Intelligent terminal data acquisition control system and method based on electric power environment

Abstract

The invention belongs to the technical field of electric power environment monitoring, and particularly relates to an intelligent terminal data acquisition control system and method based on an electric power environment. The intelligent terminal data acquisition control method based on the electric power environment comprises the following steps that S10, an intelligent terminal acquires environment data and operation data in a target object, associates the environment data with the operation data according to acquisition time, generates an associated data set with a timestamp and stores the associated data set to the local; and S20, the intelligent terminal calls the local historical fault data and the associated data set, and an analysis model is constructed based on a multi-factor coupling analysis model. According to the scheme, the problem that the comprehensive assessment accuracy of the overall operation risk of the target object is low due to the fact that the intelligent terminal cannot capture the hidden risk formed by multi-factor coupling is solved through multi-factor coupling analysis, and meanwhile, the fault probability is cooperatively reduced, the environmental adaptability is enhanced, decision-making intelligence is achieved by combining hierarchical response and dynamic self-adaptive adjustment, and the risk assessment efficiency is improved. And full-life-cycle intelligent management of the power equipment is achieved.

Description

Technical Field

[0001] This solution belongs to the technical field of power environment monitoring, and specifically relates to an intelligent terminal data acquisition control system and method based on the power environment. Background Art

[0002] In power systems, target objects (such as power cabinets and power station buildings) serve as core hubs for power transmission and distribution, and their safe and stable operation is crucial. Existing technologies deploy smart terminals within these targets, connecting them to various sensors and collectors. These terminals capture data from these sensors and collectors and upload it to a central server, enabling remote monitoring of the target objects.

[0003] For example, Chinese patent CN110989476A discloses a monitoring system for remotely monitoring the environment of a target object. This system uses a monitoring center (smart terminal) to receive internal environmental and operational data from various sensors (such as temperature, humidity, smoke, current, and voltage sensors). Upon detecting an anomaly, the system communicates to a remote terminal (a central server or a maintenance personnel terminal) to achieve remote monitoring. The patent also incorporates a cooling module (an emergency response device, which in prior art also includes power control devices and sound and light alarm devices). This module, formed by a heat absorber, a radiator, and a fan, forms a closed heat transfer circuit. This module dissipates heat when temperatures exceed a certain limit, enabling intelligent data collection and control by the monitoring center.

[0004] However, while this patent collects multiple data points, such as temperature, humidity, current, voltage, and smoke, through sensor modules, its core data processing involves comparing each data point individually with a preset standard value to determine whether a single indicator is abnormal. While its monitoring center collects multiple data points, such as temperature, humidity, current, and voltage, each data point is calculated independently. There is no mechanism for multi-data correlation analysis, and it is even more incapable of calculating the correlation between multiple data points (such as the impact of coordinated changes in temperature and humidity on insulation performance). Furthermore, this patent's monitoring center is primarily responsible for data reception, anomaly determination, and command forwarding. It lacks the ability to perform joint calculations, such as weighted fusion and trend prediction, on multiple data points, making it difficult to convert multi-dimensional data into a comprehensive risk value. This design results in the system being able to identify only explicit risks arising from the exceeding of a single indicator, and failing to capture implicit risks arising from the coupling of multiple factors. This limits the comprehensive assessment of the overall operational risk of the target object. Summary of the Invention

[0005] The purpose of this solution is to provide an intelligent terminal data acquisition control system and method based on the power environment to solve the problem that the intelligent terminal is unable to capture the hidden risks formed by the coupling of multiple factors, resulting in low accuracy in the comprehensive assessment of the overall operational risk of the target object.

[0006] To achieve the above objectives, this solution provides a smart terminal data acquisition and control method based on the power environment, comprising the following steps: S10: The intelligent terminal collects environmental data and operating data from the target object, associates the environmental data with the operating data based on the collection time, generates a timestamp-containing associated data set, and stores it locally; S20: The intelligent terminal retrieves local historical fault data and associated data sets, constructs an analysis model based on a multi-factor coupling analysis model, and through training and fitting the data change function before and after the fault, identifies the dynamic correlation between the historical fault data and the associated data sets and calculates the impact ratio of each data on the fault. The currently collected data is used as the initial data, combined with the continuous data segments in the associated data sets, and input into the analysis model. The data after the preset time period is calculated as the predicted data. Based on the predicted data and the impact ratio, a weighted algorithm is used to obtain a comprehensive risk value. S30: The intelligent terminal classifies the comprehensive risk value into different risk levels, obtains the impact ratio from the analysis model according to the risk level, and activates different emergency devices according to the impact ratio and risk level; S40: The intelligent terminal adjusts the collection frequency of environmental data and operating data according to the risk level, sets the verification interval according to the collection frequency, uses the analysis model to calculate the comprehensive risk value as the adjusted risk based on the verification interval, adjusts the working power of the emergency device according to the changing trend of the adjusted risk obtained multiple times, and adjusts the collection frequency at the same time.

[0007] This solution also includes an intelligent terminal data acquisition and control system based on the power environment that uses the intelligent terminal data acquisition and control method based on the power environment.

[0008] The principle and technical effect of this solution are as follows: First, the intelligent terminal conducts in-depth correlation training on historical fault data and real-time collected environmental and operating data based on a multi-factor coupling analysis model. By identifying the dynamic correlation between different data dimensions, the influence of each factor on the fault is calculated, and then the data change function before and after the fault is fitted. In this way, this solution can capture hidden risks that cannot be discovered by traditional single indicator monitoring, such as the impact of the coordinated changes in temperature and humidity on the insulation performance of equipment, and realize the accurate identification of multi-factor coupling risks. It directly improves the comprehensiveness and accuracy of risk assessment, enabling this solution to provide early warning of potential fault hazards and avoid ignoring the risk accumulation caused by the superposition of multiple factors due to the normality of a single indicator. At the same time, it can also provide early warning of potential fault hazards and avoid ignoring the risk accumulation caused by the superposition of multiple factors due to the normality of a single indicator, thereby effectively preventing equipment failures.

[0009] Secondly, the intelligent terminal divides the comprehensive risk value into different levels and arranges the factors according to their impact ratio within each risk level, activating the corresponding emergency devices. In the low-risk stage, this solution implements preventative control measures for factors with higher impact ratios, such as activating local cooling or dehumidification devices. As the risk level increases, more key emergency devices are gradually activated, forming a gradient response mechanism from early warning to emergency disposal. This hierarchical control strategy ensures the targeted and effective risk response, not only delaying the development of risks through low-cost measures in the early stages, but also quickly activating comprehensive protection mechanisms in emergency situations to minimize the probability of equipment damage.

[0010] Furthermore, the intelligent terminal dynamically adjusts the data collection frequency and the operating power of the emergency device according to the risk level, forming a closed-loop feedback mechanism. In high-risk situations, this solution automatically increases the frequency of data collection and strengthens emergency measures to ensure real-time monitoring and rapid response to equipment status. In low-risk situations, it reduces the collection frequency and energy consumption to optimize resource allocation. By continuously monitoring risk trends, this solution can optimize control strategies in real time, such as increasing the power of emergency devices in advance based on rising risk trends, or reducing energy consumption in a timely manner after the risk is alleviated. This adaptive mechanism significantly improves the flexibility and energy efficiency of the solution, reducing operating costs while ensuring equipment safety.

