows: in, It is the mean of variable X. is the mean of variable Y, and a is the observation number, a = 1, 2, ..., b.

[0031] In multi-parameter fusion technology, the Pearson correlation coefficient can be used to screen parameters relevant to fault or scheduling decisions. By setting a correlation threshold (e.g., |r|≥0.6), when the Pearson correlation coefficient between two parameters meets this threshold, it is considered that there is a strong linear correlation between the two parameters, and they are incorporated into subsequent analysis and decision-making processes. For example, if the calculated Pearson correlation coefficient r = 0.9 between ambient temperature and inverter temperature, it indicates that these two parameters are highly positively correlated, and ambient temperature is an important reference parameter in inverter fault diagnosis.

[0032] In one specific embodiment, assuming that in a photovoltaic power station, the ambient temperature X (unit: °C) and inverter temperature Y (unit: °C) data for 5 consecutive days were recorded as shown in Table 1:

[0033] Table 1: Correlation data between ambient temperature and inverter temperature.

[0034]

[0035]

[0036] First, calculate the mean value of the ambient temperature X. Mean value of inverter temperature Y Then calculate the numerator of the above formula: And calculate the denominator of the above formula: Then Pearson correlation coefficient This indicates a completely positive linear correlation between ambient temperature and inverter temperature.

[0037] S1.1.3. Generate a fault rule layer by associating fault types, fault-related parameter sets, and fault handling methods.

[0038] S1.2. Based on the historical fault data of new energy power plants, analyze the connection relationships of power generation equipment, energy storage equipment, power transmission equipment, meteorological elements and power grid nodes, and then construct the entity relationship layer of new energy power plants.

[0039] Furthermore, the entity relationship layer for constructing the new energy power station includes: S1.2.1, taking the power generation equipment, energy storage equipment and power transmission equipment of the new energy power station as operating entities, taking the meteorological elements and meteorological areas of the new energy power station as meteorological entities, and taking the dispatching instructions and grid nodes of the new energy power station as dispatching entities.

[0040] It should be added that power generation equipment includes, but is not limited to: wind turbines, photovoltaic modules, hydroelectric generators, and biomass generators; energy storage equipment includes, but is not limited to: battery energy storage systems, which have attributes such as battery type, capacity, and charge / discharge status; power transmission equipment includes, but is not limited to: transformers, transmission lines, switchgear, and inverters; meteorological elements include, but are not limited to: wind speed, wind direction, light intensity, temperature, and humidity; meteorological regions include, but are not limited to: geographical coordinates, altitude, area, topography, and climate type; dispatching instructions include, but are not limited to: active power instructions, reactive power instructions, power generation plan instructions, equipment start / stop instructions, and voltage adjustment instructions; and grid nodes include, but are not limited to: voltage amplitude, voltage phase, frequency, injected power, and short-circuit capacity.

[0041] S1.2.2 Based on the data association analysis method, the connection relationship between the operation entity, meteorological entity and scheduling entity is obtained, and then the entity relationship layer of the new energy power station is obtained.

[0042] It should be added that data association analysis is a technique used to discover potential relationships between variables in a dataset, with its core objective being to identify frequently co-occurring patterns or rules. In new energy power plants, this algorithm constructs a multi-dimensional association network by analyzing data interactions between operating entities (power generation equipment, energy storage equipment, power transmission equipment), meteorological entities (irradiance, temperature, wind speed), and scheduling entities (scheduling instructions, grid nodes), thereby forming an entity relationship layer.

[0043] S1.3. Based on historical fault data of new energy power plants, establish a dynamic weight layer for new energy power plants, which includes meteorological-output correlation degree and dispatch priority coefficient.

[0044] Furthermore, the establishment of the dynamic weight layer for the new energy power station includes: S1.3.1, extracting the irradiance-power curve from historical fault data and calculating the meteorological-output correlation weight through a random forest regression model.

[0045] It should be added that the formula for the random forest regression model is: Where η is the meteorological-output correlation weight, σ i Let f be the weight of the i-th tree. iThe predicted value for the i-th tree is used because the relationship between meteorological parameters and power output is complex (such as the irradiance saturation effect), and traditional linear regression cannot fit the nonlinear curve. The formula's function is to generate dynamic meteorological-power output correlation weights.

[0046] In one specific embodiment, if the tree depth is set to 5, to prevent excessive depth from causing noise fitting, 70% of the features are randomly selected for training. The specific data is shown in Table 2.

