A charging pile safety monitoring method and system

Abstract

The present application relates to the technical field of monitoring or controlling charging stations, and in particular to a charging pile safety monitoring method and system, comprising the following steps: step 1, monitoring phase; step 2, primary feedback; step 3, traceability analysis phase, determining the fire point and the fire cause; step 3.1, on-site situation extraction; step 3.2, fire point analysis, based on image features, when the pixels in the image satisfy the condition of brightness L>L thresh and color in C range , the area where the pixels are located is determined as a suspected flame area F; the present application realizes accurate and comprehensive monitoring and safety warning through the real-time collection of data by multiple sensors in all directions. Through traceability analysis, the fire point and the cause can be quickly and clearly determined, and the responsibility can be defined. The primary feedback timely alarms and assists in judgment, and the secondary feedback is targeted based on the traceability results, the double feedback mechanism optimizes the emergency process, improves the user experience, and comprehensively safeguards the operation safety of the charging pile.

Description

Technical Field

[0001] This invention relates to the field of monitoring or controlling charging stations, and in particular to a method and system for monitoring the safety of charging piles. Background Technology

[0002] Charging pile safety monitoring is a technical means to ensure the safety of charging piles during operation. It involves real-time monitoring of various operating parameters of the charging piles and the surrounding environment, timely detection and handling of potential safety hazards, and ensuring the stable and safe operation of the charging piles and providing reliable charging services to users.

[0003] Traditional monitoring methods and systems have many shortcomings. They often only monitor simple electrical parameters and lack comprehensive or accurate monitoring of key safety factors such as temperature, smoke, flame, and charging gun insertion / removal force. After an accident, it is difficult to quickly and accurately determine the ignition point and cause of the fire, and there is a lack of effective tracing mechanisms. At the same time, the feedback mechanism is imperfect, failing to promptly transmit alarm information to relevant personnel and failing to provide targeted handling based on the cause of the accident, resulting in low emergency response efficiency and a poor user experience.

[0004] To address this, the present invention proposes a charging pile safety monitoring method and system. This system achieves comprehensive and accurate monitoring through multiple types of high-precision sensors, utilizes advanced source tracing analysis to identify the ignition point and cause of the fire, and employs a robust dual feedback mechanism to promptly issue alarms and take targeted actions based on the source tracing results. This enables comprehensive, efficient, and intelligent monitoring of charging pile safety, thereby improving the safety, reliability, and user experience of charging pile operation. Summary of the Invention

[0005] The technical problem to be solved: often only simple electrical parameters can be monitored, and the monitoring of key safety factors such as temperature, smoke, flame and charging gun insertion and extraction force is not comprehensive or accurate enough.

[0006] To address the shortcomings of existing technologies, this invention provides a method and system for monitoring the safety of charging piles, thereby solving the technical problems mentioned in the background section.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for safety monitoring of charging piles includes the following steps:

[0009] Step 1, the monitoring phase, uses sensors and thermal imaging binocular cameras to monitor leakage, overload, short circuit, smoke, flame and the surrounding environment in real time, providing a basis for subsequent anomaly judgment;

[0010] Step 2: Initial feedback, issuing an alert for any abnormalities detected during the monitoring phase in Step 1;

[0011] Step 3: Source tracing analysis phase, determining the ignition point and cause of the fire;

[0012] Step 3.1: Extracting on-site information;

[0013] Step 3.2, Ignition Point Analysis: Based on the acquired on-site video data and image features, when the pixels in the image satisfy the brightness L > L... thresh And the color is in C range Within the frame, the region where the pixel is located is identified as a suspected flame region F. Simultaneously, an optical flow algorithm is used to calculate the motion vector V of the smoke region between adjacent frames. t The ignition point is determined by combining the analysis results of flame morphology and smoke trajectory.

[0014] Step 3.3, Analysis of the cause of the fire;

[0015] Step 3.3.1, Internal Factor Analysis: Examine the leakage current I detected by the leakage current sensor. leak Based on the charging line current I monitored by the current transformer and the temperature T of key components monitored by the temperature sensor, data from the vehicle battery management system is obtained, and the current I during the charging process is analyzed. thaege Battery temperature T battery Voltage V battery The state of charge parameter indicates an abnormality in the monitored area or is the cause of a fire when the monitored value is greater than the set threshold.

[0016] Step 3.3.2, External Factor Analysis: The half-hour thermal imaging video data extracted by the thermal imaging binocular camera in the environmental visual monitoring module is segmented into a two-dimensional image matrix I(x, y, t), and abnormal temperature points are screened for abnormal temperature analysis. At the same time, the image sequence is processed using a deep learning-based behavior analysis algorithm, and a convolutional neural network is used to learn from the training data to extract the features of different behaviors for abnormal behavior analysis. Then, the video segment from the appearance of the abnormal point to the issuance of the fire alarm is extracted.

