Method and system for predicting remaining useful life of traction transformer
By combining fuzzy inference system and neural network in the rail transit power supply system, analyzing the top oil temperature data of the traction transformer and building a combined prediction model, the short-term and coarse-grained problems of traction transformer life prediction in the existing technology are solved, and long-term high Accurate oil temperature prediction and life expectancy ensure the safety of train operation.
Patent Information
- Application Number
- CN202210031252.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-12
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-01-12
AI Technical Summary
In the rail transit power supply system, the remaining service life prediction of the traction transformer depends on historical fault data and probability distribution model, and lacks attention to the impact of the internal thermal state of the transformer, resulting in short-term and coarse-grained problems in the prediction model and cannot be realized. Long-term high accuracy prediction.
Using a combined prediction model based on type-2 type fuzzy inference system and long-term short-term memory neural network, a long-term prediction model of dynamic health status is established by acquiring and analyzing the top oil temperature data of the traction transformer, and a long-term prediction model of dynamic health status is extracted, and the trend and random change time series are combined. CEEMDAN decomposition algorithm constructs a combined prediction model to predict oil temperature and calculate the remaining service life.
It realizes long-term stable and high-accuracy prediction of the oil temperature of the traction transformer in the rail transit power supply system, improves the prediction accuracy and width, guides status monitoring and maintenance, and ensures the safety and reliability of train operation.
Smart Images

Figure CN114372416B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of rail transit technology, and in particular to a method and system for predicting the remaining service life of a traction transformer in a rail transit power supply system. Background Art
[0002] As a typical electronic control device, oil-immersed transformer can use the principle of electromagnetic induction to change AC voltage. It plays the role of voltage and current conversion, isolation, impedance transformation, magnetic saturation transformation, etc. in the circuit. It has the advantages of stable structure and strong anti-interference ability. It is widely used in railway transportation, power transmission and other fields.
[0003] As a typical oil-immersed power transformer, the traction transformer is a core component of the traction power supply system. The reliability of transformer operation is an important basis for ensuring the efficiency and stability of the power supply system and thus ensuring the safe operation of high-speed railways. Accurate prediction of the remaining service life of the transformer is a prerequisite for making maintenance decisions based on the situation and ensuring driving safety. Temperature is the main factor affecting the life of the traction transformer. The load capacity and insulation aging rate of the transformer are mainly determined by its internal thermal state. The top oil temperature of the transformer is a basic parameter that effectively reflects the thermal condition of the traction transformer. Fully exploring the evolution law of the time series of the top oil temperature of the transformer and establishing an accurate oil temperature prediction model to support the estimation of the remaining service life are of great practical significance for improving the reliability of the traction transformer and reducing the failure rate of the power supply system.
[0004] At present, the existing prediction of the remaining life of transformers has not completely broken through the traditional life prediction method based on "historical fault data" and "probability distribution model", and insufficient attention has been paid to the impact of the internal thermal state of the transformer on equipment aging. At the same time, the current prediction model for the top oil temperature of the transformer presents the characteristics of "short prediction" and "coarse granularity", and insufficient attention has been paid to the long-term accurate prediction model with hours as the basic unit. Summary of the invention
[0005] The object of the present invention is to provide a life estimation model that constructs the top oil temperature of a traction transformer and the remaining useful life (Remaining Useful Life, RUL), establishes a long-term prediction model of the dynamic health status based on the top oil temperature of the traction transformer, and adopts a combined prediction model based on a type-2 fuzzy inference system and a long short-term memory neural network to achieve long-term, stable, and high-accuracy prediction of the transformer oil temperature, thereby achieving online prediction of the trend of changes in the health of the traction transformer, to provide a method and system for predicting the remaining useful life of a traction transformer in a rail transit power supply system, so as to solve at least one technical problem existing in the above-mentioned background technology.
[0006] In order to achieve the above object, the present invention adopts the following technical solutions:
[0007] In one aspect, the present invention provides a method for predicting the remaining service life of a traction transformer, comprising:
[0008] Obtain the top oil temperature data of the traction transformer of the rail transit power supply system during operation;
[0009] Establish a mathematical model of the remaining service life of the traction transformer based on the internal thermal change process of the transformer;
[0010] A time series model is established based on the top oil temperature data during the operation of the traction transformer, the change trend of the top oil temperature is analyzed and features are extracted to obtain a trend change time series and a random change time series of the top oil temperature;
[0011] Based on the trend change time series and the random change time series, the evolution trend of the transformer top oil temperature in the next period is predicted, and the prediction result is substituted into the remaining service life mathematical model of the traction transformer to obtain the remaining service life of the transformer.