[0047] Table 2: Irradiance-Power Prediction Table for Random Forest Regression Model

[0048] <![CDATA[Irradiance (W / m 2 )]]> Actual power (kW) Predicted power (kW) 800 150 148 1000 180 178 1200 195 192

[0049] Note: The training data contains 70% features (irradiance, temperature, humidity), and the root mean square error (RMSE) of the validation set is 2.3kW.

[0050] This table is used to verify the predictive effect of the random forest regression model on the power output of photovoltaic power plants. It shows the comparison between the actual power and the model-predicted power under different irradiance conditions, reflecting the accuracy and applicability of the model.

[0051] S1.3.2 Extract grid security parameters, equipment status parameters, economic benefit parameters, and environmental constraint parameters of new energy power plants from historical fault data, and determine the dispatch priority coefficient of new energy power plants through the analytic hierarchy process.

[0052] It should be added that the power grid safety parameters include: real-time operation indicators and regulation capability indicators. Real-time operation indicators include: frequency deviation: the difference between the current grid frequency and the rated value (e.g., ±0.2Hz); voltage over-limit risk: the probability that the node voltage exceeds ±10% of the rated value; harmonic distortion rate: the total harmonic distortion rate of the inverter output current / voltage (e.g., >3% alarm). Regulation capability indicators include: frequency regulation response speed: the delay time from the issuance of the command to the achievement of power adjustment target (e.g., ≤2 seconds); reactive power regulation margin: the available dynamic reactive power capacity (e.g., ±10MVar); black start capability: the time to restore power supply after system failure (e.g., ≤30 minutes).

[0053] Equipment status parameters include: health indicators and operating load indicators. Health indicators include: Photovoltaic modules: efficiency degradation rate: the percentage decrease in efficiency compared to the initial efficiency (e.g., annual degradation > 0.5% warning); hot spot temperature gradient: the temperature difference on the module surface (e.g., > 5℃ triggers protection); Wind turbines: gearbox vibration amplitude: the effective value of acceleration (e.g., > 7.1 mm / s²). 2Alarm), blade crack index: damage quantification value based on acoustic emission detection, energy storage system: health status: percentage of remaining capacity (e.g., <80% power limit), internal resistance change rate: average daily increase in single-cell internal resistance (e.g., >0.1mΩ / day warning); operating load indicators include: equipment utilization rate: ratio of actual output to rated capacity (e.g., >90% triggers load reduction), cumulative losses: number of switching / temperature cycles of key components (e.g., converter IGBT).

[0054] Economic benefit parameters include: market revenue indicators and cost control indicators. Market revenue indicators include: spot electricity price sensitivity: revenue per unit output (RMB / MWh), ancillary service revenue: frequency regulation mileage pricing (RMB / MW·time), reserve capacity compensation (RMB / MW·h), and carbon emission rights revenue: carbon emission reduction per unit of electricity generated (tCO2 / MWh). Cost control indicators include: equipment loss cost: the impact of charge-discharge cycles on energy storage life (RMB / cycle), operation and maintenance cost: expected costs of fault repair and preventive maintenance, and wind and solar curtailment loss: energy loss due to power curtailment (MWh / day).

[0055] Environmental constraints include meteorological condition indicators and geographical environment indicators. Meteorological condition indicators include: irradiance fluctuation rate: the range of irradiance change within 5 minutes (e.g., >200W / m²). 2 Triggering adjustments), wind speed prediction error: the deviation between actual wind speed and predicted value (e.g., >3m / s, model correction), extreme weather warning: the probability and intensity level of events such as typhoons and sandstorms; geographical environmental indicators include: topographic shading effect: the shadow coverage of mountains and buildings on the photovoltaic array, atmospheric transmittance: the radiation attenuation coefficient caused by haze and dust (e.g., <0.85, cleaning is required), salt spray corrosion index: the salt deposition rate in coastal areas (mg / m³). 2 ·day).

[0056] Furthermore, the step of determining the scheduling priority coefficient of the new energy power station through the analytic hierarchy process includes: S1.3.2.1, setting the target layer as the layer for determining the scheduling priority coefficient of the new energy power station.

[0057] S1.3.2.2, take the grid security, equipment status, economic benefits and environmental constraints of new energy power plants as the criteria of the criterion layer.

[0058] S1.3.2.3 Construct sub-indicator layers based on the specific parameters of each criterion.

[0059] In one specific embodiment, the sub-indicator layer includes: voltage stability and frequency deviation for grid security; equipment temperature and failure rate for equipment status; power generation cost and electricity sales revenue for economic benefits; and carbon emissions and land use efficiency for environmental constraints.

