Locomotive high-voltage side electric energy quality evaluation method, device and equipment, storage medium and program product

Abstract

The invention discloses an electric energy quality evaluation method, device and equipment of a locomotive high-voltage side, a storage medium and a program product. The method comprises the following steps: acquiring network voltage and primary side current corresponding to a target locomotive in real time; constructing a data set corresponding to the target locomotive based on the network voltage and the primary side current; calling a pre-trained electric energy quality evaluation model, and determining a target electric energy quality evaluation result of the target locomotive based on the data set; wherein the power quality evaluation model is obtained by training a pre-constructed transfer learning model by using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the data volume in the first historical data set is far smaller than the data volume in the second historical data set. By adopting the method, the accuracy of the electric energy quality evaluation result can be improved for the electric locomotive with less data volume of the operation data.

Description

Technical Field

[0001] The present application relates to the technical field of electric locomotives, and in particular to a method, apparatus, device, storage medium, and program product for evaluating the power quality of a locomotive high-voltage side. Background Art

[0002] Currently, power quality assessment for electric locomotives primarily relies on acquiring locomotive operating data using sensors or other means, correlating this data with useful information, and then using intelligent algorithms and models for detection, analysis, application, and evaluation to generate power quality assessment results. However, for locomotives with limited operating data, this approach can result in lower accuracy in power quality assessment results.

[0003] Therefore, how to improve the accuracy of power quality assessment results for electric locomotives with less operating data has become an urgent problem to be solved. Summary of the Invention

[0004] The embodiments of the present application provide a method, apparatus, device, storage medium and program product for evaluating the power quality of a locomotive on the high-voltage side, which can improve the accuracy of the power quality evaluation results for electric locomotives with a small amount of operating data.

[0005] In a first aspect, an embodiment of the present application provides a method for evaluating the power quality of a locomotive high-voltage side, the method comprising:

[0006] Obtain the grid voltage and primary current corresponding to the target locomotive in real time;

[0007] Based on the grid voltage and primary current, a data set corresponding to the target locomotive is constructed;

[0008] Call the pre-trained power quality assessment model and determine the target power quality assessment result of the target locomotive based on the data set;

[0009] The power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set;

[0010] The first historical data set includes a first historical grid voltage and a first historical primary current corresponding to the target locomotive; the second historical data set includes a second locomotive grid voltage and a second historical primary current corresponding to the source locomotive.

[0011] In one embodiment, a pre-trained power quality assessment model is called to determine a target power quality assessment result for a target locomotive based on a data set, including: normalizing the data set to obtain a normalized data set; inputting the normalized data set into the pre-trained power quality assessment model to obtain an initial power quality assessment result for the target locomotive; and denormalizing the initial power quality assessment result to obtain a target power quality assessment result for the target locomotive.

[0012] In one embodiment, the transfer learning model is constructed in the following manner: constructing a first input layer for inputting data corresponding to the target locomotive, and constructing a second input layer for inputting data corresponding to the source locomotive; constructing L latent layers, wherein the first L-1 latent layers are shared latent layers, and the Lth latent layer includes a first latent layer and a second latent layer, the first latent layer is associated with the data corresponding to the target locomotive, and the second latent layer is associated with the data corresponding to the source locomotive; constructing a first output layer for outputting the power quality of the target locomotive, and constructing a second output layer for outputting the power quality of the source locomotive; constructing a transfer learning model based on the first input layer, the second input layer, the L latent layers, the first output layer, and the second output layer.

