CN-121984107-A - Photovoltaic power station assembly power scheduling optimization method and system

CN121984107ACN 121984107 ACN121984107 ACN 121984107ACN-121984107-A

Abstract

The embodiment of the application provides a photovoltaic power station assembly power scheduling optimization method and a system, and relates to the technical field of photovoltaic power station power scheduling, wherein the method comprises the steps of obtaining real-time capability information of each inverter, wherein the real-time capability information comprises maximum available power and maximum response rate boundaries of the inverters after the inverters are adjusted according to deviation between current actual physical manifestations and theoretical physical limits; the method comprises the steps of obtaining health state and risk assessment information of each inverter, wherein the health state and risk assessment information comprises real-time health scores and short-term performance deterioration risks generated by early signs of performance decay identified by the inverters according to variation trends of key operation parameters, and carrying out differential power distribution on a plurality of inverters according to maximum available power, maximum response rate boundaries, real-time health scores and short-term performance deterioration risks so as to match real physical output limits of the inverters. The application can improve the overall operation efficiency and economic benefit of the photovoltaic power station.

Inventors

HU RUI
WEI SHILEI
Mo Wangyi
CHENG XIAOFEI
Ji Peichen
LI JIAN
LI YUAN
QIN LIUYUN
ZENG JINFU

Assignees

南方电网互联网服务有限公司

Dates

Publication Date: 20260505
Application Date: 20251209

Claims (10)

