Modularized portable transformer multi-parameter detection device

Abstract

The invention provides a portable transformer multi-parameter detection device with a modular design, and relates to the technical field of transformer detection devices.According to the portable transformer multi-parameter detection device with the modular design, by arranging multiple sets of modular parameter detection devices, comprehensive, efficient and automatic detection of electrical parameters of a transformer can be achieved; an automatic identification mechanism can be utilized to ensure that the detection module is strictly matched with the work order requirement, and the effect of improving the detection accuracy and the operation safety is achieved; intelligent path planning comprehensively considering the path length and the road risk is combined, so that the inspection route is effectively optimized, resources are saved, and potential safety hazards are reduced; and a risk prediction model based on a Bayesian network is introduced, so that combined risk assessment and real-time early warning of multiple electrical indexes of the transformer are realized, and the equipment fault prevention capability is improved. According to the overall scheme, the intelligent level and the working efficiency in the transformer detection process are greatly improved, and the applicability in a multi-transformer detection scene is enhanced.

Description

Technical Field

[0001] This invention relates to the field of transformer testing device technology, specifically a modularly designed portable transformer multi-parameter testing device. Background Technology

[0002] With the continuous expansion of power system scale and the increasing complexity of operating environments, transformers, as key equipment in power systems, directly affect the safety and stability of the power grid. Traditional transformer testing equipment mostly performs single-parameter testing, and the testing process relies on manual operation, resulting in problems such as low testing efficiency, susceptibility to errors, and fragmented data management. During regular inspections of plant areas or large facilities, due to the numerous testing items and the wide distribution of transformers, how to achieve efficient multi-parameter testing, accurate identification, and intelligent scheduling has become an urgent technical challenge to be solved.

[0003] In the prior art, CN117408513A discloses a method and system for fault risk detection across the entire power grid link. The method includes: acquiring power grid link information to obtain power grid equipment information; obtaining equipment information characteristics based on the power grid equipment information; inputting the power grid link information into a preset link prediction model to obtain information on substations and transmission lines with potential risks; constructing a power grid equipment risk prediction model based on the equipment information characteristics; obtaining the predicted risk status of the entire power grid link based on the information on substations and transmission lines with potential risks through the power grid equipment risk prediction model; and arranging a power grid equipment maintenance plan based on the predicted risk status of the entire power grid link. While this solution can identify risks in electrical equipment, it does not solve the problem of low efficiency caused by the large number of detection items. Furthermore, this problem is amplified in scenarios where multiple transformers need to be detected, and there are also risks of false positives and false negatives, resulting in poor overall applicability of the solution.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a modularly designed portable transformer multi-parameter detection device to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The modularly designed portable transformer multi-parameter testing device specifically includes: The data transmission module is used to construct a communication network to receive preset work order information and send it to the control module; Several sets of parameter detection modules are provided, which are used to detect the electrical parameters of the transformer and send them to the control module. The control module includes a path planning unit, an identification unit, and a risk prediction unit, wherein: The path planning unit has an ant colony algorithm and an electronic map built in, which is used to plan the movement path of the detection device according to the work order information to obtain the optimal path, and generate a path table based on the optimal path and the preset transformer number. The identification unit has a built-in GPS element, which is used to collect the real-time position of the detection device after each movement, and compare the number of the transformer closest to the real-time position with the transformer number in the path table to determine whether the detection device has deviated from the optimal path. The risk prediction unit has a built-in Bayesian network model, which is used to determine the risk of each group of transformers in turn based on the electrical parameters of the transformers. An alarm module is used to issue an alarm signal based on the judgment result of the identification unit and the risk judgment of the transformer by the risk prediction unit.

[0007] Preferably, the work order information includes, but is not limited to, the location information of the transformer to be tested and the category of electrical parameters; The parameter detection module is electrically connected to the control module through a standardized interface, including but not limited to a current tester, voltage tester, and resistance tester; The electrical parameters include, but are not limited to, the transformer's current, voltage, and resistance.

[0008] Preferably, the working logic of the path planning unit is as follows: The transformers to be inspected are numbered, and the location information of the transformers to be inspected in the work order information is categorized into a set form; Based on the preset electronic map and location information, the road width at the location of the transformer and the map distance between different transformers are obtained; The cost function is constructed by taking road width and map distance as risk and path terms, respectively. The goal is to optimize the path by using the shortest path and lowest risk, and the optimal path is generated. The detection sequence of the transformer is determined based on the optimal path. The transformer numbers are sorted according to the detection sequence, and a path table is generated based on the detection sequence and the transformer numbers.

