A charging pile multi-modal load prediction intelligent edge host

Abstract

The utility model relates to edge host field especially, and it is a kind of charging pile multi-modal load prediction intelligent edge host.The technical problem that the utility model wants to solve is to provide a kind of charging pile multi-modal load prediction intelligent edge host that can fuse multi-source heterogeneous data, realize load real-time accurate prediction and high reliability operation.A kind of charging pile multi-modal load prediction intelligent edge host, including host shell and integrated encapsulation in host shell: main control unit, as the core processing and control component of the edge host;Multi-modal data acquisition interface, for accessing heterogeneous data from charging pile and its surrounding environment;Edge computing module, for the data of the multi-modal data acquisition interface transmission is imported and is localized real-time calculation, generates load prediction result.The utility model reaches the effect of fusion multi-source heterogeneous data, realizes load real-time accurate prediction and high reliability operation.

Description

Technical Field

[0001] This utility model relates to the field of edge hosts, and in particular to a smart edge host for multimodal load prediction of charging piles. Background Technology

[0002] With the rapid development of the new energy vehicle industry, charging piles, as a key infrastructure, are being built on an increasingly large scale and deployed in greater density. Charging loads are characterized by intermittency and randomness, and their large-scale, disorderly charging behavior poses a serious challenge to the stable operation of local power distribution networks, potentially leading to problems such as transformer overload, increased line losses, and degraded power quality. Therefore, achieving accurate prediction of charging pile loads is crucial for optimizing power grid dispatch, implementing demand-side response, and improving the operational efficiency of charging stations.

[0003] Traditional load forecasting solutions rely on centralized cloud platforms. Firstly, the continuous transmission of massive amounts of data places high demands on network bandwidth and incurs considerable communication costs. Secondly, there is a significant delay in data transmission to the cloud and the return of forecast results; in the event of network outages or other anomalies, the forecasting service will be interrupted, making it difficult to meet real-time requirements. Edge computing technology offers a new approach to address these issues by bringing computing power closer to charging stations, enabling localized data processing and real-time response. However, existing edge computing devices still have shortcomings in handling charging load forecasting tasks. Charging load is not only affected by charging behavior itself but also closely related to various temporal and environmental factors such as temperature, date type, and electricity pricing policies. Existing systems often have a single data dimension and fail to fully integrate multi-source heterogeneous data for in-depth analysis, resulting in limited forecast accuracy. They also struggle to meet the comprehensive requirements of diverse data access, real-time computation, model specialization, and operational reliability in charging pile load forecasting scenarios. Therefore, there is an urgent need to develop an intelligent edge host for multi-modal load forecasting of charging piles that can integrate multi-source heterogeneous data and achieve real-time, accurate load forecasting with high reliability. Utility Model Content

[0004] In order to overcome the shortcomings of existing intelligent edge hosts for load forecasting, which have a single data dimension and fail to fully integrate multi-source heterogeneous data for in-depth analysis, resulting in limited prediction accuracy and difficulty in meeting the reliability requirements of charging pile load forecasting, this utility model aims to provide a multi-modal intelligent edge host for charging pile load forecasting that can integrate multi-source heterogeneous data and achieve real-time accurate load forecasting and high-reliability operation.

[0005] This utility model is achieved by the following specific technical means: A smart edge host for multimodal load prediction in charging piles includes a host housing and components integrated and packaged within the host housing. The main control unit serves as the core processing and control component of the edge host; A multimodal data acquisition interface is electrically connected to the main control unit and is used to access heterogeneous data from the charging pile and its surrounding environment. An edge computing module is electrically connected to the main control unit. The edge computing module has a pre-built load forecasting algorithm model, which is used to perform localized real-time calculations on the data input from the multimodal data acquisition interface to generate load forecasting results. The data storage module is electrically connected to the main control unit and is used to cache the real-time data collected by the multimodal data acquisition interface and the prediction results generated by the edge computing module. The communication module is electrically connected to the main control unit and is used to upload the load forecast results and key data to the cloud platform, and to receive instructions or model updates from the cloud platform. The power management module is electrically connected to the main control unit and other modules, and is used to provide stable power supply and power protection for the entire edge host. The security protection module is electrically connected to the main control unit and is used to encrypt and securely isolate the data communication and data storage of the edge host.

[0006] Furthermore, the multimodal data acquisition interface includes: The charging pile operation data interface is used to access the charging pile's real-time power, voltage, current, charging status, and vehicle identification information; An environmental sensing data interface is used to access environmental data collected by sensors deployed at charging stations. The environmental data includes at least temperature, humidity, and weather conditions. The time-series data interface is used to access calendar information and real-time clock signals. The calendar information includes at least date type, holiday information, and peak-valley electricity price periods.

[0007] Furthermore, the edge computing module is a computing card that integrates an AI acceleration chip, which is one of a GPU, NPU, or FPGA.

[0008] Furthermore, the communication module is a multi-mode communication module, supporting at least two communication methods including G / G cellular networks, Ethernet, and Wi-Fi.

[0009] Furthermore, it also includes a local human-machine interaction module, which is electrically connected to the main control unit and is used to realize local status display and command input; the human-machine interaction module includes status indicator lights, physical buttons and / or touch screen.

[0010] Compared with the prior art, the present invention has the following beneficial effects: This invention achieves the effect of integrating multi-source heterogeneous data, realizing real-time accurate load prediction and high-reliability operation. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of the system framework structure of this utility model.