[0011] In addition, this solution uses multi-factor coupling analysis to predict the risk evolution trend in advance, combines the hierarchical response mechanism to initiate precise prevention and control at the budding stage of the risk, and dynamically adaptively adjusts to continuously optimize the prevention and control strategy based on real-time risk changes. The three together significantly reduce the probability of failure. This collaborative mechanism also enhances the flexibility of this solution to cope with complex environments, allowing it to maintain a stable prevention and control effect in the face of uncertainties such as operating condition fluctuations and equipment aging, effectively extending the service life of the equipment. The implementation of intelligent decision-making further improves management efficiency. The optimal prevention and control method independently generated by this solution reduces the risk of manual intervention, and continuously improves the strategy through continuous learning, ultimately achieving the goal of intelligent management of the entire life cycle of power equipment.

[0012] In summary, this solution solves the problem of intelligent terminals being unable to capture the hidden risks caused by multi-factor coupling, resulting in low accuracy in the comprehensive assessment of the overall operational risk of the target object, through multi-factor coupling analysis. At the same time, combined with hierarchical response and dynamic adaptive adjustment, it synergistically reduces the probability of failure, enhances environmental adaptability, realizes intelligent decision-making, and achieves intelligent management of the entire life cycle of power equipment.

[0013] Furthermore, when constructing the analysis model in step S20, the following steps are included: S21: Cleans historical fault data and related datasets with timestamps, fills in missing values, and standardizes them. It extracts the collaborative features of environmental data and operational data at the same timestamp, as well as the correlation between changes in data at different timestamps. It also mines dynamic correlations from the collaborative features and correlations of data, including the coupling effects of data with the same timestamp and the evolutionary correlations of data across timestamps. S22: Based on the multi-factor coupling analysis theory, an architecture with a time correlation layer is constructed. The time correlation layer is used to process the temporal correlation relationship of various data at different timestamps. S23: Iteratively train the analysis model using historical fault data and associated data sets, adjust the associated weight parameters in the time dimension through cross-validation, optimize the model's predictive performance, and prevent overfitting through regularization techniques.

[0014] This solution significantly improves risk assessment accuracy and model adaptability by deeply mining the dynamic correlations of timestamped data. For example, during substation equipment operation, simultaneous high temperature and high humidity can synergistically accelerate insulation aging, and even small fluctuations in temperature and humidity over several hours can accumulate into hidden risks. Existing technologies, which analyze data only in a single context or ignore temporal correlations, are unable to identify these dynamic risks arising from the coupling of multiple factors. However, this solution, by extracting collaborative features from same-timestamp data and evolving correlations across timestamps, can provide early warning of such hidden dangers. Furthermore, an architecture with a temporal correlation layer can capture the temporal patterns of parameter changes during long-term equipment operation. For example, the gradual anomaly of transformer oil temperature as it changes with load is difficult to detect with existing technologies due to their lack of temporal analysis. However, this solution identifies trends through the temporal correlation layer, enhancing model robustness. After iterative training and parameter optimization, the model becomes more flexible in responding to unexpected risks. For example, it can quickly adjust policies based on historical data from similar operating conditions when a load surge occurs. Existing technologies, due to their fixed models, cannot dynamically adapt. However, this solution significantly improves the targeted and effective risk prevention and control through closed-loop optimization.

[0015] Furthermore, when calculating the comprehensive risk value in step S20, the initial weight is set according to the influence ratio of each data, the deviation between the current data and the predicted data is obtained, and the weighted weight is adjusted according to the deviation amplitude; when an abnormal trend similar to that before the historical failure appears in the predicted data, the weight of the corresponding data is increased; combined with the change in risk level, when the risk level increases, the weight of the data associated with the predicted data is increased, and when the risk level decreases, the weight is reduced according to the changes in each data.

[0016] By dynamically adjusting the weighting of the comprehensive risk value, the accuracy and real-time responsiveness of risk assessments can be significantly improved. For example, during substation equipment operation, if current temperature data deviates significantly from predicted data and shows a persistent upward trend, this solution immediately increases the temperature weight. Furthermore, if the predicted data exhibits current fluctuations similar to those seen before a historical short-circuit fault, the current weight is further increased, ensuring that high-risk factors are prioritized. As the risk level decreases, the weight of non-critical data (data with minimal fluctuations) is gradually reduced to avoid wasted resources. Existing technologies often use fixed weights or simply adjust them based on risk level. This approach fails to dynamically refine weight assignments based on data deviations, abnormal trends, and risk level, as this solution does. For example, in the case of equipment insulation aging failures, existing technologies may treat temperature and humidity equally. However, this solution increases the humidity weighting when humidity anomalies closely align with historical failure trends, accurately capturing hidden risks. This is difficult to achieve with existing technologies due to their lack of dynamic adaptation mechanisms. Furthermore, by regularly calibrating weights with actual fault data, the assessment model continuously adapts to the actual operating patterns of the equipment, addressing the issue of existing technologies where weights become rigid, leading to a decrease in assessment accuracy over time.

[0017] Furthermore, in steps S30-S40, based on the degree of impact of historical faults and the safe operation boundaries of each device within the target object, multiple risk levels are defined, and each risk level corresponds to a safe state range and abnormal trend characteristics of the device parameters; the intelligent terminal dynamically adjusts the data collection frequency according to the current risk level, including adjusting the collection frequency of environmental data and operation data, as well as the difference in collection accuracy of different data; different prediction interval durations are set for different risk levels. When the risk level changes, the data collection frequency and prediction interval duration are gradually transitioned to maintain the continuity of data processing and ensure real-time matching of risk assessment with the actual operating status of the equipment.

[0018] This solution dynamically adjusts data collection frequency and the operating power of emergency devices, forming a closed-loop feedback mechanism that significantly improves flexibility, energy efficiency, and risk response capabilities. For example, in high-risk situations, intelligent terminals automatically increase data collection frequency and strengthen emergency measures to ensure real-time monitoring and rapid response to equipment status. In low-risk situations, they reduce data collection frequency and energy consumption, optimizing resource allocation. This dynamic adjustment mechanism enables the solution to continuously optimize prevention and control strategies based on real-time risk changes. For example, it can proactively increase emergency device power based on rising risk trends or promptly reduce energy consumption after risk mitigation, significantly improving the solution's flexibility and energy efficiency. Furthermore, this solution uses multi-factor coupled analysis to predict risk trends in advance, combines a graded response mechanism to initiate precise prevention and control measures at the initial stage of risk development, and dynamically adapts to continuously optimize prevention and control strategies based on real-time risk changes. These three combined measures significantly reduce the probability of failure. This synergistic mechanism also enhances the solution's resilience in complex environments, enabling it to maintain stable prevention and control effectiveness despite uncertainties such as fluctuating operating conditions and aging equipment, effectively extending equipment life.

[0019] In contrast, the existing technologies in the background art are unable to optimize the frequency of data collection and the operating power of emergency devices according to real-time risk changes due to the lack of dynamic adjustment capabilities. For example, in high-risk situations, the existing technologies may miss key data due to insufficient collection frequency, resulting in response delays; in low-risk situations, they may consume excessive energy. In addition, the existing technologies lack multi-factor coupling analysis and hierarchical response mechanisms, making it impossible to predict the evolution trend of risks in advance, nor to initiate precise prevention and control at the budding stage of risks, resulting in a higher probability of failure and a shortened equipment life. Therefore, the existing technologies are unable to achieve the effects of dynamic adjustment, risk prediction, and hierarchical response mentioned in this solution.

