Transformer oil level abnormity identification method and device, storage medium and equipment

Abstract

The invention discloses a transformer oil level abnormity recognition method and device, a storage medium and computer equipment, and the method comprises the steps: obtaining an oil level gauge image of an oil level gauge on an oil conservator of a target transformer, and carrying out the image recognition processing of the oil level gauge image, and obtaining an actual oil level indicated by the oil level gauge; obtaining target data corresponding to the target transformer, and processing the target data through a target oil temperature prediction model to obtain a predicted oil temperature corresponding to the target transformer, the target data comprises the current load current, the current environment temperature, the accumulated operation duration and the historical average load current of the target transformer; determining an ideal oil level corresponding to the predicted oil temperature based on a preset fitting relationship between the oil level and the oil temperature; and an oil level difference value between the actual oil level and the ideal oil level is calculated, and when the oil level difference value is larger than a preset difference value threshold value, it is determined that the actual oil level of the target transformer is abnormal.

Description

Technical Field

[0001] The present application relates to the technical field of power system equipment fault diagnosis, and in particular to a method and device for identifying abnormal oil level of a transformer, a storage medium, and a computer device. Background Art

[0002] With the continuous development of mankind, the current power system has become an indispensable part of human life. As an important part of the power system, the insulating oil inside the transformer is one of the key factors to ensure the normal operation of the transformer. If the oil level of the insulating oil in the transformer is unreasonable, it is likely to cause unplanned shutdown of the transformer, thus affecting power supply. Therefore, it is extremely important to monitor the oil level of the insulating oil in the transformer.

[0003] Most current transformer oil level gauges do not have the remote transmission function of smart instruments due to cost and historical reasons, and the oil level can only be monitored by manually checking the transformer oil level gauge on site. However, manual on-site checking of the oil level gauge is inefficient, and manual checking is often done periodically, which cannot detect abnormal oil levels in a timely manner. Summary of the invention

[0004] In view of this, the present application provides a method and device for identifying abnormal oil level of a transformer, a storage medium, and a computer device. The predicted oil temperature is obtained through the target data of the target transformer, and then the ideal oil level is obtained according to the predicted oil temperature. The determination of the ideal oil level is based on the target data, which fully considers the specificity of each transformer and improves the accuracy of the ideal oil level calculation. In addition, the actual oil level is obtained by collecting the oil level gauge image, and the actual oil level and the ideal oil level are compared to determine whether the target transformer oil level is abnormal, which can realize the online monitoring of the transformer oil level, improve the automation and timeliness of transformer oil level monitoring, and improve the operation and maintenance efficiency.

[0005] According to one aspect of the present application, a method for identifying abnormal oil level in a transformer is provided, comprising:

[0006] Acquire an oil level gauge image of an oil level gauge on an oil conservator of a target transformer, and perform image recognition processing on the oil level gauge image to obtain an actual oil level indicated by the oil level gauge;

[0007] Obtain target data corresponding to the target transformer, and process the target data through a target oil temperature prediction model to obtain a predicted oil temperature corresponding to the target transformer, wherein the target data includes a current load current, a current ambient temperature, a cumulative operating time, and a historical average load current of the target transformer;

[0008] Determining an ideal oil level corresponding to the predicted oil temperature based on a preset fitting relationship between the oil level and the oil temperature;

[0009] An oil level difference between the actual oil level and the ideal oil level is calculated, and when the oil level difference is greater than a preset difference threshold, it is determined that an actual oil level of the target transformer is abnormal.

[0010] According to another aspect of the present application, a device for identifying abnormal oil level of a transformer is provided, comprising:

[0011] An actual oil level determination module is used to obtain an oil level gauge image of an oil level gauge on an oil conservator of a target transformer, and perform image recognition processing on the oil level gauge image to obtain an actual oil level indicated by the oil level gauge;

[0012] An oil temperature prediction module is used to obtain target data corresponding to the target transformer, and process the target data through a target oil temperature prediction model to obtain a predicted oil temperature corresponding to the target transformer, wherein the target data includes a current load current, a current ambient temperature, a cumulative operating time, and a historical average load current of the target transformer;

[0013] An ideal oil level determination module, used to determine the ideal oil level corresponding to the predicted oil temperature based on a preset fitting relationship between the oil level and the oil temperature;

[0014] The oil level abnormality identification module is used to calculate the oil level difference between the actual oil level and the ideal oil level, and when the oil level difference is greater than a preset difference threshold, determine that the actual oil level of the target transformer is abnormal.

[0015] According to another aspect of the present application, a storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned method for identifying abnormal oil level of a transformer is implemented.

[0016] According to another aspect of the present application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the above-mentioned method for identifying abnormal oil level of a transformer when executing the program.

[0017] By means of the above technical scheme, the present application provides a method and device for identifying abnormal oil level of a transformer, a storage medium, and a computer device. First, the oil level gauge image of the transformer can be obtained, and the oil level gauge image can be subjected to image recognition processing to determine the actual oil level indicated by the oil level gauge. The target data of the target transformer when the oil level gauge image is taken can also be obtained. Here, the target data can be the current load current, current ambient temperature, cumulative operating time, historical average load current, etc. of the target transformer. After that, these data can be input into the target oil temperature prediction model, and the target oil temperature prediction model can process these data to output the predicted oil temperature corresponding to the target transformer. After the predicted oil temperature is determined, the ideal oil level corresponding to the predicted oil temperature can be determined according to the preset fitting relationship between the predicted oil temperature, the oil level, and the oil temperature. After obtaining the actual oil level and the ideal oil level, the oil level difference between the actual oil level and the ideal oil level can be calculated, and then the oil level difference can be compared with the preset difference threshold. If the comparison finds that the oil level difference is less than or equal to the preset difference threshold, then it means that the actual oil level of the target transformer is normal, otherwise, it means that the actual oil level of the target transformer is abnormal. The embodiment of the present application obtains the predicted oil temperature through the target data of the target transformer, and then obtains the ideal oil level according to the predicted oil temperature. The determination of the ideal oil level is based on the target data, which fully considers the specificity of each transformer and improves the accuracy of the ideal oil level calculation. In addition, the actual oil level is obtained by collecting the oil level gauge image, and the actual oil level and the ideal oil level are compared to determine whether the target transformer oil level is abnormal, which can realize the online monitoring of the transformer oil level, improve the automation and timeliness of transformer oil level monitoring, and improve the operation and maintenance efficiency.