[0009] Preferably, the path term in the cost function is the sum of the map distances between all transformers; The risk term in the cost function is the sum of the inverse of the road width at all transformer locations multiplied by a preset additional risk factor. The additional risk factor is used to reflect the road conditions at the transformer locations, and its value is inversely proportional to the quality of the road conditions.

[0010] Preferably, the working logic of the identification unit is as follows: After each movement of the detection device, the real-time position of the detection device is collected, and the real-time position is compared with the position information of the transformer. The transformer closest to the real-time position and its corresponding transformer number are identified, and the map distance between the real-time position and the closest transformer is obtained. The path table is queried to obtain the corresponding transformer number based on the real-time detection sequence, and then compared with the transformer number corresponding to the nearest transformer: If and only if two sets of transformers have the same number and the map distance between the real-time location and the nearest transformer does not exceed the preset distance error, the detection device is considered to have moved according to the path table and has not deviated from the optimal path, and the judgment result is normal. In all other cases, the judgment result is abnormal, and an alarm signal is issued through the alarm module.

[0011] Preferably, the working logic of the risk prediction unit is as follows: The electrical parameters of the transformer are represented in vector form to generate an index vector, which is then used as the input to the Bayesian network model. Based on the indicator vector, the corresponding risk event and its posterior probability are inferred using a Bayesian network model. The risk events and posterior probabilities corresponding to the indicator vectors are arranged in descending order of probability, and then set as the output of the Bayesian network model. The transformer is deemed to be at risk if and only if the posterior probability of the risk event ranked first exceeds a preset probability threshold.

[0012] Preferably, the data used for optimizing the Bayesian network model during training is obtained from the historical fault data of each transformer.

[0013] Preferably, after the risk prediction unit determines the risk, it considers the group of transformers to have been inspected. The inspection device then moves along the optimal path to inspect the next group of transformers in the path table. After the inspection device has inspected all the transformers in sequence, it returns to the initial position along the original path.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention, by setting up multiple modular parameter detection devices, not only achieves comprehensive, efficient, and automated detection of transformer electrical parameters, but also utilizes an automatic identification mechanism to ensure strict matching between detection modules and work order requirements, thereby improving detection accuracy and operational safety. Combined with intelligent path planning that comprehensively considers path length and road risks, it effectively optimizes inspection routes, saves resources, and reduces safety hazards. Furthermore, by introducing a risk prediction model based on Bayesian networks, it achieves joint risk assessment and real-time early warning of multiple transformer electrical indicators, improving equipment fault prevention capabilities. The overall solution significantly enhances the intelligence level and work efficiency in the transformer inspection process and strengthens its applicability in multi-transformer inspection scenarios. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall modular structure of the present invention; Figure 2 This is a schematic diagram of the workflow of the path planning unit in this invention; Figure 3 This is a schematic diagram of the workflow of the identification unit in this invention; Figure 4 This is a schematic diagram of the workflow of the risk prediction unit in this invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0017] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0018] Example: Please see Figures 1-4 The present invention provides a technical solution: The modularly designed portable transformer multi-parameter detection device specifically includes: a data transmission module, several sets of parameter detection modules, a control module, an alarm module, and a display module.

[0019] The data transmission module is used to transmit data via the network, receive preset work order information and send it to the control module. The work order information includes, but is not limited to, the location information of the transformer to be tested and the category of electrical parameters. The parameter detection module is electrically connected to the control module through a standardized interface, including, but not limited to, a current tester, a voltage tester and a resistance tester, for detecting the electrical parameters of the transformer, including, but not limited to, the transformer's current, voltage and resistance, and sending the electrical parameters to the control module. The control module includes an identification unit, a path planning unit, and a risk prediction unit, which can be implemented using an MCU microcontroller chip. The path planning unit has built-in ant colony algorithm and electronic map, which are used to plan the movement path of the detection device according to the work order information to obtain the optimal path, and generate a path table based on the optimal path and the preset transformer number. The identification unit has a built-in GPS component, which is used to collect the real-time position of the detection device after each movement. The number of the transformer closest to the real-time position is compared with the transformer number in the path table to determine whether the detection device has deviated from the optimal path. The risk prediction unit has a built-in Bayesian network model, which is used to determine the risk of each group of transformers in turn based on the electrical parameters of the transformers. The alarm module is used to issue alarm signals based on the detection results of the identification unit and the risk assessment of the transformer by the risk prediction unit.