[0013] In one embodiment, the method further includes: obtaining a first historical grid voltage and a first historical primary current of the target locomotive in each first historical period, and obtaining a second historical grid voltage and a second historical primary current of the source locomotive in each second historical period; wherein the number of first historical periods is much smaller than the number of second historical periods; determining a first grid voltage characteristic parameter corresponding to each first historical grid voltage, and determining a second grid voltage characteristic parameter corresponding to each second historical grid voltage; for each first historical grid voltage, adding a first label to the first historical grid voltage based on the first grid voltage characteristic parameter corresponding to the first historical grid voltage, and, for each second historical grid voltage, adding a first label to the first historical grid voltage based on the second historical grid voltage corresponding to the second historical grid voltage. Grid voltage characteristic parameters, adding a second label to the second historical grid voltage; the first label is used to characterize the actual power quality corresponding to the first historical grid voltage, and the second label is used to characterize the actual power quality corresponding to the second historical grid voltage; based on multiple first historical grid voltages, multiple first historical primary currents and the first label of each first historical grid voltage, a first historical data set is constructed, and, based on multiple second historical grid voltages, multiple second historical primary currents and the second label of each second historical grid voltage, a second historical data set is constructed; based on the first historical data set and the second historical data set, a transfer learning model is trained to obtain a trained transfer learning model; and the trained transfer learning model is used as a power quality assessment model.

[0014] In one embodiment, determining the first grid voltage characteristic parameter corresponding to each first historical grid voltage, and determining the second grid voltage characteristic parameter corresponding to each second historical grid voltage, includes: for each first historical grid voltage, determining multiple first historical grid voltage signals, multiple first fundamental grid voltage signals, and first fundamental amplitude of the first historical grid voltage within the first historical period; based on the multiple first historical grid voltage signals, multiple first fundamental grid voltage signals, and first fundamental amplitude, determining the first harmonic content corresponding to the first historical grid voltage within the first historical period, and using the first harmonic content as the first grid voltage characteristic parameter corresponding to the first historical grid voltage; for each second historical grid voltage, determining multiple second historical grid voltage signals, multiple second fundamental grid voltage signals, and second fundamental amplitude of the second historical grid voltage within the second historical period; based on the multiple second historical grid voltage signals, multiple second fundamental grid voltage signals, and second fundamental amplitude, determining the second harmonic content corresponding to the second historical grid voltage within the second historical period, and using the second harmonic content as the second grid voltage characteristic parameter corresponding to the second historical grid voltage.

[0015] In one embodiment, a transfer learning model is trained based on a first historical data set and a second historical data set to obtain a trained transfer learning model, including: normalizing the first historical data set to obtain a first training set, and normalizing the second historical data set to obtain a second training set; inputting the first training set and the second training set into the transfer learning model to obtain a first predicted power quality corresponding to each normalized first historical grid voltage in the first training set and a second predicted power quality corresponding to each normalized second historical grid voltage in the second training set; performing denormalization processing on multiple first predicted power qualities to obtain multiple processed first predicted power qualities, and performing denormalization processing on multiple second predicted power qualities to obtain multiple processed second predicted power qualities; training the transfer learning model in a direction of reducing the difference between each processed first predicted power quality and the corresponding first label, and in a direction of reducing the difference between each processed second predicted power quality and the corresponding second label to obtain a trained transfer learning model.

[0016] In a second aspect, the present application provides a device for evaluating the power quality of a locomotive high-voltage side, the device comprising:

[0017] Acquisition module, used to obtain the grid voltage and primary current corresponding to the target locomotive in real time;

[0018] A construction module is used to construct a data set corresponding to the target locomotive based on the grid voltage and primary current;

[0019] An evaluation module is used to call a pre-trained power quality evaluation model and determine a target power quality evaluation result for a target locomotive based on a data set;

[0020] The power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set;

[0021] The first historical data set includes a first historical grid voltage and a first historical primary current corresponding to the target locomotive; the second historical data set includes a second locomotive grid voltage and a second historical primary current corresponding to the source locomotive.

[0022] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are performed:

[0023] Obtain the grid voltage and primary current corresponding to the target locomotive in real time;

[0024] Based on the grid voltage and primary current, a data set corresponding to the target locomotive is constructed;

[0025] Call the pre-trained power quality assessment model and determine the target power quality assessment result of the target locomotive based on the data set;

[0026] The power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set;

[0027] The first historical data set includes a first historical grid voltage and a first historical primary current corresponding to the target locomotive; the second historical data set includes a second locomotive grid voltage and a second historical primary current corresponding to the source locomotive.

