Transformer construction period environmental protection risk assessment method and system

Abstract

The invention provides a transformer construction period environmental protection risk assessment method and system, and the method comprises the steps: inputting transformer image data of a construction site in a construction period into an image recognition model, and recognizing the current abnormal coordinate information of a transformer; obtaining current abnormal state data based on the current abnormal coordinate information; inputting the current abnormal state data, the obtained current operation state parameters of the transformer and the obtained current environment parameters into a pre-constructed dynamic evaluation model for evaluation to obtain the current risk degree of the transformer; based on the current risk degree, outputting an environmental protection risk assessment result and an equipment safety monitoring result of the transformer; wherein the dynamic evaluation model is obtained by training an adaptive neural network by adopting historical abnormal state data, historical operation state parameters, historical environment parameters and historical risk degrees of the transformer. The current environmental protection risk degree of the transformer can be output in real time, and the dynamic monitoring requirement of the transformer in the construction period is met.

Description

Technical Field

[0001] The present invention belongs to the technical field of power equipment monitoring, and in particular relates to a method and system for assessing environmental risks during transformer construction. Background Art

[0002] During the construction and operation of pole-mounted transformers, environmental factors, the equipment's condition, and external influences can all impact its performance, stability, and safety. Therefore, to ensure transformer reliability during the construction phase, existing technologies typically employ monitoring and early warning methods to identify equipment anomalies and take appropriate action. These conventional methods primarily rely on on-site inspections, remote monitoring, and early warning systems based on fixed thresholds. These methods aim to promptly identify issues and mitigate potential safety risks during the transformer's construction and initial operation phases.

[0003] The correlation analysis between the existing transformer anomaly detection and operation status monitoring relies on an evaluation mechanism with fixed parameter settings. It has poor adaptability in complex environments and cannot meet the dynamic monitoring needs during the construction period. Summary of the Invention

[0004] To overcome the above-mentioned deficiencies of the prior art, in a first aspect, the present invention proposes a method for environmental risk assessment during transformer construction, comprising:

[0005] Input the transformer image data at the construction site during the construction period into the image recognition model to identify the current abnormal coordinate information of the transformer; based on the current abnormal coordinate information, obtain the current abnormal state data;

[0006] Input the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters into a pre-built dynamic assessment model for evaluation to obtain the current risk of the transformer, which includes the current environmental risk and the current equipment failure risk;

[0007] Based on the current risk level, output the environmental risk assessment result and equipment safety monitoring result of the transformer;

[0008] The dynamic assessment model is obtained by training an adaptive neural network using historical abnormal state data, historical operating state parameters, historical environmental parameters and historical risk levels of the transformer.

[0009] Preferably, the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters are input into a pre-built dynamic assessment model for assessment to obtain the current risk of the transformer, including:

[0010] The input layer of the adaptive neural network based on the dynamic evaluation model receives the current abnormal state data, the current operating state parameters of the transformer and the current environmental parameters;

[0011] Based on the hidden layer of the adaptive neural network, nonlinear mapping is performed on the current abnormal state data, the current operating state parameters of the transformer, and the current environmental parameters to obtain an operating state impact contribution factor and an environmental impact contribution factor;

[0012] Based on the actual risk and reference risk of the transformer's historical operating conditions, the operating state parameter weight is obtained using the operating state parameter weight adjustment formula; based on the actual value and historical average of the transformer's operating state parameter, the environmental parameter weight is obtained using the environmental parameter weight adjustment formula;

[0013] Based on the output layer of the adaptive neural network, the current risk is calculated according to the operating state impact contribution factor, the environmental impact contribution factor, the operating state parameter weight and the environmental parameter weight.

[0014] Preferably, the current risk satisfies the following formula:

[0015]

[0016] In the above formula, R(t) is the current risk at time t, α is the contribution factor of the operating state, N is the total number of operating parameters in the operating state parameters, and W i is the i-th operating parameter weight in the operating state parameter weight, P i (t) is the real-time measurement value of the i-th operating parameter at time t, is the historical reference value of the i-th operating parameter, β is the environmental impact contribution factor, M is the total number of environmental parameters in the environmental parameter, V j is the jth environmental parameter weight in the environmental parameter weights, E j (t) is the measured value of the jth environmental parameter among the environmental parameters at time t, is the average value of the jth environmental parameter, γ is the weight of the sudden risk factor, and dR / dt is the instantaneous change rate of the sudden risk.

[0017] Preferably, the operating state parameter weight satisfies the following formula:

[0018]

[0019] In the above formula, W i (t) is the weight of the i-th operating parameter at time t, is the initial weight of the i-th operating parameter, is the historical operating conditions adaptation factor, R k(t) is the actual risk of the kth historical working condition at time t, is the reference risk of the kth historical operating condition, and K is the total number of historical operating conditions;

[0020] The environmental parameter weights satisfy the following formula:

[0021]

[0022] In the above formula, V j (t) is the weight of the jth environmental parameter at time t, is the initial weight of the jth environmental parameter, μ is the environmental adaptation factor, P m (t) is the actual value of the mth operating parameter involved in the calculation at time t, is the historical mean of the mth operating parameter involved in the calculation, and S is the number of operating parameters involved in the calculation.

[0023] Preferably, the pre-construction process of the dynamic assessment model includes:

[0024] Determine the historical risk of the transformer based on the acquired historical abnormal state data, historical operating state parameters, and historical environmental parameters; determine the historical environmental parameter weight and historical environmental parameter measurement value based on the historical abnormal state data; determine the historical abnormal state data weight, the historical construction stage of the historical abnormal state data, and the historical abnormal state data measurement value based on the historical abnormal state data;

[0025] The historical environmental parameter weights, the historical environmental parameter measurements, the historical abnormal state data measurements, the historical construction stages, and the historical abnormal state data weights are input into an adaptive neural network to output a predicted environmental risk degree; the difference between the historical risk degree and the predicted environmental risk degree is used as a prediction error; with the goal of minimizing the prediction error, the adaptive neural network is trained to obtain a dynamic evaluation model.

[0026] Preferably, obtaining current abnormal state data based on the current abnormal coordinate information includes:

[0027] The acquired standard transformer spatial position information is associated with the current abnormal coordinate information using coordinate mapping technology to obtain abnormal components and areas;

[0028] Based on the abnormal components and areas, determining multiple abnormal points of the transformer;

[0029] Cluster analysis is used to analyze the multiple abnormal points to obtain current abnormal state data.

[0030] Preferably, the acquired standard transformer spatial position information is associated with the current abnormal coordinate information using a coordinate mapping technique to obtain abnormal components and areas, including:

[0031] The acquired standard transformer spatial position information is associated with the current abnormal coordinate information using coordinate mapping technology to obtain a three-dimensional coordinate system of the abnormal point;

[0032] Based on the three-dimensional coordinate system of the abnormal points, a point cloud registration algorithm or a nearest neighbor search algorithm is used to determine the abnormal parts and areas.

[0033] Preferably, the cluster analysis is used to analyze the multiple abnormal points to obtain current abnormal state data, including:

[0034] Analyzing the multiple abnormal points using cluster analysis to obtain dense distribution probabilities of the multiple abnormal points in the transformer space, and converting the dense distribution probability into a spatial abnormality distribution map;

[0035] Based on the spatial anomaly distribution map, current abnormal state data is determined.

[0036] Preferably, after outputting the environmental risk assessment result and the equipment safety monitoring result of the transformer based on the current risk level, the method further includes:

[0037] If the current risk level is higher than a preset threshold, an early warning message is sent to a construction monitoring center via a remote terminal;

[0038] The warning information includes: abnormal spatial location information of the transformer, abnormal risk category of the transformer, current risk level of the transformer, level of current risk level, emergency notification of abnormal risk category and recommended handling measures for abnormal risk category.