[0020] Furthermore, the intelligent terminal constructs a topological network based on the physical location relationship between the target objects. The intelligent terminal regards the adjacent intelligent terminals in the topological network as adjacent terminals, and integrates the associated data sets and historical fault data of the local and adjacent terminals through the topological network. Based on the spatiotemporal correlation characteristics of the data, the causal relationship pattern between the adjacent target object data and the local data is identified, and an adjacent impact model reflecting the degree and direction of data impact is constructed; the intelligent terminal generates verification information according to the risk level and sends it to the adjacent terminal. After receiving the verification information, the adjacent terminal inputs the currently collected data into the adjacent impact model to infer the expected data of the local target object, and then compares it with the actually collected data. According to the comparison result, feedback information of the verification information is generated and sent to the intelligent terminal; the intelligent terminal summarizes the feedback information, infers the confidence of the current risk level according to the summarized result, and adjusts the collection frequency of environmental data and operation data according to the confidence.

[0021] By building a topological network to enable data linkage and verification between intelligent terminals, this solution significantly improves the accuracy and dynamic adaptability of risk assessments. For example, when a switchgear intelligent terminal in a substation detects a medium risk, it sends a verification request to an adjacent transformer terminal via the topological network. The adjacent terminal uses a constructed neighboring impact model to estimate the potential changes in transformer data caused by the switchgear. After comparing the data with the actual collected data, it provides feedback on any deviations. The switchgear terminal then adjusts its data collection frequency accordingly—if the deviation is significant, it increases the frequency of key data collection to accurately capture the risk's spread. Furthermore, through multi-terminal linkage confidence assessment, this solution enables intelligent terminals to more accurately determine the authenticity of risks and dynamically optimize collection strategies. For example, when a line terminal detects an overcurrent risk, it combines load data from adjacent transformers for cross-validation, avoiding risk assessment bias caused by misjudgment of data from a single device. Furthermore, by continuously updating the neighboring impact model, the intelligent terminal can better adapt to long-term factors such as equipment aging and environmental changes, improving the overall stability of the system.

[0022] In complex power environments, the mutual influence of devices is inevitable. This solution, through a cross-device data association and verification mechanism, can achieve more accurate and efficient risk management and control. In addition, the smart terminal sends a verification signal to the adjacent smart terminal based on the risk level, and corrects its own collection frequency and interval time according to the risk level of the adjacent terminal, achieving dynamic adjustment and coordinated response. This mechanism can ensure the real-time and accuracy of risk assessment, and at the same time, through the coordinated response mechanism, it can more effectively reduce the probability of failure. For example, when a device in a certain area fails, the adjacent terminals can increase the collection frequency, respond quickly, and take measures to jointly address the risk.

[0023] In existing technologies, most smart terminals operate independently and lack a cross-device data association verification mechanism, which often leads to risk assessment bias due to misjudgment of data from a single device. Furthermore, the acquisition frequency adjustment of existing technologies relies solely on local data, making it difficult to achieve dynamic optimization of acquisition strategies through multi-terminal linkage confidence assessments like this solution, and even more difficult to adapt to long-term factors such as equipment aging and environmental changes through continuously updated adjacent impact models. Therefore, in complex power environments, existing technologies have difficulty achieving accurate and efficient risk management and control effects. However, through the above-mentioned optimization features, this solution can better meet the risk management and control needs in complex power environments and achieve intelligent management of power equipment throughout its life cycle.

[0024] Furthermore, when there are multiple adjacent terminals in a smart terminal, the adjacent impact models are first prioritized based on the physical distance between each adjacent target object and the local target object; each adjacent impact model is superimposed in order of priority, and the expected data in the same time window is weightedly fused according to the preset window time to obtain the expected data distribution under the comprehensive impact, and then the expected data of each adjacent terminal in the same time and space area are compared, the overlapping or conflicting expected data are identified and the valid expected data are screened out; the valid expected data are spatiotemporally integrated to generate the final expected data, the final expected data is compared with the actual collected data, and the weight distribution when the adjacent impact models are superimposed is adjusted according to the comparison results.

[0025] This solution significantly improves risk assessment accuracy and dynamic adaptability through collaborative verification among intelligent terminals. For example, in a substation, when a switchgear terminal detects a risk, it sorts and superimposes adjacent impact models based on physical distance, generating a comprehensive expected data distribution through weighted fusion. The expected data from adjacent terminals is compared, valid data is filtered and integrated, and finally compared with actual data to adjust weight distribution. This process integrates multi-source data, avoids single-source misjudgment, eliminates inaccurate data, ensures reliable results, and adapts to long-term factors such as equipment aging and environmental changes. Existing technologies treat intelligent terminals as independent entities, lack cross-device data correlation verification, and are unable to integrate data from adjacent devices for cross-validation. This can easily lead to biased risk assessments due to single-source misjudgment. Existing technologies rely solely on local data for collection frequency adjustment, making it impossible to optimize collection strategies through multi-terminal collaboration or adapt to long-term factors through continuously updated models. This makes it difficult to accurately and efficiently manage risks. Through data linkage verification and dynamic adjustment of weight distribution, this solution better addresses risk management needs in complex power environments, enabling intelligent management of power equipment throughout its lifecycle, and significantly improving the system's intelligence and stability.

[0026] First, in complex scenarios with multiple adjacent terminals, this solution effectively integrates the impact of multiple devices by overlaying adjacent impact models sorted by physical distance and combining them with a filtering mechanism for spatially and temporally overlapping data. For example, when processing voltage data at an industrial park distribution terminal, models of adjacent distribution rooms, transformer banks, and cable lines are prioritized by distance, filtering for highly accurate overlapping data. This ensures that the predicted voltage is more realistic, resolving the bias caused by simple averaging in existing technologies. Second, when some adjacent terminals generate abnormal data due to faults or interference, the multi-model overlay and filtering mechanism automatically identifies and removes outliers. For example, in wind power scenarios, the collector line terminal can eliminate abnormal wind speed data from faulty wind turbines, ensuring stable load forecasts based on the consensus of other terminals and preventing single abnormal data from influencing overall predictions. Furthermore, existing smart terminals typically use fixed time windows and weight adjustment frequencies, processing data uniformly regardless of peak or off-peak demand. This results in insufficient computing power during peak periods and wasted resources during off-peak periods. This solution dynamically adjusts the time window and weight adjustment frequency to achieve on-demand computing power allocation. During low-power periods, the window duration is extended and the frequency of adjustments is reduced to reduce consumption. During peak periods, the window duration is shortened and the weights of key models are increased, concentrating resources on processing core data. This is more efficient than the fixed model of existing technologies. Furthermore, as equipment ages or the environment changes, continuous weight adjustments can automatically adapt to changes in influencing relationships. For example, after years of operation, a substation bus terminal can autonomously reduce the transformer model weight and increase the cable line weight without manual intervention. Existing technologies rely on manual reconfiguration and struggle to cope with long-term dynamic changes, highlighting the superiority of this solution.

[0027] Furthermore, the intelligent terminal calculates the expected data based on the verification information and performs a risk assessment on the expected data, enables or adjusts the access current and working power of the emergency device based on the risk assessment results, and generates a maintenance signal based on the risk assessment results and the verification information and sends it to the central server.

[0028] When a target object experiences a serious problem, this solution derives expected data by overlaying adjacent impact models. This approach then performs a tiered risk assessment based on deviation levels, parameter correlations, and historical similarities. This approach not only accurately determines the fault type and scope, but also predicts electromagnetic shielding risks based on data characteristics. For example, when a substation switchgear shorts, the terminal utilizes models of adjacent transformers and busbars to identify the short circuit location and inform transmission strategies. Furthermore, this solution utilizes the model-derived results to determine the risk level of electromagnetic shielding that may be triggered. This allows verification information to provide both computational assistance and transmission support, avoiding transmission delays caused by existing technologies that fail to predict shielding. Furthermore, for predicted electromagnetic shielding risks, this solution utilizes frequency hopping transmission and appends corroborating data from adjacent terminals to ensure that the central server receives the verification information completely. For example, when a line is overloaded and potentially shielding is present, the intelligent terminal transmits information via frequency hopping, along with load data from adjacent towers, addressing the issue of single signal loss.