[0017] Step 4: Secondary Feedback. After completing the source tracing analysis in Step 3, relevant information needs to be processed and fed back in a targeted manner based on the cause of the fire.

[0018] In one possible implementation, the monitoring phase is specifically as follows:

[0019] The system monitors the current data of the charging pile using leakage current sensors to capture leakage information; it monitors the current changes in the charging lines using current transformers to detect overload or short circuit anomalies; it monitors the temperature data of key heat-generating components inside the charging pile using temperature sensors; it monitors the fire situation in the area surrounding the charging pile by installing ionization smoke sensors and infrared flame detectors; it acquires video information and temperature changes in the area surrounding the charging pile using thermal imaging binocular cameras; and it obtains key data of the vehicle battery through communication with the vehicle battery management system.

[0020] In one possible implementation, step 2 utilizes the ionization smoke sensor and infrared flame detector from step 1 to trigger an alarm mechanism when the smoke concentration exceeds a set threshold or a flame spectrum is detected, and then feeds the alarm information back to the remote monitoring center and the client.

[0021] In one possible implementation, step 4 specifically includes the following cases:

[0022] 4.1 If the cause of the fire is internal, relevant data involving abnormal circuits in the charging pile or vehicle will be thoroughly processed and fed back to the remote control center via network communication technology. This information will be displayed in the form of intuitive charts and detailed data reports, and detailed fault information and maintenance suggestions will be fed back to the operation and maintenance personnel's client. The cause of the accident and the estimated time for service restoration will be fed back to the user's client.

[0023] 4.2 If the cause of the fire is external, the video clips of abnormal temperature analysis and abnormal behavior analysis extracted from the external cause analysis will be organized and fed back to the remote control center. The staff of the remote control center will view the video on the monitoring platform, combine it with the analysis report, judge the severity of the accident and potential risks, and promptly connect any illegal information involved with law enforcement departments.

[0024] In one possible implementation, the monitoring system implementing the above-mentioned charging pile safety monitoring method includes a perception layer, a data transmission layer, a data processing layer, and a user interaction layer.

[0025] The perception layer is responsible for collecting various types of data, covering multiple sensors and devices; the data transmission layer undertakes the task of transmitting data from the perception layer; the data processing layer receives data from the perception layer and processes and analyzes it; the user interaction layer provides different services to maintenance personnel and ordinary users.

[0026] Beneficial effects compared to existing technologies:

[0027] 1. This solution utilizes high-definition cameras to review surveillance videos, combining image feature analysis and smoke motion vector calculation to determine the ignition point. By comprehensively considering flame patterns, smoke trajectories, and data from charging piles and vehicle sensors, it accurately determines whether the fire originated from the vehicle, the charging pile, or a nearby foreign object. Simultaneously, it deeply analyzes the causes of the fire from both internal and external perspectives. Internal causes are identified through electrical parameter analysis to determine faults in the charging pile and vehicle's internal circuitry, while external causes are explored through video analysis of abnormal temperatures and behaviors to investigate human factors. This comprehensive and in-depth source tracing analysis not only quickly clarifies the cause of the fire after an incident, providing a basis for subsequent maintenance and improvements, but also accurately defines responsibility, avoids disputes, and is of great significance for improving the charging pile safety management system.

[0028] 2. In this solution, a primary feedback mechanism rapidly triggers multiple alarm methods upon detecting abnormal smoke or flames. Local audible and visual alarms alert nearby personnel to evacuate, and alarm information is simultaneously transmitted to the remote monitoring center and the client. The remote monitoring center automatically retrieves historical data to aid in judgment and uses artificial intelligence algorithms to classify and prioritize alarm information, addressing emergencies first. The client provides different information based on user roles: maintenance personnel receive detailed fault analysis and handling suggestions, while ordinary users receive concise safety reminders and information on nearby charging points, along with an added voice alarm function. Secondary feedback addresses issues based on the source tracing results, providing detailed fault and maintenance information for internal causes and video and risk warnings for external causes. This dual feedback mechanism works in tandem to achieve efficient connection from incident occurrence to handling, optimizes emergency response processes, ensures the safety of personnel and property, and enhances user experience, increasing user confidence and satisfaction with charging station usage. Attached Figure Description

[0029] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0030] Figure 1 A flowchart illustrating the steps of a safety monitoring method for charging piles;

[0031] Figure 2 This is a system framework diagram of a charging pile safety monitoring system. Detailed Implementation

[0032] Preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the present invention can also be implemented in various different forms, and therefore the present invention is not limited to the embodiments described below.

