Distributed photovoltaic power distribution network intelligent regulation and control method and system

Abstract

The invention discloses a distributed photovoltaic power distribution network intelligent regulation and control method and system, and belongs to the technical field of intelligent power grids. The intelligent regulation and control method comprises the following steps: for the whole power distribution network, carrying out holographic monitoring of distributed photovoltaic and energy storage based on a power distribution network PWU to obtain holographic monitoring data; identifying the holographic monitoring data, predicting the distributed photovoltaic power generation power based on the holographic monitoring data, and generating an identification result and a prediction result; and generating an intelligent regulation and control strategy for the power distribution network based on the identification result and the prediction result, and carrying out intelligent regulation and control on the power distribution network according to the intelligent regulation and control strategy. According to the invention, high-quality data support is provided for deep data analysis and intelligent regulation and control of the power distribution network.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and more specifically, to a method and system for intelligent control of distribution networks for distributed photovoltaic systems. Background Technology

[0002] Distribution network PMU technology:

[0003] A phasor measurement unit (PMU) is an advanced power system condition monitoring device that can measure the precise phasors of key quantities such as grid voltage, current, frequency, and phase angle in real time. In recent years, PMUs have begun to be used in distribution networks, namely distribution network PMUs, which enable high-precision, high-resolution, and panoramic monitoring of the distribution network. However, current distribution network PMU data is mainly used for visualization, lacking in-depth data analysis and intelligent applications.

[0004] Impact of distributed photovoltaic (PV) grid integration:

[0005] The large-scale, high-density integration of distributed photovoltaic (PV) power has significantly impacted indicators such as voltage, frequency, harmonics, and reactive power in the distribution network. The volatility and intermittency of PV output also exacerbate the difficulty of operating and controlling the distribution network. Traditional single-point, low-frequency monitoring methods are insufficient to comprehensively assess the impact of PV integration, necessitating the development of a comprehensive monitoring capability with "wide coverage, multiple measurement points, and high resolution."

[0006] Applications of energy storage systems in power distribution networks:

[0007] To mitigate the impact of renewable energy integration on the distribution network, various types of energy storage systems are being deployed extensively on the distribution network side. Energy storage systems can improve power quality by smoothing photovoltaic output, regulating active and reactive power, and improving overall power quality through proper scheduling. However, the complex operating conditions and variable parameters of energy storage systems place higher demands on monitoring, analysis, and scheduling control. Therefore, there is an urgent need to develop technologies for state awareness and active management throughout the entire lifecycle of energy storage.

[0008] Big data analytics and artificial intelligence technologies for power distribution networks:

[0009] With the advancement of new infrastructure such as the Internet of Things for distribution and smart distribution rooms, the distribution network has entered the era of big data. New artificial intelligence technologies such as machine learning and deep learning provide new approaches for the in-depth value mining of massive amounts of data in the distribution network. However, current applications mainly focus on single scenarios such as load forecasting and fault diagnosis, lacking integrated analysis and multi-scenario collaboration. There is a need to explore cross-domain technology integration to improve the overall intelligence level of the distribution network.

[0010] In summary, traditional distribution network monitoring and control technologies are no longer adequate for the large-scale integration of distributed photovoltaic and energy storage systems. How to fully utilize advanced technologies such as distribution network power management units (PMUs), big data analytics, and artificial intelligence to construct a comprehensive monitoring and intelligent control system for distributed energy has become a research hotspot and a critical issue requiring breakthroughs in the field of smart grids. Summary of the Invention

[0011] To address the above problems, this invention proposes a smart control method for distribution networks oriented towards distributed photovoltaic power, comprising:

[0012] For the entire distribution network, holographic monitoring of distributed photovoltaic and energy storage is carried out based on the distribution network PWU to obtain holographic monitoring data;

[0013] The holographic monitoring data is identified, and based on the holographic monitoring data, the distributed photovoltaic power generation is predicted, generating identification results and prediction results;

[0014] Based on the identification and prediction results, an intelligent control strategy for the distribution network is generated, and the distribution network is intelligently controlled using the intelligent control strategy.

[0015] Optionally, holographic monitoring of distributed photovoltaic and energy storage based on distribution network PWU can be performed, including:

[0016] Based on the distribution network topology and the access locations of distributed photovoltaic and energy storage in the distribution network, the layout of the distribution network PMU is adjusted to achieve full-coverage holographic monitoring of distributed photovoltaic and energy storage.

[0017] Optionally, acquire holographic monitoring data, including:

[0018] Set the acquisition frequency of the distribution network PMU, and collect multi-dimensional parameter data of distributed photovoltaic and energy storage and operating status data of distribution network equipment in real time under the set acquisition frequency. Then, fuse the multi-dimensional parameter data and the operating status data to obtain holographic monitoring data.

[0019] Optionally, the holographic monitoring data is identified, including:

[0020] The holographic monitoring data is input into the abnormal operating condition intelligent identification model, and the abnormal operating condition intelligent identification model identifies the holographic monitoring data to determine suspected abnormal operating conditions in the power distribution network.

[0021] Optional, abnormal operating conditions are defined as follows: based on the actual operating experience of distributed photovoltaic power grid connection, sudden drop in photovoltaic power generation, voltage exceeding limits, harmonic exceedance, and islanding operation are defined as abnormal operating conditions.

[0022] Optionally, the abnormal operating condition intelligent identification model is used to identify the holographic monitoring data, including:

[0023] Using the aforementioned intelligent identification model for abnormal operating conditions, abnormal features of photovoltaic power generation are extracted from the holographic monitoring data. These abnormal features are then identified to determine suspected abnormal operating conditions in the power distribution network.

[0024] Optional anomalous characteristics of photovoltaic power generation include:

[0025] Abnormal power fluctuation rate, abnormal voltage deviation rate, and abnormal harmonic content.

[0026] Optionally, the distributed photovoltaic power generation can be predicted, including:

[0027] Based on meteorological characteristics, a multi-timescale power prediction model is constructed, and based on the power prediction model, predictions are made using holographic monitoring data.

[0028] Optionally, generate intelligent control strategies for the distribution network, including:

[0029] Based on the constraints of photovoltaic power generation forecast, distribution network real-time operation status, load forecast curve, and available energy storage capacity, a mathematical model for the coordinated optimization of power generation, grid, load, and storage is established. Based on the mathematical model, multi-scenario optimization objectives and market trading mechanisms are generated at two time scales: day-ahead and intraday.

[0030] For day-ahead optimization, guided by the multi-scenario optimization objectives of sources, grids, loads and storage, algorithms such as mixed integer programming are used to formulate the day-ahead optimal operation plan for the distribution network, and determine the distributed photovoltaic output forecast, distribution network topology pre-construction, energy storage charging and discharging strategy and interruptible load response strategy.