[0039] In a second aspect, the present invention also proposes a transformer construction period environmental risk assessment system, comprising:

[0040] The abnormal state data acquisition module is used to input the transformer image data at the construction site during the construction period into the image recognition model to identify the current abnormal coordinate information of the transformer; based on the current abnormal coordinate information, obtain the current abnormal state data;

[0041] A risk determination module is configured to input the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters into a pre-built dynamic assessment model for evaluation to obtain the current risk of the transformer, wherein the current risk includes the current environmental risk and the current equipment failure risk; wherein the dynamic assessment model is obtained by training an adaptive neural network using the historical abnormal state data, historical operating state parameters, historical environmental parameters, and historical risk of the transformer;

[0042] The result acquisition module is used to output the environmental risk assessment result and equipment safety monitoring result of the transformer based on the current risk level.

[0043] In a third aspect, the present invention also proposes an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0044] The memory is used to store one or more programs;

[0045] When the one or more programs are executed by the at least one processor, the method for assessing environmental risks during the transformer construction period is implemented.

[0046] In a fourth aspect, the present invention application further proposes a readable storage medium having an execution program stored thereon, which, when executed, implements the aforementioned method for environmental risk assessment during the transformer construction period.

[0047] Compared with the closest prior art, the present invention has the following beneficial effects:

[0048] The present invention provides a method and system for assessing environmental risks during the construction period of a transformer, comprising: inputting transformer image data from a construction site during the construction period into an image recognition model to identify current abnormal coordinate information of the transformer; obtaining current abnormal state data based on the current abnormal coordinate information; inputting the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters into a pre-built dynamic assessment model for assessment to obtain the current risk of the transformer, the current risk including the current environmental risk and the current equipment failure risk; and outputting the environmental risk assessment result of the transformer and the equipment safety monitoring result based on the current risk; wherein the dynamic assessment model is obtained by training an adaptive neural network using the historical abnormal state data, historical operating state parameters, historical environmental parameters, and historical risk of the transformer. By combining the adaptive neural network with multiple historical operating data and multiple historical environmental parameters of the transformer to construct a dynamic assessment model, and then inputting the current operating state parameters and current environmental parameters of the transformer to be assessed into the dynamic assessment model, the current environmental risk of the transformer can be output in real time, thereby meeting the dynamic monitoring requirements during the transformer construction period. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A flowchart of a method for environmental risk assessment during transformer construction provided by the present invention;

[0050] Figure 2 This is an architecture diagram of a transformer construction period environmental risk assessment system provided by the present invention;

[0051] Figure 3 The present invention provides an environmental risk assessment system for transformer construction period, which is used for monitoring the transformer construction period.

[0052] Figure 4 This is a schematic diagram of the operation of an electronic device provided in the present invention application. DETAILED DESCRIPTION

[0053] The specific implementation methods of the present invention will be further described in detail below with reference to the accompanying drawings.

[0054] Example 1:

[0055] like Figure 1 As shown, the present invention proposes a method for environmental risk assessment during transformer construction period, which may include:

[0056] Step 1: Input the transformer image data at the construction site during the construction period into the image recognition model to identify the current abnormal coordinate information of the transformer; based on the current abnormal coordinate information, obtain the current abnormal state data;

[0057] Step 2: Input the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters into a pre-built dynamic assessment model for evaluation to obtain the current risk of the transformer, wherein the current risk includes the current environmental risk and the current equipment failure risk; wherein the dynamic assessment model is obtained by training an adaptive neural network using the historical abnormal state data, historical operating state parameters, historical environmental parameters, and historical risk of the transformer;

[0058] Step 3: Based on the current risk level, output the environmental risk assessment result and equipment safety monitoring result of the transformer.

[0059] In steps 1 to 3 above, the pole-mounted transformer can be monitored, and multi-source sensors can be deployed during its construction period to obtain transformer image data at the construction site during the construction period. The spatial information, appearance status, and operating data of the transformer and the construction site can then be obtained from the transformer image data. Combined with the intelligent analysis of the dynamic assessment model, the dynamic assessment model can dynamically adjust the early warning strategy during the intelligent analysis to ensure accurate assessment and response of the current risk and the current equipment failure risk.

[0060] First, during the transformer construction phase, image acquisition devices, point cloud ranging equipment, and environmental and operational data sensors were installed. These devices work together to obtain multi-dimensional information about the transformer and its surroundings. The image acquisition device captures images of the construction site and the transformer's exterior, the point cloud ranging equipment measures the spatial coordinates of transformer components, and the environmental and operational data sensors collect parameters such as temperature, humidity, wind speed, current, and voltage. This data serves as input for subsequent analysis.

[0061] Secondly, the multi-dimensional information obtained about the transformer and its surroundings provides foundational support for subsequent data processing and analysis. Various sensors were installed according to construction site layout requirements, and the collection range was adjusted based on the transformer's structural characteristics and construction progress to ensure the integrity and accuracy of the monitoring data.

[0062] In the above step 1, the transformer image data at the construction site during the construction period is input into the image recognition model to identify the current abnormal coordinate information of the transformer, which may include the following steps:

[0063] Step S1: Collect image data of the transformer and the construction site, and use point cloud ranging technology to obtain the spatial position information of the transformer. The image data is collected by an image acquisition device, which may include a high-definition industrial camera, an infrared camera, or a multispectral imaging device to adapt to different lighting environments and construction conditions. The image acquisition device is fixedly installed on the transformer bracket, the monitoring pillar at the construction site, or the drone platform, and can cover the transformer body and key areas of the construction site. The resolution of the image data is adjusted according to the monitoring needs of the construction site. For example, during the critical installation phase, the image acquisition frequency can be increased to capture dynamic changes during the construction process. The image acquisition device is used to obtain real-time images of the construction site and transmit the data to the image processing unit. The point cloud ranging equipment uses laser scanning technology to perform three-dimensional modeling of the transformer and its surrounding environment, obtain the spatial position relationship of each component, and store it in the data storage unit.

[0064] After image data is collected, it is transmitted to the image processing unit via the data transmission module. The image processing unit includes an image preprocessing module that performs image preprocessing, such as noise reduction, deblurring, and contrast enhancement, to improve data quality. To ensure stable data transmission, image data can be transmitted using wired or wireless networks, such as those based on 5G (5th Generation Mobile Communication Technology) or Wi-Fi (Wireless Fidelity) to meet the real-time monitoring needs of construction sites.

[0065] When collecting image data, point cloud ranging equipment can be used using laser scanning technology to create a three-dimensional model of the transformer and its surroundings. This equipment emits a high-frequency laser beam and calculates its return time and angular change to construct spatial position information for the transformer. The accuracy of point cloud ranging is affected by the laser wavelength, measurement distance, and ambient light interference. Therefore, during the measurement process, the appropriate ranging mode should be selected based on the size of the transformer and the complexity of the construction site. For example, for smaller transformers, a short-range, high-precision mode can be used, while for larger transformers, a long-range scanning mode can be used to ensure data integrity.

[0066] The acquired point cloud data is stored in a data storage unit and time-stamped synchronously with the image data for subsequent fusion analysis. The point cloud data includes the surface morphology of the transformer itself, the relative positions of key components, and the spatial layout of the surrounding construction environment. The data storage unit can utilize high-capacity storage media to support the storage of massive amounts of point cloud data and be equipped with an indexing mechanism to improve data retrieval efficiency. The data storage unit can be a database. During subsequent analysis, the point cloud data can be used to construct a digital model of the transformer, providing a high-precision spatial reference for condition monitoring and anomaly detection.

[0067] Step S2: Using an image recognition model, the transformer's exterior is detected for abnormalities and the coordinates of the abnormalities are obtained. This image recognition model uses deep learning methods to extract features from the collected image data, identifying abnormalities such as burns, deformation, and oil leaks, and recording the current coordinates of the abnormalities.