[0033] The technical solution in this application embodiment is to solve the problems mentioned in the background art, and the overall idea is as follows:

[0034] Example 1:

[0035] Please refer to Figure 1 As shown in the figure, this embodiment introduces a safety monitoring method for charging piles, including the following steps:

[0036] Step 1: Monitoring Phase

[0037] The core of this stage is to use various sensors and devices to collect comprehensive and real-time data on the charging piles and their surrounding environment, so as to provide a basis for subsequent anomaly judgment.

[0038] 1.1 Electrical Parameter Monitoring

[0039] Leakage current monitoring: High-sensitivity leakage current sensors are installed in the main circuit and branch circuits of the charging pile to collect leakage current data every 10 milliseconds; this is because leakage current may cause electrical fires, and timely capture of leakage current information is crucial.

[0040] Current change monitoring: Using a current transformer with a suitable ratio, the current change data of the charging line is collected every 5 milliseconds; by monitoring the current in real time, abnormal conditions such as overload or short circuit can be detected in time.

[0041] Temperature monitoring: High-precision temperature sensors are installed in key heat-generating components inside the charging pile, such as power modules and transformers, to collect temperature data once per minute; excessively high temperatures may be a sign of equipment malfunction or abnormal operation.

[0042] 1.2 Smoke and Flame Monitoring

[0043] Ionization smoke sensors and infrared flame detectors are installed inside and around the charging pile; the smoke sensor detects the smoke concentration once per second, and the flame detector scans the flame spectrum once every 0.5 seconds; the smoke sensor and flame detector are cleaned and calibrated regularly to ensure their stable performance.

[0044] Multiple data transmission channels, such as 4G, 5G, Wi-Fi, and Bluetooth, are set up for smoke and flame monitoring equipment. When the main transmission channel fails, it automatically switches to the backup channel to ensure timely transmission of alarm signals. At the same time, smoke and flame monitoring data is stored locally so that critical information can be saved even in the event of network failure.

[0045] 1.3 Environmental Visual Monitoring

[0046] Thermal imaging binocular cameras are installed above and around key locations of the charging pile. They combine video monitoring and thermal imaging functions, and can display the temperature changes of target objects in the monitoring environment in real time. They can monitor and record 24 hours a day without interruption, with the frame rate set to 25fps. The camera housing has automatic defogging and dustproof functions to reduce the impact of bad weather and environment on the camera.

[0047] The video data from the thermal imaging binocular camera is preprocessed locally, and image enhancement algorithms are used to improve image clarity and recognition, reducing monitoring errors caused by insufficient light or complex backgrounds. At the same time, the video transmission strategy is optimized, and the video frame rate and resolution are dynamically adjusted according to network conditions to ensure stable transmission of real-time images to the remote monitoring center. The monitoring video is stored for at least 2 weeks, and data is cleaned and backed up regularly to prevent data loss.

[0048] 1.4 Vehicle Battery Data Monitoring

[0049] When the charging pile is connected to the vehicle, a communication connection is established with the vehicle's battery management system. Key data of the vehicle battery is acquired in real time through standard data transmission protocols (such as CAN bus protocol), with the data acquisition frequency set to once per second. A data verification mechanism is added to perform CRC verification and other checks on the acquired battery data to ensure the accuracy and integrity of the data.

[0050] When the network fails and battery data cannot be transmitted in real time, the battery data is cached locally on the charging pile and automatically retransmitted after the network is restored. At the same time, the battery data is analyzed in real time. If abnormal battery data is found, in addition to sending alarm information to the remote monitoring center and the client, the abnormal battery situation is also notified to the user through the charging pile display screen.

[0051] Step 2, First Feedback

[0052] When the smoke sensor detects that the smoke concentration exceeds the set smoke alarm threshold, or when the infrared flame detector scans a flame spectrum that meets the flame alarm conditions, the alarm mechanism is immediately triggered. In addition to being transmitted to the local alarm device and the remote monitoring center via wired or wireless communication, the alarm signal will also be sent to other nearby charging piles to achieve regional joint alarm, expand the alarm range, and remind more people to pay attention to safety.

[0053] In the remote monitoring center, in addition to displaying the location of the charging pile, the alarm type, and the alarm time on the pop-up alarm interface, the monitoring data of the charging pile over a period of time will be automatically retrieved, including electrical parameters, temperature, vehicle battery data, etc., to help staff quickly judge the accident situation; at the same time, artificial intelligence algorithms are used to classify and prioritize alarm information, and high-urgency alarm events are handled first.