[0020] When conducting transformer testing, a wide variety of testing items are required, resulting in a diverse range of testing equipment. This approach modularizes these testing devices, standardizing the interfaces of each parameter testing module. These modules can be added, replaced, or called upon based on pre-defined work orders. This significantly reduces equipment relocation and assembly time and automatically identifies the type of parameter testing module, preventing human error during module additions or replacements, thus further improving operational safety and efficiency.

[0021] The route planning unit has a pre-installed electronic map, and its working logic is as follows: The transformers to be inspected are numbered, and the location information of the transformers to be inspected in the work order information is categorized into a set form; Based on the preset electronic map and location information, the road width at the location of the transformer and the map distance between different transformers are obtained; The cost function is constructed by treating road width and map distance as risk and path terms, respectively. The goal is to optimize the path by using the shortest path and lowest risk, and the optimal path is generated.

[0022] Taking regular factory inspections as an example, the location information of each transformer, after being categorized into sets, can be represented as follows:

[0023] In the formula Indicates the first Transformer assembly, Indicates the first The location information of the transformer group can be latitude and longitude, or two-dimensional coordinates from an electronic map, depending on the type of electronic map or the project requirements. Indicates the transformer index. This indicates the total number of transformers to be tested. The transformer indexes can be manually numbered according to actual testing needs.

[0024] With the electronic map and the location information of each transformer, the map distance between the transformers and the road width at their locations can be obtained. The map distance refers to the distance between two transformers along accessible routes such as roads or paths on the electronic map, not the straight-line distance between two points. This value can be obtained from the electronic map. Similarly, the road width can be obtained by combining the electronic map with the map scale (i.e., the road width on the electronic map multiplied by the map scale). Specifically: If the electronic map pre-set in the path planning unit is not connected to an external network, the road network of the electronic map can be abstracted into a graph structure, with nodes as intersections, edges as roads, and edge weights as road lengths or travel times. Classic algorithms such as Dijkstra's algorithm or A* algorithm can be used to calculate the shortest path distance between two points, and the output result is the map distance. If the route planning unit is connected to an external network via a data transmission module, it can directly call the API interface of a third-party electronic map and output the map distance directly.

[0025] Understandably, connecting only to the factory's IoT network without an external network provides more secure monitoring of stored data and is suitable for scenarios with high confidentiality requirements; connecting to an external network is more convenient and suitable for other scenarios with less stringent confidentiality requirements. The choice between the two methods depends on the specific needs.

[0026] A cost function is constructed using road width and map distance as the risk and path terms, respectively. The path term in the cost function is the sum of the map distances between all transformers, and the risk term is the sum of the reciprocal of the road width at each transformer location multiplied by a preset additional risk factor. This additional risk factor reflects the road conditions at the transformer locations, and its value is inversely proportional to the quality of the road conditions. The expression for the cost function is:

[0027] In the formula This represents the function value of the cost function. Indicates from the first Group of transformers to the first Map distance of the transformer group Indicates the index of a transformer that has not been visited. Indicates the first The width of the road where the transformer is located. Indicates the first Additional risk factors related to the location of the transformer group. , All of these represent weighting coefficients.

[0028] The risk factor is determined by both road width and additional risk factors, with the additional risk factor being inversely proportional to the quality of the road conditions. Taking the regular inspection of the plant area as an example, if the road where a transformer is located is under construction, the additional risk factor for that location can be modified to a larger constant beforehand, increasing the risk factor. This allows the algorithm to avoid repeatedly traversing that road in subsequent optimizations, thus mitigating road risks and protecting equipment safety. For roads not under construction, the additional risk factor can be set to 1, having no impact on the risk factor. In other words, the wider the road, the less likely a collision will occur, and the lower the risk.

[0029] Specifically, when using the ant colony algorithm for path optimization, the number of ants can be initialized to 50 to ensure search diversity, and the pheromone matrix can be initialized to 1.