[0060] S1.3.2.4 Based on historical fault data of new energy power plants, construct judgment matrices for each criterion layer, and calculate the weight of each criterion using the eigenvector method.

[0061] It should be added that the construction of the judgment matrix for each criterion layer includes: denoting power grid security, equipment status, economic benefits, and environmental constraints as C1, C2, C3, and C4 respectively, and constructing the criterion layer judgment matrix as shown in Table 3.

[0062] Table 3: Example diagram of the criterion-level judgment matrix.

[0063] C1 C2 C3 C4 C1 1 3 5 7 C2 1 / 3 1 3 5 C3 1 / 5 1 / 3 1 3 C4 1 / 7 1 / 5 1 / 3 1

[0064] Matrix element interpretation: Row C1: Importance of grid security to other criteria, C1 / C2=3: Grid security is slightly more important than equipment condition, C1 / C3=5: Grid security is significantly more important than economic benefits, C1 / C4=7: Grid security is strongly more important than environmental constraints; Row C2: Importance of equipment condition to other criteria, C2 / C3=3: Equipment condition is slightly more important than economic benefits, C2 / C4=5: Equipment condition is significantly more important than environmental constraints; Row C3: Importance of economic benefits to environmental constraints, C3 / C4=3: Economic benefits are slightly more important than environmental constraints. Weights are calculated using the eigenvector method. For example, grid security (C1) weight 0.637, equipment condition (C2) weight 0.258, economic benefits (C3) weight 0.105, and environmental constraints (C4) weight 0.005.

[0065] S1.3.2.5 Construct a judgment matrix for each indicator under each criterion, and then calculate the weight of each sub-indicator corresponding to each criterion layer using the eigenvector method.

[0066] S1.3.2.6. Perform minimum-maximum normalization on the parameter values ​​in the historical fault data, and sum the weights of each sub-indicator with the normalized values ​​to obtain the scores of each criterion layer.

[0067] S1.3.2.7. The scheduling priority coefficient of the new energy power plant is obtained by weighting and summing the scores of each criterion layer with their respective weights.

[0068] S1.3.3. The meteorological-output correlation coefficient and the dispatch priority coefficient are used as the dynamic weighting layer of the new energy power plant.

[0069] This invention, by introducing a meteorological-output correlation weight and a scheduling priority coefficient, quantifies the impact of environmental and equipment status on output in real time, ensuring the dynamic adaptability of the control strategy and thereby improving the control response speed of new energy power plants in variable environments.

[0070] S1.4. The fault rule layer, entity relationship layer and dynamic weight layer of the new energy power station are used as the knowledge graph for power station equipment scheduling.

[0071] This invention, through the construction of a power plant equipment scheduling knowledge graph comprising a fault rule layer, an entity relationship layer, and a dynamic weight layer, achieves deep fusion and intelligent reasoning of multi-dimensional data, thereby improving the diagnostic efficiency of complex faults.

[0072] S2. Real-time collection of electrical parameters, meteorological parameters and equipment status parameters of the new energy power station through sensor network technology, input of the parameters into the fault rule layer and matching with the preset fault rule condition threshold to generate anomaly judgment results.

[0073] It should be added that the electrical parameters of a new energy power plant include voltage, current, power, and frequency. These parameters are typically monitored in a new energy power plant by sensors or instruments installed on power generation equipment (such as photovoltaic inverters and wind turbines), energy storage equipment (such as battery packs), and power transmission lines. For example, photovoltaic inverters have built-in sensors that monitor the voltage and current of the DC input and AC output, while current transformers and voltage transformers on power transmission lines monitor the current and voltage in the lines.

[0074] Meteorological parameters include ambient temperature, humidity, irradiance, wind speed, and wind direction. These parameters are obtained by meteorological stations or sensors installed around the power station. For example, irradiance sensors are typically installed in unobstructed locations to accurately measure solar radiation intensity, while wind speed and direction sensors are installed at higher locations, such as the top of a tower or on a support structure, to obtain accurate meteorological data.

[0075] Equipment status parameters include the operating status of power generation equipment, energy storage equipment, and power transmission equipment. For example, the temperature and fan speed of photovoltaic modules, the vibration and oil temperature of wind turbines, and the individual voltage and temperature of battery cells. These parameters are acquired through specialized sensors installed on the equipment. For instance, temperature sensors are installed on the back of the modules or inside the equipment, vibration sensors are installed near rotating parts, and voltage sensors are installed on the individual battery cells.