[0012] Preferably, a model RUL is established between the remaining service life of the transformer and the hot spot temperature of the transformer winding point. For the working loss of the traction transformer on a certain day, after running for 24 hours, the remaining effective service life is defined as the difference between the ideal RUL obtained by the Arrhenius law and the loss on that day.
[0013] Preferably, obtaining the trend change time series and random change time series of the top oil temperature includes:
[0014] Define the original rail transit power supply system traction transformer top oil temperature time series;
[0015] Use EMD to decompose the original time series and define the IMF signal of the first node after decomposition:
[0016] Define the signal residual after the first EMD decomposition;
[0017] Decompose the remaining signal after the first EMD decomposition to obtain the second node signal:
[0018] Repeat the above steps for each node to obtain the final remaining semaphore.
[0019] Preferably, a long short-term memory neural network LSTM prediction model is used for trend change time series, and a type-2FIS method is used for random change time series, and a combined prediction model is established.
[0020] Preferably, the type-2FIS method is used to predict the random variation time series, including:
[0021] Input the random item time series of the transformer top oil temperature, establish a rule base based on IF-THEN fuzzy rules; define a database using fuzzy set membership functions; make inference decisions based on fuzzy rules, and introduce particle swarm optimization algorithm to find the optimal fuzzy rules; fuzzy reasoning; defuzzification, and display clear results.
[0022] Preferably, the CEEMDAN decomposition algorithm is used to extract the top oil temperature time series features, and the decomposition is performed to obtain chaotic time series nodes and monotonic trend sequence nodes; a combined prediction model is constructed, and type-2FIS is used to predict the chaotic time series nodes, and LSTM is used to predict the monotonic trend sequence nodes; the CEEMDAN reconstruction superposition principle is used to obtain the predicted oil temperature, the predicted oil temperature is divided into hours, and is substituted into the remaining service life mathematical model of the transformer to obtain the remaining service life of the traction transformer.
[0023] In a second aspect, the present invention provides a system for predicting the remaining useful life of a traction transformer, comprising:
[0024] An acquisition module, used to acquire top oil temperature data of a traction transformer of a rail transit power supply system during operation;
[0025] A building module is used to establish a mathematical model of the remaining useful life of the traction transformer based on the internal thermal change process of the transformer;
[0026] An extraction module is used to establish a time series model based on the top oil temperature data during the operation of the traction transformer, analyze the change trend of the top oil temperature and perform feature extraction to obtain a trend change time series and a random change time series of the top oil temperature;
[0027] The prediction module is used to predict the evolution trend of the transformer top oil temperature in the next period based on the trend change time series and the random change time series, and substitute the prediction result into the remaining service life mathematical model of the traction transformer to obtain the remaining service life of the transformer.
[0028] In a third aspect, the present invention provides a computer device, comprising a memory and a processor, wherein the processor and the memory communicate with each other, the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the remaining useful life prediction method for a traction transformer as described above.
[0029] In a fourth aspect, the present invention provides an electronic device, comprising a memory and a processor, wherein the processor and the memory communicate with each other, the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the remaining useful life prediction method for a traction transformer as described above.
[0030] In a fifth aspect, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the method for predicting the remaining useful life of a traction transformer as described above.
[0031] The beneficial effects of the present invention are as follows: the CEEMDAN time series decomposition model is adopted to extract time series features to the maximum extent and improve the accuracy and precision of the prediction; the combined prediction method of LSTM and type-2FIS is adopted, which integrates the advantages of fuzzy reasoning system and neural network for time series prediction, greatly improves the prediction accuracy and prediction width, and can be used to predict the top oil temperature of traction transformers in rail transit power supply systems and then estimate the remaining life, guide the status monitoring and condition-based maintenance of traction transformers in rail transit power supply systems, and ensure the safe and reliable operation of trains.
[0032] Additional aspects and advantages of the present invention will be given in part in the following description, which will become obvious from the following description, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.
[0034] Figure 1 This is a schematic diagram of the combined prediction principle of the remaining service life of the traction transformer of the rail transit power supply system according to an embodiment of the present invention.
[0035] Figure 2 It is a schematic diagram of the modal components obtained after CEEMDAN decomposition of the top oil temperature of the traction transformer of the rail transit power supply system according to an embodiment of the present invention. DETAILED DESCRIPTION
[0036] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below by the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.
[0037] It should be understood by those skilled in the art that unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those skilled in the art to which this invention belongs.
[0038] It should also be understood that terms, such as those defined in commonly used dictionaries, should be understood to have a meaning consistent with that in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless as defined herein.
[0039] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements and / or groups thereof.
[0040] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.
[0041] To facilitate understanding of the present invention, the present invention is further explained below with reference to specific embodiments in conjunction with the accompanying drawings, and the specific embodiments do not constitute a limitation on the embodiments of the present invention.
[0042] Those skilled in the art should understand that the drawings are merely schematic diagrams of embodiments, and the components in the drawings are not necessarily necessary for implementing the present invention.