[0030] For each ant, it will be the first Starting from the transformer group, the transition probability of successively selecting the next access point (i.e., the location of any transformer not yet reached) is:

[0031] In the formula Indicates the first During the nth iteration, the 1st Only one ant from the first The transformer group moved to the first The probability of forming a transformer. Indicates the first In the nth iteration, from the 1st... The transformer group moved to the first The pheromone matrix of the transformer group, with an initial value of 1, i.e. , Indicates the first The transformer group moved to the first The heuristic information for the transformer group is calculated using the following formula:

[0032] This value indicates that shorter, less risky paths are more attractive, reducing the probability of choosing high-risk points.

[0033] , These represent the importance of pheromones and the importance of heuristic functions, respectively. Both are preset constants that can be set to 1 and 2, respectively. This represents the set of indices for all transformers that have not yet been visited.

[0034] The ant randomly selects the next point according to the probability distribution, adds it to the path sequence, and repeats this process until all points have been visited (i.e., all transformers have been reached), and calculates the path cost of the entire movement path based on the cost function.

[0035] During the iteration process, to prevent getting trapped in local optima, the pheromone matrix needs to be evaporated. Then, based on the paths chosen by each ant in the previous search round, the pheromone matrix is ​​updated, with the evaporation rate... Set to 0.1, that is:

[0036] In the formula Indicates the first During the nth iteration, the 1st The contribution of a single ant to the pheromone matrix is ​​calculated as follows:

[0037] In the formula Indicates the first During the nth iteration, the 1st The path cost of the path taken by an ant. This is a preset constant, typically taken as 100. Represents an indicator function, when the first... The path taken by the ant includes the edges Time (i.e., from the first) The transformer group moved to the first (group transformer) The index of the unreached transformer is 1, and the index of the pheromone matrix from transformer 1 to transformer 2 is 0. Conversely, the index of the unreached transformer is 0. For example, if there are 5 transformers and the ant starts from transformer 1, the possible paths include 1-2-3-4-5, 1-3-2-4-5, 1-5-4-2-3, and so on. If the ant's final path includes the segment 1-2, the corresponding index function is 1, and the pheromone matrix from transformer 1 to transformer 2 will increase in the next iteration, increasing the probability of the ant transitioning from transformer 1 to transformer 2. Conversely, if the ant's final path does not include the segment 1-2, the corresponding index function is 0, and the pheromone matrix from transformer 1 to transformer 2 remains unchanged in the next iteration. This ensures that the pheromone matrix of the path with the lowest cost increases during iteration, making it more likely to be chosen by the ant. Thus, the path with the lowest cost is selected as the optimal path during iteration, completing path optimization.

[0038] The detection sequence of transformers is determined based on the optimal path. The transformer numbers are then sorted according to the detection sequence, and a path table is generated based on both the detection sequence and the transformer numbers. Assuming the optimal path selected in the above example is 2-5-3-1-4, the corresponding path table would be as follows: Table 1: Path Table

[0039] By optimizing the paths between the various transformers to be tested, unnecessary detours and repetitions can be avoided, saving manpower, material resources and energy consumption. This not only reduces the overall travel distance and time, increasing the speed of completing the testing task, but also avoids high-risk areas, improving the safety of the testing operation.

[0040] The working logic of the recognition unit is as follows: After each movement of the detection device, the real-time position of the detection device is collected, and the real-time position is compared with the position information of the transformer. The transformer closest to the real-time position and its corresponding transformer number are identified, and the map distance between the real-time position and the closest transformer is obtained. The path table is queried to obtain the corresponding transformer number based on the real-time detection sequence, and then compared with the transformer number corresponding to the nearest transformer: If and only if two sets of transformers have the same number and the map distance between the real-time location and the nearest transformer does not exceed the preset distance error, the detection device is considered to have moved according to the path table and has not deviated from the optimal path, and the judgment result is normal. In all other cases, the judgment result is abnormal, and an alarm signal is issued through the alarm module.

[0041] Assuming a preset distance error of 10 meters, after the detection device completes the detection of transformer No. 2 at the starting point (detection sequence 1), the next transformer to be detected (detection sequence 2) is No. 5. Upon reaching the target location, the detection device acquires its real-time position, identifies the transformer closest to its real-time position and its corresponding transformer number, and obtains the map distance between its real-time position and the closest transformer. If the nearest transformer is No. 5, and the map distance between the detection device and No. 5 does not exceed 10 meters, then the detection device is considered to have reached the correct target location and moved along the optimal path according to the detection sequence in the path table. Conversely, if the nearest transformer is not No. 5, or the map distance between the detection device and No. 5 exceeds 10 meters, then the detection device is considered not to have moved along the optimal path, and an alarm signal is issued through the alarm module.