[0018] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0020] Figure 1 A schematic diagram of a flow chart of a method for identifying abnormal oil level in a transformer provided in an embodiment of the present application is shown;

[0021] Figure 2 A schematic flow chart of another method for identifying abnormal transformer oil level provided in an embodiment of the present application is shown;

[0022] Figure 3A schematic diagram of the structure of an initial oil temperature prediction model provided in an embodiment of the present application is shown;

[0023] Figure 4 A schematic flow chart of a method for optimizing a first weight parameter and a threshold parameter provided in an embodiment of the present application is shown;

[0024] Figure 5 A schematic diagram of the structure of a transformer oil level abnormality identification device provided in an embodiment of the present application is shown;

[0025] Figure 6 A schematic diagram of the device structure of a computer device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0026] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0027] In this embodiment, a method for identifying abnormal oil level in a transformer is provided. Figure 1 As shown, the method includes:

[0028] Step 101, obtaining an oil level gauge image of an oil level gauge on an oil conservator of a target transformer, and performing image recognition processing on the oil level gauge image to obtain an actual oil level indicated by the oil level gauge.

[0029] The embodiment of the present application provides a method for identifying abnormal transformer oil level, which can realize online real-time monitoring of transformer oil level, so that abnormal transformer oil level can be discovered in time. First, an oil level gauge image of the oil level gauge on the transformer oil storage cabinet can be obtained, and the oil level gauge image also includes an image of the oil level gauge. Here, the oil level gauge image can be obtained by shooting with a shooting device at a preset frequency, such as shooting every minute; it can also be obtained by real-time recording of a video with a shooting device and processing the video, such as intercepting a video frame from the recorded video at a preset time interval as an oil level gauge image. After obtaining the oil level gauge image, the oil level gauge image can be subjected to image recognition processing to determine the actual oil level indicated by the oil level gauge.

[0030] Step 102, obtaining target data corresponding to the target transformer, and processing the target data through a target oil temperature prediction model to obtain a predicted oil temperature corresponding to the target transformer, wherein the target data includes the current load current, current ambient temperature, accumulated operating time and historical average load current of the target transformer.

[0031] In this embodiment, the target data of the target transformer when the oil level gauge image is taken can also be obtained. Here, the target data can be the current load current, current ambient temperature, cumulative operation time, historical average load current, etc. of the target transformer. Specifically, these target data can be obtained from the background machine corresponding to the target transformer. For example, the current load current can be transmitted to the background machine through the current transformer detection of the target transformer, and then the current load current can be obtained from the background machine. The current ambient temperature can be acquired by the temperature sensor, the cumulative operation time can be determined according to the data recorded by the timer, and the historical average load current can be dynamically updated. The acquisition method of the above data is not limited here. Afterwards, these data can be input into the target oil temperature prediction model, and the target oil temperature prediction model can process these data to output the predicted oil temperature corresponding to the target transformer. In fact, the oil temperature predicted by the target oil temperature prediction model is the ideal oil temperature corresponding to the target data. That is, under normal conditions, under the environment indicated by the target data, the oil temperature in the target transformer is normal only when this oil temperature is normal.

[0032] Step 103: determining an ideal oil level corresponding to the predicted oil temperature based on a preset fitting relationship between the oil level and the oil temperature.

[0033] In this embodiment, there may be a preset fitting relationship between the oil level and the oil temperature, and the preset fitting relationship may be a fitting curve, a fitting formula, etc. After the predicted oil temperature is determined, the ideal oil level corresponding to the predicted oil temperature may be determined according to the preset fitting relationship between the predicted oil temperature, the oil level, and the oil temperature. It should be noted that the preset fitting relationship between the oil level and the oil temperature is an existing one, which is set for each transformer when it leaves the factory.

[0034] Step 104, calculating the oil level difference between the actual oil level and the ideal oil level, and when the oil level difference is greater than a preset difference threshold, determining that the actual oil level of the target transformer is abnormal.

[0035] In this embodiment, after obtaining the actual oil level and the ideal oil level, the oil level difference between the actual oil level and the ideal oil level can be calculated, and then the oil level difference is compared with a preset difference threshold. If the comparison finds that the oil level difference is less than or equal to the preset difference threshold, then it means that the actual oil level of the target transformer is normal, otherwise, it means that the actual oil level of the target transformer is abnormal.

[0036] By applying the technical solution of this embodiment, first, the oil level gauge image of the transformer can be obtained, and the oil level gauge image can be processed by image recognition to determine the actual oil level indicated by the oil level gauge. The target data of the target transformer when the oil level gauge image is taken can also be obtained. Here, the target data can be the current load current, current ambient temperature, cumulative operating time, historical average load current, etc. of the target transformer. After that, these data can be input into the target oil temperature prediction model, and the target oil temperature prediction model can process these data to output the predicted oil temperature corresponding to the target transformer. After the predicted oil temperature is determined, the ideal oil level corresponding to the predicted oil temperature can be determined according to the preset fitting relationship between the predicted oil temperature, the oil level and the oil temperature. After obtaining the actual oil level and the ideal oil level, the oil level difference between the actual oil level and the ideal oil level can be calculated, and then the oil level difference can be compared with the preset difference threshold. If the comparison finds that the oil level difference is less than or equal to the preset difference threshold, then it means that the actual oil level of the target transformer is normal, otherwise, it means that the actual oil level of the target transformer is abnormal. The embodiment of the present application obtains the predicted oil temperature through the target data of the target transformer, and then obtains the ideal oil level according to the predicted oil temperature. The determination of the ideal oil level is based on the target data, which fully considers the specificity of each transformer and improves the accuracy of the ideal oil level calculation. In addition, the actual oil level is obtained by collecting the oil level gauge image, and the actual oil level and the ideal oil level are compared to determine whether the target transformer oil level is abnormal, which can realize the online monitoring of the transformer oil level, improve the automation and timeliness of transformer oil level monitoring, and improve the operation and maintenance efficiency.