[0043] Example 1
[0044] This embodiment 1 provides a system for predicting the remaining service life of a traction transformer, the system comprising:
[0045] An acquisition module, used to acquire top oil temperature data of a traction transformer of a rail transit power supply system during operation;
[0046] A building module is used to establish a mathematical model of the remaining useful life of the traction transformer based on the internal thermal change process of the transformer;
[0047] An extraction module is used to establish a time series model based on the top oil temperature data during the operation of the traction transformer, analyze the change trend of the top oil temperature and perform feature extraction to obtain a trend change time series and a random change time series of the top oil temperature;
[0048] The prediction module is used to predict the evolution trend of the transformer top oil temperature in the next period based on the trend change time series and the random change time series, and substitute the prediction result into the remaining service life mathematical model of the traction transformer to obtain the remaining service life of the transformer.
[0049] In this embodiment 1, the above system is used to implement a method for predicting the remaining useful life of a traction transformer, including:
[0050] Obtain the top oil temperature data of the traction transformer of the rail transit power supply system during operation;
[0051] Establish a mathematical model of the remaining service life of the traction transformer based on the internal thermal change process of the transformer;
[0052] A time series model is established based on the top oil temperature data during the operation of the traction transformer, the change trend of the top oil temperature is analyzed and features are extracted to obtain a trend change time series and a random change time series of the top oil temperature;
[0053] Based on the trend change time series and the random change time series, the evolution trend of the transformer top oil temperature in the next period is predicted, and the prediction result is substituted into the remaining service life mathematical model of the traction transformer to obtain the remaining service life of the transformer.
[0054] Among them, a model RUL is established between the remaining service life of the transformer and the hot spot temperature of the transformer winding point. For the working loss of the traction transformer on a certain day, after 24 hours of operation, the remaining effective service life is defined as the difference between the ideal RUL obtained by the Arrhenius law and the loss on that day.
[0055] Specifically, a mathematical model of the remaining service life of the traction transformer is established according to the change process of the internal thermal process of the transformer, including:
[0056] The Arrhenius law can be used to establish a model between the remaining service life of the transformer and the hot spot temperature of the winding:
[0057] RUL=Cexp(B / 273+θ HST )
[0058] Among them, θ HST =T OT +T Δ , T OT is the top oil temperature, T Δ is the local temperature rise caused by the load current; the unit of RUL calculation is hour, C and B are given insulation material coefficients or system constants, which are empirical constants. Usually the value of B is given as 15000, and the value of C will change with the change of running time, so it needs to be given in combination with the internal thermal analysis of the real transformer.
[0059] Considering the temperature difference between the transformer winding point and the oil, the top oil temperature is usually monitored to replace the winding temperature as a performance indicator for real-time evaluation of the transformer operating conditions, ignoring the slight temperature change caused by the load current, that is, the remaining service life RUL of the traction transformer is defined as:
[0060] RUL=Cexp(15000 / 273+T OT )
[0061] It can be seen from the above formula that the higher the transformer top oil temperature, the lower its remaining service life value. The oil temperature measurement of traction transformers usually records data in hours, defining the top oil temperature of the transformer within 24 hours on a certain day:
[0062] U=T1,T2,...,T 23 ,T 24
[0063] Where T i It indicates the oil temperature at the i-th hour. At the same time, the Mönsinger 6-degree aging rule stipulates that in the range of 80-140℃, the internal insulation aging rate of the transformer is twice the previous aging rate for every 6℃ increase in transformer temperature. At the same time, it is stipulated that the aging rate is 1 at 98℃. When the top oil temperature does not exceed 98℃, it can work normally for about 20 years.
[0064] As the service life increases, the top insulation material gradually deteriorates, the temperature change of the top oil temperature becomes more and more drastic, and a large amount of additional loss is generated. The additional life loss of one day of operation is defined as:
[0065]
[0066] Where P = {T i ≥80|T i ∈U,1≤i≤24}, j is the number of elements in the set P, and “ / / ” indicates integer divisibility.
[0067] For the working loss of the traction transformer on a certain day, after running for 24 hours, the remaining effective service life is defined as the difference between the ideal RUL obtained by the Arrhenius law and the loss on that day:
[0068]
[0069] Among them, T O ″ T It is the average oil temperature at the top layer of the transformer on that day.
[0070] In this embodiment 1, the trend change time series and random change time series of the top oil temperature are obtained, including:
[0071] Define the original rail transit power supply system traction transformer top oil temperature time series;
[0072] Use EMD to decompose the original time series and define the IMF signal of the first node after decomposition:
[0073] Define the signal residual after the first EMD decomposition;
[0074] Decompose the remaining signal after the first EMD decomposition to obtain the second node signal:
[0075] Repeat the above steps for each node to obtain the final remaining semaphore.