[0028] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0029] Obtain the grid voltage and primary current corresponding to the target locomotive in real time;

[0030] Based on the grid voltage and primary current, a data set corresponding to the target locomotive is constructed;

[0031] Call the pre-trained power quality assessment model and determine the target power quality assessment result of the target locomotive based on the data set;

[0032] The power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set;

[0033] The first historical data set includes a first historical grid voltage and a first historical primary current corresponding to the target locomotive; the second historical data set includes a second locomotive grid voltage and a second historical primary current corresponding to the source locomotive.

[0034] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the following steps:

[0035] Obtain the grid voltage and primary current corresponding to the target locomotive in real time;

[0036] Based on the grid voltage and primary current, a data set corresponding to the target locomotive is constructed;

[0037] Call the pre-trained power quality assessment model and determine the target power quality assessment result of the target locomotive based on the data set;

[0038] The power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set;

[0039] The first historical data set includes a first historical grid voltage and a first historical primary current corresponding to the target locomotive; the second historical data set includes a second locomotive grid voltage and a second historical primary current corresponding to the source locomotive.

[0040] The above-mentioned power quality assessment method, device, equipment, storage medium and program product for the high-voltage side of the locomotive, the computer equipment can obtain the grid voltage and primary current corresponding to the target locomotive in real time; based on the grid voltage and primary current, construct a data set corresponding to the target locomotive; call a pre-trained power quality assessment model, and determine the target power quality assessment result of the target locomotive based on the data set; wherein the power quality assessment model is obtained by training a pre-constructed transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set; the first historical data set includes the first historical grid voltage and the first historical primary current corresponding to the target locomotive; the second historical data set includes the second locomotive grid voltage and the second historical primary current corresponding to the source locomotive. Using this method, since the power quality assessment model is obtained by training a pre-constructed transfer learning model using a second historical data set with a large amount of operating data (i.e., the historical data set corresponding to the source locomotive) and a first historical data set with a small amount of operating data (i.e., the historical data set corresponding to the target locomotive), the pre-trained power quality assessment model has already learned the data characteristics of each data in the second historical data set as well as the data characteristics of each data in the first historical data set. Therefore, for electric locomotives with a small amount of operating data, the pre-trained power quality assessment model can be used to improve the accuracy of power quality assessment results. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 This is a schematic diagram of an application scenario of a method for evaluating the power quality on the high-voltage side of a locomotive provided in an embodiment of the present application;

[0043] Figure 2 This is a flow chart of a method for evaluating the power quality of a locomotive high-voltage side provided in an embodiment of the present application;

[0044] Figure 3 This is a schematic diagram of the structure of a transfer learning model provided in an embodiment of the present application;

[0045] Figure 4 This is a flow chart of another method for evaluating the power quality of a locomotive high-voltage side provided in an embodiment of the present application;

[0046] Figure 5This is a schematic structural diagram of a power quality assessment device for the high-voltage side of a locomotive provided in an embodiment of the present application;

[0047] Figure 6 It is a structural diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0049] The following introduces the application scenarios of the power quality assessment method for the high-voltage side of a locomotive provided in the embodiment of the present application.

[0050] See Figure 1 , Figure 1 This is a schematic diagram of an application scenario of a method for evaluating the power quality of a locomotive high-voltage side provided in an embodiment of the present application. Figure 1 As shown, it includes a computer device 101 and a database server 102.

[0051] The database server 102 stores the grid voltage and primary current corresponding to the target locomotive, as well as the grid voltage and primary current corresponding to the source locomotive. The computer device 101 can obtain the grid voltage and primary current corresponding to the target locomotive from the database server 102; construct a data set corresponding to the target locomotive based on the grid voltage and primary current; call a pre-trained power quality assessment model, and determine the target power quality assessment result of the target locomotive based on the data set; wherein the power quality assessment model is obtained by training a pre-constructed transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set; the first historical data set includes the first historical grid voltage and the first historical primary current corresponding to the target locomotive; the second historical data set includes the second locomotive grid voltage and the second historical primary current corresponding to the source locomotive. Using this method, since the power quality assessment model is obtained by training a pre-constructed transfer learning model using a second historical data set with a large amount of operating data (i.e., the historical data set corresponding to the source locomotive) and a first historical data set with a small amount of operating data (i.e., the historical data set corresponding to the target locomotive), the pre-trained power quality assessment model has already learned the data characteristics of each data in the second historical data set as well as the data characteristics of each data in the first historical data set. Therefore, for electric locomotives with a small amount of operating data, the pre-trained power quality assessment model can be used to improve the accuracy of power quality assessment results.