[0031] For intraday optimization, a rolling optimization method is adopted, which is based on real-time forecasting of photovoltaic power generation, real-time scheduling of energy storage and real-time response of demand side measures to correct the day-ahead plan and formulate a smart control strategy for the distribution network.

[0032] Furthermore, this invention also proposes an intelligent control system for power distribution networks, comprising:

[0033] The monitoring unit is used to perform holographic monitoring of distributed photovoltaic and energy storage based on the distribution network PWU for the entire distribution network and to acquire holographic monitoring data;

[0034] The identification and prediction unit is used to identify the holographic monitoring data and predict the distributed photovoltaic power generation based on the holographic monitoring data, and generate identification results and prediction results.

[0035] The control unit is used to generate an intelligent control strategy for the distribution network based on the identification results and prediction results, and to perform intelligent control of the distribution network using the intelligent control strategy.

[0036] Optionally, holographic monitoring of distributed photovoltaic and energy storage based on distribution network PWU can be performed, including:

[0037] Based on the distribution network topology and the access locations of distributed photovoltaic and energy storage in the distribution network, the layout of the distribution network PMU is adjusted to achieve full-coverage holographic monitoring of distributed photovoltaic and energy storage.

[0038] Optionally, acquire holographic monitoring data, including:

[0039] Set the acquisition frequency of the distribution network PMU, and collect multi-dimensional parameter data of distributed photovoltaic and energy storage and operating status data of distribution network equipment in real time under the set acquisition frequency. Then, fuse the multi-dimensional parameter data and the operating status data to obtain holographic monitoring data.

[0040] Optionally, the holographic monitoring data is identified, including:

[0041] The holographic monitoring data is input into the abnormal operating condition intelligent identification model, and the abnormal operating condition intelligent identification model identifies the holographic monitoring data to determine suspected abnormal operating conditions in the power distribution network.

[0042] Optional, abnormal operating conditions are defined as follows: based on the actual operating experience of distributed photovoltaic power grid connection, sudden drop in photovoltaic power generation, voltage exceeding limits, harmonic exceedance, and islanding operation are defined as abnormal operating conditions.

[0043] Optionally, the abnormal operating condition intelligent identification model is used to identify the holographic monitoring data, including:

[0044] Using the aforementioned intelligent identification model for abnormal operating conditions, abnormal features of photovoltaic power generation are extracted from the holographic monitoring data. These abnormal features are then identified to determine suspected abnormal operating conditions in the power distribution network.

[0045] Optional anomalous characteristics of photovoltaic power generation include:

[0046] Abnormal power fluctuation rate, abnormal voltage deviation rate, and abnormal harmonic content.

[0047] Optionally, the distributed photovoltaic power generation can be predicted, including:

[0048] Based on meteorological characteristics, a multi-timescale power prediction model is constructed, and based on the power prediction model, predictions are made using holographic monitoring data.

[0049] Optionally, generate intelligent control strategies for the distribution network, including:

[0050] Based on the constraints of photovoltaic power generation forecast, distribution network real-time operation status, load forecast curve, and available energy storage capacity, a mathematical model for the coordinated optimization of power generation, grid, load, and storage is established. Based on the mathematical model, multi-scenario optimization objectives and market trading mechanisms are generated at two time scales: day-ahead and intraday.

[0051] For day-ahead optimization, guided by the multi-scenario optimization objectives of sources, grids, loads and storage, algorithms such as mixed integer programming are used to formulate the day-ahead optimal operation plan for the distribution network, and determine the distributed photovoltaic output forecast, distribution network topology pre-construction, energy storage charging and discharging strategy and interruptible load response strategy.

[0052] For intraday optimization, a rolling optimization method is adopted, which is based on real-time forecasting of photovoltaic power generation, real-time scheduling of energy storage and real-time response of demand side measures to correct the day-ahead plan and formulate a smart control strategy for the distribution network.

[0053] In another aspect, the present invention also provides a computing device, comprising: one or more processors;

[0054] A processor is used to execute one or more programs;

[0055] When the one or more programs are executed by the one or more processors, the method described above is implemented.

[0056] In another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the method described above.

[0057] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0058] This invention proposes an intelligent control method for distribution networks, comprising: performing holographic monitoring of distributed photovoltaic (PV) and energy storage based on the distribution network power unit (PWU) across the entire distribution network to acquire holographic monitoring data; identifying the holographic monitoring data and predicting the distributed PV power generation based on the holographic monitoring data to generate identification results and prediction results; generating an intelligent control strategy for the distribution network based on the identification results and prediction results, and using the intelligent control strategy to intelligently control the distribution network. This invention provides high-quality data support for in-depth data analysis and intelligent control of distribution networks. Attached Figure Description

[0059] Figure 1 This is a flowchart of the method of the present invention;

[0060] Figure 2 This is a structural diagram of an embodiment of the method of the present invention;

[0061] Figure 3This is a flowchart illustrating the distributed photovoltaic holographic monitoring data fusion and analysis process according to an embodiment of the method of the present invention.

[0062] Figure 4 This is a roadmap for intelligent identification technology of abnormal operating conditions in distributed photovoltaic systems, as exemplified by the method of the present invention.

[0063] Figure 5 This is a block diagram of a distributed photovoltaic multi-timescale power prediction model according to an embodiment of the method of the present invention.

[0064] Figure 6 This is a schematic diagram of a distribution network intelligent control strategy for distributed photovoltaic access, as described in an embodiment of the method of the present invention. Detailed Implementation

[0065] Exemplary embodiments of the invention will now be described with reference to the accompanying drawings. However, the invention may be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to fully and completely disclose the invention and to fully convey its scope to those skilled in the art. The terminology used in the exemplary embodiments illustrated in the drawings is not intended to limit the invention. In the drawings, the same units / elements are referred to by the same reference numerals.

[0066] Unless otherwise stated, the terms used herein (including technical terms) have their common meaning as understood by one of ordinary skill in the art. Furthermore, it is understood that terms defined in commonly used dictionaries should be understood to have a meaning consistent with the context of their relevant field, and not to be interpreted as having an idealized or overly formal meaning.

[0067] Example 1:

[0068] This invention proposes an intelligent control method for power distribution networks, such as... Figure 1 As shown, it includes:

[0069] Step 1: For the entire distribution network, perform holographic monitoring of distributed photovoltaic and energy storage based on the distribution network PWU to obtain holographic monitoring data;

[0070] Step 2: Identify the holographic monitoring data, and based on the holographic monitoring data, predict the distributed photovoltaic power generation, and generate identification results and prediction results;

[0071] Step 3: Based on the identification and prediction results, generate an intelligent control strategy for the distribution network, and use the intelligent control strategy to intelligently control the distribution network.