[0054] On the client side, different levels of detail in alarm information are displayed based on the user's role (operation and maintenance personnel, ordinary users). For operation and maintenance personnel, detailed alarm data and fault analysis suggestions are provided. For ordinary users, users are reminded to pay attention to safety in a concise and clear manner, and information on nearby safe charging points is provided. At the same time, the client side adds a voice alarm function so that users can receive alarm prompts even when they cannot check their mobile phone screen in time.

[0055] Step 3: Source tracing analysis phase

[0056] This stage is crucial to the entire method, as it involves in-depth analysis of the on-site situation to determine the ignition point and cause of the fire.

[0057] 3.1 On-site situation extraction

[0058] Starting from the moment the fire occurred, the monitoring video of the past 30 minutes was reviewed using high-definition camera units to comprehensively extract on-site information, including the location and status of vehicles, the appearance of charging piles, and the activities of people and the distribution of objects in the surrounding environment.

[0059] 3.2 Ignition Point Analysis

[0060] 3.2.1 Judgment based on image features

[0061] Suspected flame area identified: Using surveillance video images acquired from high-definition cameras, a brightness threshold L was set. thresh and color range C range To determine suspected flame areas, pixels in the image must satisfy the brightness L > L. thresh And the color is in C range When the area containing the pixel is defined as a suspected flame region F, it can be expressed by the formula F={(x,y)|L(x,y)>L thresh ∧C(x, y)∈C range For example, in practical applications, after extensive analysis and research on flame images, a specific brightness threshold is set, and the color range covers the common red, yellow and other hues of flames, so as to accurately filter out areas that may be flames.

[0062] Smoke motion vector calculation: The motion vector V of the smoke region between adjacent frames is calculated using an optical flow algorithm. t The optical flow algorithm calculates motion vectors based on pixel brightness changes caused by the movement of objects in an image. It assumes that the movement of objects is continuous in a short period of time. By comparing the brightness changes of the same pixel in adjacent frames, it calculates the direction and speed of the movement of that pixel, thereby obtaining the motion vector of the smoke region. The ignition point is usually located in the starting direction of the smoke motion vector because the smoke starts to spread from the ignition point, and its starting direction of movement often points to the ignition point.

[0063] Ignition point location determination: The ignition point is determined by combining the analysis results of flame shape and smoke movement trajectory; if the flame area and the starting point of smoke movement are highly overlapping in spatial position, it can be basically determined that the location is the ignition point; if there are multiple flame areas, the center of the flame area to which the smoke movement vector is most concentrated is selected as the ignition point; this is because the area to which the smoke movement vector is most concentrated is likely the source of smoke generation, which is the location of the ignition point.

[0064] 3.2.2 Ignition point determination: If the flames are mainly concentrated in key locations such as the vehicle's battery area and engine compartment, and the vehicle exhibits abnormal conditions such as smoke or sparks during charging, the ignition point is determined to be the vehicle.

[0065] When the flames mainly start from the electrical connection parts inside or on the surface of the charging pile, and the leakage current sensor, current transformer or temperature sensor on the charging pile has abnormal data records before the fire occurs, the point of origin of the fire is determined to be the charging pile.

[0066] If the flame originates from a foreign object near the charging station, and the foreign object has no direct electrical connection with the charging station and the vehicle, the point of ignition is determined to be a foreign object in the vicinity.

[0067] 3.3 Analysis of the cause of the fire

[0068] 3.3.1 Internal cause analysis mainly involves analyzing whether there are abnormalities in the charging pile or the vehicle's internal circuitry that could have caused the fire;

[0069] 3.3.1.1 Analysis of the causes of the charging pile fire

[0070] Check the leakage current I detected by the leakage current sensor. leak , when I leak The leakage current is greater than the set threshold I. leak-thresh (e.g., 30mA) indicates a leakage fault, which may cause a fire;

[0071] Based on the charging line current I monitored by the current transformer, combined with the rated current I of the charging pile rated Set overload current threshold I rated-thresh =1.2I rated and short-circuit current threshold I short-thresh , when I>I over-thresh And the duration exceeds the overload duration threshold T over (e.g., 6 seconds) is considered overload; when I > I short-thresh The circuit was determined to be short-circuited; both overload and short circuit can cause a fire.

[0072] Based on the temperature T of the key components monitored by the temperature sensor, the normal operating temperature range [T] is set. min T max(e.g., [70℃, 80℃]) and the danger temperature threshold T danger (e.g., 90℃), when T is outside the normal operating temperature range, it indicates a temperature abnormality; when T > T danger It may cause a fire due to overheating;

[0073] 3.3.1.2 Analysis of the cause of vehicle fire

[0074] Obtain the current I during the charging process charge Voltage V charge Data, setting the normal charging current range [I c-min I c-max ] and voltage range [V c-min V c-max ], when I charge Not within the normal charging current range or V charge If the voltage is outside the normal range, it indicates abnormal charging parameters, which may have caused the vehicle to catch fire due to abnormal charging process.