[0076] The long short-term memory neural network LSTM prediction model is used for trend change time series, and the type-2FIS method is used for random change time series, and a combined prediction model is established.
[0077] The type-2FIS method is used to predict random change time series, including:
[0078] Input the random item time series of the transformer top oil temperature, establish a rule base based on IF-THEN fuzzy rules; define a database using fuzzy set membership functions; make inference decisions based on fuzzy rules, and introduce particle swarm optimization algorithm to find the optimal fuzzy rules; fuzzy reasoning; defuzzification, and display clear results.
[0079] The CEEMDAN decomposition algorithm is used to extract the top oil temperature time series features, and the chaotic time series nodes and monotonic trend sequence nodes are decomposed. A combined prediction model is constructed, and type-2FIS is used to predict the chaotic time series nodes, and LSTM is used to predict the monotonic trend sequence nodes. The CEEMDAN reconstruction superposition principle is used to obtain the predicted oil temperature, which is divided into hours and substituted into the remaining service life mathematical model of the transformer to obtain the remaining service life of the traction transformer.
[0080] In summary, in this embodiment 1, a mathematical model of the remaining life of the traction transformer is established according to the internal thermal process of the transformer; according to the remaining life model, the top oil temperature sequence is used to predict the remaining life; the monitored oil temperature is decomposed into trend terms and random terms using CEEMDAN decomposition, a Type-2 FIS prediction model is established for the random terms, and an LSTM prediction model is established for the trend terms, and finally, the signal is reconstructed by a combined prediction method to obtain the predicted value of the top oil temperature of the transformer, thereby predicting the remaining service life of the traction transformer. The method of the present invention adopts a combined prediction model, which can maximize the use of the original time series information, eliminate the influence of noise, and improve the prediction accuracy; it can be used for train operation control systems to guide the status monitoring and condition-based maintenance of traction transformers in rail transit power supply systems.
[0081] Example 2
[0082] In this embodiment 2, a combined prediction method for the remaining life of a traction transformer in a rail transit power supply system is provided to guide the status monitoring and condition-based maintenance of the traction transformer in the rail transit power supply system. In this embodiment 2, the remaining life prediction method for the traction transformer in a rail transit power supply system based on the combination of FIS and LSTM includes:
[0083] According to the change process of the internal thermal process of the transformer, a mathematical model of the remaining service life of the traction transformer is established, and the corresponding parameters to be solved are listed: the top oil temperature T ot ;
[0084] According to the above, a time series model is established for the top oil temperature during the operation of the traction transformer, its change trend is analyzed and the CEEMDAN method is used for feature extraction to obtain its trend change time series and random change time series; a long short-term memory neural network (LSTM) prediction model is used for the trend change time series, and a type-2FIS method is used for prediction of the random change time series, and a combined prediction model is established to accurately predict the evolution trend of the top oil temperature of the transformer in the next period, and substitute it into the remaining service life mathematical model of the traction transformer to obtain the remaining service life of the transformer;
[0085] According to the remaining service life of the traction transformer of the rail transit power supply system, the condition monitoring and condition-based maintenance are guided
[0086] Preferably, a mathematical model of the remaining service life of the traction transformer is established according to the change process of the internal thermal process of the transformer, including:
[0087] The Arrhenius law can be used to establish a model between the remaining service life of the transformer and the hot spot temperature of the winding:
[0088] RUL=Cexp(B / 273+θ HST )
[0089] Among them, θ HST =T OT +T Δ , T OT is the top oil temperature, T Δ is the local temperature rise caused by the load current; the unit of RUL calculation is hour, C and B are given insulation material coefficients or system constants, which are empirical constants. Usually the value of B is given as 15000, and the value of C will change with the change of running time, so it needs to be given in combination with the internal thermal analysis of the real transformer.
[0090] Considering the temperature difference between the transformer winding point and the oil, the top oil temperature is usually monitored to replace the winding temperature as a performance indicator for real-time evaluation of the transformer operating conditions, ignoring the slight temperature change caused by the load current, that is, the remaining service life RUL of the traction transformer is defined as:
[0091] RUL=Cexp(15000 / 273+T OT )
[0092] It can be seen from the above formula that the higher the transformer top oil temperature, the lower its remaining service life value. The oil temperature measurement of traction transformers usually records data in hours, defining the top oil temperature of the transformer within 24 hours on a certain day:
[0093] U=T1,T2,...,T 23 ,T 24
[0094] Where T i It indicates the oil temperature at the i-th hour. At the same time, the Mönsinger 6-degree aging rule stipulates that in the range of 80-140℃, the internal insulation aging rate of the transformer is twice the previous aging rate for every 6℃ increase in transformer temperature. At the same time, it is stipulated that the aging rate is 1 at 98℃. When the top oil temperature does not exceed 98℃, it can work normally for about 20 years.
