TWI796047B - System and method for the analysis and prediction of transformer and computer program product thereof - Google Patents

Abstract

A system, a method and a computer program product for analysis and prediction of a transformer are disclosed. The system includes a load prediction unit for establishing a load prediction model of the transformer based on oil temperature data, load variation dada, and/or current load variation data of the transformer to predict a predicted load percentage and/or predicted oil temperature based on the load prediction model. The system includes a temperature abnormality determination unit for establishing a standard temperature rise model based on ambient temperature data and oil temperature data of the transformer, temperature difference data between the ambient temperature data and the oil temperature data, and load percentage data of the transformer to determine whether the predicted oil temperature of the transformer is abnormal based on the standard temperature rise model of the transformer and present ambient temperature of the transformer, thereby effectively analyzing and predicting the abnormalities in the transformer.

Description

Transformer Analysis and Prediction System, Method and Computer Program Product

The present invention relates to a transformer technology, in detail, to a transformer abnormality analysis and prediction system, method and computer program product thereof.

The transformer is a device that uses the principle of electromagnetic induction to change the AC voltage, and is also an important component of the power system.

When the working temperature of the transformer is too high, the insulation protection of each component is easily decomposed or damaged due to high temperature, which will greatly increase the possibility of transformer failure, fire and explosion. Therefore, it is necessary to check the temperature and load from time to time to reduce fire or explosion accidents. danger of occurrence.

In addition, the transformer oil that fills the entire transformer box usually has the functions of insulation, heat dissipation, and arc suppression. Therefore, in order to ensure the safe and reliable operation of the transformer, it is often necessary to monitor the oil temperature in the transformer. There are many factors, such as external ambient temperature, multiple sets of transformers, etc., which also need to be considered together.

However, currently there is no technology for rapid and accurate analysis of various factors that affect the abnormal temperature rise of transformers. In addition to abnormal analysis of transformers, the load prediction of transformers is also an urgent issue in the industry.

In order to solve the above problems and other problems, the present invention discloses a transformer analysis and prediction system, method and computer program product thereof.

The system for transformer analysis and prediction of the present invention includes: a load prediction unit, which establishes a load prediction model of the transformer according to the transformer's oil temperature data, load change data, and/or current load change data, so as to The load prediction model of the transformer predicts the predicted load percentage of the transformer and/or the predicted oil temperature; Establish the standard temperature rise model of the transformer based on the temperature difference data between them and the load percentage data of the transformer, so as to judge whether the predicted oil temperature of the transformer is abnormal based on the standard temperature rise model of the transformer and the current ambient temperature of the transformer .

In addition, the system for transformer analysis and prediction of the present invention further includes a sensor data collection unit, which sends a command to query the transformer terminal component coupled with the transformer, so that the transformer terminal component returns the oil temperature of the transformer Data, load change data, current load change data, ambient temperature data of the transformer, temperature difference data between the ambient temperature data and the oil temperature data, load percentage data of the transformer, for the load prediction unit and The abnormal temperature rise judging unit uses artificial intelligence to process and establish the load prediction model and the standard temperature rise model respectively.

In addition, the abnormal temperature rise judgment unit classifies the established standard temperature rise model according to the brand, capacity, and/or heat dissipation method of different transformers.

In one embodiment, the abnormal temperature rise judging unit judges whether the current oil temperature rise of the transformer is abnormal according to the standard temperature rise model of the transformer and the current ambient temperature of the transformer. When the current oil temperature rise is abnormal, the abnormal temperature rise judging unit is to continue to judge whether the current oil temperature rise of other transformers in the same environment is abnormal, and after judging the current oil temperature of the other transformers in the same environment When there is no abnormality in the temperature rise, the temperature rise abnormality judging unit continues to judge whether the temperature of the upper layer oil of the transformer and/or the temperature rise of the upper layer oil exceeds the limit.

In addition, the system for transformer analysis and prediction of the present invention further includes an active alarm unit, wherein, when the abnormal temperature rise judging unit judges that the upper layer oil temperature of the transformer and/or the upper layer oil temperature rise exceeds the limit, the active alarm unit will be activated. The alarm unit provides alarms and/or suggestions for the abnormal transformer.