[0068] After the image data from step S1 is preprocessed by the image preprocessing module, including image denoising, brightness normalization, contrast enhancement, and edge sharpening, it can reduce recognition errors caused by ambient lighting changes, interference from stains, or blur. The preprocessed image is then fed into a deep learning model, which can be trained using convolutional neural networks (CNNs) or other deep learning architectures to produce a trained model. The trained model is then used to extract feature information about the transformer's surface structure and detect any abnormal patterns.

[0069] During feature extraction, the deep learning model uses convolution kernels to extract image features at different levels. For example, shallow-level features capture basic shapes and edge information, mid-level features identify local texture and material variations, and deep-level features analyze overall structure and abnormal patterns. The model's training samples include both normal transformer images and images of transformers with various defects. These images are optimized using supervised learning to enhance the model's detection capabilities.

[0070] When detecting anomalies, the input image is matched against the characteristic patterns in the trained model, and the confidence level of the anomaly area is calculated. When the confidence level exceeds a set threshold, an anomaly is detected, and the anomaly category, severity, and coordinate information are output. This coordinate information is obtained through image analysis methods and combined with the transformer's three-dimensional spatial position information to determine the specific component where the anomaly is located. For example, if an oil leak is detected, the pixel distribution of the anomaly area is analyzed and its relationship to the transformer interface position is calculated to determine whether it involves a seal failure in a critical component. The detected anomaly data is then stored in a database.

[0071] In the above step 1, obtaining the current abnormal state data based on the current abnormal coordinate information may include:

[0072] Step 1.1: Correlate the acquired standard transformer spatial position information with the current abnormal coordinate information using coordinate mapping technology to obtain abnormal components and areas;

[0073] Step 1.2: Based on the abnormal components and areas, determine multiple abnormal points of the transformer;

[0074] Step 1.3: Analyze the multiple abnormal points using cluster analysis to obtain current abnormal state data.

[0075] As mentioned above, when the standard transformer is selected as a qualified transformer for comparison, the point cloud ranging equipment can be used to perform digital three-dimensional modeling of the standard transformer and its surrounding environment, and the spatial position information of the standard transformer can be obtained; the image data and the point cloud ranging data can be stored in the data storage unit and time synchronization marked for subsequent fusion analysis.

[0076] After collecting image data of the transformer and the construction site, pre-processing the collected image data, and using point cloud ranging technology to obtain spatial position information of the transformer; the processed image can be input into a deep learning model (for example, an image recognition model), feature information of the transformer surface structure can be extracted, and whether there is an abnormality in the appearance of the transformer can be detected based on the image recognition model. When an abnormality is detected, the category, severity and coordinate information of the abnormality are obtained, and the specific component where the abnormality is located is determined in combination with the three-dimensional spatial position information of the transformer, and then the coordinate information of the abnormality can be obtained;

[0077] In the above step 1.1, the acquired standard transformer spatial position information is associated with the current abnormal coordinate information using a coordinate mapping technique to obtain abnormal components and areas, which may include:

[0078] Step 1.1.1: Correlate the acquired standard transformer spatial position information with the current abnormal coordinate information using coordinate mapping technology to obtain a three-dimensional coordinate system of the abnormal point;

[0079] Step 1.1.2: Based on the three-dimensional coordinate system of the abnormal point, use a point cloud registration algorithm or a nearest neighbor search algorithm to determine the abnormal parts and areas.

[0080] In the above, after obtaining the abnormal coordinate information provided by the image recognition model and the point cloud ranging data, the abnormal coordinate information is matched with the point cloud ranging data to determine the abnormal component or area; the coordinate transformation technology can be used to associate the current abnormal coordinate information with the standard transformer spatial position information through coordinate mapping, and the image coordinates of the current abnormal coordinate information can be converted to the point cloud coordinate system; based on the matching results, spatial indexing can be performed in combination with the three-dimensional model of the transformer to determine the corresponding component of the abnormal point, thereby accurately identifying the abnormal component of the transformer body;

[0081] First, the image anomaly coordinate information includes the 2D pixel coordinates of the anomaly area and the anomaly category. Since the image data is captured by an image acquisition device, its coordinate system belongs to the image coordinate system, while the point cloud ranging data uses a 3D coordinate system to represent the transformer structure. Therefore, for matching, a coordinate transformation is required to convert the image coordinate information into the point cloud coordinate system. This transformation can be accomplished through camera calibration parameters, including the camera's intrinsic parameter matrix, extrinsic parameter matrix, and perspective projection relationship, to ensure the accurate mapping of the anomaly points in the 3D coordinate system.

[0082] Secondly, during the coordinate matching process, a 3D model of the transformer is first built based on the point cloud ranging data, and each component is spatially indexed. Then, using the projective transformation relationship between image coordinates and point cloud coordinates, the outliers are converted from the image coordinate system to a 3D coordinate system, and their corresponding positions on the transformer surface are calculated. This matching process uses a nearest neighbor search or point cloud registration algorithm to ensure that the outliers are accurately mapped to the specific components of the transformer.

[0083] In the above step 1.3, the cluster analysis is used to analyze the multiple abnormal points to obtain the current abnormal state data, which may include:

[0084] Step 1.3.1: Analyze the multiple abnormal points using cluster analysis to obtain dense distribution probabilities of the multiple abnormal points in the transformer space, and convert the dense distribution probability into a spatial abnormality distribution map;

[0085] Step 1.3.2: Based on the spatial anomaly distribution map, determine current abnormal state data.

[0086] As described above, after obtaining the dense distribution probability of the plurality of abnormal points in the transformer space, if the abnormal points are densely distributed, they may be marked as abnormal areas, and a spatial abnormality distribution map may be generated.

[0087] Abnormal components or regions are identified based on cluster analysis of multiple matching points. If abnormal points are densely distributed in the three-dimensional space of the transformer, the region is judged to have a high probability of abnormality and is marked as an abnormal region. Furthermore, combined with the transformer's structural information, the specific component of the abnormality can be further determined, for example, determining whether the abnormality occurs in the insulating bushing, transformer tank, or terminal block. For multiple abnormal points, a density clustering algorithm (such as DBSCAN, Density-Based Spatial Clustering of Applications with Noise) can be used to analyze the distribution of the abnormal points to enhance the accuracy of abnormal component identification.

[0088] A spatial anomaly distribution map is then generated based on the identified abnormal components and areas. Based on the 3D point cloud model, this map can be overlaid with highlighted abnormal areas to visually demonstrate the transformer's abnormal conditions. This map can be used to inform subsequent operation and maintenance decisions and serves as input data for further analysis in the dynamic assessment model, enhancing overall anomaly detection and risk assessment capabilities.

[0089] In step 2 above, the pre-construction process of the dynamic assessment model may include:

[0090] Step a: Determine the historical risk of the transformer based on the acquired historical abnormal state data, historical operating state parameters, and historical environmental parameters; determine the historical environmental parameter weight and historical environmental parameter measurement value based on the historical abnormal state data; determine the historical abnormal state data weight, the historical construction stage of the historical abnormal state data, and the historical abnormal state data measurement value based on the historical abnormal state data;

[0091] Step b: Input the historical environmental parameter weights, the historical environmental parameter measurements, the historical abnormal state data measurements, the historical construction stages, and the historical abnormal state data weights into an adaptive neural network, and output a predicted environmental risk; use the difference between the historical risk and the predicted environmental risk as a prediction error; and train the adaptive neural network with the goal of minimizing the prediction error to obtain a dynamic evaluation model.

[0092] In step 2 above, after collecting the transformer's operating and environmental parameters, historical data, including historical abnormal state data, operating state parameters, and environmental parameters, can be input into an adaptive neural network, and the adaptive neural network can be trained using the assessed historical risk level. In the dynamic assessment model, the adaptive neural network dynamically adjusts the weights of risk assessment factors based on real-time data to adapt to changes in construction phases. The adaptive neural network utilizes a feedback mechanism to perform weighted optimization of factors affecting different construction phases, enabling dynamic adjustment of risk assessments to specific working conditions, improving the accuracy and adaptability of early warnings. Therefore, the dynamic assessment model comprehensively considers the transformer's historical operating data, current operating status, and environmental factors at the construction site to calculate the current real-time risk level and provide risk prediction results.

