Photovoltaic low-carbon park smart energy management system based on digital twinning

Abstract

The invention relates to a photovoltaic low-carbon park smart energy management system and method based on digital twinning. Accurate mapping and simulation analysis of a park energy system are realized by constructing a digital twinborn model, dynamic optimization scheduling of energy equipment is realized by adopting a multi-agent cooperative control algorithm, an energy optimization strategy is formulated by taking low carbon as a target, the health state of the equipment is predicted by utilizing machine learning, and an intelligent operation and maintenance plan is generated. And decision support is provided through energy big data analysis. The system integrates multi-source data, and efficient utilization and low-carbon operation of park energy are realized through the steps of digital twin modeling, intelligent agent collaboration, low-carbon scheduling, equipment prediction, big data analysis and the like. Practical application shows that the system can improve the photovoltaic efficiency by 8%, reduce the energy cost by 18%, reduce the carbon emission intensity by 30%, and significantly improve the park energy management level and the low-carbon degree.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of photovoltaic application and low-carbon energy management, in particular to a photovoltaic low-carbon park intelligent energy management system and method based on digital twinning, which can realize efficient utilization and low-carbon operation of park photovoltaic energy and is suitable for scenes such as industrial parks, commercial complexes and residential areas. BACKGROUND

[0002] Insufficient photovoltaic energy fluctuation processing capability: traditional park energy management systems are difficult to effectively cope with the intermittency and volatility of photovoltaic power generation. Due to factors such as weather changes and day-night alternation, photovoltaic power generation is unstable, which can easily lead to problems such as grid voltage fluctuation and frequency deviation, affecting power supply quality. At the same time, when photovoltaic power generation is excessive, there is a lack of efficient energy storage and deployment mechanism, resulting in energy waste; and when photovoltaic power generation is insufficient, energy cannot be supplemented in time, leading to tight power supply in the park.

[0003] Low level of energy system collaborative optimization: the energy system in the park is complex and includes photovoltaic, energy storage, power grid, air conditioning, lighting and other subsystems, but the existing management system runs independently in each subsystem and lacks effective collaborative control. For example, the energy storage system and the photovoltaic power generation system cannot achieve dynamic matching, and air conditioning, lighting and other load devices cannot be intelligently adjusted according to the energy supply situation, resulting in low overall energy utilization efficiency and increased operating costs.

[0004] Lack of low-carbon operation decision support: in terms of low-carbon operation of the park, the existing system cannot accurately assess the carbon emissions of energy consumption and lacks intelligent decision support for low-carbon energy replacement strategies. It is difficult to dynamically adjust the energy structure and operation strategy according to real-time energy data and carbon emission targets, which limits the low-carbon development of the park.

[0005] Energy equipment state monitoring and maintenance lag: there are a large number of energy equipment such as photovoltaic equipment and energy storage devices in the park, which are widely distributed. Traditional monitoring methods mainly rely on manual inspection, which is difficult to monitor the equipment running state in real time, and the fault discovery and processing are not timely, which affects the service life of the equipment and the stability of energy supply. At the same time, there is a lack of prediction ability for the health status of the equipment, which cannot be maintained and replaced in advance, increasing the operation and maintenance cost. SUMMARY

[0006] The purpose of the present application is to provide a photovoltaic low-carbon park intelligent energy management system based on digital twinning, comprising:

[0007] A photovoltaic park digital twinning modeling module acquires park physical entity data through three-dimensional laser scanning and sensor networks, constructs and synchronizes a virtual model of the energy system in real time;

[0008] A multi-agent collaborative control module, which abstracts energy devices in the park as agents, realizes collaborative optimization operation of the energy system through a distributed collaborative mechanism;

[0009] A low-carbon energy optimization scheduling module, which establishes an optimization model and solves an optimal energy scheduling scheme with the goal of reducing carbon emissions;

[0010] An equipment health state prediction and intelligent operation and maintenance module, which deploys a sensor network to collect equipment data, uses machine learning algorithms to predict equipment states and generates operation and maintenance plans;

[0011] An energy big data analysis and decision support module, which integrates multi-source energy data, conducts in-depth analysis, and provides visual decision support.