[0072] Among them, holographic monitoring of distributed photovoltaic and energy storage based on distribution network PWU includes:

[0073] Based on the distribution network topology and the access locations of distributed photovoltaic and energy storage in the distribution network, the layout of the distribution network PMU is adjusted to achieve full-coverage holographic monitoring of distributed photovoltaic and energy storage.

[0074] Acquiring holographic monitoring data includes:

[0075] Set the acquisition frequency of the distribution network PMU, and collect multi-dimensional parameter data of distributed photovoltaic and energy storage and operating status data of distribution network equipment in real time under the set acquisition frequency. Then, fuse the multi-dimensional parameter data and the operating status data to obtain holographic monitoring data.

[0076] The identification of the holographic monitoring data includes:

[0077] The holographic monitoring data is input into the abnormal operating condition intelligent identification model, and the abnormal operating condition intelligent identification model identifies the holographic monitoring data to determine suspected abnormal operating conditions in the power distribution network.

[0078] Among them, abnormal operating conditions are defined as follows: based on the actual operating experience of distributed photovoltaic power grid connection, photovoltaic power generation power drop, voltage limit exceedance, harmonic exceedance and islanding operation are defined as abnormal operating conditions.

[0079] Specifically, in scenarios where a high proportion of distributed photovoltaic (PV) power is integrated into the distribution network, various abnormal operating conditions can easily occur due to weather, load disturbances, and the complexity of the electrical environment, such as sudden drops in PV power generation, voltage exceeding limits, harmonic exceedances, and islanding operation. Traditional static threshold-based judgment methods are insufficient to meet the real-time and accuracy requirements of intelligent control systems in terms of accuracy and response speed.

[0080] The anomaly identification method proposed in this invention can efficiently and intelligently identify the operating status of photovoltaic units in a power distribution network. The anomaly identification method of this invention includes the following steps:

[0081] Step 1: Holographic monitoring data acquisition:

[0082] High-frequency monitoring data, including the following, is collected by the distribution network PMU nodes:

[0083] D(t)=[P(t),Q(t),V(t),I(t),f(t),THD(t)]

[0084] in:

[0085] P(t): Active power;

[0086] Q(t): Reactive power;

[0087] V(t), I(t): voltage and current;

[0088] f(t): Frequency;

[0089] THD(t): Total harmonic distortion of voltage.

[0090] The data sampling frequency is set to 1Hz to 10Hz, and a sliding time window mechanism is used for processing.

[0091] Step 2: Multi-scale feature construction

[0092] This invention constructs two time-scale feature inputs:

[0093] (1) Features of the original sequence at a short time scale (second level):

[0094] Typically, T1 = 10 seconds, preserving high-frequency dynamic characteristics.

[0095] (2) Long-term (minute-level) statistical characteristics: X2 = [μ(D), σ(D), γ(D), SNR(D)]

[0096] in:

[0097] μ(D): mean; σ(D): standard deviation;

[0098] γ(D): Skewness; SNR(D): Signal-to-noise ratio.

[0099] The statistics window is set to 3-5 minutes to reflect the operating trend.

[0100] 3. Model structure design:

[0101] This invention employs a dual-channel deep neural network structure DFNet, comprising:

[0102] Channel 1: Short Time Series Modeling (Conv1D+LSTM)

[0103] Local variation features of temporal sequence extracted by 1D convolutional layers: F c =ReLU(Conv1D(X1))

[0104] LSTM network for modeling time series dependencies: H = LSTM(F) c )

[0105] Channel Two: Statistical Feature Modeling (MLP)

[0106] Multilayer perceptrons are used for dimensionality reduction and nonlinear modeling of X_2: F s =ReLU(W1X2+b1), F s′ =ReLU(W2F s +b2)

[0107] Feature fusion concatenates the outputs of the two channels: Z = [H, F_s']

[0108] The classification output uses a fully connected layer and a Softmax classifier to output the anomaly type: y = Softmax(W f Z+b)

[0109] 4. Output Definitions and Exception Types:

[0110] The model output y represents the predicted operating status of the photovoltaic unit, and the categories are defined as shown in Table 1 below:

[0111] Table 1

[0112] Category coding Abnormal operating condition types 0 Normal operation 1 Photovoltaic power drops sharply 2 Voltage exceeding limits (overvoltage / undervoltage) 3 Harmonic exceedance 4 Suspected behavior of operating in an isolated area

[0113] The model output includes: predicted category, anomaly type description, confidence score, etc.

[0114] 5. Model Training and Implementation Recommendations:

[0115] Model training process: Input samples should be no less than 500 per class, using the cross-entropy loss function:

[0116] Model deployment method: Model size less than 10MB;

[0117] It can be deployed at regional control master stations, feeder automation terminals, or edge gateways;

[0118] The recommended scrolling recognition cycle is 1 to 2 minutes.

[0119] Supports online updates and transfer learning to enhance adaptability.

[0120] By deploying the anomaly detection model of this invention, the following technical effects can be achieved:

[0121] Real-time recognition: Recognition latency is controlled within 500ms;

[0122] High recognition accuracy: The average accuracy rate in the experiment was ≥96%;

[0123] Highly adaptable: Supports different access scenarios and network topologies;

[0124] High deployability: Supports embedded edge deployment and online updates;

[0125] High reliability: Significantly reduces false alarm and false negative rates, enhancing the stability of regulation.

[0126] The identification of holographic monitoring data using the aforementioned abnormal operating condition intelligent identification model includes:

[0127] Using the aforementioned intelligent identification model for abnormal operating conditions, abnormal features of photovoltaic power generation are extracted from the holographic monitoring data. These abnormal features are then identified to determine suspected abnormal operating conditions in the power distribution network.

[0128] Among them, the abnormal characteristics of photovoltaic power generation include:

[0129] Abnormal power fluctuation rate, abnormal voltage deviation rate, and abnormal harmonic content.

[0130] The prediction of distributed photovoltaic power generation includes:

[0131] Based on meteorological characteristics, a multi-timescale power prediction model is constructed, and based on the power prediction model, predictions are made using holographic monitoring data.

[0132] Among them, the generation of intelligent control strategies for the distribution network includes:

[0133] Based on the constraints of photovoltaic power generation forecast, distribution network real-time operation status, load forecast curve, and available energy storage capacity, a mathematical model for the coordinated optimization of power generation, grid, load, and storage is established. Based on the mathematical model, multi-scenario optimization objectives and market trading mechanisms are generated at two time scales: day-ahead and intraday.

[0134] For day-ahead optimization, guided by the multi-scenario optimization objectives of sources, grids, loads and storage, algorithms such as mixed integer programming are used to formulate the day-ahead optimal operation plan for the distribution network, and determine the distributed photovoltaic output forecast, distribution network topology pre-construction, energy storage charging and discharging strategy and interruptible load response strategy.