[0075] Acquire data from the vehicle's battery management system and analyze the battery temperature T. battery Voltage V battery SOC (State of Charge) SOC Parameters such as [T] are used to set the normal battery temperature range. b-min T b-max Voltage range [V] b-min V b-max ] and SOC range [S s-min S s-max ], when T battery V battery or S SOC If the battery is outside the normal range, it is considered to have a risk of failure and may cause the vehicle to catch fire.

[0076] 3.3.2 External Factor Analysis: This mainly focuses on whether human factors caused the fire. By analyzing the video footage from the half hour before the anomaly was discovered, we explored potential fire-inducing factors from two aspects: abnormal temperature and abnormal behavior.

[0077] 3.3.2.1 Abnormal Temperature Analysis;

[0078] By segmenting the half-hour thermal imaging video data extracted by the thermal imaging binocular camera in the environmental visual monitoring module into a two-dimensional image matrix I(x, y, t), where x and y represent the coordinates of image pixels and t represents the time frame of the video, firstly, a suitable temperature threshold T is set. threshold Used to filter out abnormal temperature points; each pixel (x, y) in a thermal imaging image corresponds to a temperature value T(x, y, t). A pixel is considered to have a temperature value T when its temperature value satisfies T(x, y, t) > T. thresholdAt that time, the pixel was marked as a suspicious point;

[0079] For example, considering the temperature range of common fire sources (such as cigarette butts and lighter flames), and taking into account the fluctuations in ambient temperature in actual scenarios, T threshold Set to 200℃ (in actual applications, this threshold needs to be calibrated based on a large number of experiments and practical experience); starting from the beginning frame of the video, analyze the image frame by frame. Once a suspicious point is found, start to extract a video segment from that frame, and then continue to analyze the changes of these suspicious points.

[0080] To determine the continuity of suspicious points, the overlap rate O of suspicious point regions in two adjacent frames is calculated. Assume the suspicious point region in frame t is S. t The suspicious point region in frame t+1 is S. t+1 The formula for calculating the overlap rate is: When the overlap rate O is greater than the set continuity threshold O continue (For example, O) continue =0, indicating that there is at least 60% overlap between the suspicious point areas in the two frames, it is considered that the suspicious point is continuous, that is, there is a possibility of continued burning;

[0081] To determine whether the suspicious point region has expanded, calculate the area of ​​the suspicious point region. Let A be the area of ​​the suspicious point region in frame t. t The area of ​​the suspicious point region in frame t+1 is A. t+1 When satisfied (Where α is the area expansion ratio threshold, assuming α = 0.3, it means that the area of ​​the suspicious point region in the next frame is at least 30% larger than that in the previous frame), when it is judged that the combustion has an expanding trend;

[0082] Video segments with suspicious points that continue or expand are extracted and used as feedback for abnormal temperature analysis. This feedback can intuitively show the abnormal temperature change process that may cause a fire, providing important basis for subsequent handling and decision-making.

[0083] 3.3.2.2 Abnormal Behavior Analysis;

[0084] The half-hour monitoring video data extracted by the thermal imaging binocular camera in the environmental visual monitoring module is divided into a two-dimensional image matrix I(x, y, t), where x and y represent the coordinates of the image pixels and t represents the time frame of the video.

[0085] Image sequences are processed using deep learning-based behavior analysis algorithms. Taking human behavior analysis as an example, a large amount of video data containing normal and abnormal behaviors is first labeled to construct a training dataset. A convolutional neural network (CNN) is then used to learn from the training data and extract features of different behaviors.

[0086] In actual monitoring, for each frame of image I(x, y, t), the CNN model outputs a feature vector F(t), which contains the behavioral feature information of the target (such as a human body) in the current frame; the feature vector F(t) of the current frame is compared with the set of normal behavioral feature vectors N = {N1, N2, ..., N...}. m The cosine similarity S of each vector in the vector group is calculated using the following formula: Wherein, F(t)·N i Let ||F(f)|| and ||N|| represent the dot product of vectors. i || represent vectors F(t) and N respectively. i The model;

[0087] Set an abnormal behavior judgment threshold S threshold (For example, S) threshold =0.8); when for all i (1≤i≤m), S(F(t), N i ) threshold When this happens, it is determined that abnormal behavior has occurred in the current frame;

[0088] For example, in the charging pile scenario, normal behavior may include people plugging and unplugging charging guns normally and parking vehicles in an orderly manner. If the model detects that a person is running fast towards the charging pile in a certain frame, and the cosine similarity between the feature vector of the behavior and the feature vector of normal behavior is lower than the threshold, then the behavior is judged as abnormal behavior.