[0093] To improve the accuracy and adaptability of subsequent warnings, the adaptive neural network's weight optimization process is based on supervised learning, integrating historical risk assessment data for model training. When new data is input, the prediction error is automatically calculated, and the network parameters are adjusted through a backpropagation algorithm to reduce the prediction error and make the risk assessment more consistent with actual operating conditions. This optimization process can be optimized by setting convergence conditions to ensure computational efficiency while avoiding overfitting.

[0094] In the above step 2, the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters are input into a pre-built dynamic assessment model for evaluation to obtain the current risk of the transformer, which may include:

[0095] Step 2.1: The input layer of the adaptive neural network based on the dynamic assessment model receives the current abnormal state data, the current operating state parameters of the transformer, and the current environmental parameters;

[0096] Step 2.2: performing nonlinear mapping on the current abnormal state data, the current operating state parameters of the transformer, and the current environmental parameters based on the hidden layer of the adaptive neural network to obtain an operating state impact contribution factor and an environmental impact contribution factor;

[0097] Step 2.3: Based on the actual risk and reference risk of the transformer's historical operating conditions, the operating state parameter weight is obtained using the operating state parameter weight adjustment formula; based on the actual value and historical average of the transformer's operating state parameter, the environmental parameter weight is obtained using the environmental parameter weight adjustment formula;

[0098] Step 2.4: Based on the output layer of the adaptive neural network, calculate the current risk according to the operating state impact contribution factor, the environmental impact contribution factor, the operating state parameter weight and the environmental parameter weight.

[0099] As mentioned above, in the dynamic assessment model, an adaptive neural network is used to dynamically adjust the weights of risk assessment factors such as the operating status contribution factor and the environmental impact contribution factor based on real-time data. The adaptive neural network includes an input layer, a hidden layer, and an output layer. The input layer can receive current operating status parameters, current environmental parameters, and current abnormal status data; the hidden layer can perform nonlinear mapping to facilitate the identification of key influencing factors at different construction stages; and the output layer can calculate the final risk score and provide early warning recommendations.

[0100] The collected transformer's current operating and environmental parameters, combined with historical data such as actual risk levels and historical averages from historical operating conditions, are input into a dynamic assessment model to assess environmental risk and equipment failure risk. Transformer operating parameters include, but are not limited to, current, voltage, temperature, and load, while environmental parameters include construction site temperature, humidity, wind speed, and atmospheric pollution index. All of this data is collected by sensors and transmitted in real time to a data processing unit to ensure the dynamic assessment model receives complete input information.

[0101] The dynamic assessment model processes collected real-time data and analyzes it in conjunction with historical operating data. This historical data can also include operating status information for similar transformers under similar construction conditions, as well as past fault records and environmental factors. By constructing a multidimensional dataset, the current status is compared with the historical status to calculate the transformer's risk level during the current construction phase. Real-time current risk calculations can be based on data mining and statistical modeling methods, employing multivariate analysis to assess the degree of abnormality in the current operating conditions and generate risk predictions based on this information.

[0102] To improve the accuracy and adaptability of risk calculations, this implementation proposes a current risk calculation formula that comprehensively considers the transformer's operating status, environmental conditions, historical data, and real-time rate of change, enabling more accurate risk predictions. This formula, through an adaptive weighting mechanism, enables risk assessments to adapt to different construction phases and enable rapid responses to emergencies.

[0103] The above-mentioned calculation process of the current risk degree can satisfy the following formula:

[0104]

[0105] In the above formula, R(t) is the current risk at time t, α is the contribution factor of the operating state, N is the total number of operating parameters in the operating state parameters, and W i is the i-th operating parameter weight in the operating state parameter weight, P i (t) is the real-time measurement value of the i-th operating parameter at time t, is the historical reference value of the i-th operating parameter, β is the environmental impact contribution factor, M is the total number of environmental parameters in the environmental parameter, V j is the jth environmental parameter weight in the environmental parameter weights, E j (t) is the measured value of the jth environmental parameter among the environmental parameters at time t, is the average value of the jth environmental parameter, γ is the weight of the sudden risk factor, and dR / dt is the instantaneous change rate of the sudden risk.

[0106] The current risk formula consists of three parts:

[0107] Part 1: The weighted average of the operating status parameter deviations measures the deviation between the current operating status of the transformer and the historical normal operating status.

[0108] Part II: A weighted average of the deviations from environmental parameters, which measures the extent to which current environmental conditions have changed relative to historical environmental benchmarks.

[0109] Part III: Risk change rate, which measures the instantaneous trend of risk change to increase sensitivity to sudden anomalies.

[0110] The parameter analysis of the current risk formula is as follows:

[0111] The current risk is the comprehensive risk of the transformer at time t. As the final output value of the risk assessment, it reflects the health status and potential risk level of the transformer at the current construction stage.

[0112] In the first part above, the deviation contribution of the operating status parameters is as follows:

[0113]

[0114] P i (t): The real-time measurement value of the i-th operating parameter at time t can be: current, voltage, temperature, load, etc. These parameters directly affect the stable operation of the transformer and are important indicators of the health of the equipment.

[0115] The historical reference value of the i-th operating parameter can be based on past normal operating data or empirical values. This value can be calculated from historical operating data, such as a moving average over a period of time or a steady-state value during normal operation.

[0116] W i : The weight of the i-th operating parameter is dynamically adjusted by the adaptive neural network. Weights are assigned differently during different construction phases. For example, during the equipment installation phase, the weight of the structural stability parameter in the operating parameters is higher. During the equipment commissioning phase, the weight of the electrical parameters in the operating parameters is higher.

[0117] N: The total number of run parameters determines the number of parameters considered when calculating the mean deviation.

[0118] α: The operating status impact contribution factor serves as the influence weight of the operating status parameters in the risk calculation, and determines the degree of influence of the operating status deviation on the final risk assessment.

[0119] The second part mentioned above: the deviation contribution of environmental parameters is as follows:

[0120]

[0121] E j (t): The measured value of the jth environmental parameter at time t, such as temperature, humidity, wind speed, air pollution index, etc. These parameters affect the cooling efficiency, material expansion and contraction, insulation performance, etc. of the transformer.

[0122] The historical average value of the jth environmental parameter is used as the benchmark reference value, which is usually obtained by statistically analyzing historical construction environment data.

[0123] V j : The weight of the jth environmental parameter, adjusted by the adaptive neural network based on the construction phase. For example, during construction in inclement weather, the weights of temperature and humidity are higher. During indoor installation, the weights are lowered due to the smaller impact of environmental factors.

[0124] M: The total value of the environmental parameters determines the number of parameters considered when calculating the average environmental deviation.

[0125] β: Environmental impact contribution factor controls the influence of environmental factors in risk calculation and determines the degree of influence of environmental conditions on the final risk assessment.

[0126] The third part mentioned above: Risk change rate contributes as follows:

[0127]

[0128] dR / dt: The instantaneous rate of change of risk mainly calculates the rate of change of the current risk R(t) at time t: If dR / dt is large, it means that the risk level is rising rapidly, and there may be a sudden failure or emergency; if dR / dt is small, the risk changes smoothly and the operating status is relatively stable.

[0129] γ: The weight of the sudden risk factor. If the possibility of sudden failure is high, γ can be increased to improve risk response capabilities. If you want to reduce the impact of short-term fluctuations, you can appropriately reduce γ to improve stability.