[0029] Furthermore, after receiving the maintenance signal, the central server will use the intelligent terminal that generates the verification information in the maintenance signal as the starting terminal, and define the initial maintenance range including the starting terminal and the adjacent terminals of the starting terminal in the topological network; based on the risk assessment results in the maintenance signal, the corresponding risk level is marked for each location within the initial maintenance range; based on the risk level and the electromagnetic characteristics and correlation relationship of the target object corresponding to the intelligent terminal, the electromagnetic interference intensity of each location in the initial maintenance range is calculated; and the maintenance route is planned in combination with the electromagnetic interference intensity, risk level and target object distribution.

[0030] This solution combines risk assessment results with verification information. These two elements not only initiate assisted verification with adjacent smart terminals, but also, when the risk level is high and electromagnetic shielding is strong, form a multi-layered maintenance signal chain based on adjacent terminals. When strong electromagnetic shielding blocks direct signal transmission in the area where the target object is located, the smart terminal generating the maintenance signal sends the risk assessment results and verification information to the nearest adjacent terminal. After verification, the adjacent terminal continues to transmit the signal to its adjacent terminals until the signal breaks through the shielded area and reaches the central server. This avoids information loss caused by interruptions in a single transmission path and addresses the existing issues of maintenance signals being susceptible to shielding interference and low transmission reliability.

[0031] At the same time, after receiving the maintenance signal transmitted through layers, the central server can accurately trace the diffusion path of the fault impact by tracing the verification information attached to each terminal, thereby more accurately defining the maintenance scope. For example, based on the differences in verification information between adjacent terminals during the signal transmission process, the core area and edge areas affected by the fault can be identified to avoid excessive scope or omissions. At the same time, combined with the signal attenuation level feedback from each terminal in the risk assessment results, the electromagnetic interference intensity at different locations is calculated, and then a more practical maintenance route is formulated: priority is given to paths with weaker interference intensity and that can cover the core area, reducing signal loss of maintenance equipment in strong interference areas, and improving maintenance efficiency and accuracy.

[0032] Furthermore, when the smart terminal is connected to the maintenance terminal for communication, the environmental data and operating data of the target object are sent to the maintenance terminal. After the smart terminal disconnects the communication connection with the maintenance terminal, the smart terminal collects and updates the environmental data and operating data to the adjacent terminal; the central server updates the initial maintenance range and the electromagnetic interference intensity of each location based on the most recently received maintenance signal; the central server obtains the environmental data and operating data of the target object from the maintenance terminal based on the initial maintenance range.

[0033] The smart terminal establishes an encrypted communication link with the maintenance terminal and verifies identity permissions to ensure real-time and secure transmission of environmental data and operational data, avoiding the problem of open transmission being easily interfered with or stolen, allowing maintenance personnel to accurately determine the cause of the fault based on dynamic data; the maintenance terminal stores the data locally, which serves as an immediate reference for on-site maintenance and retains the original records. This solves the problem of traditional maintenance where data relies solely on central server storage and is easily lost due to transmission interruptions. Moreover, this data can be used as a reference case in subsequent similar fault maintenance to form knowledge accumulation. After the smart terminal is disconnected from the maintenance terminal, it collects data at high frequency and updates it to adjacent terminals to ensure that peripheral equipment can promptly perceive the status changes after the fault is repaired, providing a basis for preliminary adjustments to the maintenance scope. The central server updates the maintenance scope and electromagnetic interference intensity based on the latest maintenance signal, avoiding the solidification of the maintenance scope due to data lag. After the maintenance terminal leaves the electromagnetic interference range, it synchronizes the first-hand data it directly obtains to the central server. The central server compares this data with the data transmitted by the relay and corrects the range deviation caused by the accumulated errors in the transmission chain. This ensures that the maintenance range can accurately cover the actual affected area while avoiding the waste of resources caused by excessive expansion. Compared with the method of relying solely on a single data source to locate the range, the accuracy is significantly improved. At the same time, the central server integrates the complete data feedback from the maintenance terminal to form a closed-loop record from fault discovery, maintenance implementation, to effect verification. This provides a basis for subsequent risk assessment model iteration and optimization of the electromagnetic interference intensity calculation algorithm, promotes the intelligent upgrade of this solution, and solves the problem of data dispersion and difficulty in forming effective feedback in existing technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 Flowchart of a method for collecting and controlling data from a smart terminal based on an electric power environment in an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The following will clearly and completely describe the concept and technical effects of the present invention in conjunction with the embodiments to fully understand the purpose, features and effects of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present invention: like Figure 1 As shown, the intelligent terminal data acquisition and control method based on the power environment includes the following steps: S10: The intelligent terminal collects environmental data and operating data from the target object, associates the environmental data with the operating data based on the collection time, generates a timestamp-containing associated data set, and stores it locally; S20: The intelligent terminal retrieves local historical fault data and associated data sets, constructs an analysis model based on a multi-factor coupling analysis model, and through training and fitting the data change function before and after the fault, identifies the dynamic correlation between the historical fault data and the associated data sets and calculates the impact ratio of each data on the fault. The currently collected data is used as the initial data, combined with the continuous data segments in the associated data sets, and input into the analysis model. The data after the preset time period is calculated as the predicted data. Based on the predicted data and the impact ratio, a weighted algorithm is used to obtain a comprehensive risk value. S30: The intelligent terminal classifies the comprehensive risk value into different risk levels, obtains the impact ratio from the analysis model according to the risk level, and activates different emergency devices according to the impact ratio and risk level; S40: The intelligent terminal adjusts the collection frequency of environmental data and operating data according to the risk level, sets the verification interval according to the collection frequency, uses the analysis model to calculate the comprehensive risk value as the adjusted risk based on the verification interval, adjusts the working power of the emergency device according to the changing trend of the adjusted risk obtained multiple times, and adjusts the collection frequency at the same time.

[0036] Environmental data includes temperature (°C), humidity (%RH), dust concentration (mg / m³), and electromagnetic radiation intensity (μT), all collected by corresponding sensors deployed within the target object. Operational data includes current (A), voltage (V), power (kW), insulation resistance (MΩ), and device start / stop status (0 / 1), acquired in real time by the power monitoring module. Historical fault data is derived from the local logs of smart terminals and includes the fault timestamp, fault type (such as short circuit, overload, insulation aging, etc.), and environmental and operational data for the hour before and after the fault. These data are automatically recorded and stored by the terminal. In step S20, the trained multi-factor coupling analysis model is used to fit the pre- and post-fault data change function. A random forest algorithm is then used to quantify the contribution of each data point to the fault (i.e., the impact percentage). For example, for a short circuit fault, the model calculates specific values ​​such as 30% for temperature and 40% for current.

[0037] The construction of the analysis model in step S20 includes the following steps: S21: Clean historical fault data and related data sets with timestamps, fill in missing values ​​and standardize them, extract the collaborative features of environmental data and operation data at the same timestamp, and the correlation between changes in various data at different timestamps; mine dynamic associations from the collaborative features and correlations of various data, including the coupling effects of data with the same timestamp and the evolutionary correlations of data across timestamps.

[0038] Specifically, data preprocessing and dynamic association mining are used to clean the historical fault data and associated data sets with timestamps, and outliers are removed using a density-based method. Missing values ​​are interpolated using a time series interpolation method. Filling, using the temporal continuity of data to improve filling accuracy. is the missing value to be filled (at time t), is the previous timestamp ( Time) known data, is the next timestamp ( known data at that time.