[0012] Further, the virtual model established by the digital twin modeling module includes the physical structure, electrical characteristics, and operating state of the photovoltaic array, energy storage system, power grid, and load equipment, and is synchronized with real-time data of the physical system.

[0013] Further, the agents in the multi-agent collaborative control module include photovoltaic agents, energy storage agents, load agents, and power grid agents, which exchange information through a communication network and make autonomous decisions.

[0014] Further, the low-carbon energy optimization scheduling module takes the minimum carbon emissions as the objective function, considers factors such as energy cost and equipment constraints, and uses mixed integer programming or genetic algorithms to solve the scheduling scheme.

[0015] Further, the equipment health state prediction and intelligent operation and maintenance module uses algorithms such as random forest and LSTM to establish equipment performance degradation models and fault prediction models.

[0016] Further, the energy big data analysis and decision support module integrates multi-source energy data and uses data mining techniques such as association analysis and clustering analysis for in-depth analysis.

[0017] Further, the photovoltaic low-carbon park smart energy management method includes the following steps:

[0018] Construct a digital twin model of the photovoltaic park to synchronize data between the physical system and the virtual model in real time;

[0019] Abstract energy devices in the park as agents and realize collaborative control of the energy system through a distributed collaborative mechanism;

[0020] Establish an energy optimization scheduling model and solve an optimal scheduling scheme with the goal of reducing carbon emissions;

[0021] Deploy a sensor network to collect equipment data, use machine learning algorithms to predict equipment health states, and generate operation and maintenance plans;

[0022] It integrates multi-source energy data, conducts in-depth analysis, and provides visual decision support.

[0023] Furthermore, in the step of constructing the digital twin model, geometric information of the park is obtained through three-dimensional laser scanning, and operating parameters are collected using a sensor network to establish a virtual model that includes the physical characteristics and electrical parameters of the equipment.

[0024] Furthermore, in the multi-agent collaborative control step, each agent makes autonomous decisions based on global energy goals and local information, and adjusts the operating status of energy equipment through negotiation and cooperation.

[0025] Furthermore, in the low-carbon energy optimization scheduling step, time series analysis, neural networks, and other algorithms are used to predict photovoltaic power generation and load demand as inputs to the optimization model.

[0026] The present invention has the following beneficial effects:

[0027] (I) Digital Twin Modeling Technology for Photovoltaic Industrial Parks

[0028] This invention proposes a digital twin modeling technology for photovoltaic (PV) parks, establishing a real-time mapping relationship between the physical park and a virtual model. Geometric information and operational parameters of the park's physical entities are acquired through technologies such as 3D laser scanning and sensor networks. A virtual model of the park's energy system is then created using digital twin technology. This model not only includes the physical structure and electrical characteristics of equipment such as PV arrays, energy storage systems, and power grids, but also reflects the real-time operating status of the equipment and energy flow. Through the digital twin model, simulation analysis, optimized scheduling, and fault prediction of the park's energy system can be performed, improving the precision of energy management.

[0029] (II) Multi-agent cooperative control algorithm

[0030] To address the problem of coordinated optimization in energy systems, this invention designs a multi-agent cooperative control algorithm. The photovoltaic system, energy storage system, and load equipment within the park are abstracted as multiple agents, each with autonomous decision-making and communication capabilities. Through a distributed cooperative mechanism, the agents exchange information in real time and dynamically adjust their behavior based on global energy goals and local operating states. For example, the photovoltaic agent adjusts its power generation strategy based on solar irradiance forecasts, the energy storage agent controls charging and discharging based on energy supply and demand, and the load agent optimizes its electricity consumption plan based on electricity prices and energy supply conditions. Through multi-agent cooperation, the overall optimized operation of the park's energy system is achieved.

[0031] (III) Low-carbon energy optimization dispatch strategy

[0032] This invention proposes a low-carbon energy optimization scheduling strategy aimed at reducing carbon emissions in industrial parks, comprehensively considering factors such as energy costs and equipment constraints. A low-carbon energy optimization model is established to coordinate and optimize energy sources such as photovoltaic power generation, energy storage systems, and grid power purchases with load demands from air conditioning, lighting, and production equipment. Prediction technology is used to forecast photovoltaic power generation and load demand in real time, and the carbon emissions of different energy options are calculated by combining these forecasts with carbon emission factors. An optimization algorithm is then used to solve for the optimal energy scheduling scheme, minimizing carbon emissions while meeting the park's energy needs and promoting the park's low-carbon development.