[0037] Further, as a refinement and extension of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, another method for identifying abnormal transformer oil level is provided, such as Figure 2 As shown, the method includes:

[0038] Step 201, obtaining an oil level gauge image of an oil level gauge on an oil conservator of a target transformer.

[0039] Step 202, obtaining an oil level gauge template corresponding to the target transformer, wherein the oil level gauge template includes a plurality of first scale areas and a first pointer area corresponding to the oil level gauge; performing image recognition processing on the oil level gauge image, identifying a plurality of second scale areas and a second pointer area from the oil level gauge image; aligning the plurality of second scale areas with the plurality of first scale areas in the oil level gauge template, calculating the pointer spacing between a second pointer corresponding to the second pointer area and a first pointer corresponding to the first pointer area after the alignment processing, and determining the actual oil level corresponding to the second pointer according to the pointer spacing and a preset oil level corresponding to the first pointer.

[0040] In this embodiment, for the oil level gauge of the target transformer, an oil level gauge template consistent with the oil level gauge can be set first, and the oil level gauge template can include multiple first scale areas and a first pointer area, wherein each first scale area is used to indicate a scale of the oil level gauge, and the first pointer area is used to indicate the pointer of the oil level gauge. The number of first scale areas on the oil level gauge template corresponds to the number of scales in the oil level gauge. It should be noted that the oil level gauge corresponding to the oil level gauge template is completely consistent with the oil level gauge set on the target transformer.

[0041] Image recognition is performed on the oil level gauge image, and multiple second scale areas and one second pointer area are identified therefrom. Among them, the number of the second scale areas is the same as the number of the first scale areas. After determining the multiple second scale areas and one second pointer area in the oil level gauge image, the multiple first scale areas on the oil level gauge template can be matched one by one with the multiple second scale areas. For example, if the scale of the oil level gauge is in the up and down direction, then the top first scale area can be matched with the top second scale area. Under normal circumstances, the subsequent first scale area can be matched one by one with the subsequent second scale area, and the alignment of the two scale areas is completed. In order to ensure that there is no deviation during alignment, the top first scale area can be matched with the top second scale area, and the bottom first scale area can be matched with the bottom second scale area, and the complete alignment of the two scales is completed. After the alignment process, the pointer spacing between the first pointer corresponding to the first pointer area and the second pointer corresponding to the second pointer area can be determined.

[0042] For example, for a pointer-type oil level gauge, the pointer spacing can be the pointer angle between the first pointer and the second pointer. Based on the pointer angle and the relative position of the second pointer to the first pointer, the difference in the current oil level readings between the first pointer and the second pointer can be calculated. The first pointer can correspond to a preset oil level reading. Then, based on the preset oil level reading corresponding to the first pointer and the above oil level reading difference, the actual oil level indicated by the current second pointer can be calculated.

[0043] For another example, for a liquid level type oil level gauge, there is no real pointer on it, so the first pointer area, the first pointer, the second pointer area and the second pointer can be virtual, and the first scale area and the second scale area are real. The first pointer area can be the area indicated by the highest position of the liquid level on the oil level gauge template, and the preset oil level indicated by the first pointer is also the highest position of the liquid level. Then, the liquid level of the oil level gauge in the oil level gauge image can be identified in the same way, and the virtual second pointer area can be determined by the highest position of the liquid level. After determining the second pointer area, the pointer spacing (that is, the distance in the vertical direction) between the two virtual pointers and the relative position between the two virtual pointers can be calculated according to the highest position of the liquid level indicated by the first pointer area and the highest position of the liquid level indicated by the second pointer area. Since the oil level difference between two adjacent scales of the oil level gauge is known, the oil level reading difference between the current virtual first pointer and the virtual second pointer can be calculated according to the pointer spacing, and the virtual first pointer corresponds to a preset oil level reading, then the current actual oil level can be calculated according to the preset oil level reading and the oil level reading difference.

[0044] Step 203, obtaining target data corresponding to the target transformer, and processing the target data through a target oil temperature prediction model to obtain a predicted oil temperature corresponding to the target transformer, wherein the target data includes the current load current, current ambient temperature, accumulated operating time and historical average load current of the target transformer.

[0045] Step 204 , determining an ideal oil level corresponding to the predicted oil temperature based on a preset fitting relationship between the oil level and the oil temperature.

[0046] Step 205, calculating the oil level difference between the actual oil level and the ideal oil level, and when the oil level difference is greater than a preset difference threshold, determining that the actual oil level of the target transformer is abnormal.

[0047] Step 206, obtaining the acquisition time of the oil level gauge image and the transformer identifier corresponding to the target transformer, generating an alarm message according to the acquisition time, the transformer identifier, the actual oil level, the ideal oil level and the oil level difference, and sending the alarm message to a preset management terminal.

[0048] In this embodiment, after determining that the target transformer does have an abnormal oil level according to the oil level gauge image, the acquisition time of the oil level gauge image and the transformer identifier corresponding to the target transformer can be obtained. Here, the transformer identifier can be a unique identifier for each transformer. According to the acquisition time of the oil level gauge image, the transformer identifier, the actual oil level corresponding to the target transformer, the ideal oil level under the target data, and the oil level difference calculated above, an alarm message is jointly generated, and the alarm message can be sent to a preset management terminal, so that the staff corresponding to the preset management terminal can understand the occurrence time of the oil level abnormality through the acquisition time in the alarm message after seeing the alarm message; understand which transformer has an abnormal oil level through the transformer identifier in the alarm message; and understand whether the oil level abnormality of the target transformer is serious through the actual oil level, ideal oil level and oil level difference in the alarm message. By generating the above-mentioned alarm message and automatically sending the alarm message to the preset terminal, the embodiment of the present application can timely remind the staff to understand the abnormal oil level of the transformer, and deal with the abnormal oil level as soon as possible, thereby reducing the unplanned downtime of the target transformer and ensuring that power supply can be quickly restored.