[0135] For intraday optimization, a rolling optimization method is adopted, which is based on real-time forecasting of photovoltaic power generation, real-time scheduling of energy storage and real-time response of demand side measures to correct the day-ahead plan and formulate a smart control strategy for the distribution network.

[0136] The following technologies are mainly applied in this invention, including:

[0137] The application framework of distributed photovoltaic and energy storage holographic monitoring based on distribution network PMU is as follows: Figure 2 As shown, it includes:

[0138] Distribution network-level distributed PMU deployment scheme design:

[0139] Based on the distribution network topology and the location of distributed photovoltaic (PV) and energy storage connections, and taking into account factors such as observability, maintainability, and economy, the layout scheme of distribution network power measurement units (PMUs) is optimized. PMU measurement devices are widely deployed at the grid connection points of distributed PV and energy storage, and at key nodes of the distribution network (such as substations, distribution rooms, switching stations, and ring main units) to achieve full-coverage monitoring of distributed energy resources and the distribution network in their connected areas.

[0140] Real-time acquisition of PMU data with multiple parameters and multiple spatiotemporal scales:

[0141] By utilizing the distribution network PMU to collect multi-dimensional parameters in real time, including distributed photovoltaic power generation output, energy storage charging and discharging power, node voltage, line power flow, frequency, and harmonics, the data acquisition frequency (e.g., per cycle, second-level, minute-level) can be flexibly set according to the application scenarios such as real-time control of the distribution network, fault diagnosis, and power quality assessment, enabling refined monitoring at multiple spatiotemporal scales. Simultaneously, operating status data from photovoltaic inverters, energy storage controllers, and other equipment are acquired, providing a basis for comprehensively assessing the impact of distributed energy integration.

[0142] Regional centralization and cloud aggregation of multi-source heterogeneous monitoring data:

[0143] Based on the distribution network IoT communication architecture, and employing various communication technologies (such as fiber optics, 4G / 5G, and narrowband IoT), distribution network PMU data and distributed energy operation data are transmitted in real time to the regional distribution automation master station and cloud big data platform. A scheme for regional centralized and cloud-based aggregation of distributed energy monitoring data is designed to balance communication efficiency, latency jitter, and data redundancy.

[0144] Multi-temporal and spatial scale data fusion based on a distributed computing framework, such as Figure 3 As shown, it includes:

[0145] To address the characteristics of massive time-series data from distribution network power supply units (PMUs), a distributed computing framework (such as Hadoop and Spark) is employed to design an efficient method for organizing and fusing spatiotemporal data. This results in the construction of a comprehensive operational status profile of the distribution network at multiple time scales (minute, hour, and day) and spatial granularities (feeder, substation, and grid levels). The correlation between the operational characteristics of distributed photovoltaic (PV) and energy storage systems and load and grid operation is explored, characterizing the impact of renewable energy integration on indicators such as power flow distribution, voltage quality, and network loss levels.

[0146] Intelligent identification technology for abnormal operating conditions of distributed photovoltaic systems based on machine learning, such as Figure 4 As shown, it includes:

[0147] Definition of abnormal operating conditions for multiple scenarios in distributed photovoltaic power generation:

[0148] Based on the actual operating experience of distributed photovoltaic power grid integration, typical abnormal operating conditions such as sudden drop in photovoltaic power generation, voltage exceeding limits, harmonic exceedance, and islanding operation are defined, along with their judgment thresholds.

[0149] Anomaly condition classification model based on multi-timescale feature extraction:

[0150] For holographic monitoring data at different time granularities, key quantities reflecting abnormal characteristics of photovoltaic power generation, such as power fluctuation rate, voltage deviation rate, and harmonic content, are extracted. Machine learning classification algorithms, such as support vector machine (SVM) and random forest (RF), are used to construct a multi-time-scale abnormal operating condition classification model to achieve real-time identification, accurate location, and automatic classification of abnormal operating conditions.

[0151] Intelligent identification method for abnormal operating conditions:

[0152] The online application uses a pre-trained abnormal condition classification model to intelligently analyze the real-time data stream of the distribution network PMU. Once a suspected abnormal condition is detected, the system quickly determines the type of abnormality, locates the location of the abnormality, automatically generates alarm information, and triggers corresponding corrective or defensive measures.

[0153] Intelligent forecasting technology for distributed photovoltaic power generation that integrates meteorological data, such as Figure 5 As shown, it includes:

[0154] Selection of key meteorological characteristics for photovoltaic power generation:

[0155] The correlation between meteorological elements such as solar radiation intensity, temperature, wind speed, and cloud cover and photovoltaic power generation was analyzed. Key meteorological features that have a significant impact on photovoltaic power generation were screened, and corresponding numerical weather forecast data were obtained based on the region of the photovoltaic power station / module.

[0156] Multi-timescale power prediction model combining photovoltaic holographic monitoring data:

[0157] By integrating minute-level resolution photovoltaic power monitoring data and numerical weather forecast data, and employing machine learning regression algorithms such as Support Vector Regression (SVR) and Gradient Boosting Decision Tree (GBDT), an ultra-short-term photovoltaic power prediction model at the minute / hour level is established to achieve rolling forecasts for the next 0-4 hours. Simultaneously, deep learning time series models, such as Long Short-Term Memory (LSTM) networks and Gated Recurrent Unit (GRU) networks, are used to establish hourly / daily short-term photovoltaic power prediction models to achieve forecasts for the next 1-7 days.

[0158] Photovoltaic power prediction methods:

[0159] An online application of a pre-trained multi-timescale power prediction model is used to regularly update meteorological forecast data and provide rolling predictions of distributed photovoltaic (PV) output curves at the minute, hour, and day levels. Error analysis and reliability assessment are performed on the prediction results, and the prediction model parameters are dynamically adjusted. By combining PV power prediction results with distribution network load forecasts, the trend of renewable energy consumption can be predicted early, providing input for optimized control of the distribution network operation.

[0160] Intelligent control strategies for distribution networks with high penetration rates of distributed photovoltaic (PV) integration, such as... Figure 6 As shown, it includes:

[0161] Multi-timescale collaborative optimization model of source-grid-load-storage:

[0162] Taking into account constraints such as photovoltaic power generation forecast, real-time operation status of distribution network, load forecast curve, and available energy storage capacity, a mathematical model for source-grid-load-storage collaborative optimization is established to form multi-scenario optimization objectives and market trading mechanisms at two time scales: day-ahead and intraday. These objectives include maximizing new energy consumption, minimizing grid losses, minimizing operating costs, and maximizing demand-side response profits.