Secondly, the method for transformer analysis and prediction of the present invention uses a computing device or a computer. Based on the temperature difference data between the transformers and the load percentage data of the transformer, the standard temperature rise model of the transformer is established; according to the oil temperature data of the transformer, the load change data, and/or the current load change data, the transformer’s temperature rise model is established. load forecasting model; according to the load forecasting model of the transformer, predict the predicted load percentage of the transformer and/or predict the oil temperature; and judge the temperature of the transformer according to the standard model of temperature rise of the transformer and the current ambient temperature Whether the predicted oil temperature is abnormal.

Again, the computer program product of the present invention is used for loading the computer program product into a computing device or computer to execute the above-mentioned method for transformer analysis and prediction.

Therefore, the system and method for transformer analysis and prediction of the present invention promote the application of real-time monitoring of transformers through the transformer terminal unit (TTU), and perform real-time monitoring and analysis to help judge whether the transformer is normal. Changes in temperature rise to avoid unexpected failure or damage caused by overheating of the transformer. AI technology is used to establish a prediction model for different loads to predict the change of transformer load in advance and predict whether the temperature change exceeds the allowable upper limit of the transformer, so as to avoid transformer failure by scheduling in advance .

11:Transformer

12: Transformer terminal element (TTU)

2: Transformer Analysis and Prediction System

21: Sensor data collection unit

22: Load forecasting unit

23: Abnormal temperature rise judgment unit

24:Active alarm unit

S201~S205: steps

S301~S307: steps

S401~S404: steps

Fig. 1 is a schematic block diagram of the transformer analysis and prediction system of the present invention.

Fig. 2 is a schematic flow chart of the transformer analysis and prediction method of the present invention.

Fig. 3 is a schematic flow chart of abnormal temperature rise judgment of the transformer analysis and prediction method of the present invention.

FIG. 4 is a schematic flow chart of the load forecasting method of the transformer analysis and forecasting method of the present invention.

FIG. 5 is a schematic diagram of the temperature rise-load percentage of the transformer analysis and prediction method of the present invention.

The implementation of this case is described below through specific examples, and those skilled in this art can easily understand other advantages and effects of this case from the content disclosed herein. The structures, proportions, sizes, etc. shown in the drawings attached to this manual are only used to match the content disclosed in the manual, for the understanding and reading of those who are familiar with this technology, and are not used to limit the conditions that can be implemented in this case. Therefore any revision Modifications, changes or adjustments shall still fall within the scope covered by the technical content disclosed in this case, as long as they do not affect the functions and goals that can be achieved in this case.

Please refer to FIG. 1 , the transformer analysis and prediction system 2 of the present invention includes a temperature rise abnormality judgment unit 23 and a load prediction unit 22 , and can further include a sensor data collection unit 21 and an active alarm unit 24 .

In one embodiment, each unit in FIG. 1 can be software, hardware or firmware; if it is hardware, it can be a processing unit, processor, computer or server with data processing and computing capabilities; if If it is software or firmware, it may include instructions executable by a processing unit, processor, computer or server.

The sensor data collection unit 21 regularly sends commands (such as DNP 3.0) to query the transformer terminal unit (TTU) 12 through wireless communication such as NB IoT or a concentrator, and can also receive TTU to actively upload data to collect voltage, current, Actual work, virtual work, transformer (real-time) temperature, transformer (upper layer) oil temperature, transformer real-time power consumption (apparent power), rated power of the transformer, ambient temperature and other data can be stored in the system 2 or cloud database . In addition, by installing the transformer terminal unit (transformer terminal unit; TTU) 12 on the distribution transformer 11, through the relevant sensing elements on the TTU 12 (for example: current sensor (current transformer; CT), temperature sensing (transformer, voltage converter, etc.) to capture analog signals such as current, temperature, voltage, etc., calculate and return relevant real, imaginary power, apparent power and other data through TTU 12, and automatically or manually set the brand name of transformer 11 in advance in the system, Rated capacity, installation environment, installation location coordinates and other information. In other words, the sensor data collection unit 21 collects the oil temperature data of the transformer 11, the load variation data, the current load variation data, the ambient temperature data where the transformer 11 is located, the temperature difference data between the ambient temperature data and the oil temperature data, and the transformer 11 temperature data. The load percentage data are provided for the load forecasting unit 22 and the abnormal temperature rise judgment unit 23 to process and establish a load forecasting model and a standard temperature rise model respectively with artificial intelligence. For example, using ambient temperature data (such as T _e ), oil temperature data (such as upper oil temperature (such as T _top )), temperature difference data between ambient temperature data and oil temperature data (such as T _d ), transformer load percentage data As training data, input models such as DNN, CNN, and RNN for deep learning to establish a standard temperature rise model, that is, the standard temperature rise-load percentage curve, see Figure 5. Another example is to use transformer oil temperature data, load change data, and current load variable data as training data, and input DNN, CNN, and RNN models for deep learning to establish a load prediction model.