[0039] Because power data (such as temperature, current, etc.) has the characteristic of continuous fluctuation, when a data point is significantly different from the normal data of adjacent timestamps (i.e., in a low-density distribution area), it is identified as an outlier. For example, if the temperature data of a timestamp is much higher than the temperature data of the adjacent timestamps before and after, and does not conform to the physical laws of equipment operation, it is identified as an outlier. If the suspected outlier is different from the calculated interpolation result If the deviation exceeds a reasonable range (e.g., the deviation exceeds 20% of the normal fluctuation threshold), the data point is identified as an outlier and removed. After the outlier is removed, the remaining missing data points are filled using the above time series interpolation formula to ensure the continuity and integrity of the data set, providing a reliable data foundation for subsequent multi-factor coupling analysis and model training.

[0040] The standardization process uses the improved min-max formula, as shown in the following formula (1): (1), in, is the normalized data value, The original data value By scaling factor (Default 0.8, adjustable range 0.5-1.0) and offset (Default 0.1, adjustable range 0-0.3) Adjust data distribution to adapt to the fluctuation characteristics of power data.

[0041] When extracting collaborative features of the same timestamp, the synergy coefficient is used to quantify the degree of linear synergy between the environment and the operating data. The closer the value is to 1, the stronger the synergy. The calculation formula of the synergy coefficient is shown in the following formula (2): (2), in, is the synergy coefficient between the environment and the operation data at timestamp t; The i-th environmental data item at timestamp t (such as temperature, humidity); is the i-th item of operating data (e.g., current, voltage) at timestamp t. Select three items from the environmental data that are closely related to the operating data (temperature, humidity, and electromagnetic radiation intensity), and select three items from the operating data (current, voltage, and power), ensuring that the number of environmental data items and the number of operating data items involved in the synergy coefficient calculation are both 3 (i.e., n = 3).

[0042] Cross-timestamp evolution correlation is measured by evolution correlation, combining data change rate and time impact weight. The larger the value, the more significant the potential impact on the fault. The specific details are shown in the following formula (3): (3), in, For the time interval The correlation of data evolution within Timestamp and The i-th data under is the time interval (unit: minute, default is 15 minutes); is the time impact weight of the i-th data (temperature =0.3, current =0.4, others =0.1).

[0043] S22: Based on the multi-factor coupling analysis theory, an architecture including a time correlation layer is constructed, where the time correlation layer is used to process the temporal correlation relationship of various data at different timestamps.

[0044] Specifically, the architecture design of the temporal association layer uses a temporal attention mechanism, and the core calculation is shown in the following formula (4): (4), in, is the attention output vector at timestamp t; is the query vector; is the key vector; is a value vector, 、 、 The input time series data (associated data set with timestamps) is passed through three learnable weight matrices 、 、 Perform linear transformation to obtain = × , = × , = × ,in is the input data vector at time t, 、 、 The weight matrix optimized during model training (initial value is randomly initialized, dimension is d_model× ). ; It is a scaling factor to prevent the gradient from disappearing due to excessive dot product; is a normalization function that compresses the weights to the [0,1] interval.

[0045] Formula (4) is obtained by querying the vector , key vector Sum value vector It automatically focuses on key historical timestamp data and enhances the ability to capture time series correlation.

[0046] S23: Iteratively train the analysis model using historical fault data and associated data sets, adjust the associated weight parameters in the time dimension through cross-validation, optimize the model's predictive performance, and prevent overfitting through regularization techniques.

[0047] Specifically, model training and optimization training adopt a dynamic learning rate strategy , combining exponential decay and periodic fluctuations to balance stability and optimization capabilities. is the learning rate at iteration number t; is the initial learning rate (default 0.001); is the attenuation coefficient (default 0.0001); is the fluctuation amplitude (default 0.1); is the fluctuation frequency (default 0.01).

[0048] Regularization uses the elastic net regularization term , redundant features are eliminated through L1 and L2 regularization coefficients to prevent overfitting. is the regularization loss term; is the L1 regularization coefficient (default 0.0005); is the L2 regularization coefficient (default 0.001); is the weight parameter of the temporal association layer.

[0049] Cross-validation uses a hierarchical time series partitioning method to divide the data set into a ratio of 7:1:2, and the loss value of the validation set is used. Adjust the time-related weights to optimize the model generalization performance. is the average loss of the validation set; m is the number of samples in the validation set (usually m=2000); is the actual fault label (0=normal, 1=fault); is the model predicted probability value (range [0,1]).

[0050] When calculating the comprehensive risk value in step S20, the initial weight is set according to the influence ratio of each data, the deviation between the current data and the predicted data is obtained, and the weighted weight is adjusted according to the deviation amplitude; when an abnormal trend similar to that before the historical failure appears in the predicted data, the weight of the corresponding data is increased; combined with the change in risk level, when the risk level increases, the weight of the data associated with the predicted data is increased, and when the risk level decreases, the weight is reduced according to the changes in each data.

[0051] The logic behind adjusting weights based on risk level is to strengthen risk correlation. As the risk level increases, data strongly correlated with predicted data (such as temperature and insulation resistance) has a greater potential impact on failures, and increasing their weights enhances the sensitivity of the risk assessment. As the risk level decreases, the impact of non-critical data decreases, and reducing their weights reduces redundant calculations. For example, at high risk, if predicted data shows a strong correlation between temperature and historical short-circuit failures, the temperature weight is increased to ensure it accounts for a higher proportion of the overall risk value.

[0052] The analytical model generates forecast data as follows: It takes the currently collected initial data and continuous data segments from the associated dataset (e.g., time series data from the past hour) as input. The model's time association layer captures the data's temporal evolution characteristics (e.g., the trend of temperature changes with current). Combined with the synergistic features of the data output by the multi-factor coupling layer (e.g., the correlation between temperature, humidity, and insulation resistance), a long short-term memory (LSTM) network is used to predict various data items for preset future time periods (e.g., 10 minutes, 30 minutes, or 1 hour), outputting a predicted data sequence. The LSTM's hidden layer dimension is 128, and the output layer is mapped to a predicted value that matches the number of input data items through a fully connected layer.

[0053] In one implementation of this solution, initial weights are set based on the impact of each data point and are pre-set based on historical fault data. For example, in a transformer risk assessment, if winding temperature is the key factor in most faults, its initial weight will be higher than load current, providing a benchmark for subsequent adjustments.

[0054] The weight is adjusted based on the deviation between real-time and predicted data. For example, if the transformer oil temperature is predicted to be 60°C but the actual value is 75°C, exceeding the normal range, the weight of the oil temperature will be increased accordingly. If the actual ambient humidity is close to the predicted value, the weight will be maintained or slightly adjusted. This mechanism strengthens the role of abnormal data in risk assessment.

[0055] When the predicted data shows an abnormal trend similar to that before a historical fault, the corresponding data weight is increased. For example, before a switchgear fault, the trip coil current showed a trend of "fluctuation → increase → sudden rise". If the current data shows the same pattern, the weight of this current data will be immediately increased, even if the absolute value does not reach the warning threshold, thus identifying the precursor to the fault in advance.

[0056] Adjust weights as risk levels change. In high-risk situations, key data (such as current and temperature) that are strongly correlated with predicted data have a significant direct impact on faults, and their weights are maintained or increased (such as the current weight increasing from 40% to 50%). Non-critical data (such as dust concentration) contributes less to the current risk, and their weights are reduced (such as from 10% to 5%) to reduce interference from unnecessary factors. In low-risk situations, all data are in a safe range, and the weights of key data maintain their initial values ​​(based on the impact ratio setting). The weights of non-critical data are adjusted according to their fluctuations. If the data does not change significantly (such as humidity remaining stable within the normal range), the weight is appropriately reduced (such as from 10% to 8%). If the data shows a slight abnormal trend (such as a slow increase in humidity), the weight remains unchanged for continuous monitoring.