[0042] Understandably, the data collection points here are the time points after the detection device has moved, including the initial starting position and each time it reaches the transformer to be detected. Compared to constantly determining whether the detection device has deviated from the optimal path based on its real-time position, this node-based detection method can significantly reduce the amount of data processing, avoiding potential reliability risks associated with complex systems while meeting functional requirements.

[0043] The risk prediction unit is pre-installed with a Bayesian network model, and its working logic is as follows: The electrical parameters of the transformer are represented in vector form to generate an index vector, which is then used as input to the Bayesian network model. The index vector is represented as follows:

[0044] In the formula Represents an index vector. Indicates the first Class of electrical parameters, Indicates the category index of electrical parameters. This represents the total number of categories of electrical parameters. It's understandable that, because the specific values ​​of electrical parameters differ, the index vectors formed by these different values ​​are also different, i.e., index vectors... There are also multiple groups.

[0045] Based on the indicator vector, a Bayesian network model is used to infer the corresponding risk events and their posterior probabilities. Risk events include, but are not limited to, winding overheating and insulation failure. The set of risk events is represented as:

[0046] In the formula Represents a set of risk events. Indicates the first Risk events, An index representing risk events. This represents the total number of risk event categories, and its corresponding posterior probability is calculated as follows:

[0047] In the formula This represents the posterior probability, i.e., the probability of the input at the posterior position. When the group of index vectors occurs, the first occurrence The probability of risk events, For the first Prior probability of risk events, Represents conditional probability, that is: the first The first risk event observed The probability of the group index vector. The marginal probability of the index vector, used as a normalization factor, is calculated as follows:

[0048] The risk events and posterior probabilities corresponding to the indicator vectors are arranged in descending order of probability, and then set as the output of the Bayesian network model. The transformer is deemed to be at risk if and only if the posterior probability of the risk event ranked first exceeds a preset probability threshold.

[0049] The data used for optimizing the Bayesian network model during training were obtained from the historical fault data of each transformer.

[0050] For Bayesian network models, , These two sets of parameters are the core of posterior probability inference, obtained from the optimization training process. The methods for obtaining them are as follows: For prior probability In other words, risk events can be statistically analyzed from historical failure data. The frequency of occurrence across all data is used to determine the occurrence frequency of risk events within historical failure data. The ratio between the number of times a test is performed and the total number of tests.

[0051] For conditional probability In other words, if the amount of historical fault data is small, discretization modeling can be used. This involves comparing electrical parameters with preset intervals and dividing them into several intervals (e.g., low=0, medium=1, high=2). The corresponding index vector is then composed of discrete values ​​(0, 1, 2). This allows for the analysis of statistical risk events. The probability of the index vector taking each discrete value is obtained below. Assume there are 100 sets of historical fault data, and 25 risk events occurred. There were 4 instances of low current ( , in parentheses (Representing random discrete values), there were 7 instances of low voltage ( The resistance was too low 10 times. ), then risk events Under these conditions, the probability that the current, voltage, and resistance are all too low is: The corresponding conditional probability It is approximately equal to 0.018. If the amount of historical fault data is large, continuous modeling can be used. This involves assuming that the changes in electrical parameters follow a certain conditional distribution (such as a Gaussian distribution), directly substituting the corresponding probability distribution function, and using maximum likelihood estimation or kernel density estimation to train the conditional probability. Both methods can yield the conditional probability parameter; the choice depends on the amount of data.

[0052] After the risk prediction unit assesses the risk, it considers the inspection of that group of transformers complete. The inspection device then moves along the optimal path to inspect the next group of transformers in the path list. In short, the inspection process for each group of transformers includes three stages: "inspection device movement," "electrical parameter acquisition," and "risk assessment." Once the risk assessment is complete, the inspection process for that group of transformers is considered finished. At this point, the inspection sequence in the path list is incremented, indicating the start of a new inspection process for the next group of transformers. After inspecting all transformers in the order listed in the path list, the inspection device returns to its initial position along the same path to execute the instructions given in the next work order.

[0053] In this step, a Bayesian network model is used to assess the risks of each transformer. This not only enables timely risk analysis and early warning of the transformer's usage, but also identifies possible risk events, greatly improving the accuracy of risk warnings. This facilitates targeted maintenance and repairs in the future. Furthermore, by setting probability thresholds, misjudgments can be avoided, thereby further improving the overall reliability of the detection device.