[0163] Multi-timescale intelligent control methods for distribution networks:

[0164] For day-ahead optimization, guided by multi-scenario optimization objectives across power generation, grid, load, and storage, algorithms such as mixed-integer programming are employed to formulate the optimal day-ahead operation plan for the distribution network, determining distributed photovoltaic output forecasting, distribution network topology pre-construction, energy storage charging and discharging strategies, and interruptible load response strategies. For intraday optimization, a rolling optimization approach is adopted, combining real-time photovoltaic power generation forecasting, real-time energy storage scheduling, and real-time demand-side response measures to correct the day-ahead plan. Technical specifications for the distribution network intelligent controller are formulated to enable the issuance and execution of minute-level real-time control commands.

[0165] Distributed photovoltaic full life cycle operation and maintenance management methods:

[0166] Leveraging holographic monitoring data, the system provides full lifecycle management for connected distributed photovoltaic (PV) power stations, including fault early warning, performance evaluation, and operation and maintenance troubleshooting. Combined with the power station work order system, it forms a closed-loop business process encompassing visualized distributed PV access, alarm linkage processing, and intelligent operation and maintenance decision-making.

[0167] The key points of this invention are: proposing an optimized deployment method for distribution network PMUs in distributed photovoltaic (PV) grid access scenarios, and constructing a "wide-coverage, multi-dimensional, and high-resolution" holographic monitoring system; developing multi-source heterogeneous spatiotemporal data fusion technology to improve the observability and analytical granularity of new energy access in the distribution network; innovatively applying machine learning and deep learning to intelligent identification of abnormal PV operating conditions and multi-timescale power prediction; and pioneering a source-grid-load-storage multi-timescale collaborative optimization model and intelligent control method for distributed PV high-penetration scenarios.

[0168] This invention proposes a method and system for holographic monitoring of distributed photovoltaic (PV) power in distribution networks. By constructing a "wide-coverage, multi-dimensional, and high-resolution" holographic monitoring system, and integrating artificial intelligence technologies such as machine learning and deep learning, it innovatively develops intelligent analysis technologies based on the distribution network power management unit (PMU) for identifying abnormal operating conditions of distributed PV and predicting power at multiple time scales. Based on this, it forms a multi-scenario, multi-time-scale intelligent control strategy for distributed PV access to the distribution network. Compared with traditional technologies, this invention has the following significant advantages:

[0169] 1. Enable panoramic perception of distributed photovoltaic power grid integration, and improve the observability of power grid operation;

[0170] By optimizing the deployment scheme of distribution network PMU, full-coverage monitoring of distributed photovoltaic and its connected area distribution network is achieved. A panoramic database of distribution operation status with multi-temporal and spatial scales and multi-dimensional parameters is constructed. This reveals the intrinsic correlation between distributed photovoltaic power output fluctuations and distribution network power flow distribution, voltage quality, network loss level and other indicators. It effectively characterizes the impact of large-scale new energy access on the distribution network, provides high-quality data support for distribution planning, operation and control, and improves the observability and refined management level of the distribution network.

[0171] 2. Significantly improves the speed and accuracy of identifying abnormal operating conditions in the distribution network, ensuring the safe operation of the distribution network;

[0172] By integrating distribution network PMU monitoring data and machine learning technology, this intelligent identification method for abnormal operating conditions of distributed photovoltaic (PV) power generation is highly automated, real-time, and interpretable. It can quickly and accurately identify and locate abnormal PV power generation behaviors from massive amounts of data, promptly detect potential grid safety hazards, and prevent secondary accidents such as large-scale cascading failures and equipment damage. Compared with traditional manual identification and qualitative analysis methods, this invention can reduce the average identification time of abnormal operating conditions from minutes to seconds, and increase the identification accuracy from 85% to over 95%. This significantly enhances the distribution network's ability to perceive and respond to the impact of distributed PV access, effectively ensuring the safe and stable operation of the distribution network.

[0173] 3. Significantly improves the accuracy of distributed photovoltaic power generation forecasting, laying the foundation for the economical operation of the distribution network;

[0174] By integrating distribution network PMU data, meteorological forecast data, and deep learning technology, a multi-timescale photovoltaic (PV) power generation prediction model at the minute, hour, and day levels is formed. This model fully explores the nonlinear correlation between PV output and meteorological conditions and historical power generation data, improving the ability to predict the volatility and uncertainty of PV output. Compared with traditional PV power prediction methods, this invention reduces the average absolute percentage error of ultra-short-term (minute-level), short-term (hour-level), and medium-term (day-level) predictions by more than 8%, 12%, and 15%, respectively. It can provide high-confidence PV output predictions for day-ahead and intraday optimization scheduling and real-time control of distribution networks, maximizing the absorption of new energy sources while ensuring the safety margin of the distribution network and achieving economical operation of the distribution network.

[0175] 4. Support flexible and intelligent control of the distribution network, and give full play to the advantages of coordinated generation, grid, load and storage;

[0176] By constructing a multi-timescale collaborative optimization model and intelligent control strategy for high-penetration distributed photovoltaic (PV) scenarios, this invention collaboratively optimizes various resources and methods, including PV power generation access, flexible distribution network opening, demand-side response, and flexible energy storage adjustment. It achieves optimal matching of internal and external multi-energy complementarity within the distribution network on a larger spatiotemporal scale, fully leveraging the flexibility of all parties involved (source, grid, load, and storage) and enhancing the overall distribution system's ability to absorb new energy. In typical distribution network simulation analysis and field application pilots, this invention can increase distributed PV penetration by 20%, improve the non-clean energy power generation substitution rate by 15%, and enhance overall energy utilization efficiency by 10%, providing crucial technical support for achieving "dual-carbon" goals and building a clean, low-carbon, safe, and efficient modern energy system.

[0177] 5. Innovate and construct a full life-cycle management approach for distributed photovoltaic power generation to promote the healthy development of the new energy industry;

[0178] Based on holographic monitoring data of power distribution, an innovative full lifecycle management model for distributed photovoltaic (PV) has been formed, encompassing access assessment, fault early warning, performance evaluation, operation and maintenance, and coordinated dispatch. A closed-loop business process of "monitoring-supervision-evaluation-dispatch" has been established to effectively guide the standardized and orderly access of distributed PV, reduce "curtailment" caused by disorderly development, and achieve a dynamic balance between maximizing the benefits of distributed PV and optimizing the efficiency of the power distribution network, thus safeguarding the healthy and sustainable development of the new energy industry.

[0179] In summary, the holographic monitoring method and system for distributed photovoltaic (PV) power grids of this invention are based on a "wide-coverage, multi-dimensional, and high-resolution" holographic monitoring system, with intelligent analysis technologies such as intelligent identification of abnormal operating conditions and multi-timescale power prediction as the core, and with the goal of multi-timescale collaborative optimization of source-grid-load-storage operation, forming an integrated solution for panoramic perception, intelligent analysis, and flexible control of distributed PV access. It is expected to fundamentally revolutionize the planning and design concepts, operation control modes, and market-based trading mechanisms of distribution networks, and play a milestone role in promoting the energy revolution and green and low-carbon development process.