[0095] As the service life increases, the top insulation material gradually deteriorates, the temperature change of the top oil temperature becomes more and more drastic, and a large amount of additional loss is generated. The additional life loss of one day of operation is defined as:
[0096]
[0097] Where P = {T i ≥80|T i ∈U,1≤i≤24}, j is the number of elements in the set P, and “ / / ” indicates integer divisibility.
[0098] For the working loss of the traction transformer on a certain day, after running for 24 hours, the remaining effective service life is defined as the difference between the ideal RUL obtained by the Arrhenius law and the loss on that day:
[0099]
[0100] Among them, T O ″ T It is the average oil temperature at the top layer of the transformer on that day.
[0101] The method of establishing a time series model for the top oil temperature during operation of the traction transformer, analyzing its variation trend and extracting features using the CEEMDAN method includes:
[0102] Step 1. Define the original rail transit power supply system traction transformer top oil temperature time series as S i (t) = S (t) + ε0w i (t), where ε0 is the noise coefficient, i = 0, 1, ..., I, and the original time series is decomposed using EMD, and the first node IMF (Intrinsic Mode Functions, IMFs) signal after decomposition is defined as:
[0103]
[0104] Step 2. Define the signal residual after the first EMD decomposition:
[0105]
[0106] Step 3. Decompose the remaining signal after the first EMD decomposition considering white noise to obtain the node 2 signal:
[0107]
[0108] Step 4. Repeat the above steps for each node to get the final remaining semaphore:
[0109]
[0110] Among them, M is the total number of decomposed signal node IMFs. CEEMDAN decomposition is used to effectively extract the characteristics of the original signal at different time scales, improve the smoothness of the prediction, and reduce the prediction error.
[0111] The type-2FIS method is used to predict random change time series, including:
[0112] Step 1: Input the random item time series of transformer top oil temperature and establish a rule base based on IF-THEN fuzzy rules;
[0113] Step 2: Define the database using fuzzy set membership function;
[0114] Step 3: Make reasoning decisions based on fuzzy rules and introduce particle swarm optimization algorithm to find the optimal fuzzy rules;
[0115] Step 4: Fuzzy reasoning;
[0116] Step 5: Defuzzify and display the clear result.
[0117] The LSTM prediction model is used for trend change time series, and the type-2FIS method is used for random change time series. A combined prediction model is established, including:
[0118] For trend items, the type-2FIS prediction algorithm is used. Compared with type-1FIS, type-2FIS sets more membership functions to be optimized, has good chaotic sequence perception ability, and can fully extract its change characteristics using fuzzy prediction methods; LSTM has the ability to capture time series dependencies, and has good fitting ability for the monotonic trend evolution of the end nodes of CEEMDAN decomposition. Therefore, it is planned to use type-2FIS to predict the decomposed chaotic time series nodes, and use LSTM to predict the monotonic trend motion sequence. The main steps are as follows:
[0119] Step 1. Use the CEEMDAN decomposition algorithm to extract the top oil temperature time series characteristics, initialize the parameters, and decompose it into IMF1, IMF2, …, IMF according to the decomposition steps. 12 、IMF 13 , where IMF1-IMF 12 is a chaotic time series node, IMF 13 It is a monotonic trend sequence node.
[0120] Step 2. Build a combined forecasting model for IMF1-IMF 12 Use type-2FIS for prediction, and assume that the prediction results are s′1,s′2,...,s′ 12 ; IMF 13 Use LSTM for prediction, and let the prediction result be s′ 13 .
[0121] Step 3. Use CEEMDAN to reconstruct the superposition principle to obtain the predicted oil temperature:
[0122] OT p =s′1+s′2+...+s′ 13
[0123] Step 4. Convert the obtained OT p Divide it into hours and substitute it into the mathematical model of the transformer's remaining service life to obtain the remaining service life of the traction transformer.
[0124] It can be seen from the technical solution provided by the above embodiment that the method of this embodiment adopts the CEEMDAN time series decomposition model, which can maximize the extraction of time series features and improve the accuracy and precision of the prediction; it adopts the combined prediction method of LSTM and type-2FIS, which integrates the advantages of fuzzy reasoning system and neural network for time series prediction, greatly improves the prediction accuracy and prediction width, and can be used to predict the top oil temperature of traction transformers in rail transit power supply systems and then estimate the remaining life, guide the status monitoring and condition-based maintenance of traction transformers in rail transit power supply systems, and ensure the safe and reliable operation of trains.