[0033] (iv) Equipment Health Status Prediction and Intelligent Operation and Maintenance System

[0034] To address the problem of lagging condition monitoring and maintenance of energy equipment, this invention develops an equipment health status prediction and intelligent operation and maintenance system. A sensor network is deployed on key equipment such as photovoltaic devices and energy storage devices to collect real-time operating parameters (e.g., temperature, voltage, current) and environmental data (e.g., light intensity, humidity). Machine learning algorithms are used to establish an equipment health status prediction model to predict the aging degree and failure probability of the equipment. Based on the prediction results, an intelligent operation and maintenance plan is generated to schedule equipment maintenance and replacement in advance, reducing equipment failure rates and extending equipment lifespan. Simultaneously, the system can automatically identify equipment faults and issue alarms, guiding maintenance personnel to quickly locate and handle faults, improving operation and maintenance efficiency.

[0035] (V) Energy Big Data Analysis and Decision Support Platform

[0036] This invention constructs an energy big data analysis and decision support platform to deeply mine and analyze energy data in industrial parks. It integrates multi-source data, including photovoltaic power generation data, energy storage data, load data, and power grid data, to establish an energy big data center. Utilizing data mining and machine learning technologies, it analyzes energy consumption patterns, load characteristics, and equipment operating efficiency, providing data support for energy optimization scheduling, equipment operation and maintenance, and low-carbon strategy formulation. Simultaneously, a visual decision-making interface is developed to present the analysis results to managers in intuitive charts and reports, assisting them in making scientific decisions and improving the park's energy management level. Attached Figure Description

[0037] Figure 1 Overall process flow diagram. Detailed Implementation

[0038] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0039] Example 1

[0040] (I) Equipment and System Preparation

[0041] Hardware equipment includes photovoltaic arrays, energy storage systems (such as lithium battery packs), smart meters, sensor networks (including temperature sensors, light sensors, voltage sensors, etc.), edge computing devices, and communication equipment deployed within the park. The photovoltaic arrays utilize high-efficiency monocrystalline silicon solar panels, with the total installed capacity determined based on the park's electricity demand. The energy storage system capacity is configured to be 30%-50% of the peak photovoltaic power generation. The sensor network covers all key energy equipment and nodes within the park, enabling real-time data acquisition.

[0042] Software System: A cloud-based smart energy management platform is built, including digital twin modeling software, multi-agent collaborative control software, low-carbon energy optimization and scheduling software, equipment health status prediction software, and energy big data analysis software. System applications are developed using programming languages ​​such as Python and Java, integrating machine learning frameworks such as TensorFlow and PyTorch to support algorithm execution. The Neo4j graph database is used to store digital twin model data, and the InfluxDB time-series database is used to store real-time energy data.

[0043] (II) Process Steps

[0044] Construction of digital twin model of photovoltaic park

[0045] Data Acquisition: Geometric information such as the terrain and building layout of the park is obtained through a 3D laser scanner, and the operating parameters and environmental data of photovoltaic equipment, energy storage system, load equipment are collected in real time using a sensor network.

[0046] Model Building: Based on the collected data, a virtual model of the park's energy system is built using digital twin modeling software, including a photovoltaic array model, an energy storage model, a power grid model, and a load model. The model incorporates the physical characteristics, electrical parameters, and control logic of the equipment to achieve an accurate mapping of the physical system.

[0047] Model synchronization: Real-time data synchronization between the virtual model and the physical system is achieved through a data interface, ensuring that the virtual model accurately reflects the operating status of the physical system. Simulation technology is used to verify and optimize the model, improving its accuracy and reliability.

[0048] Multi-agent cooperative control implementation

[0049] Intelligent agent design: The photovoltaic system, energy storage system, air conditioning system, lighting system and other systems in the park are abstracted into different types of intelligent agents. Each intelligent agent includes a decision-making module, a communication module and an execution module.

[0050] Information exchange: Agents exchange information in real time through a communication network, including energy production data, storage status, and load demand. A distributed collaborative mechanism is established, enabling agents to make autonomous decisions based on global energy goals and local information.