1. A photovoltaic power plant module power scheduling optimization method, the method comprising: Acquiring real-time capability information of each inverter, wherein the real-time capability information comprises the maximum available power and the maximum response rate boundary of the inverters after being adjusted according to the deviation between the current actual physical expression and the theoretical physical limit; Acquiring health status and risk assessment information of each inverter, wherein the health status and risk assessment information comprises real-time health scores and short-term performance deterioration risks generated by early signs of performance decay identified by the inverters according to the change trend of key operation parameters; and performing differentiated power distribution on the inverters according to the maximum available power, the maximum response rate boundary, the real-time health score and the short-term performance degradation risk so as to match the real physical output limits of the inverters.
2. The method of claim 1, wherein the obtaining health and risk assessment information for each of the inverters, the health and risk assessment information including real-time health scores and short-term performance degradation risks generated by early signs of performance degradation identified by the inverters from trends in key operating parameters, comprises: Acquiring the proportion of the temperature rise of the IGBT module in the inverter and the temperature difference of the current environment and the reduction rate of the conversion efficiency; Identifying early signs of performance decay according to the ratio and the change trend of the rate of decline; acquiring instantaneous dynamic response data of each inverter when executing a power instruction; according to the instantaneous dynamic response data, adaptively adjusting the instantaneous stability boundary of the inverter to obtain an adjusted instantaneous stability boundary; Deducing an instantaneous limit bearable by the inverter in the first time according to the adjusted instantaneous stability boundary; Generating a real-time health score and a risk of performance deterioration in a short period based on the instantaneous limit and the early signs of performance decay.
3. The method of claim 2, wherein said obtaining transient dynamic response data for each of said inverters when executing power commands comprises: synchronously collecting the junction temperature of an IGBT module, the rotating speed of a cooling fan, the alternating current output power and the direct current-alternating current conversion efficiency in the inverter; when the inverter receives a power instruction, calculating the instantaneous change rate, the instantaneous acceleration and the response lag time of the IGBT module junction temperature, the cooling fan rotating speed, the alternating current output power and the direct current-alternating current conversion efficiency in a preset time window to obtain a current feature vector; matching the current feature vector with a preset abnormal pattern fingerprint library to obtain a matching result; and judging an abnormal level according to the matching result, dynamically adjusting the instantaneous bearable power limit and response time of the current feature vector, compressing and encoding to obtain instantaneous dynamic response data.
4. The method of claim 3, wherein calculating the instantaneous rate of change, the instantaneous acceleration, and the response delay time for the IGBT module junction temperature, the cooling fan rotational speed, the ac output power, and the dc-to-ac conversion efficiency within a preset time window when the inverter receives the power command, to obtain the current feature vector, comprises: Generating a transient response reference curve based on the current running state and the environmental condition of the inverter in a preset time window when the inverter receives a power instruction; calculating the junction temperature of the IGBT module, the rotating speed of the cooling fan, the alternating current output power, the direct current-alternating current conversion efficiency and the transient response reference curve, and calculating the transient change rate, the transient acceleration and the response lag time to obtain the transient deviation, the deviation change rate and the deviation accumulation quantity; And obtaining a current feature vector according to the instantaneous deviation, the deviation change rate and the deviation accumulation quantity.
5. The method of claim 3, wherein the matching the current feature vector with a preset abnormal pattern fingerprint library to obtain a matching result further comprises: when the similarity of the fingerprint library of the preset abnormal mode matching the current feature vector is lower than a preset threshold value, marking the current feature vector as an unidentified mode; when a plurality of current feature vectors marked as unidentified modes are continuously monitored and the current feature vectors of the unidentified modes show clustering trend in a target feature space, obtaining clustered feature vectors and triggering a target mode analysis flow; In the target pattern analysis flow, the clustering feature vector is analyzed in combination with the environmental condition of the current feature vector of the unidentified pattern and the operation working condition of the inverter so as to identify the potential physical root of the clustering feature vector and obtain an analysis result; and when the current feature vector of the unrecognized mode is determined to be a novel nonlinear instantaneous attenuation mode according to the analysis result, generating the feature vector of the novel nonlinear instantaneous attenuation mode, corresponding abnormal grade assessment and instantaneous limit adjustment strategies, and dynamically adding the feature vector, the corresponding abnormal grade assessment and the instantaneous limit adjustment strategies into the fingerprint library of the preset abnormal mode.
6. The method according to claim 5, wherein when a plurality of current feature vectors marked as unidentified patterns are continuously monitored and the plurality of current feature vectors of unidentified patterns exhibit a clustering trend in a target feature space, obtaining a clustered feature vector and triggering a target pattern analysis flow comprises: When a plurality of current feature vectors marked as unidentified modes are continuously monitored, preprocessing the plurality of current feature vectors marked as unidentified modes to obtain preprocessed current feature vectors of unidentified modes, wherein the preprocessing comprises the steps of removing environmental noise and sporadic data points; obtaining the salt spray level, the humidity change rate and the environmental temperature gradient of the geographic position of the inverter; dynamically adjusting the distance measurement weight and the clustering radius of the initial feature space according to the salt spray level, the humidity change rate and the environmental temperature gradient to obtain a target feature space; And when the current feature vectors of the plurality of the pre-processed unrecognized modes show clustering trend in the target feature space, acquiring clustering feature vectors to trigger a target mode analysis flow.
7. The method according to claim 5, wherein in the target pattern analysis process, in combination with an environmental condition under which a current feature vector of the unidentified pattern appears and an operation condition of the inverter, the cluster feature vector is analyzed to identify a potential physical root of the cluster feature vector, and an analysis result is obtained, including: Performing deep correlation on the transient variation characteristics of the internal physical parameters of the clustering characteristic vector and the environmental conditions and the operation working conditions of the inverter, and identifying the sensitivity of the internal physical parameters to the environmental conditions and the operation working conditions to obtain a sensitivity analysis result; And analyzing the clustering feature vector of the clustering trend according to the sensitivity analysis result to identify the potential physical root of the clustering feature vector, so as to obtain an analysis result.
8. The method of claim 1, wherein said differentially distributing power to the plurality of inverters to match actual physical output limits of the plurality of inverters based on the maximum available power, the maximum response rate boundary, the real-time health score, and the risk of performance degradation in the short term comprises: Constructing a multi-dimensional evaluation matrix according to the maximum available power, the maximum response rate boundary, the real-time health score and the short-term performance deterioration risk; determining the priority of the inverter according to the multi-dimensional evaluation matrix; And carrying out differential power distribution on the plurality of inverters according to the priorities of the inverters so as to match the actual physical output limits of the plurality of inverters.
9. The method of claim 8, wherein said differentially distributing power to a plurality of said inverters according to their priorities to match actual physical output limits of a plurality of said inverters further comprises: After differential power distribution is carried out on a plurality of inverters, comparing the difference value between the actual output power of each inverter and the actual physical output limit in real time to obtain a plurality of difference values; When at least one difference value is larger than a preset difference value, determining the priority of the inverter again and carrying out power distribution again; and when all the difference values are smaller than or equal to a preset difference value, maintaining the original power distribution.
10. A photovoltaic power plant assembly power scheduling optimization system, the system comprising: The first acquisition module is used for acquiring real-time capability information of each inverter, wherein the real-time capability information comprises the maximum available power and the maximum response rate boundary of the inverters after being adjusted according to the deviation between the current actual physical expression and the theoretical physical limit; The second acquisition module is used for acquiring the health state and risk assessment information of each inverter, wherein the health state and risk assessment information comprises a real-time health score and a short-term performance deterioration risk generated by early signs of performance deterioration identified by the inverter according to the change trend of the key operation parameters; And the matching module is used for carrying out differential power distribution on the inverters according to the maximum available power, the maximum response rate boundary, the real-time health score and the short-term performance deterioration risk so as to match the actual physical output limits of the inverters.