[0057] Regularly compare actual failures with model predictions to calibrate weights. If cooling system flow data repeatedly causes equipment overheating failures due to underweighting, its weight will be increased. This addresses the issue of "outdated weights" caused by equipment aging and environmental changes, ensuring the model adapts to actual risk characteristics.

[0058] Existing technologies often use fixed weights or adjust them based solely on risk level. This solution achieves precise weight adaptation through a four-tiered dynamic adjustment process: deviation → trend → level → calibration. For example, in cable fault assessment, existing technologies may assign fixed weights to cable temperature and air humidity. However, this solution increases the weight of each when humidity exhibits trends similar to historical corrosion failures, and decreases it when temperature deviations are small, thus improving the targeted nature of the assessment.

[0059] In steps S30-S40, based on the impact of historical faults and the safe operating boundaries of each device within the target object, multiple risk levels are defined. Each risk level corresponds to a safe status range and abnormal trend characteristics of the device parameters; the intelligent terminal dynamically adjusts the data collection frequency according to the current risk level, including adjusting the collection frequency of environmental data and operating data, as well as the difference in collection accuracy of different data; different prediction interval durations are set for different risk levels. When the risk level changes, the data collection frequency and prediction interval duration are gradually transitioned to maintain the continuity of data processing and ensure real-time matching of risk assessment with the actual operating status of the equipment.

[0060] In one implementation of this solution, risk levels are categorized into low, medium, and high, taking into account the impact of historical failures on equipment operation and the parameter boundaries for safe equipment operation. Low risk corresponds to equipment parameters remaining within the safe range with no abnormal fluctuations; medium risk corresponds to parameters approaching the safe range or experiencing occasional minor abnormal fluctuations; and high risk corresponds to parameters exceeding the safe range or showing a continuous abnormal deterioration trend.

[0061] When adjusting data collection frequency, the smart terminal dynamically adjusts the collection mechanism based on the current risk level. In low-risk scenarios, environmental and operational data are collected at regular, fixed intervals, while non-critical data is collected with reduced accuracy. In medium-risk scenarios, critical data is collected more frequently and at shorter intervals, while non-critical data maintains its regular frequency. In high-risk scenarios, all data is collected at the highest frequency, with critical data collected at the highest accuracy level.

[0062] Among them, key data refers to data that is strongly correlated with faults (such as current, voltage, temperature, and insulation resistance), and its abnormal changes directly lead to faults (such as current exceeding the limit causing a short circuit); non-key data refers to data that is weakly correlated with faults (such as dust concentration and environmental noise), and its abnormalities require long-term accumulation before they may affect the equipment (such as dust accumulation leading to poor heat dissipation).

[0063] In the prediction interval time setting, the corresponding prediction interval length is matched to each risk level. The prediction interval is the longest when the risk is low, the interval is shorter when the risk is medium, and the interval is the shortest when the risk is high. High-frequency prediction is used to track data changes in real time.

[0064] During the transition of risk level changes, when the risk level rises from low to medium or medium to high, the data collection frequency is gradually increased according to a fixed gradient, and the prediction interval is gradually shortened according to a fixed gradient; when it drops from high to medium or medium to low, the collection frequency and prediction interval are reversed according to the same gradient to ensure continuous and stable data processing and achieve accurate matching of risk assessment and actual equipment status.

[0065] The intelligent terminal builds a topological network based on the physical location relationship between the target objects. The intelligent terminal regards the adjacent intelligent terminals in the topological network as adjacent terminals, integrates the associated data sets and historical fault data of the local and adjacent terminals through the topological network, identifies the causal relationship pattern between the adjacent target object data and the local data based on the spatiotemporal correlation characteristics of the data, and constructs an adjacent impact model that reflects the degree and direction of data impact.

[0066] Among them, based on the spatiotemporal correlation of data, the Granger causality test is adopted: if the voltage data of adjacent terminals changes before the local current changes, and is statistically significantly correlated, then the "adjacent voltage → local current" is judged to be a causal relationship.

[0067] A data acquisition and control method for smart terminals based on the power environment uses smart terminals to collect environmental and operational data, associate them to generate and store time-stamped data sets. Historical fault data and associated data sets are retrieved to construct a multi-factor coupling analysis model. The correlation between data and faults is identified, and the impact of each data on the fault is calculated to obtain a comprehensive risk value. Risk levels are divided according to the comprehensive risk value, and different emergency devices are activated based on the level and impact ratio. The data acquisition frequency and emergency device power are dynamically adjusted. At the same time, the smart terminal constructs a topological network based on physical location, integrates data from adjacent terminals to construct an adjacent impact model, adjusts the acquisition frequency through mutual verification, generates maintenance signals, and plans maintenance routes.

[0068] The smart terminal generates verification information based on the risk level and sends it to the adjacent terminal. After receiving the verification information, the adjacent terminal inputs the currently collected data into the adjacent impact model to infer the expected data of the local target object, and then compares it with the actual collected data. Based on the comparison results, it generates feedback information of the verification information and sends it to the smart terminal.

[0069] Smart terminals aggregate feedback information and calculate the confidence level of the current risk level based on the aggregated results. The frequency of environmental and operational data collection is adjusted accordingly. The confidence level is calculated based on the consistency of feedback from adjacent terminals. Specifically, if the verification results of n adjacent terminals are aggregated and the deviation between the expected data of k terminals and the actual local data is less than 10%, the confidence level is k / n × 100%. For example, if two of three adjacent terminals verify the same results, the confidence level is 67%.

[0070] Specifically, when there are multiple adjacent terminals in a smart terminal, the adjacent impact models are first prioritized based on the physical distance between each adjacent target object and the local target object; each adjacent impact model is superimposed in order of priority, and the expected data in the same time window is weightedly fused according to the preset window time to obtain the expected data distribution under the comprehensive impact, and then the expected data of each adjacent terminal in the same time and space area are compared, the overlapping or conflicting expected data are identified and the valid expected data are screened out; the valid expected data are spatiotemporally integrated to generate the final expected data, the final expected data is compared with the actual collected data, and the weight distribution when the adjacent impact models are superimposed is adjusted according to the comparison results.

[0071] Among them, when the effective expected data is integrated in time and space to generate the final expected data, the effective expected data of each adjacent terminal are aligned according to the time window (such as 10 minutes); the weighted average (the closer the distance, the higher the weight) is used to fuse the overlapping data in space; for conflicting data (deviation > 30%), the terminal data with historical verification accuracy > 80% is retained; and the final expected data curve covering the next 10 minutes, 30 minutes, and 1 hour is generated.

[0072] More specifically, the closer the physical distance is to the adjacent target object, the higher the initial weight of its adjacent impact model; if the physical distance is the same, the priority is adjusted according to the actual impact frequency of the adjacent object on the local target object in historical data, and the model weight with a higher impact frequency has a higher priority.

[0073] The preset window time is dynamically adapted according to the operating characteristics of the equipment. For example, for switchgear with high response speed requirements, the window time is set to a shorter period. For transformer equipment with slow changes, the window time is set to a longer period, ensuring that high-frequency calculation redundancy is avoided while capturing changes in key data.