[0125] Example 3
[0126] Accurately predicting the remaining service life of the traction transformer is a prerequisite for making maintenance decisions based on the situation and ensuring driving safety. In this embodiment 3, a schematic diagram of a method for predicting the remaining service life of the traction transformer of a rail transit power supply system is provided. Figure 1 The specific processing steps include:
[0127] Step 1: Establish a mathematical model of the remaining service life of the traction transformer based on the internal thermal process of the transformer, including:
[0128] The Arrhenius law can be used to establish a model between the remaining service life of the transformer and the hot spot temperature of the winding:
[0129] RUL=C exp(B / 273+θ HST )
[0130] Among them, θ HST =T OT +T Δ , T OT is the top oil temperature, T Δ is the local temperature rise caused by the load current; the unit of RUL calculation is hour, C and B are given insulation material coefficients or system constants, which are empirical constants. Usually the value of B is given as 15000, and the value of C will change with the change of running time, so it needs to be given in combination with the internal thermal analysis of the real transformer.
[0131] Considering the temperature difference between the transformer winding point and the oil, the top oil temperature is usually monitored to replace the winding temperature as a performance indicator for real-time evaluation of the transformer operating conditions, ignoring the slight temperature change caused by the load current, that is, the remaining service life RUL of the traction transformer is defined as:
[0132] RUL=C exp(15000 / 273+T OT )
[0133] It can be seen from the above formula that the higher the transformer top oil temperature, the lower its remaining service life value. The oil temperature measurement of traction transformers usually records data in hours, defining the top oil temperature of the transformer within 24 hours on a certain day:
[0134] U=T1,T2,...,T 23 ,T 24
[0135] Where T iIt indicates the oil temperature at the i-th hour. At the same time, the Mönsinger 6-degree aging rule stipulates that in the range of 80-140℃, the internal insulation aging rate of the transformer is twice the previous aging rate for every 6℃ increase in transformer temperature. At the same time, it is stipulated that the aging rate is 1 at 98℃. When the top oil temperature does not exceed 98℃, it can work normally for about 20 years.
[0136] As the service life increases, the top insulation material gradually deteriorates, the temperature change of the top oil temperature becomes more and more drastic, and a large amount of additional loss is generated. The additional life loss of one day of operation is defined as:
[0137]
[0138] Where P = {T i ≥80|T i ∈U,1≤i≤24}, j is the number of elements in the set P, and “ / / ” indicates integer divisibility.
[0139] For the working loss of the traction transformer on a certain day, after running for 24 hours, the remaining effective service life is defined as the difference between the ideal RUL obtained by the Arrhenius law and the loss on that day:
[0140]
[0141] Among them, T″ OT It is the average oil temperature at the top layer of the transformer on that day.
[0142] Step 2: Establish a time series model for the top oil temperature during the operation of the traction transformer, analyze its changing trend and use the CEEMDAN method to extract features. Figure 2 The top oil temperature series is decomposed into 13 modes using CEEMDAN.
[0143] Define the original rail transit power supply system traction transformer top oil temperature time series as S i (t) = S (t) + ε0w i (t), where ε0 is the noise coefficient, i = 0, 1, ..., I, and the original time series is decomposed using EMD, and the first node IMF (Intrinsic Mode Functions, IMFs) signal after decomposition is defined as:
[0144]
[0145] Define the signal residual after the first EMD decomposition:
[0146]
[0147] Decompose the remaining signal after the first EMD decomposition considering white noise and get the node 2 signal:
[0148]
[0149] Repeat the above steps for each node to get the final remaining semaphore:
[0150]
[0151] Among them, M is the total number of decomposed signal node IMFs. CEEMDAN decomposition is used to effectively extract the characteristics of the original signal at different time scales, improve the smoothness of the prediction, and reduce the prediction error.
[0152] Step 3: Use the type-2FIS method to predict the random change time series. The basic steps are as follows:
[0153] Input the random item time series of transformer top oil temperature and establish a rule base based on IF-THEN fuzzy rules;
[0154] Define a database using fuzzy set membership functions;
[0155] Reasoning and decision making are carried out based on fuzzy rules, and particle swarm optimization algorithm is introduced to find the optimal fuzzy rules;
[0156] Fuzzy reasoning;
[0157] Defuzzify and display clear results.
[0158] Step 4: The LSTM prediction model is used for the trend change time series, and the type-2FIS method is used for the random change time series.
[0159] Step 5: Reconstruct the time series based on CEEMDAN decomposition and obtain the remaining service life of the transformer.
[0160] The CEEMDAN decomposition algorithm is used to extract the top oil temperature time series characteristics, initialize the parameters, and decompose it into IMF1, IMF2, …, IMF according to the decomposition steps. 12 、IMF 13 , where IMF1-IMF 12 is a chaotic time series node, IMF 13 It is a monotonic trend sequence node.