[0089] Once abnormal behavior is detected, a video segment containing the process of the abnormal behavior is extracted from the frame where the abnormal behavior begins. This video segment serves as feedback content for the abnormal behavior analysis. It can clearly show the time, location, and specific manifestations of the abnormal behavior, which helps maintenance personnel to quickly understand the situation on site and take corresponding measures to handle it, so as to avoid potential safety accidents.

[0090] Step 4, Secondary Feedback

[0091] After completing the source tracing analysis in step 3, the relevant information needs to be processed and fed back in a targeted manner, depending on whether the cause of the fire is internal or external.

[0092] 4.1 If the cause of the fire is internal.

[0093] Data processing and analysis: In-depth processing of relevant data involving abnormalities in charging piles or vehicle internal circuits, including abnormal data obtained from leakage current sensors, current transformers, and temperature sensors, as well as current, voltage, and battery management system data during vehicle charging; detailed analysis of the correlation between these data, outlining the timeline of fault occurrence and development trend;

[0094] ​Feedback to the remote control center: The compiled internal data, ignition point information, and detailed fault analysis report are sent to the remote control center via network communication technologies (such as HTTP, MQTT, etc.). On the monitoring platform of the remote control center, this information is displayed in the form of intuitive charts (such as current change curves and temperature trend graphs) and detailed data reports, which facilitates staff to fully understand the accident situation. Based on this information, staff assess the extent of damage caused by the accident to the charging pile and vehicle, and formulate corresponding repair and replacement plans.

[0095] Feedback to the client: For maintenance personnel, push detailed fault information and repair suggestions to guide them in quickly troubleshooting and repairing the problem; for example, if it is determined that a fire was caused by overheating of a power module in the charging pile, inform the maintenance personnel to replace the power module and check the surrounding circuits; for users, inform them of the cause of the accident in easy-to-understand language (e.g., "A fire was caused by a fault in the charging pile's internal equipment, and charging service is currently suspended. We are working hard to repair it"), and explain the estimated time for service restoration to reassure users and remind them to pay attention to relevant safety tips when using the charging pile in the future.

[0096] 4.2 If the cause of the fire is external.

[0097] Video organization and annotation: Further organize the video clips of abnormal temperature analysis and abnormal behavior analysis extracted from the external factor analysis, add detailed annotations to each video clip, indicating the time, location, type of abnormality (such as the approach of a high-temperature object, abnormal behavior of personnel, etc.) and its relationship with the ignition point;

[0098] Feedback to the remote control center: The annotated video clips, fire location information, and external cause analysis report are sent to the remote control center. Staff at the remote control center review the video on the monitoring platform and, in conjunction with the analysis report, determine the severity of the accident and potential risks. For example, if it is discovered that someone deliberately set fire near the charging station, the relevant law enforcement departments are notified immediately, and monitoring of the area is strengthened. Simultaneously, this accident is recorded as a case study to improve safety management plans and prevent similar incidents from recurring.

[0099] Feedback to the client: Display video clips and detailed accident analysis on the maintenance personnel's client, reminding them to strengthen daily patrols of the area and be vigilant against similar safety hazards; For the user's client, inform the user that the accident was caused by external factors (e.g., "This accident was caused by a fire caused by a foreign object in the vicinity. We have taken measures to ensure safety. We apologize for any inconvenience caused"), and remind the user to pay attention to the surrounding environment when using the charging station and to notify the maintenance personnel in a timely manner if any suspicious situations are found; At the same time, some safety knowledge can be pushed to improve the user's safety awareness.

[0100] Example 2:

[0101] like Figure 2 As shown, this embodiment is based on the monitoring method of embodiment 1 to construct a charging pile safety monitoring system, which aims to achieve comprehensive, efficient and intelligent safety monitoring of charging piles and improve the safety and reliability of charging pile operation.

[0102] 1. System Architecture

[0103] The system mainly consists of a perception layer, a data transmission layer, a data processing layer, and a user interaction layer.

[0104] 1.1 Perception Layer: Responsible for collecting various types of data, covering multiple sensors and devices;

[0105] High-sensitivity leakage current sensors are installed in the main circuit and branch circuit of the charging pile, along with backup sensors, to collect leakage current data every 10 milliseconds to ensure stable and reliable leakage current monitoring; current transformers with transformer ratio matching are used to collect charging line current change data every 5 milliseconds, and multiple sets of transformers are used for redundant monitoring.

[0106] High-precision temperature sensors are installed on key heat-generating components to collect temperature data every minute, increasing the number of sensors and adopting distributed monitoring.