[0074] When the expected data of multiple adjacent terminals overlap in the same space-time region, the consistent expected data of the majority of adjacent terminals shall be regarded as the valid data; if there is a conflict (i.e., the expected data of different terminals are significantly different), the expected data output by the model with higher accuracy shall be retained in combination with the accuracy of each adjacent impact model in historical verification.

[0075] If the final expected data deviates significantly from the actual collected data, the weight of the adjacent influencing model whose output is closer to the actual data will be increased, and the weight of the model with the larger deviation will be reduced; if the deviation remains small, the current weight distribution will be maintained to reduce unnecessary adjustment consumption.

[0076] The intelligent terminal calculates the expected data based on the verification information and performs a risk assessment on the expected data. Based on the risk assessment results, the intelligent terminal enables or adjusts the access current and working power of the emergency device, and generates a maintenance signal based on the risk assessment results and verification information and sends it to the central server.

[0077] In a specific implementation of this solution, the smart terminal receives verification information through an encrypted communication protocol. The information contains three core parts. Basic parameter range: the threshold value of the key operating parameters of the target object (such as the allowable range of voltage fluctuation, the normal temperature range, etc.), which serves as the baseline for data verification; historical normal data characteristics: the parameter fluctuation pattern and typical stable operation curve of the target object under normal operating conditions in the past three months, which presents the normal status of the equipment in a quantitative manner; abnormal sign records: recent non-fatal abnormal data (such as instantaneous voltage deviation, slight temperature exceeding the limit, etc.) and the timestamp of occurrence, which are used to associate the development trend of the current fault. During the parsing process, the smart terminal automatically extracts the numerical range and characteristic parameters in the information and converts them into a structured data format that can be directly used for calculation.

[0078] The smart terminal invokes the stacked neighboring impact models, activating each neighboring terminal's impact model sequentially based on physical distance, with the closest neighboring impact model taking precedence. Each neighboring impact model independently derives the expected future change curve for the target object's key parameters (such as current, temperature, and insulation resistance) based on historical trends and current anomaly signs in the verification information. The derived results of each neighboring impact model are weighted. For example, for every 50% decrease in distance, the corresponding model's weight increases by 30%, ultimately forming a comprehensive set of expected data covering three timeframes: 10 minutes, 30 minutes, and 1 hour.

[0079] For the integrated expected data, the deviation between the expected data and the basic parameter range in the verification information is calculated. Deviations exceeding 30% are considered high risk, 10%-30% are medium risk, and less than 10% are low risk. The consistency of the changing trends of different parameters is analyzed. If current anomalies and temperature exceed the limit occur simultaneously, the correlation is determined to be "close", while if only a single parameter is abnormal, it is determined to be "weak". The current expected data curve is compared with the curves in the historical fault database. If the match degree exceeds 80%, it is determined to be "highly similar", 60%-80% is "moderately similar", and less than 60% is "lowly similar". The final risk level is determined based on the weighted formula of "deviation degree × 40% + correlation × 30% + historical similarity × 30%".

[0080] Maintenance information includes four core categories: Abnormal location positioning: Combining the parameter gradient changes derived from multiple models, locate the physical area with the most significant abnormal parameters; Possible fault types: Based on historical similarity results in risk assessment, match typical types in the fault database (such as insulation aging, poor contact, etc.); Recommended troubleshooting tools: A preset tool list (such as insulation tester, infrared thermal imager, etc.) is associated with the fault type; Impact scope estimation: Based on parameter correlation analysis, the scope of adjacent equipment that may be affected is defined. Priority marking rules: High risk level corresponds to the "urgent" mark (requires response within 1 hour), medium risk corresponds to the "routine" mark (requires response within 4 hours), and low risk corresponds to the "planned" mark (requires response within 24 hours).

[0081] During transmission, within the preset industrial frequency band, the frequency channel is switched at intervals of 50ms to avoid known electromagnetic interference frequency bands; three sets of key parameter snapshots generated by adjacent terminals are intercepted (aligned with the current expected data timestamp) as a basis for verification; a CRC checksum is appended to each frame of data, and if the receiving end fails the check, it automatically requests retransmission. If the number of retransmissions exceeds three, it switches to the backup communication link.

[0082] The intelligent terminal adjusts emergency device parameters based on the risk level. In high-risk situations, the emergency circuit breaker's input current is increased to 120% of the rated value, operating power is increased to 80% of the maximum load, and the response delay is shortened to less than 50ms. In medium-risk situations, the input current is maintained at the rated value, operating power is adjusted to 60%-70%, and the response delay is maintained at 100ms. In low-risk situations, the input current is reduced to 80% of the rated value and operating power is reduced to below 50%, reducing unnecessary energy consumption.

[0083] Specifically, after receiving the maintenance signal, the central server will use the intelligent terminal that generates the verification information in the maintenance signal as the starting terminal, and define the initial maintenance range including the starting terminal and the adjacent terminals of the starting terminal in the topological network; based on the risk assessment results in the maintenance signal, the corresponding risk level is marked for each location within the initial maintenance range; based on the risk level and the electromagnetic characteristics and correlation relationship of the target object corresponding to the intelligent terminal, the electromagnetic interference intensity of each location in the initial maintenance range is calculated; and the maintenance route is planned in combination with the electromagnetic interference intensity, risk level and target object distribution.

[0084] More specifically, when the smart terminal is connected to the maintenance terminal for communication, the environmental data and operating data of the target object are sent to the maintenance terminal. After the smart terminal disconnects the communication connection with the maintenance terminal, the smart terminal collects and updates the environmental data and operating data to the adjacent terminal; the central server updates the initial maintenance range and the electromagnetic interference intensity of each location based on the most recently received maintenance signal; the central server obtains the environmental data and operating data of the target object from the maintenance terminal based on the initial maintenance range.

[0085] In one specific implementation of this solution, when a smart terminal establishes a communication connection with a maintenance terminal, it first verifies the maintenance terminal's identity and maintenance permissions. Once verified, an encrypted communication link is established. The smart terminal then sends the target object's environmental data (including temperature and humidity, dust concentration, and ambient electromagnetic radiation) and operational data (including equipment start and stop status, key component pressure values, and energy transmission stability parameters) in real-time priority order. The smart terminal also receives parameter retrieval requests from the maintenance terminal and supplements historical data segments for specific time periods as needed. When the smart terminal loses communication with the maintenance terminal, it immediately activates a high-frequency acquisition mode, continuously capturing trends in environmental and operational data. Once data fluctuations stabilize, it selects parameters that differ significantly from those before the disconnection and packages them for transmission to adjacent terminals to update the shared database, ensuring that adjacent terminals have the latest status.

[0086] After receiving the most recent maintenance signal, the central server parses the status update records of each terminal contained therein, compares them with the boundary parameters of the original initial maintenance range, adds newly abnormal terminals to the range or removes terminals that have returned to normal, and simultaneously recalculates and adjusts the distribution map of electromagnetic interference intensity based on the risk level changes at each location in the signal. Based on the updated initial maintenance range, data acquisition instructions are sent to maintenance terminals that are not within the range or are not within the encircled area of ​​the range. The instructions specify the time interval and parameter type of the environmental data and operating data to be returned; after receiving the data feedback from the maintenance terminal, it is compared with the historical data stored in its own database. If there is a difference, a secondary verification mechanism is triggered, requiring the maintenance terminal to supplement the environmental working conditions description at the time of data collection.

[0087] During the inspection process, the smart terminal and the maintenance terminal establish a dedicated communication link through authentication and encryption, leveraging the stability of short-range transmission to ensure the real-time and complete transmission of environmental and operational data. This close connection mechanism, based on the physical proximity of the devices, enables low-latency interaction and avoids data loss or distortion caused by electromagnetic interference during long-distance transmission.