[0180] The embodiments are detailed to support the claims in the specification and enhance the completeness of the technical disclosure.

[0181] In a preferred embodiment, an intelligent control system for a power distribution network is provided, comprising: a monitoring unit, a data processing unit, an anomaly identification unit, and a control response unit, wherein the anomaly identification unit has the following specific structure and function:

[0182] The system consists of the following components and functional modules:

[0183] 1. Monitoring Unit:

[0184] Through distribution network PMU devices deployed at distributed photovoltaic access points, energy storage grid connection nodes, and key power equipment;

[0185] Real-time acquisition of multi-dimensional electrical parameter data such as voltage, current, active power, reactive power, frequency, and harmonics;

[0186] The sampling frequency is set to 1 to 10 times per second, and dynamically adjusted according to the actual control accuracy requirements.

[0187] 2. Data Processing Unit:

[0188] The raw time-series data from the PMU is cleaned, denoised, and standardized.

[0189] Construct a second-level raw sequence window data X1 and a minute-level statistical feature vector X2, which are then input into the anomaly detection model.

[0190] The statistical characteristics include, but are not limited to, mean, standard deviation, skewness, spectral energy, and signal-to-noise ratio.

[0191] 3. Anomaly Detection Unit (DFNet Module):

[0192] It includes a neural network structure consisting of two parts: Channel 1 (sequence channel): X1 is input into a one-dimensional convolutional neural network and an LSTM network to extract short-term dynamic features; Channel 2 (statistical channel): X2 is input into a multilayer perceptron network (MLP) to model long-term running trends;

[0193] The outputs of the two channels are spliced ​​and fused, and the corresponding running status labels are output through the Softmax classifier.

[0194] The abnormal operating conditions that can be identified include: sudden drop in photovoltaic power generation; voltage exceeding limits (overvoltage or undervoltage); excessive harmonic content; suspected islanding behavior; and other custom extended abnormal types.

[0195] 4. Control and Response Unit:

[0196] Receive the output of the anomaly detection unit;

[0197] If the identification result is abnormal, an alarm message will be automatically triggered;

[0198] The distribution network control strategy module is invoked to execute the corresponding processing logic, such as:

[0199] Adjust the charging and discharging status of the energy storage system;

[0200] Removal can interrupt the load;

[0201] Modify the power factor settings of the photovoltaic inverter;

[0202] Notify the operations and maintenance platform to conduct on-site investigation.

[0203] Taking the power distribution network of a large industrial park as an example, the park has a distributed photovoltaic power station with an installed capacity of 50MW, distributed across the rooftops of 10 factory buildings, totaling more than 1,000 photovoltaic power generation units. It also has 5 centralized energy storage systems with a total capacity of 10MWh. The park's power distribution network consists of one 110kV main substation, five 10kV switching stations, 20 10kV distribution lines, and 50 10kV / 0.4kV distribution transformers. The load type is mainly industrial electricity.

[0204] 1. Distribution network PMU deployment and data acquisition:

[0205] A distribution network PMU is installed at each distributed photovoltaic grid-connected point and energy storage grid-connected point to measure in real time the active and reactive power, DC side voltage and current of each photovoltaic power generation unit, as well as the active and reactive power and charging / discharging status of the energy storage system. The data acquisition frequency is once per second. Simultaneously, distribution network PMUs are also installed at key nodes such as the 110kV main substation and 10kV switching stations in the park to monitor indicators such as voltage, current, frequency, active power, and reactive power at the beginning and end of the lines. The data sampling frequency is per electrical angle. The distribution network PMUs communicate in real time with the park's dispatch master station via a fiber optic ring network.

[0206] 2. Multi-source heterogeneous data fusion analysis:

[0207] A power distribution big data platform based on Kafka, Spark, and HDFS was built to aggregate PMU data from the park's power distribution network, as well as operational monitoring data from equipment such as photovoltaic inverters, energy storage controllers, and smart meters. Numerical weather forecast data for the next 168 hours within the park, provided by the meteorological department, was also incorporated, including multi-dimensional meteorological elements such as solar radiation intensity, temperature, wind speed, and humidity.

[0208] The power distribution big data platform performs preprocessing such as cleaning and normalization on the aggregated multi-source heterogeneous data to form a unified spatiotemporal data format. Using 1 minute, 10 minutes, and 1 hour as time scales, and photovoltaic power generation units, distribution lines, switching stations, and the entire park as spatial scales, it constructs a "one station, one file" panoramic power distribution information database.

[0209] 3. Intelligent identification of abnormal operating conditions in distributed photovoltaic systems:

[0210] Based on holographic monitoring data, typical abnormal operating conditions of distributed photovoltaic power generation in the park were summarized, including: sudden drop in local power generation caused by shading, output power oscillation caused by string-level hot spots, and photovoltaic output limitation due to continuous high temperature of the inverter. 500 samples of various abnormal operating conditions were manually labeled.

[0211] A multi-scale PMU data feature extraction method was employed to automatically construct a feature set reflecting photovoltaic anomalies, including power surge slope, fluctuation duration, and frequency abrupt change amplitude. Anomalies were classified using a random forest algorithm, and model parameters were optimized through cross-validation, achieving a classification accuracy of 98%.

[0212] The trained random forest model is deployed on the power distribution big data platform to perform real-time anomaly identification on the aggregated PMU data. When a suspected abnormal data point is detected, the abnormal operating condition type is determined, an abnormal alarm message is generated and pushed to the power distribution dispatch intelligent decision interface, and the abnormal point is accurately located through the power distribution geographic information system.

[0213] 4. Multi-timescale prediction of distributed photovoltaic power generation:

[0214] For ultra-short timescales (10 minutes to 4 hours), historical photovoltaic power generation data aggregated over 10 minutes is used as input, and meteorological numerical forecast data is used as auxiliary input. A support vector regression model is employed for rolling forecasts. The optimal parameters of the model are obtained through grid optimization, and the distributed photovoltaic power generation in the industrial park is predicted for 30 consecutive days, with the average absolute percentage error controlled within 6%.

[0215] For short-term timescales (1 to 7 days), using hourly aggregated historical photovoltaic power generation data as input and meteorological numerical forecasts as auxiliary input, a long short-term memory neural network model is employed for prediction. Transfer learning techniques are used to pre-train the model with data from photovoltaic power plants in other similar climate regions, improving prediction accuracy. The prediction error for the park's photovoltaic power generation over the next 168 hours has been reduced to within 8%.

[0216] 5. Collaborative optimization and scheduling across multiple scenarios involving power generation, grid, load, and storage:

[0217] Based on distributed photovoltaic power generation forecasting, combined with industrial load demand forecasting for industrial parks, thermal power unit output adjustment costs, and energy storage battery charging and discharging characteristics, a multi-scenario collaborative optimization model of "source-grid-load-storage" is constructed. A multi-timescale optimization scheduling strategy is designed with the objectives of maximizing renewable energy consumption, minimizing power supply costs, and optimizing power quality.