[0130] The current risk formula dynamically adjusts the weights adaptively. Compared with the traditional method using fixed weights, this formula dynamically adjusts W based on the neural network.i and V j , ensuring optimal weight distribution during the construction phase and improving adaptability. A unified mathematical framework integrates environmental and operational parameters, avoiding false alarms caused by single parameter anomalies and improving overall early warning accuracy. Quantifying risk trends through dR / dt allows this method to more quickly respond to emergencies such as equipment failures or extreme weather events.

[0131] The current risk formula comprehensively evaluates the risk level of the transformer through three core parts, combining the operating status, environmental conditions and sudden changes to achieve more accurate risk prediction. Among them: α controls the impact of operating status parameters; β controls the impact of environmental factors; γ is responsible for sudden risk response; W i and V j Dynamically adjust weights to optimize construction phase adaptability. This current risk formula enables dynamic optimization of risk assessments at different construction phases, improving responsiveness to sudden failures and extreme environments, reducing false alarms, and increasing the reliability and accuracy of early warnings.

[0132] After the risk assessment calculations are complete, the dynamic assessment model utilizes a feedback mechanism to weight and optimize the influencing factors at different construction stages. In the early stages of construction, environmental parameters may have a greater impact on the transformer, so their weighting is automatically increased. Later in the construction process, the transformer's operating status may have a more significant impact, so the weighting is gradually adjusted to prioritize the transformer's electrical and mechanical characteristics. This dynamic adjustment process is accomplished through an adaptive learning algorithm that continuously receives new data and adjusts the weighting of different factors based on their contribution to risk prediction accuracy.

[0133] To optimize the weights of risk factors at each construction stage, this paper proposes an adaptive adjustment formula for construction stage weights. This formula dynamically optimizes the weights of risk factors, enabling the weights of various parameters to be adjusted based on the actual conditions of different construction stages, thereby improving the accuracy and adaptability of risk assessments.

[0134] In the first part above, the process of determining the weight of the operating state parameter can satisfy the following formula:

[0135]

[0136] In the above formula, W i (t) is the weight of the i-th operating parameter at time t, is the initial weight of the i-th operating parameter, is the historical operating conditions adaptation factor, R k (t) is the actual risk of the kth historical working condition at time t, is the reference risk of the kth historical operating condition, and K is the total number of historical operating conditions;

[0137] The operating state parameter weight formula is used to calculate the dynamic weight W of the operating state parameter at time t. i (t), where:

[0138] W i (t): The dynamic weight of the i-th operating parameter, representing the relative importance of that operating parameter (e.g., current, voltage, temperature, etc.) in the risk assessment at time t. The importance of different operating parameters varies with the construction phase, and this weight can be dynamically adjusted based on the changing risk.

[0139] The initial weight of the i-th operating status parameter is the initial weight set during initialization, which is set based on historical experience or expert experience and represents the baseline weight of the parameter at the initial stage of construction.

[0140] The historical operating condition adaptation factor is used to control the sensitivity of weight adjustment and determines the impact of risk changes on weight adjustment. A higher value indicates that the risk is more sensitive to changes in the construction phase and the weight adjustment is larger; a smaller value indicates that the risk is more sensitive to changes in the construction phase and the weight adjustment is larger; A value of 0 indicates that the weight adjustment is relatively stable and not easily affected by short-term fluctuations.

[0141] R k (t): The actual risk under the kth historical working condition at time t. This represents the risk of the working condition in the historical data that is closest to the current working condition. This value is usually matched based on the similarity of the construction phases. For example, in the early stages of construction, this value may be calculated based on historical risks under similar construction environments.

[0142] The reference risk of the kth historical operating condition represents the average risk under similar operating conditions in past historical data and serves as a reference for calculating risk changes.

[0143] K: Number of historical conditions represents the number of historical conditions used to calculate the average risk deviation. A larger K value allows weight adjustments to be influenced by more historical data, resulting in greater stability. A smaller K value relies more on recent data, making adjustments more flexible.

[0144] In the second part above, the environmental parameter weights satisfy the following formula:

[0145]

[0146] In the above formula, V j (t) is the weight of the jth environmental parameter at time t, is the initial weight of the jth environmental parameter, μ is the environmental adaptation factor, P m(t) is the actual value of the mth operating parameter involved in the calculation at time t, is the historical mean of the mth operating parameter involved in the calculation, and S is the number of operating parameters involved in the calculation.

[0147] The environmental parameter weight formula is used to calculate the dynamic weight V of the environmental parameter j (t), where:

[0148] V j (t): The dynamic weight of the jth environmental parameter, representing the relative importance of the environmental parameter (such as temperature, humidity, and wind speed) in the risk assessment at time t. The importance of environmental factors varies at different construction stages. For example, during equipment installation, ambient humidity may have a greater impact on insulation materials and therefore receive a higher weight. During equipment commissioning, the importance of environmental parameters may decrease.

[0149] The initial weight of the jth environmental parameter is the baseline weight of the environmental parameter at the beginning of construction, which is set based on historical experience.

[0150] μ: Environmental adaptation factors control the degree to which environmental parameters influence weight changes. A larger μ results in larger adjustments to environmental parameter weights, indicating greater sensitivity to environmental changes; a smaller μ results in slower adjustments to environmental parameter weights, relying more heavily on long-term trends.

[0151] P m (t): The real-time value of the mth operating parameter at time t represents the operating status data of the transformer at the current moment, such as the current current, voltage, temperature, etc.

[0152] The historical mean of the mth operating parameter represents the average value of that parameter in historical data and is typically used to determine whether the current operating parameter has significantly deviated. For example, if the current current value deviates significantly from the historical mean, this indicates that significant changes may have occurred during the construction phase.

[0153] S: The value of the number of parameters used to calculate the average state deviation determines the range of parameters considered when calculating weight adjustments. A larger S makes the calculation more stable, while a smaller S makes this method more flexible and can quickly adapt to environmental changes.

[0154] In the first part, when the change of the operating state parameters is large, i.e. When W increases i (t) increases, which increases the weight of the operating state parameters in the risk assessment. On the contrary, when the change of the operating state parameters is small, W i (t) Gradually restore to the initial weight

[0155] In the second part, when the changes in environmental parameters are large, i.e. When it increases, V j (t) decreases, which reduces the weight of environmental parameters in risk assessment. On the contrary, when the change of environmental parameters is small, V j (t) Gradually restore to the initial weight

[0156] In the early stages of construction, environmental parameters have a greater impact on risk assessment, so the weight of environmental parameters is higher, while the weight of operating status parameters is relatively lower. In the later stages of construction, the impact of operating status parameters on risk assessment gradually increases, and the weight of environmental parameters gradually decreases.

[0157] This current risk formula automatically adjusts the weights within the dynamic assessment model based on the construction phase, improving assessment accuracy. By dynamically adjusting the influencing factors of environmental and operating parameters, false alarms caused by drastic changes in a single factor can be avoided, thereby improving the stability of early warnings. Given the complex and ever-changing construction environment, this formula allows for real-time adjustment of parameter weights, making this method adaptable to diverse construction conditions.

[0158] Through the above-mentioned adaptive adjustment formula for construction stage weights, the present invention realizes the optimization of risk assessment weights at each stage of transformer construction, ensures that the risk assessment results are more in line with the actual construction environment, and provides more reliable technical guarantees for equipment operation safety and environmental compliance.

[0159] In step 3 above, the final output of the dynamic assessment model is the current environmental risk and equipment failure risk of the transformer. The current risk can be represented by a hierarchical scoring system, for example, divided into three levels: low, medium, and high, or a numerical scoring method can be used to map the current risk to a continuous value between 0 and 1 to provide a more refined risk assessment result. After outputting the environmental risk assessment result and equipment safety monitoring result of the transformer based on the current risk, the following is also included:

[0160] If the current risk level is higher than a preset threshold, an early warning message is sent to a construction monitoring center via a remote terminal;

[0161] The warning information includes: abnormal spatial location information of the transformer, abnormal risk category of the transformer, current risk level of the transformer, level of current risk level, emergency notification of abnormal risk category and recommended handling measures for abnormal risk category.