[0088] After maintenance is completed, intelligent terminals frequently collect and update data to adjacent terminals. Adjacent terminals then dynamically adjust the maintenance scope based on this data. However, relaying data between multiple intelligent terminals can introduce data attenuation or deviation, resulting in uncertainty in the initial, updated maintenance scope. Once outside the electromagnetic interference range, maintenance terminals leverage their direct communication capabilities with the central server to synchronize the raw data acquired during the maintenance process to the server. This serves as a verification and correction benchmark for data transmitted from multiple terminals. By comparing and calibrating raw data with relayed data, accumulated errors in the transmission chain are eliminated.

[0089] The above is only an embodiment of the present invention, and the common knowledge such as the specific structure and characteristics of the scheme is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. The intelligent terminal data acquisition and control method based on the power environment is characterized by: The following steps are involved: S10: The intelligent terminal collects environmental data and operating data from the target object, associates the environmental data with the operating data based on the collection time, generates a timestamp-containing associated data set, and stores it locally; S20: The intelligent terminal retrieves local historical fault data and associated data sets, constructs an analysis model based on a multi-factor coupling analysis model, and through training and fitting the data change function before and after the fault, identifies the dynamic correlation between the historical fault data and the associated data sets and calculates the impact ratio of each data on the fault. The currently collected data is used as the initial data, combined with the continuous data segments in the associated data sets, and input into the analysis model. The data after the preset time period is calculated as the predicted data. Based on the predicted data and the impact ratio, a weighted algorithm is used to obtain a comprehensive risk value. S30: The intelligent terminal classifies the comprehensive risk value into different risk levels, obtains the impact ratio from the analysis model according to the risk level, and activates different emergency devices according to the impact ratio and risk level; S40: The intelligent terminal adjusts the collection frequency of environmental data and operating data according to the risk level, sets the verification interval according to the collection frequency, uses the analysis model to calculate the comprehensive risk value as the adjusted risk based on the verification interval, adjusts the working power of the emergency device according to the changing trend of the adjusted risk obtained multiple times, and adjusts the collection frequency at the same time.

2. The method for intelligent terminal data acquisition and control based on power environment according to claim 1 is characterized in that: When constructing the analysis model in step S20, the following steps are included: S21: Cleans historical fault data and related datasets with timestamps, fills in missing values, and standardizes them. It extracts the collaborative features of environmental data and operational data at the same timestamp, as well as the correlation between changes in data at different timestamps. It also mines dynamic correlations from the collaborative features and correlations of data, including the coupling effects of data with the same timestamp and the evolutionary correlations of data across timestamps. S22: Based on the multi-factor coupling analysis theory, an architecture with a time correlation layer is constructed. The time correlation layer is used to process the temporal correlation relationship of various data at different timestamps. S23: Iteratively train the analysis model using historical fault data and associated data sets, adjust the associated weight parameters in the time dimension through cross-validation, optimize the model's predictive performance, and prevent overfitting through regularization techniques.

3. The method for intelligent terminal data acquisition and control based on power environment according to claim 2, characterized in that: When calculating the comprehensive risk value in step S20, the initial weight is set according to the influence ratio of each data, the deviation between the current data and the predicted data is obtained, and the weighted weight is adjusted according to the deviation amplitude; when an abnormal trend similar to that before the historical failure appears in the predicted data, the weight of the corresponding data is increased; combined with the change in risk level, when the risk level increases, the weight of the data associated with the predicted data is increased, and when the risk level decreases, the weight is reduced according to the changes in each data.

4. The method for intelligent terminal data acquisition and control based on power environment according to claim 3 is characterized in that: In steps S30-S40, based on the impact of historical faults and the safe operation boundaries of each device in the target object, multiple risk levels are defined, and each risk level corresponds to a safe state range and abnormal trend characteristics of the device parameters; The intelligent terminal dynamically adjusts the data collection frequency based on the current risk level, including adjusting the collection frequency of environmental data and operating data, as well as the differences in collection accuracy of different data; different prediction intervals are set for different risk levels. When the risk level changes, the data collection frequency and prediction interval are gradually transitioned to maintain the continuity of data processing and ensure real-time matching of risk assessment with the actual operating status of the equipment.

5. The method for collecting and controlling data from smart terminals based on the power environment according to claim 4 is characterized in that: Intelligent terminals build a topological network based on the physical location relationship between target objects. They consider adjacent intelligent terminals in the topological network as neighboring terminals. Through the topological network, they fuse the associated data sets and historical fault data of local and adjacent terminals. Based on the temporal and spatial correlation characteristics of the data, they identify the causal relationship pattern between the adjacent target object data and the local data, and build a neighboring influence model that reflects the degree and direction of data influence. The smart terminal generates verification information based on the risk level and sends it to the adjacent terminal. After receiving the verification information, the adjacent terminal inputs the currently collected data into the adjacent impact model to infer the expected data of the local target object, and then compares it with the actual collected data. Based on the comparison results, the adjacent terminal generates verification information feedback information and sends it to the smart terminal; The intelligent terminal summarizes the feedback information, calculates the confidence level of the current risk level based on the summarized results, and adjusts the collection frequency of environmental data and operation data based on the confidence level.

6. The method for intelligent terminal data acquisition and control based on power environment according to claim 5, characterized in that: When there are multiple adjacent terminals for a smart terminal, the adjacent impact models are first prioritized based on the physical distance between each adjacent target object and the local target object; the adjacent impact models are superimposed in order of priority, and the expected data in the same time window are weightedly fused according to the preset window time to obtain the expected data distribution under the comprehensive impact, and then the expected data of each adjacent terminal in the same time and space area are compared to identify overlapping or conflicting expected data and filter out valid expected data; the valid expected data are spatiotemporally integrated to generate the final expected data, and the final expected data are compared with the actual collected data, and the weight distribution when the adjacent impact models are superimposed is adjusted according to the comparison results.

7. The method for collecting and controlling data from smart terminals based on power environment according to claim 6, characterized in that: The intelligent terminal calculates the expected data based on the verification information and performs a risk assessment on the expected data. Based on the risk assessment results, the intelligent terminal enables or adjusts the access current and working power of the emergency device, and generates a maintenance signal based on the risk assessment results and verification information and sends it to the central server.

8. The method for intelligent terminal data acquisition and control based on power environment according to claim 7, characterized in that: After receiving the maintenance signal, the central server will use the intelligent terminal that generated the verification information in the maintenance signal as the starting terminal, and define the initial maintenance range in the topological network, including the starting terminal and its adjacent terminals. Based on the risk assessment results in the maintenance signal, the corresponding risk level is marked for each location within the initial maintenance range. Based on the electromagnetic characteristics and correlation between the risk level and the target object corresponding to the smart terminal, the electromagnetic interference intensity at each location within the initial maintenance range is calculated; the maintenance route is planned based on the electromagnetic interference intensity, risk level and target object distribution.

9. The method for intelligent terminal data acquisition and control based on power environment according to claim 8, characterized in that: When the smart terminal is connected to the maintenance terminal, the environmental data and operating data of the target object are sent to the maintenance terminal. After the smart terminal disconnects from the maintenance terminal, the smart terminal collects and updates the environmental data and operating data to the adjacent terminal; the central server updates the initial maintenance range and the electromagnetic interference intensity of each location based on the most recently received maintenance signal; the central server obtains the environmental data and operating data of the target object from the maintenance terminal based on the initial maintenance range.

10. The intelligent terminal data acquisition and control system based on the power environment is characterized by: The smart terminal data acquisition and control method based on the power environment described in any one of claims 1 to 9 is used.

Images