[0052] Optionally, computer device 101 may be a terminal device or a server. The terminal devices mentioned herein may include, but are not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices may include smart TVs, smart air conditioners, smart car devices, and projection devices. Portable wearable devices may include smart watches, smart bracelets, and head-mounted devices. The server mentioned herein may be an independent physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server providing cloud computing services, etc., without limitation.

[0053] See Figure 2 , Figure 2 1 is a flow chart of a method for evaluating the power quality of a locomotive high voltage side provided by an embodiment of the present application. The method can be executed by a computer device (for example, the computer device 101 described above). Figure 2 As shown, the power quality assessment method for the high-voltage side of the locomotive may include but is not limited to the following steps:

[0054] S201. Obtain the grid voltage and primary current corresponding to the target locomotive in real time.

[0055] The target locomotive is a locomotive with a relatively small amount of operating data, for example, a locomotive running on a new line, or a locomotive newly put into operation.

[0056] In an optional embodiment, the computer device obtains the grid voltage and primary current corresponding to the target locomotive in real time, which may be the network voltage and primary current of the target locomotive after the pantograph is raised.

[0057] S202: Construct a data set corresponding to the target locomotive based on the grid voltage and the primary current.

[0058] In an optional implementation, the computer device constructs a data set corresponding to the target locomotive based on the grid voltage and primary current, and may use a set consisting of the grid voltage and primary current corresponding to the target locomotive as the data set corresponding to the target locomotive.

[0059] S203. Call a pre-trained power quality assessment model and determine a target power quality assessment result for the target locomotive based on the data set; wherein the power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set.

[0060] The first historical data set includes the first historical grid voltage and the first historical primary current corresponding to the target locomotive; the second historical data set includes the second locomotive grid voltage and the second historical primary current corresponding to the source locomotive.

[0061] In an optional embodiment, the target power quality assessment result of the target locomotive is the grid voltage assessment result of the target locomotive. Optionally, the target power quality assessment result may include but is not limited to normal, weakened, abnormal, faulty, etc., which are not limited here.

[0062] In an optional embodiment, before step S203, the computer device may collect the first historical grid voltage and the first historical primary current corresponding to the target locomotive, and construct a first historical data set based on the first historical grid voltage and the first historical primary current, and collect the second historical grid voltage and the second historical primary current corresponding to the source locomotive, and construct a second historical data set based on the second historical grid voltage and the second historical primary current.

[0063] In this embodiment, the computer device may also store the first historical data set and the second historical data set in a local database, or the computer device may upload the first historical data set and the second historical data set to a database server so that the database server stores the first historical data set and the second historical data set.

[0064] In an optional embodiment, after step S203, the computer device may further output the target power quality assessment result of the target locomotive, which is helpful for the staff to take corresponding treatment measures based on the target power quality assessment result.

[0065] In an embodiment of the present application, a computer device can obtain the grid voltage and primary current corresponding to the target locomotive in real time; based on the grid voltage and primary current, a data set corresponding to the target locomotive is constructed; a pre-trained power quality assessment model is called, and based on the data set, a target power quality assessment result of the target locomotive is determined; wherein the power quality assessment model is obtained by training a pre-constructed transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set; the first historical data set includes the first historical grid voltage and the first historical primary current corresponding to the target locomotive; the second historical data set includes the second locomotive grid voltage and the second historical primary current corresponding to the source locomotive. Using this method, since the power quality assessment model is obtained by training a pre-constructed transfer learning model using a second historical data set with a large amount of operating data (i.e., the historical data set corresponding to the source locomotive) and a first historical data set with a small amount of operating data (i.e., the historical data set corresponding to the target locomotive), the pre-trained power quality assessment model has already learned the data characteristics of each data in the second historical data set as well as the data characteristics of each data in the first historical data set. Therefore, for electric locomotives with a small amount of operating data, the pre-trained power quality assessment model can be used to improve the accuracy of power quality assessment results.