[0162] During early warning, if the risk exceeds a set threshold, the system triggers an early warning mechanism and outputs the risk prediction results to the monitoring platform at the construction monitoring center, allowing construction personnel and operations and maintenance personnel to take timely action. If the risk exceeds the threshold, an early warning message is generated and the abnormal component or area is located, enabling environmental risk assessment and equipment safety monitoring. This warning information can be sent to the construction monitoring center via a remote terminal, prompting operations and maintenance personnel to take appropriate measures.

[0163] When the current risk R(t) calculated by the dynamic assessment model exceeds the set warning threshold R th , when the early warning mechanism is automatically triggered. The early warning threshold R th It can be set based on the construction environment, transformer operating status, and historical risk data, and can be dynamically adjusted to meet the safety requirements of different construction stages. When an abnormal risk is detected, it combines point cloud ranging data with image recognition data to accurately identify the abnormal component or area and generate fault location information.

[0164] Warning information can also include the spatial coordinates of the abnormal component, risk category, risk level, and recommended treatment measures. The spatial coordinates are calculated based on the fusion of point cloud data and image data to ensure that the abnormal point accurately corresponds to the actual physical component of the transformer. Risk categories include but are not limited to structural anomalies, abnormal electrical parameters, or excessive environmental impacts. The risk level is categorized as low, medium, or high based on the calculated current risk R(t). Recommended treatment measures are based on historical maintenance records and expert system recommendations, providing possible fault cause analysis and corresponding treatment plans.

[0165] Warning information can be transmitted via wireless or wired networks to remote terminals, including the monitoring system at the construction monitoring center and the mobile devices of operators. Upon receiving the warning information, the remote terminal displays the transformer's risk status in real time and displays the spatial distribution of abnormal components through a visual interface, allowing operators to quickly understand the fault situation. To ensure the timeliness of warning information, a low-latency communication protocol is used during data transmission, ensuring that warning signals are transmitted to relevant personnel in the shortest possible time.

[0166] To further enhance the reliability of early warnings, a secondary confirmation mechanism can be implemented after an alert is triggered, combining historical data to review the current risk level. If the current risk level consistently exceeds the threshold for a short period of time, the alert level can be upgraded, and an emergency notification can be sent to a higher-level monitoring center or management department. This mechanism effectively reduces false alarms and ensures the accuracy and authority of early warning information.

[0167] As mentioned above, when the risk level exceeds the preset threshold, an early warning is triggered, the abnormal component or area is determined, and fault location information is generated; the early warning information is sent to the remote terminal. When the risk level continues to exceed the threshold, the early warning level is upgraded and an emergency notification is sent to the monitoring center to realize transformer environmental risk assessment and equipment safety monitoring.

[0168] The present invention's method for assessing environmental risks during the transformer construction period, based on point cloud ranging and a dynamic assessment model, includes collecting image data and point cloud ranging data to identify abnormalities in the transformer's appearance and obtain abnormal coordinate information; matching image recognition data with point cloud data to determine abnormal components or areas; collecting transformer operating status parameters and environmental parameters, and combining them with historical data to input an adaptive neural network to assess risk; dynamically adjusting the weights of risk assessment factors through the adaptive neural network to adapt to changes in the construction phase; if the risk exceeds a preset threshold, outputting an early warning message and accurately locating the abnormal component, thereby achieving environmental risk assessment and equipment safety monitoring. Therefore, the present invention can improve the accuracy of risk prediction during the transformer construction period, reduce false positives and missed positives, and enhance equipment operation and maintenance efficiency and construction safety. This embodiment, through the coordinated application of point cloud ranging, image recognition, dynamic early warning and other technologies, achieves an accurate assessment of environmental risks during the pole-mounted transformer construction period, ensures equipment safety and environmental compliance during construction, and improves monitoring efficiency and response speed.

[0169] Those skilled in the art should know that:

[0170] Existing monitoring technologies typically employ a single detection method. For example, optical imaging is used to inspect the transformer's appearance, while operational data acquisition systems are used to assess the stability of parameters such as load, temperature, and current. These methods operate independently, resulting in isolated data in practice, making comprehensive and integrated fault assessment difficult. Furthermore, optical imaging is significantly affected by factors such as ambient lighting, obstructions, and contamination, potentially leading to false or missed detections. Operational data acquisition solutions based on fixed sensor deployments struggle to adapt to the dynamic changes of complex construction environments.

[0171] In terms of data processing, existing methods often rely on traditional statistical analysis or fixed algorithmic models for early warning, relying on preset empirical thresholds for anomaly detection. However, this approach has significant limitations and cannot adaptively adjust warning criteria based on changes in the construction environment. For example, in extreme weather or complex construction conditions, a single fixed threshold can lead to an increased false alarm rate, compromising the reliability of the monitoring system. Furthermore, most current early warning systems fail to fully utilize multi-dimensional data for deep integration, resulting in the failure to promptly identify certain potential risks, impacting the effectiveness of transformer safety assessments during the construction period.

[0172] In actual engineering applications, these technical limitations may lead to multiple potential problems. First, due to the lack of unified data sources, there is a lack of effective correlation analysis between appearance anomaly detection and operating status monitoring, and it may be impossible to accurately determine the root cause of equipment anomalies. Second, early warning mechanisms that rely on fixed parameter settings have poor adaptability in complex environments and cannot meet the dynamic monitoring needs during the construction period. Finally, because most existing systems use static data processing methods and lack intelligent adaptive adjustment capabilities, certain hidden dangers during the construction process are difficult to detect and effectively address in a timely manner, which in turn affects the long-term stability of the equipment and environmental compliance.

[0173] Therefore, current technology still has many shortcomings in environmental risk assessment during the transformer construction period. There is an urgent need for a more accurate, intelligent, and adaptable monitoring and early warning method to improve the risk identification capability in complex construction environments and enhance the level of protection for equipment safety and environmental compliance.

[0174] This method combines point cloud ranging technology with a dynamic assessment model to achieve precise risk assessment during the transformer construction period and improve the accuracy of equipment anomaly detection. By integrating image data with point cloud data, it can precisely locate abnormal transformer components or areas, thereby enhancing the reliability of fault identification. This method utilizes image data acquisition, point cloud ranging, image recognition, operational status monitoring, and an adaptive early warning model to achieve environmental risk assessment during the transformer construction period, improving monitoring accuracy and real-time response capabilities.

[0175] The dynamic assessment model based on the adaptive neural network of the present invention can dynamically adjust the weights of risk assessment factors according to real-time data, can adapt to different construction stages, improve the adaptability and accuracy of risk assessment, and reduce false alarm rates and missed alarm rates.

[0176] The early warning mechanism of the present invention can trigger an alarm when the risk level exceeds a threshold, and accurately output abnormal components and environmental impact information, providing early warning for construction monitoring and equipment maintenance, ensuring safety and environmental compliance during the transformer construction period.

[0177] Example 2:

[0178] like Figure 2 As shown, the present invention also provides a transformer construction period environmental risk assessment system, comprising:

[0179] The abnormal state data acquisition module is used to input the transformer image data at the construction site during the construction period into the image recognition model to identify the current abnormal coordinate information of the transformer; based on the current abnormal coordinate information, obtain the current abnormal state data;

[0180] A risk determination module is configured to input the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters into a pre-built dynamic assessment model for evaluation to obtain the current risk of the transformer, wherein the current risk includes the current environmental risk and the current equipment failure risk; wherein the dynamic assessment model is obtained by training an adaptive neural network using the historical abnormal state data, historical operating state parameters, historical environmental parameters, and historical risk of the transformer;

[0181] The result acquisition module is used to output the environmental risk assessment result and equipment safety monitoring result of the transformer based on the current risk level.