[0049] In an embodiment of the present application, optionally, before the step 203 of "processing the target data through the target oil temperature prediction model to obtain the predicted oil temperature corresponding to the target transformer", the method also includes: constructing an initial oil temperature prediction model, wherein the initial oil temperature prediction model includes an input layer, at least one hidden layer, an output layer, a first weight parameter between the input layer and the first hidden layer and between multiple hidden layers, a threshold parameter of each hidden layer, and a second weight parameter between the output layer and the last hidden layer; optimizing the first weight parameter and the threshold parameter, and configuring the initial oil temperature prediction model according to the optimized first weight parameter and threshold parameter to obtain an intermediate oil temperature prediction model; training the intermediate oil temperature prediction model according to the first training sample set to optimize the second weight parameter to obtain the target oil temperature prediction model, wherein each sample in the first training sample set contains the same data type as the target data and carries an oil temperature label.

[0050] In this embodiment, an initial oil temperature prediction model can be constructed first. Here, the initial oil temperature prediction model can include an input layer, one or more hidden layers, and an output layer. A first weight parameter is set between each node of the input layer and each node of the first hidden layer, a first weight parameter is set between each node of every two adjacent hidden layers, a threshold parameter is set for each node of each hidden layer, and a second weight parameter is set between each node of the last hidden layer and the output layer node. When constructing the initial oil temperature prediction model, initial values ​​can be assigned to each first weight parameter, threshold parameter, and second weight parameter. For example, Figure 3 As shown, an initial oil temperature prediction model of an embodiment of the present application is shown. The initial oil temperature prediction model only contains one hidden layer. The first layer in the figure is the input layer, the second layer is the hidden layer, and the last layer is the output layer. The input layer has m nodes, the hidden layer has L nodes, and the output layer has one node. w is the first weight parameter, b is the threshold parameter, β is the second weight parameter, w 11 is the first weight parameter between the first node of the first layer and the first node of the second layer, b1 is the threshold parameter corresponding to the first node of the second layer, and β1 is the second weight parameter from the first node of the second layer to the only node of the last layer. Similarly, w Lm is the first weight parameter between the mth node in the first layer and the Lth node in the second layer, b L is the threshold parameter corresponding to the Lth node in the second layer.

[0051] Afterwards, the second weight parameter between the last hidden layer and the output layer can be fixed, and the first weight parameter and the threshold parameter can be optimized first. When the optimization is completed, the initial value of the corresponding first weight parameter can be replaced by the optimized first weight parameter, and the initial value of the corresponding threshold parameter can be replaced by the optimized threshold parameter, so that the initial oil temperature prediction model can be reconfigured to obtain the intermediate oil temperature prediction model.

[0052] After obtaining the intermediate oil temperature prediction model, the first weight parameter and the threshold parameter are fixed, and then the second weight parameter is optimized. Specifically, the first weight parameter and the threshold parameter can be kept unchanged, the intermediate oil temperature prediction model can be trained using the first training sample set, and the second weight parameter can be continuously adjusted. When the model loss value of the intermediate oil temperature prediction model is less than the preset loss threshold, the intermediate oil temperature prediction model at this time can be used as the final target oil temperature prediction model. The embodiment of the present application optimizes different model parameters in steps, which can make the model gradually adapt to the data during the training process, which is conducive to improving the stability of model training; in the initial stage, the first weight parameter between the input layer and the hidden layer, the hidden layer and the hidden layer, and the threshold parameter of the hidden layer are optimized, so that the model can better extract useful information from the input data, thereby reducing the dependence on specific training samples. On this basis, the second weight parameter between the output layer and the last hidden layer is optimized, so that the model can better fit the training data while maintaining the generalization ability. Such an optimization order helps to reduce the amount of calculation in the training process, improve the efficiency of parameter optimization, and at the same time help to improve the prediction accuracy and generalization ability of the model.

[0053] In an embodiment of the present application, optionally, the "optimizing the first weight parameter and the threshold parameter" includes: constructing a vector set according to the first weight parameter and the threshold parameter, wherein the element corresponding to each vector in the vector set is used to indicate a set of values ​​of the first weight parameter and the threshold parameter; using the second training sample set and the initial oil temperature prediction model to obtain the first oil temperature prediction value corresponding to each training sample, and calculating the first target score according to the first oil temperature prediction value corresponding to each training sample and the actual value of the oil temperature indicated by the oil temperature label through a preset scoring calculation formula; configuring the initial oil temperature prediction model according to the elements corresponding to any vector, using the second training sample set and the configured initial oil temperature prediction model to obtain the second oil temperature prediction value corresponding to each training sample, and calculating the first target score according to the first oil temperature prediction value corresponding to each training sample and the actual value of the oil temperature indicated by the oil temperature label through the preset scoring calculation formula. The second target score corresponding to any one of the vectors is calculated based on the two oil temperature prediction values ​​and the actual value of the oil temperature; the optimal score is determined based on the first target score and the second target score corresponding to each vector, and the vector update formula corresponding to each vector in the vector set is determined based on the second target score corresponding to each vector, and the element corresponding to each vector is updated based on the vector update formula; the initial oil temperature prediction model is reconfigured based on the updated vector, and the second target score of each updated vector is recalculated, and the optimal score is updated based on the recalculated second target score; when the number of updates corresponding to each vector is less than or equal to the preset number threshold, return to the step of determining the vector update formula corresponding to each vector in the vector set based on the second target score corresponding to each vector, otherwise, the element of the vector corresponding to the updated optimal score is used as the optimized first weight parameter and threshold parameter.

[0054] In this embodiment, if Figure 4 As shown, when optimizing the first weight parameter and the threshold parameter, a vector set can be constructed first. The elements of each vector in the vector set are determined according to the first weight parameter and the threshold parameter. For example, the first weight parameter includes 9 and the threshold parameter includes 3, then the number of elements in each vector can be 12, and each element corresponds to a value of the first weight parameter, or corresponds to a value of the threshold parameter. The number of vectors in the vector set can be pre-set, for example, the number of vectors in the vector set is set to 50, then 50 vectors can be randomly generated according to the first weight parameter and the threshold parameter, the dimension of each vector is 12, and the value of the element in each vector is randomly generated.