[0076] Please see Figure 3 As shown, exemplarily, the generation of anomaly judgment results includes: S2.1, matching the electrical parameters, meteorological parameters and equipment status parameters of the new energy power station with the fault rules of the corresponding fault rule layer.

[0077] S2.2. Through the query and reasoning functions of the power station equipment scheduling knowledge graph, determine whether the conditions of a certain fault rule are met. If they are met, trigger the corresponding abnormal judgment result of the new energy power station. If none of them are met, determine that the abnormal judgment result of the new energy power station is no fault.

[0078] S3. When the anomaly judgment result is an anomaly, the fault source is located based on the entity relationship layer and dynamic weight layer in the power plant equipment scheduling knowledge graph, and a control instruction containing the control strategy is generated.

[0079] For example, the step of locating the fault source and generating a control instruction containing a control strategy includes: S3.1, using the association relationship between entities in the power plant equipment scheduling knowledge graph to locate the fault and obtain each fault source of the new energy power plant.

[0080] Furthermore, the method of using the relationships between entities in the power plant equipment scheduling knowledge graph to locate faults includes: S3.1.1, matching and comparing each abnormal parameter of the new energy power plant with the abnormal parameters corresponding to each abnormal node to obtain each abnormal node of the new energy power plant.

[0081] S3.1.2 Starting from the abnormal node, the knowledge graph of power plant equipment scheduling is traversed through the depth-first search algorithm to obtain the associated nodes of each abnormal node in the new energy power plant.

[0082] S3.1.3, The meteorological-output correlation degree and scheduling priority coefficient of each abnormal node in the new energy power station corresponding to each associated node are respectively denoted as φ. jq and j is the abnormal node number, j = 1, 2, ..., n, and q is the associated node number, q = 1, 2, ..., p.

[0083] S3.1.4, According to the formula Calculate the dynamic weight values ​​of each abnormal node and its associated nodes in the new energy power plant. α1 and α2 are the weights of the meteorological-output correlation degree and the scheduling priority coefficient, respectively, with α1+α2=1 and α1>α2.

[0084] It's important to add that the weather-output correlation directly impacts power generation efficiency: meteorological parameters (such as irradiance, temperature, and wind speed) directly determine the power output capacity of new energy power plants. For example, the power output of a photovoltaic power plant is highly positively correlated with irradiance, while excessively high temperatures may lead to a decrease in inverter efficiency. The weather-output correlation weight quantifies the real-time impact of meteorological conditions on power output, and its weight should be prioritized to ensure rapid response to environmental changes. For example, if the irradiance increases from 1000 W / m² during a certain period... 2 The power dropped sharply to 500W / m 2When the weather-output correlation weight is high, the system will prioritize adjusting energy storage output to compensate for power gaps and avoid grid frequency deviations. The dispatch priority coefficient is a guarantee for grid security and equipment health. Taking into account factors such as grid urgency and remaining equipment lifespan, its second-highest weight can optimize grid operation when the weather is stable. For example, in the absence of extreme weather, increasing the dispatch priority weight can prioritize handling equipment overload or voltage exceeding limits. Therefore, α1 is set to α2. For ease of analysis, α1 can be specifically set to 0.6, and α2 can be specifically set to 0.4.

[0085] S3.1.5 Sum the dynamic weight values ​​of each abnormal node in the new energy power station corresponding to each associated node to obtain the total dynamic weight value of each abnormal node corresponding to each associated path.

[0086] S3.1.6 Sort the total dynamic weight values ​​of each abnormal node and its associated paths from largest to smallest, and extract the associated paths of the abnormal nodes with the first-ranked values ​​as the fault paths of each abnormal node in the new energy power plant.

[0087] S3.1.7 Based on the fault paths of each abnormal node in the new energy power station, extract the fault sources corresponding to the fault paths of each abnormal node in the new energy power station from the historical fault data of the new energy power station, and use them as the fault sources of the new energy power station.

[0088] S3.2 Extract reference values ​​for each fault source in the new energy power station from the power station equipment scheduling knowledge graph, and then calculate the difference between the actual value and the reference value of each fault source in the new energy power station. The ratio of the absolute value of the difference to the reference value is used as the influence degree of each fault source in the new energy power station.

[0089] S3.3 Based on the fault sources and their impact in the new energy power plant, search for historical cases similar to the fault sources and their impact in the power plant equipment scheduling knowledge graph, and then extract the control rules for each historical case.