The load prediction unit 22 establishes a load prediction model of the transformer 11 based on the oil temperature data, load variation data, and/or current load variation data of the transformer 11, so as to predict the predicted load percentage and/or forecast of the transformer 11 according to the load prediction model oil temperature. In one embodiment, the load forecasting model can be, for example, an RNN forecasting model, which actively predicts the load percentage in 15-30 minutes and predicts the corresponding oil temperature.

The abnormal temperature rise judging unit 23 establishes a standard temperature rise model of the transformer 11 based on the ambient temperature data and the oil temperature data where the transformer 11 is located, the temperature difference data between the ambient temperature data and the oil temperature data, and the load percentage of the transformer 11. According to the standard temperature rise model of the transformer 11 and the current ambient temperature where the transformer is located, it is judged whether the predicted oil temperature of the transformer is abnormal. In addition, the abnormal temperature rise judging unit 23 judges whether the current oil temperature rise of the transformer 11 is abnormal according to the standard temperature rise model and the current ambient temperature where the transformer 11 is located. When it is judged that the current oil temperature rise of the transformer 11 is abnormal, the abnormal temperature rise judging unit 23 continues to judge whether the current oil temperature rise of other transformers under the same environment is abnormal. When judging that the current oil temperature rise of other transformers under the same environment is not abnormal, the abnormal temperature rise judging unit 23 continues to judge whether the upper layer oil temperature of the transformer 11 and/or whether the upper layer oil temperature rise exceeds the limit.

In one embodiment, the abnormal temperature rise judgment unit 23 compares whether the current oil temperature rise or the predicted oil temperature rise is abnormal according to the standard temperature rise model, that is, the temperature rise-load percentage curve shown in FIG. The predicted oil temperature is subtracted from the current ambient temperature to obtain the current oil temperature rise or the predicted oil temperature rise. If currently If the oil temperature rise or the predicted oil temperature rise is in the area below the oblique line in the figure, it means that there is no abnormality in the temperature rise; if not, it means that the temperature rise is abnormal. Then, by grouping the transformers 11 of the same area and the same brand, the problems caused by the environment are eliminated, and the transformers 11 with abnormal temperature rise are included in the subsequent active alarm group. In addition, for the established standard temperature rise model, the temperature rise abnormal judgment unit 23 classifies the transformer 11 according to its brand, capacity, and heat dissipation method (such as self-cooling, general forced oil circulation, and guided forced oil circulation), so as to be used in the future. When monitoring other transformers, the corresponding standard temperature rise model can be selected, instead of suggesting a standard temperature rise model for each transformer to be monitored one by one.

In addition, the temperature rise judgment depends on the insulation level of the transformer 11 and the heat dissipation form of the transformer 11. There are different upper oil temperature upper limits and temperature rise limits (as shown in the table below). When the oil temperature exceeds the limit or the temperature rise exceeds the limit, the active alarm unit 24 Provide warnings or related suggestions.

In addition, when the abnormal temperature rise judging unit 23 judges that the predicted oil temperature predicted by the load predicting unit 22 is abnormal according to the standard temperature rise model and the ambient temperature where the transformer 11 is located, an active alarm will be issued. Yuan 24 took the initiative to issue an alarm. Alternatively, the active alarm unit 24 actively issues an alarm for the transformers 11 included in the active alarm group by the abnormal temperature rise judgment unit 23 .

Referring to Fig. 2, the transformer analysis and prediction method of the present invention comprises the following steps:

In step S201, the ambient temperature data of the transformer, the oil temperature data of the transformer, the temperature difference data between the ambient temperature data and the oil temperature data, the load percentage data of the transformer, the load variation data, and the current load variation data are collected. Then enter steps S202 and S203, and there is no order limitation for these two steps.

In step S202, a standard temperature rise model is established according to the ambient temperature data where the transformer is located, the oil temperature data, the temperature difference data between the ambient temperature data and the oil temperature data, and the load percentage data of the transformer. In step S203, a load prediction model is established according to the transformer oil temperature data, load variation data, and current load variation data. Then enter step S204. In step S204, the predicted load percentage of the transformer and/or the predicted oil temperature are predicted according to the load prediction model of the transformer. Then enter step S205.