[0218] For the day-ahead time scale, with an hourly time step, a mixed-integer linear programming method is used to obtain the optimal operation plan of the distribution network in the park for the day-ahead 24 hours, including the ideal output of each photovoltaic power station, the charging and discharging power of the energy storage system, and the demand response plan for interruptible loads. For the intraday scale, with a 10-minute time step, a rolling optimization strategy is used, combined with the ultra-short-term photovoltaic power forecast results, to continuously revise the day-ahead plan of the power grid, and form real-time dispatch instructions for the energy storage system and interruptible loads.

[0219] By deploying the system and method of this invention, the power distribution network in the park has achieved remarkable results: the utilization efficiency of photovoltaic power generation has increased by 30% compared with the traditional management model, and the curtailment rate has decreased from 5% to less than 1%; the photovoltaic penetration rate in the area has increased from 20% to 35%, maximizing the absorption of renewable energy while ensuring the safe and stable operation of the power grid; by tapping the potential of demand-side response, the average electricity price for users in the park has decreased by 0.15 yuan / kWh, and the operating cost has been reduced by 15%.

[0220] This embodiment fully demonstrates that the distributed photovoltaic holographic monitoring system and intelligent control technology of the present invention can significantly improve the sensing, analysis, and control capabilities of new energy access to the distribution network. While ensuring the safe and economical operation of the distribution network, it maximizes the absorption of clean energy, balancing grid benefits, user benefits, and environmental benefits, providing important technical support for achieving the "dual carbon" goal. The technical solution proposed in this invention has good prospects for widespread application.

[0221] Example 2:

[0222] This invention also proposes a smart distribution network control system 200 for distributed photovoltaic systems, comprising:

[0223] The monitoring unit 201 is used to perform holographic monitoring of distributed photovoltaic and energy storage based on the distribution network PWU for the entire distribution network and to acquire holographic monitoring data.

[0224] The identification and prediction unit 202 is used to identify the holographic monitoring data and predict the distributed photovoltaic power generation based on the holographic monitoring data, and generate identification results and prediction results.

[0225] The control unit 203 is used to generate an intelligent control strategy for the distribution network based on the identification results and prediction results, and to perform intelligent control on the distribution network using the intelligent control strategy.

[0226] Among them, holographic monitoring of distributed photovoltaic and energy storage based on distribution network PWU includes:

[0227] Based on the distribution network topology and the access locations of distributed photovoltaic and energy storage in the distribution network, the layout of the distribution network PMU is adjusted to achieve full-coverage holographic monitoring of distributed photovoltaic and energy storage.

[0228] Acquiring holographic monitoring data includes:

[0229] Set the acquisition frequency of the distribution network PMU, and collect multi-dimensional parameter data of distributed photovoltaic and energy storage and operating status data of distribution network equipment in real time under the set acquisition frequency. Then, fuse the multi-dimensional parameter data and the operating status data to obtain holographic monitoring data.

[0230] The identification of the holographic monitoring data includes:

[0231] The holographic monitoring data is input into the abnormal operating condition intelligent identification model, and the abnormal operating condition intelligent identification model identifies the holographic monitoring data to determine suspected abnormal operating conditions in the power distribution network.

[0232] Among them, abnormal operating conditions are defined as follows: based on the actual operating experience of distributed photovoltaic power grid access, sudden drop in photovoltaic power generation, voltage exceeding the limit, harmonic exceedance, and islanding operation are defined as abnormal operating conditions.

[0233] The identification of holographic monitoring data using the aforementioned abnormal operating condition intelligent identification model includes:

[0234] Using the aforementioned intelligent identification model for abnormal operating conditions, abnormal features of photovoltaic power generation are extracted from the holographic monitoring data. These abnormal features are then identified to determine suspected abnormal operating conditions in the power distribution network.

[0235] Among them, the abnormal characteristics of photovoltaic power generation include:

[0236] Abnormal power fluctuation rate, abnormal voltage deviation rate, and abnormal harmonic content.

[0237] The prediction of distributed photovoltaic power generation includes:

[0238] Based on meteorological characteristics, a multi-timescale power prediction model is constructed, and based on the power prediction model, predictions are made using holographic monitoring data.

[0239] Among them, the generation of intelligent control strategies for the distribution network includes:

[0240] Based on the constraints of photovoltaic power generation forecast, distribution network real-time operation status, load forecast curve, and available energy storage capacity, a mathematical model for the coordinated optimization of power generation, grid, load, and storage is established. Based on the mathematical model, multi-scenario optimization objectives and market trading mechanisms are generated at two time scales: day-ahead and intraday.

[0241] For day-ahead optimization, guided by the multi-scenario optimization objectives of sources, grids, loads and storage, algorithms such as mixed integer programming are used to formulate the day-ahead optimal operation plan for the distribution network, and determine the distributed photovoltaic output forecast, distribution network topology pre-construction, energy storage charging and discharging strategy and interruptible load response strategy.

[0242] For intraday optimization, a rolling optimization method is adopted, which is based on real-time forecasting of photovoltaic power generation, real-time scheduling of energy storage and real-time response of demand side measures to correct the day-ahead plan and formulate a smart control strategy for the distribution network.

[0243] This invention provides high-quality data support for in-depth data analysis and intelligent control of power distribution networks.

[0244] Example 3:

[0245] Based on the same inventive concept, this invention also provides a computer device, which includes a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to implement corresponding method flows or corresponding functions, thereby implementing the steps of the methods in the above embodiments.

[0246] Example 4:

[0247] Based on the same inventive concept, this invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the steps of the method in the above embodiments.

[0248] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0249] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0250] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0251] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0252] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0253] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for intelligent control of distribution networks for distributed photovoltaic power generation, characterized in that, The intelligent control method includes: For the entire distribution network, holographic monitoring of distributed photovoltaic and energy storage is carried out based on the distribution network PWU to obtain holographic monitoring data; The holographic monitoring data is identified, and based on the holographic monitoring data, the distributed photovoltaic power generation is predicted, generating identification results and prediction results; Based on the identification and prediction results, an intelligent control strategy for the distribution network is generated, and the distribution network is intelligently controlled using the intelligent control strategy.

2. The intelligent control method according to claim 1, characterized in that, The holographic monitoring of distributed photovoltaic and energy storage based on distribution network PWU includes: Based on the distribution network topology and the access locations of distributed photovoltaic and energy storage in the distribution network, the layout of the distribution network PMU is adjusted to achieve full-coverage holographic monitoring of distributed photovoltaic and energy storage.