[0055] After constructing the initial oil temperature prediction model, the first weight parameter and the threshold parameter in the initial oil temperature prediction model are set with initial values. Each training sample in the second training sample set can be input into the initial oil temperature prediction model to obtain the first oil temperature prediction value corresponding to each training sample, and the first target score is calculated by a preset score calculation formula according to the actual value of the oil temperature corresponding to each training sample. Here, the preset score calculation formula can be a formula for evaluating the difference between the oil temperature prediction value and the actual value of the oil temperature. The second training sample set can be the same sample set as the first training sample set, or it can be a different sample set, which is not required here.

[0056] For each vector in the vector set, the following operation can be performed: the elements contained in any vector are used as the values ​​of the first weight parameter and the threshold parameter to reconfigure the initial oil temperature prediction model. Specifically, the initial values ​​of the first weight parameter and the threshold parameter in the initial oil temperature prediction model can be replaced with the numerical values ​​corresponding to each element in the vector to obtain a new initial oil temperature prediction model. Each training sample in the second training sample set is input into the new initial oil temperature prediction model, and the second oil temperature prediction value corresponding to the training sample is output. Using the second oil temperature prediction value and the corresponding true oil temperature value, the second target score corresponding to the vector can be calculated through the above target score calculation formula.

[0057] In this way, after each vector in the vector set has performed the above operation, each vector can obtain a corresponding second target score. Put the above first target score and multiple second target scores together, find the optimal score, that is, find a set of first weight parameters and threshold parameters with the best oil temperature prediction effect, and record the optimal score and the corresponding set of first weight parameters and threshold parameters. Specifically, the better the oil temperature prediction effect corresponding to the vector, the higher the corresponding second target score can be. The second target scores can be sorted, and the first vector with the second target score ranking in the top A can be found, and B second vectors can be randomly determined, and the remaining vectors can be used as the third vector. Specifically, A and B can be pre-set values. Each vector corresponds to a different vector update formula, which is used to update the elements in the vector, that is, to update a set of values ​​of the first weight parameter and the threshold parameter.

[0058] Among them, the vector update formula of the first vector can be expressed as:

[0059]

[0060] Where t represents the number of updates corresponding to any current vector; N represents the preset number threshold; Ψ m,nrepresents the value corresponding to the nth element of the mth vector; α∈(0,1] is a random number; U and V are random numbers, U∈[0.5,1] and V∈[0,1] respectively; W is a random number that follows a normal distribution; X represents a 1×d matrix, d represents the number of elements in each vector, and all elements in the matrix are 1.

[0061] The vector update formula of the third vector can be expressed as:

[0062]

[0063] Among them, o represents the first vector with the highest current second target score; Ψ′ represents the vector with the lowest current second target score in the vector set; D is a 1×d matrix, in which each element is randomly assigned a value of 1 or -1, and D′=D T (DD T ) -1 ; Z represents the number of vectors contained in the vector set.

[0064] The vector update formula of the second vector can be expressed as:

[0065]

[0066] in, represents a set of values ​​of the first weight parameter and threshold parameter corresponding to the current optimal score; λ1 is the step size control parameter, which is a random number that obeys the N(0,1) distribution; λ2∈[-1,1] is a random number and also a step size control parameter; F i represents the second target score of the current vector; F g and F w Respectively represent a set of values ​​of the first weight parameter and the threshold parameter corresponding to the current best score, and a set of values ​​of the first weight parameter and the threshold parameter corresponding to the minimum score; is the smallest constant that avoids zero in the denominator.

[0067] After each vector determines its corresponding vector update formula, the vector update formula can be used to update the values ​​of the elements inside the vector, thereby obtaining the updated vector. After that, the initial oil temperature prediction model is reconfigured using each updated vector, and the second target score corresponding to each new vector is recalculated after configuration, and these newly obtained second target scores are compared with the optimal score of the previous round. If any of the newly obtained second target scores is higher than the optimal score, then the optimal score of the previous round is updated using the highest second target score, otherwise, the original optimal score remains unchanged. In addition, after the update, the number of updates of each current vector can also be determined. It should be noted that at a certain moment, the number of updates of each vector is the same. If the number of updates of any vector is less than or equal to the preset number threshold, it means that the vector update operation has not yet ended, and the vector update formula corresponding to each vector can continue to be determined according to the new second target score, and a new round of updates can be performed, and the above operations are continued to be repeated; if the number of updates of any vector is greater than the preset number threshold, it means that the vector update operation is over, and the value of the element in the vector corresponding to the updated optimal score can be used as the first weight parameter and threshold parameter after optimization. During the whole process, the vector update formula corresponding to each vector is dynamically determined according to its calculated second target score.

[0068] The embodiment of the present application can effectively optimize the first weight parameter and threshold parameter of the initial oil temperature prediction model by constructing a vector set, using a scoring mechanism to evaluate vector performance, and solving the optimal score at each step to avoid local optimality, thereby improving the prediction performance of the model and enhancing its reliability and accuracy in practical applications.

[0069] In an embodiment of the present application, optionally, the step 202 of "performing image recognition processing on the oil level gauge image, and identifying multiple second scale areas and a second pointer area from the oil level gauge image" includes: graying the oil level gauge image to obtain a target grayscale image, and calculating the first gradient value in the horizontal direction and the second gradient value in the vertical direction corresponding to each pixel point based on the target grayscale image, and calculating the gradient amplitude corresponding to each pixel point based on the first gradient value and the second gradient value; comparing the gradient amplitude corresponding to each pixel point with a preset amplitude threshold, and determining the target edge point from the pixel point corresponding to the target grayscale image based on the comparison result; connecting the target edge points to obtain multiple target areas corresponding to the oil level gauge image, and identifying the second pointer area and multiple second scale areas from the multiple target areas based on a morphological analysis method.

[0070] In this embodiment, graying is a process of converting a color image into a gray image. After acquiring the oil level gauge image, firstly, the oil level gauge image can be grayed to obtain a target gray image.