[0012] Figure 2 This is a schematic diagram of the multimodal data acquisition interface of this utility model. Detailed Implementation

[0013] The present invention will be further described below with reference to the accompanying drawings: Example

[0014] A smart edge host for multimodal load prediction of charging piles, such as Figure 1-2 As shown, it includes a main unit housing and components integrated and packaged within the main unit housing: The main control unit serves as the core processing and control component of the edge host; A multimodal data acquisition interface is electrically connected to the main control unit and is used to access heterogeneous data from the charging pile and its surrounding environment. An edge computing module is electrically connected to the main control unit. The edge computing module has a pre-built load forecasting algorithm model, which is used to perform localized real-time calculations on the data input from the multimodal data acquisition interface to generate load forecasting results. The data storage module is electrically connected to the main control unit and is used to cache the real-time data collected by the multimodal data acquisition interface and the prediction results generated by the edge computing module. The communication module is electrically connected to the main control unit and is used to upload the load forecast results and key data to the cloud platform, and to receive instructions or model updates from the cloud platform. The power management module is electrically connected to the main control unit and other modules, and is used to provide stable power supply and power protection for the entire edge host. The security protection module is electrically connected to the main control unit and is used to encrypt and securely isolate the data communication and data storage of the edge host.

[0015] Working principle: This utility model's intelligent edge host for multimodal load prediction of charging piles achieves intelligent prediction and edge management of charging pile loads through the collaborative work of highly integrated hardware modules.

[0016] First, the multimodal data acquisition interface receives heterogeneous data in real time; the charging pile operation data interface collects real-time power, voltage, current, charging status, and vehicle identification information of the charging pile; the environmental perception data interface acquires environmental data such as temperature, humidity, and weather conditions collected by sensors deployed at the charging station; the time-series data interface receives calendar information (such as date type and holidays) and real-time clock signals, including peak and off-peak electricity price periods; this multimodal data is transmitted to the main control unit, which, as the core processing component, coordinates the data flow and initializes the processing; subsequently, the edge computing module calls a pre-set load forecasting algorithm model to process the input... Multimodal data is processed locally in real time, without relying on the cloud, to quickly generate short-term or long-term load forecast results, thereby achieving accurate prediction of charging pile demand fluctuations. At the same time, the data storage module is used to cache the collected real-time data and the forecast results generated by the edge computing module, ensuring data reliability and fast access. After generating the load forecast results, the communication module uploads the forecast results and key data to the cloud platform for cloud monitoring and analysis; and receives instructions or model updates from the cloud platform to achieve remote optimization. In addition, the local human-machine interaction module provides local status display and instruction input functions, facilitating on-site operation and monitoring.

[0017] Throughout the process, the power management module provides stable power supply and power protection for each module, ensuring reliable operation of the equipment in complex environments; the safety protection module ensures the security of data communication and data storage through encryption and security isolation technologies, preventing data leakage; through the above workflow, real-time prediction of charging pile load, edge intelligent decision-making and cloud collaboration are realized, improving the operating efficiency of charging piles and the stability of the power grid.

[0018] Although this disclosure has been described in detail with reference to exemplary embodiments, it is not limited thereto, and it will be apparent to those skilled in the art that various modifications and changes may be made thereto without departing from the scope of this disclosure.

Claims

1. A smart edge host for multimodal load prediction of charging piles, characterized in that, Includes a main unit housing and components integrated and packaged within the main unit housing: The main control unit serves as the core processing and control component of the edge host; A multimodal data acquisition interface is electrically connected to the main control unit and is used to access heterogeneous data from the charging pile and its surrounding environment. An edge computing module is electrically connected to the main control unit. The edge computing module has a pre-built load forecasting algorithm model, which is used to perform localized real-time calculations on the data input from the multimodal data acquisition interface to generate load forecasting results. The data storage module is electrically connected to the main control unit and is used to cache the real-time data collected by the multimodal data acquisition interface and the prediction results generated by the edge computing module. The communication module is electrically connected to the main control unit and is used to upload the load forecast results and key data to the cloud platform, and to receive instructions or model updates from the cloud platform. The power management module is electrically connected to the main control unit and each module, and is used to provide stable power supply and power protection for the entire edge host. The security protection module is electrically connected to the main control unit and is used to encrypt and securely isolate the data communication and data storage of the edge host.

2. The intelligent edge host for multimodal load prediction of charging piles according to claim 1, characterized in that, The multimodal data acquisition interface includes: The charging pile operation data interface is used to access the charging pile's real-time power, voltage, current, charging status, and vehicle identification information; An environmental sensing data interface is used to access environmental data collected by sensors deployed at charging stations. The environmental data includes at least temperature, humidity, and weather conditions. The time-series data interface is used to access calendar information and real-time clock signals. The calendar information includes at least date type, holiday information, and peak-valley electricity price periods.

3. The intelligent edge host for multimodal load prediction of charging piles according to claim 1, characterized in that, The edge computing module is a computing card that integrates an AI acceleration chip, which is one of a GPU, NPU, or FPGA.

4. The intelligent edge host for multimodal load prediction of charging piles according to claim 1, characterized in that, The communication module is a multi-mode communication module, which supports at least two of the following communication methods: G / G cellular network, Ethernet, and Wi-Fi.

5. The intelligent edge host for multimodal load prediction of charging piles according to claim 1, characterized in that, It also includes a local human-machine interaction module, which is electrically connected to the main control unit and is used to realize local status display and command input; the human-machine interaction module includes status indicator lights, physical buttons and / or touch screen.

Images