3. The intelligent control method according to claim 1, characterized in that, The acquisition of holographic monitoring data includes: Set the acquisition frequency of the distribution network PMU, and collect multi-dimensional parameter data of distributed photovoltaic and energy storage and operating status data of distribution network equipment in real time under the set acquisition frequency. Then, fuse the multi-dimensional parameter data and the operating status data to obtain holographic monitoring data.

4. The intelligent control method according to claim 1, characterized in that, Identifying the holographic monitoring data includes: The holographic monitoring data is input into the abnormal operating condition intelligent identification model, and the abnormal operating condition intelligent identification model identifies the holographic monitoring data to determine suspected abnormal operating conditions in the power distribution network.

5. The intelligent control method according to claim 4, characterized in that, The abnormal operating conditions are specifically defined as follows: based on the actual operating experience of distributed photovoltaic power grid connection, sudden drop in photovoltaic power generation, voltage exceeding limits, harmonic exceedance, and islanding operation are defined as abnormal operating conditions.

6. The intelligent control method according to claim 4, characterized in that, The abnormal operating condition intelligent identification model is used to identify holographic monitoring data, including: Using the aforementioned intelligent identification model for abnormal operating conditions, abnormal features of photovoltaic power generation are extracted from the holographic monitoring data. These abnormal features are then identified to determine suspected abnormal operating conditions in the power distribution network.

7. The intelligent control method according to claim 6, characterized in that, The abnormal characteristics of photovoltaic power generation include: Abnormal power fluctuation rate, abnormal voltage deviation rate, and abnormal harmonic content.

8. The intelligent control method according to claim 1, characterized in that, The prediction of distributed photovoltaic power generation includes: Based on meteorological characteristics, a multi-timescale power prediction model is constructed, and predictions are made based on holographic monitoring data.

9. The intelligent control method according to claim 1, characterized in that, Generate intelligent control strategies for distribution networks, including: Based on the constraints of photovoltaic power generation forecast, distribution network real-time operation status, load forecast curve, and available energy storage capacity, a mathematical model for the coordinated optimization of power generation, grid, load, and storage is established. Based on the mathematical model, multi-scenario optimization objectives and market trading mechanisms are generated at two time scales: day-ahead and intraday. For day-ahead optimization, guided by the multi-scenario optimization objectives of sources, grids, loads and storage, algorithms such as mixed integer programming are used to formulate the day-ahead optimal operation plan for the distribution network, and determine the distributed photovoltaic output forecast, distribution network topology pre-construction, energy storage charging and discharging strategy and interruptible load response strategy. For intraday optimization, a rolling optimization method is adopted, which is based on real-time forecasting of photovoltaic power generation, real-time scheduling of energy storage and real-time response of demand side measures to correct the day-ahead plan and formulate a smart control strategy for the distribution network.

10. A smart control system for distribution networks oriented towards distributed photovoltaic power, characterized in that, The intelligent control system includes: The monitoring unit is used to perform holographic monitoring of distributed photovoltaic and energy storage based on the distribution network PWU for the entire distribution network and to acquire holographic monitoring data; The identification and prediction unit is used to identify the holographic monitoring data and predict the distributed photovoltaic power generation based on the holographic monitoring data, and generate identification results and prediction results. The control unit is used to generate an intelligent control strategy for the distribution network based on the identification results and prediction results, and to perform intelligent control of the distribution network using the intelligent control strategy.

11. The intelligent control system according to claim 10, characterized in that, The holographic monitoring of distributed photovoltaic and energy storage based on distribution network PWU includes: Based on the distribution network topology and the access locations of distributed photovoltaic and energy storage in the distribution network, the layout of the distribution network PMU is adjusted to achieve full-coverage holographic monitoring of distributed photovoltaic and energy storage.

12. The intelligent control system according to claim 10, characterized in that, The acquisition of holographic monitoring data includes: Set the acquisition frequency of the distribution network PMU, and collect multi-dimensional parameter data of distributed photovoltaic and energy storage and operating status data of distribution network equipment in real time under the set acquisition frequency. Then, fuse the multi-dimensional parameter data and the operating status data to obtain holographic monitoring data.

13. The intelligent control system according to claim 10, characterized in that, Identifying the holographic monitoring data includes: The holographic monitoring data is input into the abnormal operating condition intelligent identification model, and the abnormal operating condition intelligent identification model identifies the holographic monitoring data to determine suspected abnormal operating conditions in the power distribution network.

14. The intelligent control system according to claim 13, characterized in that, The abnormal operating conditions are specifically defined as follows: based on the actual operating experience of distributed photovoltaic power grid connection, sudden drop in photovoltaic power generation, voltage exceeding limits, harmonic exceedance, and islanding operation are defined as abnormal operating conditions.

15. The intelligent control system according to claim 13, characterized in that, The abnormal operating condition intelligent identification model is used to identify holographic monitoring data, including: Using the aforementioned intelligent identification model for abnormal operating conditions, abnormal features of photovoltaic power generation are extracted from the holographic monitoring data. These abnormal features are then identified to determine suspected abnormal operating conditions in the power distribution network.

16. The intelligent control system according to claim 15, characterized in that, The abnormal characteristics of photovoltaic power generation include: Abnormal power fluctuation rate, abnormal voltage deviation rate, and abnormal harmonic content.

17. The intelligent control system according to claim 16, characterized in that, The prediction of distributed photovoltaic power generation includes: Based on meteorological characteristics, a multi-timescale power prediction model is constructed, and predictions are made based on holographic monitoring data.

18. The intelligent control system according to claim 10, characterized in that, Generate intelligent control strategies for distribution networks, including: Based on the constraints of photovoltaic power generation forecast, distribution network real-time operation status, load forecast curve, and available energy storage capacity, a mathematical model for the coordinated optimization of power generation, grid, load, and storage is established. Based on the mathematical model, multi-scenario optimization objectives and market trading mechanisms are generated at two time scales: day-ahead and intraday. For day-ahead optimization, guided by the multi-scenario optimization objectives of sources, grids, loads and storage, algorithms such as mixed integer programming are used to formulate the day-ahead optimal operation plan for the distribution network, and determine the distributed photovoltaic output forecast, distribution network topology pre-construction, energy storage charging and discharging strategy and interruptible load response strategy. For intraday optimization, a rolling optimization method is adopted, which is based on real-time forecasting of photovoltaic power generation, real-time scheduling of energy storage and real-time response of demand side measures to correct the day-ahead plan and formulate a smart control strategy for the distribution network.

19. A computer device, characterized in that, include: One or more processors; A processor is used to execute one or more programs; When the one or more programs are executed by the one or more processors, the method described in any one of claims 1-9 is implemented.

20. A computer-readable storage medium, characterized in that, It contains a computer program, which, when executed, implements the method as described in any one of claims 1-9.

Images