[0071] Gradient is a quantity that describes the direction and intensity of brightness changes in an image, and gradient amplitude is the size of the gradient, which describes the degree of brightness change at the pixel point. In the oil level gauge image, edges (such as the boundaries of the pointer and the scale) often correspond to areas with large brightness changes. Therefore, for each pixel in the target grayscale image, its first gradient value in the horizontal direction (left and right) and the second gradient value in the vertical direction (up and down) can be calculated. These two gradient values ​​represent the intensity of brightness change of the pixel point in two directions respectively. By combining the gradient values ​​in the horizontal and vertical directions, the gradient amplitude corresponding to each pixel point can be calculated.

[0072] Edge points are points in an image with large changes in brightness, usually corresponding to the boundaries of objects. Here, we want to identify the edges of the pointer and scale of the oil level gauge. Therefore, the gradient amplitude corresponding to each pixel can be compared with a preset amplitude threshold. If the gradient amplitude exceeds this threshold, the pixel is considered to be a target edge point. By connecting the identified target edge points, multiple continuous edge areas in the oil level gauge image can be obtained, which are called target areas. These areas correspond to the outlines of the pointer and scale. It should be noted that the preset amplitude threshold can be set based on experience.

[0073] After obtaining multiple target areas, these target areas are further analyzed to identify the second pointer area (i.e., the target area where the oil level gauge pointer is located) and multiple second scale areas (i.e., the target area where the oil level gauge scale is located). Specifically, this can be achieved through a morphological analysis method, such as identifying the pointer and scale by finding an area with specific shape and size characteristics, and then the second pointer area and multiple second scale areas can be extracted from the oil level gauge image. The embodiment of the present application can quickly identify the second pointer area and the second scale area from the oil level gauge image, which is beneficial to improving the efficiency of determining the actual oil level while having high accuracy.

[0074] In an embodiment of the present application, optionally, after step 201, the method further includes: dividing the oil level gauge image into regions, and determining a target sub-region from the divided sub-regions based on a reference image corresponding to the oil level gauge, wherein the reference image is an image marked with a region of interest after being captured by a shooting device at the same shooting position; performing feature extraction on the target sub-region to obtain target local features corresponding to each of the target sub-regions, and calculating enhancement parameters corresponding to each of the target sub-regions based on the target local features, wherein the local features include at least one of brightness mean, brightness variance, contrast, and edge strength; performing enhancement processing on each of the target sub-regions based on the enhancement parameters to obtain an enhanced oil level gauge image.

[0075] In this embodiment, before using the oil level gauge image to identify the actual oil level of the target transformer, the oil level gauge image can be enhanced so that the quality of the oil level gauge image can be further improved, thereby improving the accuracy of subsequent identification. First, the oil level gauge image is divided into multiple smaller sub-regions. Each part in the oil level gauge image may have different brightness, contrast, etc. By dividing the image into regions, each sub-region can be more conveniently processed separately. After obtaining the sub-regions, the current oil level gauge image can be compared with the reference image, and which sub-regions contain relevant information of the oil level gauge from multiple sub-regions can be determined, and these sub-regions are determined as target sub-regions. Here, the reference image is a pre-shot and processed image, which contains a marked region of interest (i.e., the key part of the oil level gauge), and its shooting angle is consistent with the shooting angle corresponding to the oil level gauge image. By comparing the oil level gauge image and the reference image, the region of interest can be quickly determined from the oil level gauge image. Feature extraction is a key step in image processing, which involves extracting useful information from an image. For the target sub-region, the local features extracted therefrom may include brightness mean (average brightness of the region), brightness variance (degree of change in regional brightness), contrast (degree of change in light and dark in the region), and edge strength (clarity of the edge in the region), etc. Based on the extracted target local features, the enhancement parameters required for each target sub-region can be calculated. These parameters may include brightness adjustment amount, contrast adjustment amount or adjustment values ​​of other image processing parameters. Finally, the calculated enhancement parameters can be used to perform corresponding enhancement processing on the target sub-region. For example, operations such as increasing brightness, increasing contrast, and sharpening edges are performed on each target sub-region. Through enhancement processing, the oil level gauge image can be made clearer and easier to identify. The embodiment of the present application divides the oil level gauge image into regions to obtain multiple sub-regions, finds the target sub-region from the sub-regions, and performs enhancement processing on each target sub-region respectively, and finally obtains the enhanced oil level gauge image, which can greatly improve the quality of the oil level gauge image and is conducive to improving the efficiency and accuracy of subsequent oil level recognition.

[0076] Further, as Figure 1 The specific implementation of the method, the embodiment of the present application provides a device for identifying abnormal oil level of a transformer, such as Figure 5 As shown, the device comprises:

[0077] An actual oil level determination module is used to obtain an oil level gauge image of an oil level gauge on an oil conservator of a target transformer, and perform image recognition processing on the oil level gauge image to obtain an actual oil level indicated by the oil level gauge;

[0078] An oil temperature prediction module is used to obtain target data corresponding to the target transformer, and process the target data through a target oil temperature prediction model to obtain a predicted oil temperature corresponding to the target transformer, wherein the target data includes a current load current, a current ambient temperature, a cumulative operating time, and a historical average load current of the target transformer;

[0079] An ideal oil level determination module, used to determine the ideal oil level corresponding to the predicted oil temperature based on a preset fitting relationship between the oil level and the oil temperature;

[0080] The oil level abnormality identification module is used to calculate the oil level difference between the actual oil level and the ideal oil level, and when the oil level difference is greater than a preset difference threshold, determine that the actual oil level of the target transformer is abnormal.

[0081] Optionally, the device further comprises:

[0082] A model building module, used for constructing an initial oil temperature prediction model before processing the target data through the target oil temperature prediction model to obtain the predicted oil temperature corresponding to the target transformer, wherein the initial oil temperature prediction model includes an input layer, at least one hidden layer, an output layer, a first weight parameter between the input layer and the first hidden layer and between multiple hidden layers, a threshold parameter of each hidden layer, and a second weight parameter between the output layer and the last hidden layer;

[0083] A parameter optimization module is used to optimize the first weight parameter and the threshold parameter, and configure the initial oil temperature prediction model according to the optimized first weight parameter and threshold parameter to obtain an intermediate oil temperature prediction model; train the intermediate oil temperature prediction model according to the first training sample set to optimize the second weight parameter to obtain the target oil temperature prediction model, wherein each sample in the first training sample set contains the same data type as the target data and carries an oil temperature label.