[0051] Collaborative optimization: When energy supply and demand are unbalanced, each intelligent agent adjusts its behavior through negotiation and cooperation. For example, when photovoltaic power generation is in surplus, the energy storage agent increases its charging power, and the load agent appropriately increases unnecessary loads; when photovoltaic power generation is insufficient, the energy storage agent releases energy, the load agent reduces unnecessary loads, and at the same time, the grid agent increases its power purchase capacity, thereby achieving collaborative optimization of the energy system.

[0052] Low-carbon energy optimization dispatch execution

[0053] Data forecasting: Utilizing historical data and weather forecast information, forecasts are made regarding photovoltaic power generation and load demand. Time series analysis and neural network algorithms are employed to improve forecast accuracy, providing a reliable basis for energy dispatch.

[0054] Model establishment: Taking the minimum carbon emissions as the objective function, and considering constraints such as energy cost, equipment charging and discharging efficiency, and power balance, a low-carbon energy optimization scheduling model is established.

[0055] Optimization Solution: The model is solved using optimization algorithms such as mixed-integer programming and genetic algorithms to obtain the optimal energy dispatch scheme. The scheme includes photovoltaic power output plans, energy storage charging and discharging strategies, and grid power purchase plans.

[0056] Implementation and Adjustment of the Plan: Control the operation of energy equipment according to the optimized plan and monitor the status of the energy system in real time. When the actual operation does not match the forecast, adjust the scheduling plan in a timely manner to ensure that the energy system always operates in a low-carbon and efficient state.

[0057] Equipment health status prediction and intelligent operation and maintenance

[0058] Data acquisition and preprocessing: Real-time acquisition of equipment operating parameters and environmental data through sensor networks, followed by preprocessing such as cleaning and filtering to remove noise and outliers.

[0059] Model training: Use historical data to train the equipment health status prediction model, and use machine learning algorithms such as random forest and long short-term memory network (LSTM) to establish equipment performance degradation model and fault prediction model.

[0060] Condition assessment and prediction: Input the real-time collected data into the trained model to assess the health status of the equipment and predict the remaining service life and failure probability of the equipment.

[0061] Operation and Maintenance Decision-Making and Execution: Based on equipment health status assessment and prediction results, an intelligent operation and maintenance plan is generated, including maintenance time, maintenance content, and replacement parts. When a fault warning is issued, operation and maintenance personnel are promptly notified to handle the situation and ensure the safe and reliable operation of the equipment.

[0062] Energy Big Data Analysis and Decision Support

[0063] Data Integration and Storage: Data from multiple sources, including photovoltaic power generation, energy storage, load, power grid, and meteorological data, is integrated and stored in the energy big data center. Distributed storage and data indexing technologies are employed to improve data storage and retrieval efficiency.

[0064] Data Analysis and Mining: Utilizing data mining algorithms to analyze big data on energy, uncovering knowledge such as energy consumption patterns, load characteristics, and equipment operating efficiency. For example, correlation analysis can identify key factors affecting photovoltaic power generation efficiency, and cluster analysis can categorize loads into different types and formulate corresponding control strategies.

[0065] Decision Support and Visualization: Based on data analysis results, provide decision support for park energy management, such as energy optimization scheduling suggestions, equipment upgrade and renovation plans, and low-carbon development strategies. Develop a visual decision-making interface to present the analysis results to managers in intuitive charts, reports, and other formats, assisting them in making informed decisions.

[0066] (III) Specific Case Analysis

[0067] Case 1: After implementing the smart energy management system of this invention, an industrial park monitored the photovoltaic array's operating status in real time using a digital twin model. It was discovered that some photovoltaic modules were experiencing a decrease in power generation efficiency due to dust accumulation. The system promptly generated cleaning work orders, guiding maintenance personnel to perform cleaning, which improved the photovoltaic array's power generation efficiency by 8%. Simultaneously, a multi-agent collaborative control algorithm was used to achieve dynamic matching between photovoltaic, energy storage, and load. During periods of sufficient sunlight, the energy storage system promptly stores excess electricity; during peak electricity consumption periods, it releases stored energy to meet some load demand, reducing grid electricity purchase costs. Statistics show that the park's daily grid electricity purchases decreased by 25%, and energy costs decreased by 18%.