[0090] S3.4 Based on the control rules of each historical case, generate corresponding control instructions for each abnormal node of the new energy power station.

[0091] S4. Execute control commands and collect operational response data after control through sensor networks to evaluate the control effect.

[0092] For example, the evaluation of the control effect of the new energy power plant includes: S4.1, extracting the operation data of the new energy power plant after executing the control command from the control command response data of the new energy power plant.

[0093] S4.2 Compare the running data with the set reference running data baseline range.

[0094] S4.3 If all operating data are within the set reference range, the control effect of the new energy power station is deemed qualified; otherwise, the control effect of the new energy power station is deemed unqualified.

[0095] In this embodiment of the invention, after executing the control command, response data is collected in real time and compared with the qualified operating range. If the limit is exceeded, the strategy is automatically regenerated, forming a closed loop of "monitoring-control-evaluation". This achieves synergistic optimization of anomaly detection accuracy and control efficiency, and reduces operation and maintenance costs.

[0096] S5. If the control effect is satisfactory, the process ends; otherwise, S2 is executed and the control command is regenerated.

[0097] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for intelligent monitoring, analysis, and control of new energy dispatch, characterized in that: The method includes: S1. Construct a knowledge graph for power plant equipment scheduling based on historical fault data of new energy power plants, which includes a fault rule layer, an entity relationship layer and a dynamic weight layer; S2. Real-time collection of electrical parameters, meteorological parameters and equipment status parameters of new energy power plants through sensor network technology, input of parameters into the fault rule layer and matching with preset fault rule condition thresholds to generate anomaly judgment results; S3. When the anomaly judgment result is an anomaly, the fault source is located based on the entity relationship layer and dynamic weight layer in the power plant equipment scheduling knowledge graph, and a control instruction containing the control strategy is generated. S4. Execute the control command and collect the operation response data after control through the sensor network, and then evaluate the control effect; S5. If the control effect is satisfactory, the process ends; otherwise, S2 is executed and the control command is regenerated. Establish a dynamic weighting layer for new energy power plants, including: extracting features such as irradiance-power curves from historical fault data and calculating the meteorological-output correlation weights using a random forest regression model; extracting grid security parameters, equipment status parameters, economic benefit parameters, and environmental constraint parameters of new energy power plants from historical fault data and determining the dispatch priority coefficients of new energy power plants using the analytic hierarchy process; and using the meteorological-output correlation and dispatch priority coefficients as the dynamic weighting layer for new energy power plants. The method of determining the scheduling priority coefficient of new energy power plants using the analytic hierarchy process includes: setting the target layer as the layer for determining the scheduling priority coefficient of new energy power plants; using grid security, equipment status, economic benefits, and environmental constraints of new energy power plants as criteria for the criterion layer; constructing sub-index layers based on the specific parameters of each criterion; constructing judgment matrices for each criterion layer based on historical fault data of new energy power plants, and calculating the weights of each criterion using the eigenvector method; constructing judgment matrices for each index under each criterion, and then calculating the weights of each sub-index corresponding to each criterion layer using the eigenvector method; performing minimum-maximum normalization on the parameter values ​​in the historical fault data, and weighting and summing the weights of each sub-index with the normalized values ​​to obtain the scores of each criterion layer; and weighting and summing the scores of each criterion layer with their weights to obtain the scheduling priority coefficient of the new energy power plants.

2. The intelligent monitoring, analysis, and control method for new energy dispatching according to claim 1, characterized in that: The construction of the power plant equipment scheduling knowledge graph includes: The environmental parameters, equipment status, performance indicators, time series, and fault handling methods of each fault in the new energy power plant are extracted from the historical fault data of the new energy power plant. Then, the fault rule layer of the new energy power plant is generated through multi-parameter fusion technology. Based on the historical fault data of new energy power plants, we analyze the connection relationships of power generation equipment, energy storage equipment, power transmission equipment, meteorological elements and power grid nodes, and then construct the entity relationship layer of new energy power plants. Based on historical fault data of new energy power plants, a dynamic weight layer for new energy power plants is established, which includes meteorological-output correlation degree and dispatch priority coefficient. The fault rule layer, entity relationship layer, and dynamic weight layer of the new energy power plant are used as the knowledge graph for power plant equipment scheduling.