[0084] Optionally, the parameter optimization module is used to:

[0085] Constructing a vector set according to the first weight parameter and the threshold parameter, wherein an element corresponding to each vector in the vector set is used to indicate a set of values ​​of the first weight parameter and the threshold parameter;

[0086] Using the second training sample set and the initial oil temperature prediction model, a first oil temperature prediction value corresponding to each training sample is obtained, and a first target score is calculated according to the first oil temperature prediction value corresponding to each training sample and the actual value of the oil temperature indicated by the oil temperature label through a preset score calculation formula;

[0087] The initial oil temperature prediction model is configured according to the elements corresponding to any vector, and the second oil temperature prediction value corresponding to each training sample is obtained by using the second training sample set and the configured initial oil temperature prediction model, and the second target score corresponding to any vector is calculated according to the second oil temperature prediction value corresponding to each training sample and the actual oil temperature value through the preset score calculation formula;

[0088] Determine an optimal score according to the first target score and the second target score corresponding to each vector, determine a vector update formula corresponding to each vector in the vector set according to the second target score corresponding to each vector, and update the element corresponding to each vector based on the vector update formula;

[0089] reconfiguring the initial oil temperature prediction model according to the updated vectors, recalculating the second target score of each updated vector, and updating the optimal score based on the recalculated second target score;

[0090] When the number of updates corresponding to each vector is less than or equal to the preset number threshold, return to the step of determining the vector update formula corresponding to each vector in the vector set according to the second target score corresponding to each vector; otherwise, the elements of the vector corresponding to the updated optimal score are used as the optimized first weight parameter and threshold parameter.

[0091] Optionally, the actual oil level determination module is used to:

[0092] Acquire an oil level gauge template corresponding to the target transformer, wherein the oil level gauge template includes a plurality of first scale areas and a first pointer area corresponding to the oil level gauge;

[0093] Performing image recognition processing on the oil level gauge image, and identifying a plurality of second scale areas and a second pointer area from the oil level gauge image;

[0094] The multiple second scale areas are aligned with the multiple first scale areas in the oil level gauge template, and after the alignment, the pointer distance between the second pointer corresponding to the second pointer area and the first pointer corresponding to the first pointer area is calculated, and the actual oil level corresponding to the second pointer is determined based on the pointer distance and the preset oil level corresponding to the first pointer.

[0095] Optionally, the actual oil level determination module is further used to:

[0096] Graying the oil level gauge image to obtain a target grayscale image, and calculating a first gradient value in the horizontal direction and a second gradient value in the vertical direction corresponding to each pixel point according to the target grayscale image, and calculating a gradient amplitude corresponding to each pixel point according to the first gradient value and the second gradient value;

[0097] Comparing the gradient amplitude corresponding to each pixel point with a preset amplitude threshold, and determining the target edge point from the pixel points corresponding to the target grayscale image based on the comparison result;

[0098] The target edge points are connected to obtain a plurality of target areas corresponding to the oil level gauge image, and a second pointer area and a plurality of second scale areas are identified from the plurality of target areas according to a morphological analysis method.

[0099] Optionally, the device further comprises:

[0100] A region division module, configured to divide the oil level gauge image of the target transformer into regions after acquiring the oil level gauge image, and determine a target sub-region from the divided sub-regions based on a reference image corresponding to the oil level gauge, wherein the reference image is an image marked with an area of ​​interest after being photographed by a photographing device at the same photographing position;

[0101] A feature extraction module, used to extract features from the target sub-regions, obtain target local features corresponding to each target sub-region, and calculate enhancement parameters corresponding to each target sub-region according to the target local features, wherein the local features include at least one of brightness mean, brightness variance, contrast, and edge strength;

[0102] The image enhancement module is used to perform enhancement processing on each of the target sub-areas according to the enhancement parameters to obtain an enhanced oil level gauge image.

[0103] Optionally, the device further comprises:

[0104] The alarm information generation module is used to obtain the acquisition time of the oil level gauge image and the transformer identification corresponding to the target transformer after determining that the actual oil level of the target transformer is abnormal, generate alarm information according to the acquisition time, the transformer identification, the actual oil level, the ideal oil level and the oil level difference, and send the alarm information to a preset management terminal.

[0105] It should be noted that for other corresponding descriptions of the functional units involved in the device for identifying abnormal oil level in a transformer provided in the embodiment of the present application, reference can be made to Figures 1 to 4 The corresponding description in the method will not be repeated here.

[0106] The present application also provides a computer device, which may be a personal computer, a server, a network device, etc. Figure 6 As shown, the computer device includes a bus, a processor, a memory and a communication interface, and may also include an input and output interface and a display device. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store location information. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the steps in each method embodiment are implemented.

[0107] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0108] In one embodiment, a computer-readable storage medium is provided. The computer-readable storage medium may be non-volatile or volatile, and stores a computer program thereon. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0109] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0110] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0111] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0112] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0113] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for identifying abnormal oil level in a transformer, characterized in that: include: Acquire an oil level gauge image of an oil level gauge on an oil conservator of a target transformer, and perform image recognition processing on the oil level gauge image to obtain an actual oil level indicated by the oil level gauge; Obtain target data corresponding to the target transformer, and process the target data through a target oil temperature prediction model to obtain a predicted oil temperature corresponding to the target transformer, wherein the target data includes a current load current, a current ambient temperature, a cumulative operating time, and a historical average load current of the target transformer; Determining an ideal oil level corresponding to the predicted oil temperature based on a preset fitting relationship between the oil level and the oil temperature; An oil level difference between the actual oil level and the ideal oil level is calculated, and when the oil level difference is greater than a preset difference threshold, it is determined that an actual oil level of the target transformer is abnormal.