[0107] Meanwhile, ionization smoke sensors and infrared flame detectors are installed inside and around the charging pile to detect smoke concentration and flame spectrum every second and every 0.5 seconds, respectively, and are calibrated regularly.

[0108] Thermal imaging binocular cameras are deployed above and around key locations of charging piles. They have automatic defogging and dust prevention functions, and can continuously shoot at 25fps for 24 hours to obtain real-time monitoring images and temperature change data of target objects.

[0109] When the charging station is connected to the vehicle, it acquires key data such as the vehicle battery temperature, voltage, SOC, and internal resistance every second through standard data transmission protocols such as the CAN bus protocol, and adds a data verification mechanism.

[0110] 1.2 Data Transmission Layer: Responsible for transmitting data from the sensing layer. The smoke and flame monitoring equipment is equipped with multiple data transmission channels, including 4G, 5G, Wi-Fi, and Bluetooth. In case of a primary channel failure, it automatically switches to a backup channel to ensure timely transmission of alarm signals, while simultaneously storing monitoring data locally.

[0111] The video data collected by the thermal imaging binocular camera is pre-processed locally and the frame rate and resolution are dynamically adjusted according to the network conditions before being transmitted to the remote monitoring center via the network. The video data is stored for at least two weeks and backed up regularly.

[0112] Vehicle battery data is transmitted in real time when the network is normal, and cached locally when the network fails, and retransmitted after the network is restored.

[0113] 1.3 Data Processing Layer: Receives data from the Perception Layer and processes and analyzes it;

[0114] Once the smoke or flame detection equipment triggers an alarm, the system will activate the local audible and visual alarm device on the one hand, and transmit the alarm information to the remote monitoring center on the other hand.

[0115] After receiving the alarm information, the remote monitoring center automatically retrieves the monitoring data of the charging pile over a period of time and uses artificial intelligence algorithms to classify and prioritize the alarm information to help staff quickly determine the accident situation.

[0116] In terms of source tracing analysis, starting from the moment the fire occurred, the monitoring video within 30 minutes was reviewed using high-definition cameras to extract on-site information; suspected flame areas were identified by setting brightness thresholds and color ranges, and smoke movement vectors were calculated using optical flow algorithms. The ignition point was determined by combining flame morphology and smoke movement trajectory; the cause of the fire was analyzed from both internal and external factors. Internal factors were determined by analyzing relevant parameters of the charging pile and the vehicle's internal circuitry, while external factors were explored by analyzing abnormal temperature and abnormal behavior in the video of the half hour before the anomaly was discovered.

[0117] 1.4 User Interaction Layer: Provides different services to maintenance personnel and ordinary users;

[0118] The client displays alarm information of varying levels of detail based on the user's role. Operations and maintenance personnel can obtain detailed alarm data and fault analysis suggestions, while ordinary users receive concise safety reminders and information on nearby safe charging points. The client also adds a voice alarm function.

[0119] During the secondary feedback phase, if the cause of the fire is internal, detailed fault information and maintenance suggestions are pushed to the maintenance personnel's client, the cause of the accident and the estimated recovery time are informed to the user's client, and safety tips are reminded to pay attention; if the cause is external, video clips and accident analysis are displayed on the maintenance personnel's client, reminding them to strengthen inspections, the user's client is informed that the accident was caused by external factors, reminding them to observe the surrounding environment, and safety knowledge is pushed to them.

[0120] 2. System Workflow

[0121] 2.1 After the system starts up, various sensors and devices in the perception layer continuously collect data and transmit the data to the data processing layer through the data transmission layer;

[0122] 2.2 The data processing layer analyzes data in real time. When abnormal data is detected and alarm conditions are triggered, the alarm feedback mechanism is activated to notify relevant local and remote personnel.

[0123] 2.3 In the event of a fire or other accident, the source tracing analysis process shall be initiated to determine the ignition point and cause of the fire;

[0124] 2.4. Based on the source tracing analysis results, targeted feedback processing is carried out at the user interaction layer to provide maintenance personnel with maintenance basis and users with safety prompts.