In step S205, according to the standard temperature rise model of the transformer and the current ambient temperature where the transformer is located, when it is judged that the current oil temperature rise of the transformer is abnormal or the predicted oil temperature is abnormal, an alarm and/or suggestion is given.

In the embodiment shown in FIG. 3 , steps S301 to S307 represent the process of judging and warning of abnormal temperature rise.

In step S301, according to the standard temperature rise model of the transformer and the current ambient temperature where the transformer is located, it is judged whether the current oil temperature rise of the transformer (i.e., the current oil temperature of the transformer minus the current ambient temperature) is abnormal, that is, whether the current oil temperature rise has fallen into the area above the oblique line in Figure 5, if it falls into the area above the oblique line If it falls into the area below the slash (including the slash), then it is normal. If it is abnormal, go to step S302, and if not abnormal, go to step S307 and continue monitoring.

In step S302, it is judged whether the current oil temperature rise of other transformers in the same environment is abnormal, that is, whether the current oil temperature rise of transformers with a proportion above a preset value (such as 50%) exceeds the standard curve, and if exceeded, enter step S305, and if not exceeded, go to step S303.

In step S303, it is judged whether the upper layer oil temperature and/or oil temperature rise of the transformer exceeds the limit, if so, go to step S304, and if not, go to step S307, and continue monitoring.

In step S304, abnormal warnings and/or suggestions are provided.

In addition, in step S305, it is determined whether the environment is abnormal (for example: whether there is a fire or the air conditioner fails to cause the ambient temperature to be too high), if yes, go to step S306, and if not, go back to step S303. In step S306, solve the problem of abnormal environment, then enter step S307, continue monitoring.

In the embodiment shown in FIG. 4, steps S401-404 represent the process of load prediction and abnormal temperature rise judgment.

In step S401, according to the load prediction model of the transformer, the predicted load percentage and/or the predicted oil temperature of the transformer after a period of time (such as 15-30 minutes) are predicted. Then enter step S402.

In step S402, according to the standard temperature rise model of the transformer and the current ambient temperature where the transformer is located, it is judged whether the predicted oil temperature of the transformer is abnormal, and the predicted oil temperature is subtracted from the current ambient temperature to obtain the predicted oil temperature rise, and compared with the standard temperature rise For the model (such as the temperature rise-load percentage shown in Figure 5), if the predicted oil temperature rise falls in the area above the slanted line in Figure 5, it is abnormal, and if the predicted oil temperature rise falls in the area below the slashed line in Figure 5, it is normal. If abnormal, proceed to step S403 to raise an alarm and/or suggestion, and if not abnormal, proceed to step S404 to continue monitoring.

In addition, the present invention also discloses a computer program product for loading the computer program product into a computing device or computer, such as a computing device or computer having a processor (such as CPU, GPU, etc.) and/or memory, And the computing device or computer can be used to execute the computer program product through the processor and/or memory, so as to execute the above-mentioned method and each step when executing the computer program product.

Therefore, through the transformer analysis and prediction system and method of the present invention, its computer program product, through the introduction of the transformer terminal unit (TTU) application and the development of LPWAN (Low-Power Wide-Area Network) technology, the real-time monitoring application of the transformer is promoted, Through real-time monitoring and analysis to help judge whether the transformer is normal, grasp the temperature rise of the transformer, and avoid unexpected failure or damage caused by the overheating of the transformer. In addition, AI technology is used to establish a prediction model for different loads, predict the change of transformer load in advance, and predict whether the temperature change exceeds the allowable upper limit of the transformer, schedule in advance to avoid transformer failure, and then effectively analyze and predict the transformer.

In summary, the system, method and computer program product for transformer analysis and prediction of the present invention, through the introduction of the transformer terminal unit (TTU), increase the real-time voltage, current, power consumption, oil temperature, etc. of the transformer. Detection can help to grasp the actual operation of the transformer. Through the load percentage, oil temperature detection and the temperature rise-load percentage curve established in the factory, it can compare with the current temperature rise to analyze the abnormal temperature rise, and compare it with similar conditions in the vicinity For the temperature rise of transformers of the same brand and model, find out the transformers with abnormal temperature rise changes, carry out load control, and through the classification of transformer loads, establish AI prediction models for different load types, and predict whether the power consumption will exceed the rated power after 15 to 30 minutes Capacity, and whether the corresponding temperature rise exceeds the allowable upper limit of the transformer, so as to schedule the feeder load in advance to avoid transformer failure.