2. The method according to claim 1, characterized in that Before the target data is processed by the target oil temperature prediction model to obtain the predicted oil temperature corresponding to the target transformer, the method further includes: Constructing an initial oil temperature prediction model, wherein the initial oil temperature prediction model includes an input layer, at least one hidden layer, an output layer, a first weight parameter between the input layer and a first hidden layer and between multiple hidden layers, a threshold parameter of each hidden layer, and a second weight parameter between the output layer and a last hidden layer; Optimizing the first weight parameter and the threshold parameter, and configuring the initial oil temperature prediction model according to the optimized first weight parameter and threshold parameter to obtain an intermediate oil temperature prediction model; The intermediate oil temperature prediction model is trained according to the first training sample set to optimize the second weight parameter to obtain the target oil temperature prediction model, wherein each sample in the first training sample set contains the same data type as the target data and carries an oil temperature label.

3. The method according to claim 2, characterized in that The optimizing the first weight parameter and the threshold parameter includes: Constructing a vector set according to the first weight parameter and the threshold parameter, wherein an element corresponding to each vector in the vector set is used to indicate a set of values ​​of the first weight parameter and the threshold parameter; Using the second training sample set and the initial oil temperature prediction model, a first oil temperature prediction value corresponding to each training sample is obtained, and a first target score is calculated according to the first oil temperature prediction value corresponding to each training sample and the actual value of the oil temperature indicated by the oil temperature label through a preset score calculation formula; The initial oil temperature prediction model is configured according to the elements corresponding to any vector, and the second oil temperature prediction value corresponding to each training sample is obtained by using the second training sample set and the configured initial oil temperature prediction model, and the second target score corresponding to any vector is calculated according to the second oil temperature prediction value corresponding to each training sample and the actual oil temperature value through the preset score calculation formula; Determine an optimal score according to the first target score and the second target score corresponding to each vector, determine a vector update formula corresponding to each vector in the vector set according to the second target score corresponding to each vector, and update the element corresponding to each vector based on the vector update formula; reconfiguring the initial oil temperature prediction model according to the updated vectors, recalculating the second target score of each updated vector, and updating the optimal score based on the recalculated second target score; When the number of updates corresponding to each vector is less than or equal to the preset number threshold, return to the step of determining the vector update formula corresponding to each vector in the vector set according to the second target score corresponding to each vector; otherwise, the elements of the vector corresponding to the updated optimal score are used as the optimized first weight parameter and threshold parameter.

4. The method according to claim 1, characterized in that: The performing image recognition processing on the oil level gauge image to obtain the actual oil level indicated by the oil level gauge includes: Acquire an oil level gauge template corresponding to the target transformer, wherein the oil level gauge template includes a plurality of first scale areas and a first pointer area corresponding to the oil level gauge; Performing image recognition processing on the oil level gauge image, and identifying a plurality of second scale areas and a second pointer area from the oil level gauge image; The multiple second scale areas are aligned with the multiple first scale areas in the oil level gauge template, and after the alignment, the pointer distance between the second pointer corresponding to the second pointer area and the first pointer corresponding to the first pointer area is calculated, and the actual oil level corresponding to the second pointer is determined based on the pointer distance and the preset oil level corresponding to the first pointer.

5. The method according to claim 4, characterized in that The performing image recognition processing on the oil level gauge image to recognize a plurality of second scale areas and a second pointer area from the oil level gauge image includes: Graying the oil level gauge image to obtain a target grayscale image, and calculating a first gradient value in the horizontal direction and a second gradient value in the vertical direction corresponding to each pixel point according to the target grayscale image, and calculating a gradient amplitude corresponding to each pixel point according to the first gradient value and the second gradient value; Comparing the gradient amplitude corresponding to each pixel point with a preset amplitude threshold, and determining the target edge point from the pixel points corresponding to the target grayscale image based on the comparison result; The target edge points are connected to obtain a plurality of target areas corresponding to the oil level gauge image, and a second pointer area and a plurality of second scale areas are identified from the plurality of target areas according to a morphological analysis method.

6. The method according to any one of claims 1 to 5, characterized in that After acquiring the oil level gauge image of the oil level gauge of the target transformer, the method further includes: Divide the oil level gauge image into regions, and determine a target sub-region from the divided sub-regions based on a reference image corresponding to the oil level gauge, wherein the reference image is an image marked with an area of ​​interest after being photographed by a photographing device at the same photographing position; Performing feature extraction on the target sub-regions to obtain target local features corresponding to each target sub-region, and calculating enhancement parameters corresponding to each target sub-region according to the target local features, wherein the local features include at least one of brightness mean, brightness variance, contrast, and edge strength; According to the enhancement parameters, each of the target sub-regions is enhanced to obtain an enhanced oil level gauge image.

7. The method according to claim 1, characterized in that After determining that the actual oil level of the target transformer is abnormal, the method further includes: The acquisition time of the oil level gauge image and the transformer identification corresponding to the target transformer are obtained, and an alarm message is generated according to the acquisition time, the transformer identification, the actual oil level, the ideal oil level and the oil level difference, and the alarm message is sent to a preset management terminal.

8. A device for identifying abnormal oil level in a transformer, characterized in that: include: An actual oil level determination module is used to obtain an oil level gauge image of an oil level gauge on an oil conservator of a target transformer, and perform image recognition processing on the oil level gauge image to obtain an actual oil level indicated by the oil level gauge; An oil temperature prediction module is used to obtain target data corresponding to the target transformer, and process the target data through a target oil temperature prediction model to obtain a predicted oil temperature corresponding to the target transformer, wherein the target data includes a current load current, a current ambient temperature, a cumulative operating time, and a historical average load current of the target transformer; An ideal oil level determination module, used to determine the ideal oil level corresponding to the predicted oil temperature based on a preset fitting relationship between the oil level and the oil temperature; The oil level abnormality identification module is used to calculate the oil level difference between the actual oil level and the ideal oil level, and when the oil level difference is greater than a preset difference threshold, determine that the actual oil level of the target transformer is abnormal.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. A computer device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images