[0182] The above-mentioned abnormal state data acquisition module can be a multi-source sensor. Multi-source sensors are deployed during the transformer construction period to obtain transformer image data at the construction site during the construction period; then, the spatial information, appearance status and operation data of the transformer and the construction site can be obtained from the transformer image data.

[0183] like Figure 3 As shown in the figure, during the transformer construction phase, image acquisition devices, point cloud ranging equipment, and environmental and operational data sensors were installed. These devices work together to obtain multi-dimensional information about the transformer and its surroundings. The image acquisition device captures images of the construction site and the transformer's exterior, the point cloud ranging equipment measures the spatial coordinates of transformer components, and the environmental and operational data sensors collect operating parameters such as temperature, humidity, wind speed, current, and voltage. This data serves as input for subsequent analysis.

[0184] Image data is collected using an image acquisition device, which can include high-definition industrial cameras, infrared cameras, or multispectral imaging equipment to adapt to varying lighting environments and construction conditions. The image acquisition device is fixedly mounted on the transformer support, a monitoring support at the construction site, or a drone platform, and can cover the transformer itself and key areas of the construction site. The image data resolution is adjusted according to the monitoring requirements of the construction site. The image acquisition device captures real-time images of the construction site and transmits the data to the image processing unit. Point cloud ranging equipment uses laser scanning technology to create a three-dimensional model of the transformer and its surroundings, determining the spatial relationships of each component and storing it in a data storage unit. After image data is collected, it is transmitted to the image processing unit via a data transmission module. The image processing unit includes an image preprocessing module that performs image processing, such as noise reduction, deblurring, and contrast enhancement, to improve data quality. To ensure stable data transmission, image data can be transmitted via either wired or wireless networks.

[0185] When collecting image data, point cloud ranging equipment can be used using laser scanning technology to create a three-dimensional model of the transformer and its surroundings. This equipment emits a high-frequency laser beam and calculates its return time and angular change to construct spatial position information for the transformer. The accuracy of point cloud ranging is affected by the laser wavelength, measurement distance, and ambient light interference. Therefore, during the measurement process, the appropriate ranging mode should be selected based on the size of the transformer and the complexity of the construction site. For example, for smaller transformers, a short-range, high-precision mode can be used, while for larger transformers, a long-range scanning mode can be used to ensure data integrity.

[0186] The acquired point cloud data is stored in a data storage unit and time-stamped synchronously with the image data for subsequent fusion analysis. The point cloud data includes the surface morphology of the transformer itself, the relative positions of key components, and the spatial layout of the surrounding construction environment. The data storage unit can utilize high-capacity storage media to support the storage of massive amounts of point cloud data and be equipped with an indexing mechanism to improve data retrieval efficiency. The data storage unit can be a database.

[0187] Furthermore, the risk determination module includes:

[0188] A parameter determination submodule, configured to receive the current abnormal state data, the current operating state parameters of the transformer and the current environmental parameters based on the input layer of the adaptive neural network of the dynamic evaluation model;

[0189] a contribution factor determination submodule, configured to perform nonlinear mapping on the current abnormal state data, the current operating state parameters of the transformer, and the current environmental parameters based on the hidden layer of the adaptive neural network to obtain an operating state impact contribution factor and an environmental impact contribution factor;

[0190] The weight determination submodule is used to obtain the operating state parameter weight based on the actual risk degree and reference risk degree of the historical operating conditions of the transformer using the operating state parameter weight adjustment formula; and to obtain the environmental parameter weight based on the actual value and historical average value of the operating state parameter of the transformer using the environmental parameter weight adjustment formula;

[0191] The risk determination submodule is used to calculate the current risk based on the output layer of the adaptive neural network according to the operating state impact contribution factor, the environmental impact contribution factor, the operating state parameter weight and the environmental parameter weight.

[0192] Furthermore, the current risk level satisfies the following formula:

[0193]

[0194] In the above formula, R(t) is the current risk at time t, α is the contribution factor of the operating state, N is the total number of operating parameters in the operating state parameters, and W i is the i-th operating parameter weight in the operating state parameter weight, P i (t) is the real-time measurement value of the i-th operating parameter at time t, is the historical reference value of the i-th operating parameter, β is the environmental impact contribution factor, M is the total number of environmental parameters in the environmental parameter, V j is the jth environmental parameter weight in the environmental parameter weights, E j (t) is the measured value of the jth environmental parameter among the environmental parameters at time t, is the average value of the jth environmental parameter, γ is the weight of the sudden risk factor, and dR / dt is the instantaneous change rate of the sudden risk.

[0195] Furthermore, the operating state parameter weight satisfies the following formula:

[0196]

[0197] In the above formula, W i (t) is the weight of the i-th operating parameter at time t, is the initial weight of the i-th operating parameter, is the historical operating conditions adaptation factor, R k (t) is the actual risk of the kth historical working condition at time t, is the reference risk of the kth historical operating condition, and K is the total number of historical operating conditions;

[0198] The environmental parameter weights satisfy the following formula:

[0199]

[0200] In the above formula, V j (t) is the weight of the jth environmental parameter at time t, is the initial weight of the jth environmental parameter, μ is the environmental adaptation factor, P m (t) is the actual value of the mth operating parameter involved in the calculation at time t, is the historical mean of the mth operating parameter involved in the calculation, and S is the number of operating parameters involved in the calculation.

[0201] Furthermore, the system also includes a dynamic evaluation model construction module, which is used to:

[0202] Determine the historical risk of the transformer based on the acquired historical abnormal state data, historical operating state parameters, and historical environmental parameters; determine the historical environmental parameter weight and historical environmental parameter measurement value based on the historical abnormal state data; determine the historical abnormal state data weight, the historical construction stage of the historical abnormal state data, and the historical abnormal state data measurement value based on the historical abnormal state data;

[0203] The historical environmental parameter weights, the historical environmental parameter measurements, the historical abnormal state data measurements, the historical construction stages, and the historical abnormal state data weights are input into an adaptive neural network to output a predicted environmental risk degree; the difference between the historical risk degree and the predicted environmental risk degree is used as a prediction error; with the goal of minimizing the prediction error, the adaptive neural network is trained to obtain a dynamic evaluation model.

[0204] Furthermore, the abnormal state data acquisition module includes:

[0205] A coordinate mapping submodule is used to associate the acquired standard transformer spatial position information with the current abnormal coordinate information using a coordinate mapping technique to obtain abnormal components and areas;

[0206] An abnormal point determination submodule, configured to determine multiple abnormal points of the transformer based on the abnormal components and areas;

[0207] The abnormal state data determination submodule is used to analyze the multiple abnormal points using cluster analysis to obtain current abnormal state data.

[0208] Furthermore, the coordinate mapping submodule is specifically used to:

[0209] The acquired standard transformer spatial position information is associated with the current abnormal coordinate information using coordinate mapping technology to obtain a three-dimensional coordinate system of the abnormal point;

[0210] Based on the three-dimensional coordinate system of the abnormal points, a point cloud registration algorithm or a nearest neighbor search algorithm is used to determine the abnormal parts and areas.

[0211] Furthermore, the abnormal state data determination submodule is specifically configured to:

[0212] Analyzing the multiple abnormal points using cluster analysis to obtain dense distribution probabilities of the multiple abnormal points in the transformer space, and converting the dense distribution probability into a spatial abnormality distribution map;

[0213] Based on the spatial anomaly distribution map, current abnormal state data is determined.

[0214] Furthermore, the system also includes an early warning module for:

[0215] If the current risk level is higher than a preset threshold, an early warning message is sent to a construction monitoring center via a remote terminal;

[0216] The warning information includes: abnormal spatial location information of the transformer, abnormal risk category of the transformer, current risk level of the transformer, level of current risk level, emergency notification of abnormal risk category and recommended handling measures for abnormal risk category.