[0161] Construct a combined forecasting model for IMF1-IMF 12 Use type-2FIS for prediction, and assume that the prediction results are s′1,s′2,...,s′ 12 ; IMF 13Use LSTM for prediction, and let the prediction result be s′ 13 .
[0162] The predicted oil temperature is obtained by reconstructing the superposition principle using CEEMDAN:
[0163] OT p =s′1+s′2+...+s′ 13
[0164] The obtained OT p Divide it into hours and substitute it into the mathematical model of the transformer's remaining service life to obtain the remaining service life of the traction transformer.
[0165] In summary, this embodiment 3 is used to optimize the reliability assurance process of the rail transit power supply system, and has the following advantages: a mathematical model of the top oil temperature and the remaining service life is established according to the relationship between the internal temperature changes of the transformer, thereby avoiding additional monitoring times; the top oil temperature of the transformer is predicted by combining Type-2 FIS and LSTM to improve the prediction accuracy; it can be used to predict the remaining service life of the traction transformer of the rail transit power supply system to ensure the safe and efficient operation of the rail transit power supply system.
[0166] Example 4
[0167] Embodiment 4 of the present invention provides an electronic device, including a memory and a processor, wherein the processor and the memory communicate with each other, the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute a method for predicting the remaining useful life of a traction transformer, the method comprising the following process steps:
[0168] Obtain the top oil temperature data of the traction transformer of the rail transit power supply system during operation;
[0169] Establish a mathematical model of the remaining service life of the traction transformer based on the internal thermal change process of the transformer;
[0170] A time series model is established based on the top oil temperature data during the operation of the traction transformer, the change trend of the top oil temperature is analyzed and features are extracted to obtain a trend change time series and a random change time series of the top oil temperature;
[0171] Based on the trend change time series and the random change time series, the evolution trend of the transformer top oil temperature in the next period is predicted, and the prediction result is substituted into the remaining service life mathematical model of the traction transformer to obtain the remaining service life of the transformer.
[0172] Example 5
[0173] Embodiment 5 of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, a method for predicting the remaining useful life of a traction transformer is implemented. The method includes the following steps:
[0174] Obtain the top oil temperature data of the traction transformer of the rail transit power supply system during operation;
[0175] Establish a mathematical model of the remaining service life of the traction transformer based on the internal thermal change process of the transformer;
[0176] A time series model is established based on the top oil temperature data during the operation of the traction transformer, the change trend of the top oil temperature is analyzed and features are extracted to obtain a trend change time series and a random change time series of the top oil temperature;
[0177] Based on the trend change time series and the random change time series, the evolution trend of the transformer top oil temperature in the next period is predicted, and the prediction result is substituted into the remaining service life mathematical model of the traction transformer to obtain the remaining service life of the transformer.
[0178] Example 6
[0179] Embodiment 6 of the present invention provides a computer device, including a memory and a processor, the processor and the memory communicate with each other, the memory stores program instructions executable by the processor, the processor calls the program instructions to execute a method for predicting the remaining useful life of a traction transformer, the method comprising the following steps:
[0180] Obtain the top oil temperature data of the traction transformer of the rail transit power supply system during operation;
[0181] Establish a mathematical model of the remaining service life of the traction transformer based on the internal thermal change process of the transformer;
[0182] A time series model is established based on the top oil temperature data during the operation of the traction transformer, the change trend of the top oil temperature is analyzed and features are extracted to obtain a trend change time series and a random change time series of the top oil temperature;
[0183] Based on the trend change time series and the random change time series, the evolution trend of the transformer top oil temperature in the next period is predicted, and the prediction result is substituted into the remaining service life mathematical model of the traction transformer to obtain the remaining service life of the transformer.