[0068] Case Study 2: After implementing the low-carbon energy optimization scheduling strategy of this invention, a commercial complex could plan energy dispatch schemes in advance based on weather forecasts and historical load data. On sunny days with ample sunlight, photovoltaic power generation was prioritized to meet internal load demands, and excess energy was stored in an energy storage system. On cloudy or rainy days or at night, energy storage discharge and grid power purchases were rationally scheduled. Through this optimized scheduling, the commercial complex reduced its carbon emission intensity by 30%, resulting in an annual reduction of approximately 500 tons of carbon dioxide emissions. Simultaneously, the equipment health status prediction and intelligent operation and maintenance system predicted the performance degradation of some battery modules in the energy storage system in advance, allowing for timely replacement and preventing energy storage system failures, extending equipment lifespan, and reducing operation and maintenance costs.

[0069] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A smart energy management system for photovoltaic low-carbon industrial parks based on digital twins, characterized in that, include: The S1 photovoltaic park digital twin modeling module acquires physical entity data of the park through 3D laser scanning and sensor networks, and constructs and synchronizes a virtual model of the energy system in real time. The S2 multi-agent collaborative control module abstracts energy equipment in the park into intelligent agents and realizes the collaborative optimization of the energy system through a distributed collaborative mechanism. The S3 low-carbon energy optimization scheduling module aims to reduce carbon emissions, establishes an optimization model, and solves for the optimal energy scheduling scheme. The S4 device health status prediction and intelligent operation and maintenance module deploys a sensor network to collect device data, uses machine learning algorithms to predict device status and generate operation and maintenance plans. The S5 Energy Big Data Analysis and Decision Support Module integrates multi-source energy data, performs in-depth analysis, and provides visualized decision support.

2. The system according to claim 1, characterized in that, The virtual model established by the digital twin modeling module includes the physical structure, electrical characteristics, and operating status of the photovoltaic array, energy storage system, power grid, and load equipment, and is synchronized with the real-time data of the physical system.

3. The system according to claim 1, characterized in that, The agents in the multi-agent collaborative control module include photovoltaic agents, energy storage agents, load agents, and grid agents. Each agent exchanges information and makes autonomous decisions through a communication network.

4. The system according to claim 1, characterized in that, The low-carbon energy optimization scheduling module takes the minimum carbon emissions as the objective function, considers factors such as energy costs and equipment constraints, and uses mixed integer programming or genetic algorithms to solve the scheduling scheme.

5. The system according to claim 1, characterized in that, The equipment health status prediction and intelligent operation and maintenance module uses algorithms such as random forest and LSTM to establish equipment performance degradation models and fault prediction models.

6. The system according to claim 1, characterized in that, The energy big data analysis and decision support module integrates multi-source energy data and uses data mining techniques such as correlation analysis and cluster analysis for in-depth analysis.

7. A smart energy management method for photovoltaic low-carbon industrial parks based on digital twins and multi-agent collaboration, characterized in that, Includes the following steps: S1 constructs a digital twin model of the photovoltaic park, synchronizing data between the physical system and the virtual model in real time; S2 abstracts energy equipment in the park into intelligent agents and realizes collaborative control of the energy system through a distributed collaborative mechanism. S3 aims to reduce carbon emissions by establishing an energy optimization scheduling model and solving for the optimal scheduling scheme; S4 deploys a sensor network to collect equipment data, uses machine learning algorithms to predict equipment health status, and generates operation and maintenance plans. S5 integrates multi-source energy data, performs in-depth analysis, and provides visual decision support.

8. The method according to claim 7, characterized in that, In the step of constructing the digital twin model, geometric information of the park is obtained through three-dimensional laser scanning, and operating parameters are collected using a sensor network to establish a virtual model that includes the physical characteristics and electrical parameters of the equipment.

9. The method according to claim 7, characterized in that, In the multi-agent collaborative control step, each agent makes autonomous decisions based on global energy goals and local information, and adjusts the operating status of energy equipment through negotiation and cooperation.

10. The method according to claim 7, characterized in that, In the low-carbon energy optimization scheduling step, time series analysis, neural networks and other algorithms are used to predict photovoltaic power generation and load demand as inputs to the optimization model.

Images