[0217] The system of the present invention realizes closed-loop control of transformer environmental risk assessment and equipment safety monitoring, forming a complete monitoring system from risk identification to early warning output to operation and maintenance disposal, thereby improving the safety and environmental compliance of transformers during construction.

[0218] Example 3:

[0219] like Figure 4 As shown, the present invention also provides an electronic device, which may be a computer, a single-chip microcomputer, a smart mobile device, or the like. The electronic device in this embodiment may include a processor, a memory, a transceiver component, and the like. The memory, processor, and transceiver component are connected via a bus; the memory may be used to store an execution program, which may include instructions; and the processor may be used to execute the instructions stored in the memory. The memory may also be used to store data, which may be accessed and / or modified during the execution of the instructions.

[0220] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of a transformer construction period environmental risk assessment method in the above embodiment.

[0221] Example 4:

[0222] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in the electronic device for storing programs and data. It can be understood that the storage medium here can include both the built-in storage medium in the electronic device and the extended storage medium supported by the electronic device. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory (non-volatile memory), such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a transformer construction period environmental risk assessment method in the above embodiment.

[0223] Those skilled in the art will appreciate that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0224] The present invention is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0225] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1The function specified in one or more boxes.

[0226] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0227] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the scope of protection of the pending claims.

Claims

1. A method for environmental risk assessment during transformer construction period, characterized in that: include: Input the transformer image data at the construction site during the construction period into the image recognition model to identify the current abnormal coordinate information of the transformer; Based on the current abnormal coordinate information, obtaining current abnormal state data; Input the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters into a pre-built dynamic assessment model for evaluation to obtain the current risk of the transformer, which includes the current environmental risk and the current equipment failure risk; Based on the current risk level, output the environmental risk assessment result and equipment safety monitoring result of the transformer; The dynamic assessment model is obtained by training an adaptive neural network using historical abnormal state data, historical operating state parameters, historical environmental parameters and historical risk levels of the transformer.

2. The method according to claim 1, characterized in that The current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters are input into a pre-built dynamic assessment model for assessment to obtain the current risk level of the transformer, including: The input layer of the adaptive neural network based on the dynamic evaluation model receives the current abnormal state data, the current operating state parameters of the transformer and the current environmental parameters; Based on the hidden layer of the adaptive neural network, nonlinear mapping is performed on the current abnormal state data, the current operating state parameters of the transformer, and the current environmental parameters to obtain an operating state impact contribution factor and an environmental impact contribution factor; Based on the actual risk and reference risk of the transformer's historical operating conditions, the operating state parameter weight is obtained using the operating state parameter weight adjustment formula; based on the actual value and historical average of the transformer's operating state parameter, the environmental parameter weight is obtained using the environmental parameter weight adjustment formula; Based on the output layer of the adaptive neural network, the current risk is calculated according to the operating state impact contribution factor, the environmental impact contribution factor, the operating state parameter weight and the environmental parameter weight.

3. The method according to claim 2, characterized in that The current risk level satisfies the following formula: In the above formula, R(t) is the current risk at time t, α is the contribution factor of the operating state, N is the total number of operating parameters in the operating state parameters, and W i is the i-th operating parameter weight in the operating state parameter weight, P i (t) is the real-time measurement value of the i-th operating parameter at time t, is the historical reference value of the i-th operating parameter, β is the environmental impact contribution factor, M is the total number of environmental parameters in the environmental parameter, V j is the jth environmental parameter weight in the environmental parameter weights, E j (t) is the measured value of the jth environmental parameter among the environmental parameters at time t, is the average value of the jth environmental parameter, γ is the weight of the sudden risk factor, and dR / dt is the instantaneous change rate of the sudden risk.

4. The method according to claim 2, characterized in that The operating status parameter weight satisfies the following formula: In the above formula, W i (t) is the weight of the i-th operating parameter at time t, is the initial weight of the i-th operating parameter, is the historical operating conditions adaptation factor, R k (t) is the actual risk of the kth historical working condition at time t, is the reference risk of the kth historical operating condition, and K is the total number of historical operating conditions; The environmental parameter weights satisfy the following formula: In the above formula, V j (t) is the weight of the jth environmental parameter at time t, is the initial weight of the jth environmental parameter, μ is the environmental adaptation factor, P m (t) is the actual value of the mth operating parameter involved in the calculation at time t, is the historical mean of the mth operating parameter involved in the calculation, and S is the number of operating parameters involved in the calculation.

5. The method according to claim 2, characterized in that The pre-construction process of the dynamic assessment model includes: Determine the historical risk of the transformer based on the acquired historical abnormal state data, historical operating state parameters, and historical environmental parameters; determine the historical environmental parameter weight and historical environmental parameter measurement value based on the historical abnormal state data; determine the historical abnormal state data weight, the historical construction stage of the historical abnormal state data, and the historical abnormal state data measurement value based on the historical abnormal state data; The historical environmental parameter weights, the historical environmental parameter measurements, the historical abnormal state data measurements, the historical construction stages, and the historical abnormal state data weights are input into an adaptive neural network to output a predicted environmental risk degree; the difference between the historical risk degree and the predicted environmental risk degree is used as a prediction error; with the goal of minimizing the prediction error, the adaptive neural network is trained to obtain a dynamic evaluation model.

6. The method according to claim 1, wherein The acquiring of current abnormal state data based on the current abnormal coordinate information includes: The acquired standard transformer spatial position information is associated with the current abnormal coordinate information using coordinate mapping technology to obtain abnormal components and areas; Based on the abnormal components and areas, determining multiple abnormal points of the transformer; Cluster analysis is used to analyze the multiple abnormal points to obtain current abnormal state data.

7. The method according to claim 6, characterized in that The obtained standard transformer spatial position information is associated with the current abnormal coordinate information using a coordinate mapping technique to obtain abnormal components and areas, including: The acquired standard transformer spatial position information is associated with the current abnormal coordinate information using coordinate mapping technology to obtain a three-dimensional coordinate system of the abnormal point; Based on the three-dimensional coordinate system of the abnormal points, a point cloud registration algorithm or a nearest neighbor search algorithm is used to determine the abnormal parts and areas.

8. The method according to claim 6, characterized in that The cluster analysis is used to analyze the multiple abnormal points to obtain current abnormal state data, including: Analyzing the multiple abnormal points using cluster analysis to obtain dense distribution probabilities of the multiple abnormal points in the transformer space, and converting the dense distribution probability into a spatial abnormality distribution map; Based on the spatial anomaly distribution map, current abnormal state data is determined.

9. The method according to claim 1, characterized in that After outputting the environmental risk assessment result and the equipment safety monitoring result of the transformer based on the current risk level, the method further includes: If the current risk level is higher than a preset threshold, an early warning message is sent to a construction monitoring center via a remote terminal; The warning information includes: abnormal spatial location information of the transformer, abnormal risk category of the transformer, current risk level of the transformer, level of current risk level, emergency notification of abnormal risk category and recommended handling measures for abnormal risk category.

10. A transformer construction period environmental risk assessment system, characterized in that: include: The abnormal state data acquisition module is used to input the transformer image data at the construction site during the construction period into the image recognition model to identify the current abnormal coordinate information of the transformer; Based on the current abnormal coordinate information, obtaining current abnormal state data; A risk determination module is configured to input the current abnormal state data, the obtained current operating state parameters of the transformer, and the current environmental parameters into a pre-built dynamic assessment model for evaluation to obtain the current risk of the transformer, wherein the current risk includes the current environmental risk and the current equipment failure risk; wherein the dynamic assessment model is obtained by training an adaptive neural network using the historical abnormal state data, historical operating state parameters, historical environmental parameters, and historical risk of the transformer; The result acquisition module is used to output the environmental risk assessment result and equipment safety monitoring result of the transformer based on the current risk level.

Images