3. The intelligent monitoring, analysis, and control method for new energy dispatching according to claim 2, characterized in that: The fault rule layer for generating new energy power plants through multi-parameter fusion technology includes: Based on the equipment status during each fault of the new energy power plant, the fault types are divided into three categories: electrical overload, mechanical wear and tear and environmental anomaly. The correlation of each parameter is calculated using the Pearson correlation coefficient, and parameters with a correlation greater than or equal to the set reference correlation are selected as the set of fault-related parameters. The fault rule layer is generated by associating fault types, fault-related parameter sets, and fault handling methods.

4. The intelligent monitoring, analysis and control method for new energy dispatching according to claim 2, characterized in that: The entity relationship layer for constructing the new energy power station includes: The power generation equipment, energy storage equipment, and power transmission equipment of the new energy power station are taken as the operating entities, the meteorological elements and meteorological areas of the new energy power station are taken as the meteorological entities, and the dispatching instructions and grid nodes of the new energy power station are taken as the dispatching entities. Based on data association analysis, the connection relationships between operating entities, meteorological entities, and scheduling entities are obtained, thereby obtaining the entity relationship layer of new energy power plants.

5. The intelligent monitoring, analysis, and control method for new energy dispatching according to claim 1, characterized in that: The generation of anomaly judgment results includes: The electrical parameters, meteorological parameters, and equipment status parameters of the new energy power plant are matched with the fault rules of the corresponding fault rule layer. By using the query and reasoning functions of the power plant equipment scheduling knowledge graph, it is determined whether the conditions of a certain fault rule are met. If they are met, the corresponding abnormal judgment result of the new energy power plant is triggered. If none of them are met, the abnormal judgment result of the new energy power plant is determined to be fault-free.

6. The intelligent monitoring, analysis and control method for new energy dispatching according to claim 2, characterized in that: The process of locating the fault source and generating control instructions containing control strategies includes: Fault location is achieved by utilizing the relationships between entities in the power plant equipment scheduling knowledge graph, thus identifying the various fault sources in the new energy power plant. Reference values ​​for each fault source in the new energy power plant are extracted from the power plant equipment scheduling knowledge graph. Then, the actual values ​​of each fault source in the new energy power plant are subtracted from the reference values, and the ratio of the absolute value of the difference to the reference value is used as the influence degree of each fault source in the new energy power plant. Based on the fault sources and their impact in new energy power plants, historical cases similar to the fault sources and their impact are searched in the power plant equipment scheduling knowledge graph, and then the control rules of each historical case are extracted. Based on the control rules of various historical cases, corresponding control instructions are generated for each abnormal node of the new energy power plant.

7. The intelligent monitoring, analysis, and control method for new energy dispatching according to claim 6, characterized in that: The method of fault location using the relationships between entities in the power plant equipment scheduling knowledge graph includes: By matching and comparing the abnormal parameters of the new energy power plant with the abnormal parameters corresponding to each abnormal node, the abnormal nodes of the new energy power plant can be obtained. Starting from the abnormal phenomenon node, the knowledge graph of power plant equipment scheduling is traversed by the depth-first search algorithm to obtain the associated nodes of each abnormal node in the new energy power plant; The meteorological-output correlation degree and scheduling priority coefficient of each abnormal node in the new energy power station corresponding to each associated node are respectively denoted as: and , The abnormal node number, , Number the associated nodes. ; According to the formula Calculate the dynamic weight values ​​of each abnormal node corresponding to each associated node in the new energy power plant. and These are the weights for setting the reference meteorological-output correlation and the scheduling priority coefficient, respectively. , ; The dynamic weight values ​​of each abnormal node in the new energy power plant corresponding to each associated node are summed to obtain the total dynamic weight value of each abnormal node corresponding to each associated path. The total dynamic weight values ​​of each abnormal node and its associated paths are sorted from largest to smallest, and the associated paths of the abnormal nodes with the first-ranked values ​​are extracted as the fault paths of each abnormal node in the new energy power plant. Based on the fault paths of each abnormal node in the new energy power plant, the fault sources corresponding to the fault paths of each abnormal node in the new energy power plant are extracted from the historical fault data of the new energy power plant, and these sources are used as the fault sources of the new energy power plant.

8. The intelligent monitoring, analysis and control method for new energy dispatching according to claim 1, characterized in that: To evaluate the effectiveness of the regulation, including: Extract operational data of new energy power plants after executing control commands from the control command response data; Compare the running data with the set reference running data range; If all operating data are within the set reference range, the control effect of the new energy power plant is deemed satisfactory; otherwise, the control effect is deemed unsatisfactory.

Images