[0125] Finally, it should be noted that the above embodiments are merely examples for clearly illustrating the present invention and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A charging pile safety monitoring method, characterized in that, Comprise the following steps: Step 1, monitoring stage, by means of sensors and installation of thermal imaging binocular camera, real-time monitoring of leakage, overload, short circuit, smoke, flame and the surrounding environment, provide the basis for subsequent abnormal judgment; Step 2, one feedback, the abnormal situation monitored in step 1 monitoring stage is alarmed; Step 3, traceability analysis stage, determine the fire point and fire cause; Step 3.1, on-site situation extraction; Step 3.2, fire point analysis, based on image features, when the pixels in the image meet the brightness and color within , the area where the pixel is located is determined as a suspected fire area At the same time, the motion vector of the smoke area between adjacent frames is calculated by means of the optical flow algorithm , the analysis results of flame shape and smoke motion trajectory are combined to determine the fire point; Step 3.3, fire cause analysis; Step 3.3.1, Intrinsic analysis, see the leakage current detected by the leakage sensor , according to the charging line current monitored by the current transformer , according to the temperature of the key part monitored by the temperature sensor , obtain the vehicle battery management system data, analyze the current during charging , battery temperature , voltage , state of charge parameters, when the monitored value is greater than the set threshold value, it indicates that there is an abnormality in the monitored area, or it is the cause of the fire; Step 3.3.2, external cause analysis, mainly focus on whether there is human factor to cause fire, through the analysis of the video before the abnormal discovery for half an hour, from the abnormal temperature and abnormal behavior two aspects, explore the potential fire inducing factors; 3.3.2.1, Abnormal temperature analysis, by dividing the half-hour thermal imaging video data extracted by the thermal imaging binocular camera in the environmental visual monitoring module into a two-dimensional image matrix , setting a suitable temperature threshold for screening abnormal temperature points; each pixel point in the thermal imaging image corresponds to a temperature value When the temperature value of the pixel point meets , the pixel point is marked as a suspicious point; for the continuity judgment of the suspicious point, the overlap rate of the suspicious point regions of two adjacent frames is calculated , assuming that the suspicious point region in the frame is , and the suspicious point region in the frame is , the calculation formula of the overlap rate is: ; when the overlap rate is greater than the set continuity threshold , it is considered that the suspicious point has continuity, i.e. there is a possibility of continuous burning; for the judgment of whether the suspicious point region is expanding, the area of the suspicious point region is calculated; let the area of the suspicious point region in the t frame be , and the area of the suspicious point region in the frame be , when is satisfied, it is judged that the burning has an expanding trend;​ 3.3.2.2 Abnormal behavior analysis, for each frame of image The CNN model will output a feature vector. This vector contains behavioral feature information of the target in the current frame; calculate the feature vector of the current frame. Set of feature vectors of normal behavior Cosine similarity of each vector The formula for calculating cosine similarity is: ;in, Represents the dot product of vectors. and Representing vectors respectively and The modulus; set an abnormal behavior judgment threshold. When for all , When an abnormal behavior is detected in the current frame, it is determined that such behavior has occurred. Step 4, secondary feedback, after completing the traceability analysis of step 3, the relevant information needs to be processed and fed back according to the fire cause; 4.1, if the fire cause is internal, the related data involving charging pile or vehicle internal circuit abnormal are deeply sorted and fed back to the remote control center through network communication technology, in the form of intuitive charts and detailed data reports, and the detailed fault information and maintenance suggestions are fed back to the operation and maintenance personnel client; the accident cause, the expected recovery service time is fed back to the user client; 4.2, if the fire cause is external, the video clips of abnormal temperature analysis and abnormal behavior analysis intercepted in external cause analysis are sorted and fed back to the remote control center, the staff of the remote control center views the video on the monitoring platform, judges the severity of the accident and the potential risk, and the illegal information involved is timely connected to the law enforcement department.

2. The charging pile safety monitoring method of claim 1, wherein, The monitoring stage is as follows: The leakage information is captured by monitoring the current data of the charging pile through the leakage sensor; the overload or short circuit abnormal situation is found by monitoring the current change data of the charging line through the current transformer; the temperature data of the key heating components inside the charging pile are monitored through the temperature sensor; the fire in the surrounding area of the charging pile is monitored through the installation of ion type smoke sensor and infrared flame detector; the video information and temperature change of the surrounding area of the charging pile are obtained through the thermal imaging binocular camera; the key data of the vehicle battery are obtained through the communication connection with the vehicle battery management system.

3. The charging pile safety monitoring method of claim 1, characterized in that: The step 2 triggers the alarm mechanism and feeds back the alarm information to the remote monitoring center and the client when the ion type smoke sensor and the infrared flame detector in step 1 monitor that the smoke concentration exceeds the set threshold or detect the flame spectrum.

4. A monitoring system for performing the method of claim 1 to 3, characterized in that The system comprises a perception layer, a data transmission layer, a data processing layer and a user interaction layer; The perception layer is responsible for collecting various data, covering various sensors and devices; the data transmission layer undertakes the transmission task of the perception layer data, the data processing layer receives the data transmitted by the perception layer and processes and analyzes; the user interaction layer provides different services to operation and maintenance personnel and ordinary users.

Images