Description

Photovoltaic power station assembly power scheduling optimization method and system Technical Field The application relates to the technical field of power scheduling of photovoltaic power stations, in particular to a power scheduling optimization method and system for a photovoltaic power station assembly. Background In the related art, a photovoltaic power station is an important component of clean energy, and stable and efficient operation of the photovoltaic power station is important to a power grid. However, during long-term operation, critical equipment inside the plant, in particular the inverter, can be continuously affected by environmental factors, resulting in an imperceptible slow decay of its performance. The implicit performance degradation does not trigger the traditional fault alarm, but can silently influence the actual power generation capacity of the power station and the response precision to the power grid dispatching instruction, thereby causing energy waste and power grid operation risk. The conventional power scheduling method often performs power distribution based on inaccurate equipment capability assessment due to insufficient consideration of subtle and progressive performance changes of equipment, and especially in emergency peak shaving situations where a power grid needs to respond quickly, the mismatch becomes more prominent. In actual operation of a photovoltaic power station, an Insulated Gate Bipolar Transistor (IGBT) power module inside an inverter may cause a hidden drop in heat dissipation efficiency due to microscopic wear and corrosion of a heat dissipation system (e.g., a heat dissipation fan). The drop is an extremely slow and gradual process, the sudden failure or performance dip of the fan cannot be caused in a short period, the generated running noise increase is often lower than the alarm threshold set by the automatic monitoring system of the power station, and the running noise increase is difficult to be perceived by operation and maintenance personnel in daily inspection. Due to such a recessive drop in heat dissipation efficiency, the actual operating temperature of the IGBT power module may be slightly higher than the expected value under its design conditions within a portion of the affected inverter during a specific period of operation of the power station, such as during the daytime when the solar irradiation intensity is highest and the ambient temperature is also high in the summer afternoon. Although such a temperature rise is generally still within the absolute maximum allowable junction temperature range of the IGBT module, over-temperature protection tripping is not immediately induced, the sustained, slightly elevated temperature may affect the electrical characteristics of the semiconductor device, resulting in a slight increase in the internal on-resistance of the IGBT module, resulting in higher on-losses when current passes. Meanwhile, the switching characteristics of the IGBT may also be slightly affected by temperature, resulting in a slight increase in energy loss at each switching action. Together, these cumulative effects result in a slight but sustained decrease in overall dc to ac conversion efficiency of the affected inverter. The magnitude of this efficiency degradation may be so small that it is difficult to distinguish from power fluctuations caused by aging of the photovoltaic module itself, surface dust accumulation, or other known factors such as ambient temperature variations. Disclosure of Invention The present application aims to solve at least one of the technical problems existing in the prior art. Therefore, the application provides a photovoltaic power station assembly power scheduling optimization method and system, and aims to improve the overall operation efficiency and economic benefit of a photovoltaic power station. In a first aspect, an embodiment of the present application provides a method for optimizing power scheduling of a photovoltaic power station assembly, including: Acquiring real-time capability information of each inverter, wherein the real-time capability information comprises the maximum available power and the maximum response rate boundary of the inverters after being adjusted according to the deviation between the current actual physical expression and the theoretical physical limit; Acquiring health status and risk assessment information of each inverter, wherein the health status and risk assessment information comprises real-time health scores and short-term performance deterioration risks generated by early signs of performance decay identified by the inverters according to the change trend of key operation parameters; and performing differentiated power distribution on the inverters according to the maximum available power, the maximum response rate boundary, the real-time health score and the short-term performance degradation risk so as to match the real physical output limits of the inver