[0054] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0055] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0056] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0057] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A modularly designed portable transformer multi-parameter testing device, characterized in that, Specifically, it includes: The data transmission module is used to construct a communication network to receive preset work order information and send it to the control module; Several sets of parameter detection modules are provided, which are used to detect the electrical parameters of the transformer and send them to the control module. The control module includes a path planning unit, an identification unit, and a risk prediction unit, wherein: The path planning unit has an ant colony algorithm and an electronic map built in, which is used to plan the movement path of the detection device according to the work order information to obtain the optimal path, and generate a path table based on the optimal path and the preset transformer number. The identification unit has a built-in GPS element, which is used to collect the real-time position of the detection device after each movement, and compare the number of the transformer closest to the real-time position with the transformer number in the path table to determine whether the detection device has deviated from the optimal path. The risk prediction unit has a built-in Bayesian network model, which is used to determine the risk of each group of transformers in turn based on the electrical parameters of the transformers. An alarm module is used to issue an alarm signal based on the judgment result of the identification unit and the risk judgment of the transformer by the risk prediction unit.

2. The portable transformer multi-parameter detection device with modular design according to claim 1, characterized in that: The work order information includes, but is not limited to, the location information of the transformer to be tested and the category of electrical parameters; The parameter detection module is electrically connected to the control module through a standardized interface, including but not limited to a current tester, voltage tester, and resistance tester; The electrical parameters include, but are not limited to, the transformer's current, voltage, and resistance.

3. The modularly designed portable transformer multi-parameter detection device according to claim 2, characterized in that: The working logic of the path planning unit is as follows: The transformers to be inspected are numbered, and the location information of the transformers to be inspected in the work order information is categorized into a set form; Based on the preset electronic map and location information, the road width at the location of the transformer and the map distance between different transformers are obtained; The cost function is constructed by taking road width and map distance as risk and path terms, respectively. The goal is to optimize the path by using the shortest path and lowest risk, and the optimal path is generated. The detection sequence of the transformer is determined based on the optimal path. The transformer numbers are sorted according to the detection sequence, and a path table is generated based on the detection sequence and the transformer numbers.

4. The portable transformer multi-parameter detection device with modular design according to claim 3, characterized in that: The path term in the cost function is the sum of the map distances between all transformers; The risk term in the cost function is the sum of the inverse of the road width at all transformer locations multiplied by a preset additional risk factor. The additional risk factor is used to reflect the road conditions at the transformer locations, and its value is inversely proportional to the quality of the road conditions.

5. The portable transformer multi-parameter detection device with modular design according to claim 3, characterized in that: The working logic of the identification unit is as follows: After each movement of the detection device, the real-time position of the detection device is collected, and the real-time position is compared with the position information of the transformer. The transformer closest to the real-time position and its corresponding transformer number are identified, and the map distance between the real-time position and the closest transformer is obtained. The path table is queried to obtain the corresponding transformer number based on the real-time detection sequence, and then compared with the transformer number corresponding to the nearest transformer: If and only if two sets of transformers have the same number and the map distance between the real-time location and the nearest transformer does not exceed the preset distance error, the detection device is considered to have moved according to the path table and has not deviated from the optimal path, and the judgment result is normal. In all other cases, the judgment result is abnormal, and an alarm signal is issued through the alarm module.

6. The portable transformer multi-parameter detection device with modular design according to claim 5, characterized in that: The working logic of the risk prediction unit is as follows: The electrical parameters of the transformer are represented in vector form to generate an index vector, which is then used as the input to the Bayesian network model. Based on the indicator vector, the corresponding risk event and its posterior probability are inferred using a Bayesian network model. The risk events and posterior probabilities corresponding to the indicator vectors are arranged in descending order of probability, and then set as the output of the Bayesian network model. The transformer is deemed to be at risk if and only if the posterior probability of the risk event ranked first exceeds a preset probability threshold.

7. The portable transformer multi-parameter detection device with modular design according to claim 6, characterized in that: The data used for optimizing the Bayesian network model during training is obtained from the historical fault data of each transformer.

8. The portable transformer multi-parameter detection device with modular design according to claim 6, characterized in that: After the risk prediction unit determines the risk, it considers the group of transformers to have been inspected. The inspection device then moves along the optimal path to inspect the next group of transformers in the path table. After the inspection device has inspected all the transformers in sequence, it returns to the initial position along the original path.

Images