[0066] In an optional embodiment, Figure 2 In the power quality assessment method for the high-voltage side of a locomotive shown, a computer device calls a pre-trained power quality assessment model and determines a target power quality assessment result for a target locomotive based on a data set. The method may include: normalizing the data set to obtain a normalized data set; inputting the normalized data set into the pre-trained power quality assessment model to obtain an initial power quality assessment result for the target locomotive; and denormalizing the initial power quality assessment result to obtain a target power quality assessment result for the target locomotive.

[0067] For example, suppose the dataset is X 当前 , X 当前 = , where l represents the target locomotive; It represents the grid pressure of the target locomotive at the current moment; It represents the primary current of the target locomotive at the current moment, so the computer equipment can 当前 Perform normalization to obtain a normalized data set , , where l represents the target locomotive; It represents the normalized grid pressure of the target locomotive; It represents the normalized primary current of the target locomotive.

[0068] By adopting this embodiment, the computer device can obtain a more accurate initial power quality assessment result by inputting the normalized data set corresponding to the target locomotive into a pre-trained power quality assessment model. Thus, by performing denormalization processing on the more accurate initial power quality assessment result, a more accurate target power quality assessment result can be obtained. That is, by adopting this embodiment, the accuracy of the power quality assessment can be improved.

[0069] In an optional embodiment, Figure 2 In the power quality assessment method for the high-voltage side of a locomotive shown, the transfer learning model can be constructed by a computer device in the following manner: constructing a first input layer for inputting data corresponding to a target locomotive, and constructing a second input layer for inputting data corresponding to a source locomotive; constructing L latent layers, wherein the first L-1 latent layers are shared latent layers, and the Lth latent layer includes a first latent layer and a second latent layer, the first latent layer is associated with the data corresponding to the target locomotive, and the second latent layer is associated with the data corresponding to the source locomotive; constructing a first output layer for outputting the power quality of the target locomotive, and constructing a second output layer for outputting the power quality of the source locomotive; constructing a transfer learning model based on the first input layer, the second input layer, the L latent layers, the first output layer, and the second output layer.

[0070] See Figure 3 , Figure 3 This is a schematic diagram of the structure of a transfer learning model provided in the embodiment of this application. Figure 3 As shown, it includes input layer, latent layer and output layer.

[0071] The input layer includes a first input layer 2 and a second input layer. The first input layer is used to input data corresponding to the target locomotive, and the second input layer is used to input data corresponding to the source locomotive. The hidden layer includes the first L-1 shared hidden layers and the Lth independent hidden layer. The Lth independent hidden layer includes the first hidden layer and the second hidden layer. The first hidden layer is used to reprocess the data corresponding to the target locomotive output by the first L-1 hidden layers, and the second hidden layer is used to reprocess the data corresponding to the source locomotive output by the first L-1 hidden layers. The output layer includes a first output layer 2 and a second output layer. The first output layer is used to output the power quality assessment result of the target locomotive based on the data corresponding to the target locomotive output by the first hidden layer in the Lth independent hidden layer. The second output layer is used to output the power quality assessment result of the source locomotive based on the data corresponding to the source locomotive output by the second hidden layer in the Lth independent hidden layer.

[0072] By adopting this embodiment, the computer device can adopt the above-mentioned transfer learning model. When the amount of locomotive data is insufficient or the data conditions are poor, the computer device can use the data of locomotives with sufficient data (i.e., the data corresponding to the source locomotive) for sufficient training to obtain a trained transfer learning model. Thus, the trained transfer learning model (i.e., the trained power quality assessment model) can be used to accurately assess the power quality of locomotives (target locomotives) with insufficient locomotive data or poor data conditions.