[0184] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0185] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0186] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0187] These computer program instructions can also be loaded onto a computer or other programmable data processing device, and a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0188] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative labor on the basis of the technical solution disclosed in the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for predicting the remaining useful life of a traction transformer. It is characterized in that include: Obtain the top oil temperature data of the traction transformer of the rail transit power supply system during operation; The mathematical model of the remaining service life of the traction transformer is established based on the internal thermal change process of the transformer; the model between the remaining service life of the transformer and the hot spot temperature of the winding point is established based on the Arrhenius law: RUL=Cexp(B / 273+θ HST ); Among them, θ HST =T OT +T Δ , T OT is the top oil temperature, T Δ is the local temperature rise caused by the load current; the unit of RUL calculation is hour, C and B are the given insulation material coefficients or system constants; A time series model is established according to the top oil temperature data during the operation of the traction transformer, the change trend of the top oil temperature is analyzed and features are extracted to obtain the trend change time series and random change time series of the top oil temperature; including: defining the top oil temperature time series of the original rail transit power supply system traction transformer; using EMD to decompose the original time series, defining the first node IMF signal after decomposition: defining the signal residual amount after the first EMD decomposition; decomposing the residual signal amount after the first EMD decomposition to obtain the second node signal: repeating the above steps for each node to obtain the final residual signal amount; Based on the trend change time series and the random change time series, the evolution trend of the transformer top oil temperature in the next period is predicted, and the prediction results are substituted into the mathematical model of the remaining service life of the traction transformer to obtain the remaining service life of the transformer; including: using the long short-term memory neural network LSTM prediction model for the trend change time series, using the type-2FIS method to predict the random change time series, and establishing a combined prediction model; using the type-2FIS method to predict the random change time series, including: inputting the random item time series of the transformer top oil temperature, establishing a rule base based on IF-THEN fuzzy rules; defining the use of fuzzy set membership function database; make inference decisions based on fuzzy rules, and introduce particle swarm optimization algorithm to find the optimal fuzzy rules; fuzzy reasoning; defuzzify and display clear results; use CEEMDAN decomposition algorithm to extract the top oil temperature time series characteristics, decompose to obtain chaotic time series nodes and monotonic trend series nodes; build a combined prediction model, use type-2FIS to predict chaotic time series nodes, and use LSTM to predict monotonic trend series nodes; use CEEMDAN to reconstruct the superposition principle to obtain the predicted oil temperature, divide the predicted oil temperature into hours, and substitute it into the mathematical model of the remaining service life of the transformer to obtain the remaining service life of the traction transformer.
2. The method for predicting the remaining useful life of a traction transformer according to claim 1, It is characterized in that A model RUL is established between the remaining service life of the transformer and the hot spot temperature of the transformer winding point. For the working loss of the traction transformer on a certain day, after 24 hours of operation, the remaining effective service life is defined as the difference between the ideal RUL obtained by the Arrhenius law and the loss on that day.
3. A system for predicting the remaining useful life of a traction transformer, It is characterized in that include: An acquisition module, used to acquire top oil temperature data of a traction transformer of a rail transit power supply system during operation; The construction module is used to establish a mathematical model of the remaining service life of the traction transformer according to the internal thermal change process of the transformer; the model between the remaining service life of the transformer and the hot spot temperature of the winding point is established according to the Arrhenius law: RUL=Cexp(B / 273+θ HST ); Among them, θ HST =T OT +T Δ , T OT is the top oil temperature, T Δ is the local temperature rise caused by the load current; the unit of RUL calculation is hour, C and B are the given insulation material coefficients or system constants; The extraction module is used to establish a time series model according to the top oil temperature data during the operation of the traction transformer, analyze the change trend of the top oil temperature and perform feature extraction to obtain the trend change time series and random change time series of the top oil temperature; including: defining the original rail transit power supply system traction transformer top oil temperature time series; using EMD to decompose the original time series, defining the first node IMF signal after decomposition: defining the signal residual amount after the first EMD decomposition; decomposing the residual signal amount after the first EMD decomposition to obtain the second node signal: repeating the above steps for each node to obtain the final residual signal amount; The prediction module is used to predict the evolution trend of the transformer top oil temperature in the next period based on the trend change time series and the random change time series, and substitute the prediction results into the mathematical model of the remaining service life of the traction transformer to obtain the remaining service life of the transformer; including: using the long short-term memory neural network LSTM prediction model for the trend change time series, using the type-2FIS method to predict the random change time series, and establishing a combined prediction model; using the type-2FIS method to predict the random change time series, including: inputting the random item time series of the transformer top oil temperature, establishing a rule base based on IF-THEN fuzzy rules; defining the use of fuzzy set membership The database of degree function is constructed; reasoning and decision making is performed based on fuzzy rules, and the particle swarm optimization algorithm is introduced to find the optimal fuzzy rules; fuzzy reasoning; defuzzification and clear display of results; the top oil temperature time series features are extracted using the CEEMDAN decomposition algorithm, and chaotic time series nodes and monotonic trend series nodes are decomposed; a combined prediction model is constructed, and type-2FIS is used to predict the chaotic time series nodes, and LSTM is used to predict the monotonic trend series nodes; the predicted oil temperature is obtained by reconstructing the superposition principle of CEEMDAN, and the predicted oil temperature is divided into hours, and substituted into the remaining service life mathematical model of the transformer to obtain the remaining service life of the traction transformer.
4. A computer device, comprising a memory and a processor, wherein the processor and the memory communicate with each other, the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the remaining service life prediction method for a traction transformer as described in claim 1 or 2.
5. An electronic device, It is characterized in that It includes a memory and a processor, the processor and the memory communicate with each other, the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the remaining service life prediction method of the traction transformer as claimed in claim 1 or 2.
6. A computer-readable storage medium, It is characterized in that It stores a computer program, which, when executed by a processor, implements the remaining service life prediction method of the traction transformer as claimed in claim 1 or 2.