The above-mentioned embodiments are only illustrative of the effects of this case, and are not intended to limit this case. Any person familiar with this technology can modify and change the above-mentioned implementations without violating the spirit and scope of this case. Therefore, the scope of protection of rights in this case should be listed in the scope of patent application described later.

S201~S205: steps

Claims (9)

A system for transformer analysis and prediction, including: a load prediction unit, which establishes a load prediction model of the transformer according to the oil temperature data, load change data, and/or current load change data of the transformer, and uses the transformer The load prediction model predicts the predicted load percentage of the transformer and/or the predicted oil temperature; and the abnormal temperature rise judgment unit, according to the ambient temperature data and oil temperature data where the transformer is located, and the relationship between the ambient temperature data and the oil temperature data The temperature difference data and the load percentage data of the transformer establish the standard temperature rise model of the transformer, so as to judge whether the predicted oil temperature of the transformer is abnormal according to the standard temperature rise model of the transformer and the current ambient temperature of the transformer. , the abnormal temperature rise judging unit classifies the established standard temperature rise model according to the brand, capacity, and/or heat dissipation method of different transformers. The system as described in claim item 1 further includes a sensor data collection unit, which sends commands to inquire about the transformer terminal component coupled with the transformer, so that the transformer terminal component returns the oil temperature data and load changes of the transformer data, current load variation data, ambient temperature data of the transformer, temperature difference data between the ambient temperature data and the oil temperature data, load percentage data of the transformer, for the load prediction unit and the abnormal temperature rise The judging units use artificial intelligence to process and establish the load prediction model and the standard temperature rise model respectively. The system as described in claim 1, wherein the abnormal temperature rise judging unit judges whether the current oil temperature rise of the transformer is abnormal according to the standard temperature rise model of the transformer and the current ambient temperature of the transformer. When it is judged that the current oil temperature rise of the transformer is abnormal, the abnormal temperature rise judging unit is to continue to judge whether the current oil temperature rise of other transformers in the same environment is abnormal, and then When it is determined that the current oil temperature rise of the other transformers in the same environment is not abnormal, the temperature rise abnormality judging unit continues to judge whether the upper layer oil temperature of the transformer and/or the upper layer oil temperature rise exceeds a limit. The system as described in claim 3 further includes an active alarm unit, wherein when the abnormal temperature rise judging unit determines that the temperature of the transformer’s upper layer oil and/or the temperature rise of the upper layer oil exceeds the limit, the active alarm unit is instructed to respond to the abnormality warnings and/or recommendations for that transformer. A method for analyzing and predicting a transformer, comprising: according to the ambient temperature data of the transformer, the oil temperature data of the transformer, the temperature difference data between the ambient temperature data and the oil temperature data, and the load percentage data of the transformer , establish the standard temperature rise model of the transformer, wherein, classify the established standard temperature rise model according to the brand, capacity, and/or heat dissipation method of different transformers; according to the oil temperature data and load change data of the transformer , and/or current load variation data, to establish a load prediction model for the transformer; to predict the load percentage of the transformer and to predict the oil temperature based on the load prediction model of the transformer; and to predict the temperature rise standard model of the transformer and the current ambient temperature where the transformer is located, to determine whether the predicted oil temperature of the transformer is abnormal. The method as described in claim 5, further comprising: sending a command to inquire about the transformer terminal component coupled with the transformer, so that the transformer terminal component returns the oil temperature data of the transformer, the load change data, the current The load change data, the ambient temperature data where the transformer is located, the temperature difference data between the ambient temperature data and the oil temperature data, and the load percentage data of the transformer are used for artificial intelligence processing and establishment of the load prediction model and The standard temperature rise model. The method as described in claim item 5, further comprising: According to the standard temperature rise model of the transformer and the current ambient temperature of the transformer, it is judged whether the current oil temperature rise of the transformer is abnormal; Check whether the current oil temperature rise of other transformers is abnormal. The method as described in claim item 7 further includes: when it is judged that the current oil temperature rise of the other transformers in the same environment is not abnormal, continue to judge whether the upper layer oil temperature of the transformer and/or the upper layer oil temperature rise exceed limit; and when it is judged that the temperature of the upper layer of the transformer and/or the temperature rise of the upper layer of oil exceeds the limit, an alarm and/or suggestion is given for the abnormal transformer. A computer program product for loading the computer program product into a computing device or computer to execute the method for transformer analysis and prediction as described in any one of Claims 5 to 8.

Images