[0073] In an optional embodiment, Figure 2 In the power quality assessment method for the high-voltage side of a locomotive shown, the power quality assessment model can be obtained by a computer device through the following steps:

[0074] Step 1: Obtain the first historical grid voltage and the first historical primary current of the target locomotive in each first historical period, and obtain the second historical grid voltage and the second historical primary current of the source locomotive in each second historical period; wherein the number of first historical periods is much smaller than the number of second historical periods.

[0075] The target locomotive refers to a locomotive with a smaller amount of operating data; the source locomotive refers to a locomotive with a larger amount of operating data.

[0076] Optionally, the first historical grid voltage of the target locomotive (denoted as l) in each first historical period can be expressed as , the first historical primary current can be expressed as . Where t=1, 2, ..., N l , N l It represents the number of sampling moments corresponding to the target locomotive l.

[0077] Optionally, the second historical grid pressure of the source locomotive (denoted as s) in each second historical period can be expressed as , the second historical primary current can be expressed as . Where t=1, 2, ..., N s , N s It represents the number of sampling moments corresponding to the source locomotive s, N s >>N l .

[0078] For example, the computer device may collect the first historical grid voltage, the first historical primary current, the second historical grid voltage, and the second historical primary current every 122 ns, i.e., the sampling period is 122 ns. The computer device may collect the grid voltage and primary current data after the target locomotive is raised for 5 days, and collect the grid voltage and primary current data after the source locomotive is raised for 30 days.

[0079] Step 2: Determine a first grid voltage characteristic parameter corresponding to each first historical grid voltage, and determine a second grid voltage characteristic parameter corresponding to each second historical grid voltage.

[0080] In some embodiments, the computer device determines the first grid voltage characteristic parameter corresponding to each first historical grid voltage, and determines the second grid voltage characteristic parameter corresponding to each second historical grid voltage, which may include: for each first historical grid voltage, determining multiple first historical grid voltage signals, multiple first fundamental grid voltage signals and first fundamental amplitude of the first historical grid voltage within the first historical period; based on the multiple first historical grid voltage signals, multiple first fundamental grid voltage signals and first fundamental amplitude, determining the first harmonic content corresponding to the first historical grid voltage within the first historical period, and using the first harmonic content as the first grid voltage characteristic parameter corresponding to the first historical grid voltage; for each second historical grid voltage, determining multiple second historical grid voltage signals, multiple second fundamental grid voltage signals and second fundamental amplitude of the second historical grid voltage within the second historical period; based on the multiple second historical grid voltage signals, multiple second fundamental grid voltage signals and second fundamental amplitude, determining the second harmonic content corresponding to the second historical grid voltage within the second historical period, and using the second harmonic content as the second grid voltage characteristic parameter corresponding to the second historical grid voltage.

[0081] Optionally, the computer device determines, for each first historical network voltage, a plurality of first historical network voltage signals, a plurality of first fundamental wave network voltage signals and a first fundamental wave amplitude within the first historical period of the first historical network voltage. , extract the first historical network pressure A plurality of first historical grid voltage signals (denoted as , wherein 1≤i≤N); the first historical network pressure Perform bandpass filtering to obtain the first historical network voltage A plurality of first fundamental grid voltage signals (denoted as , where 1≤i≤N), and based on multiple first fundamental wave grid voltage signals, determine the first fundamental wave amplitude (denoted as ). The frequency of the first historical grid voltage signal is 50 Hz; the first historical period N=0.02s.

[0082] Optionally, when the computer device determines the first harmonic content corresponding to the first historical grid voltage in the first historical period based on multiple first historical grid voltage signals, multiple first fundamental wave grid voltage signals and the first fundamental wave amplitude, the following formula (1) may be used.

[0083] (1)

[0084] In formula (1), It represents the harmonic content of the target locomotive in the first historical period N; It represents the i-th first historical grid voltage signal within the first historical period N of the first historical grid voltage; It represents the first fundamental wave grid voltage signal of the first historical grid voltage within the first historical period N; It represents the first fundamental wave amplitude of the first historical grid voltage within the first historical period N.

[0085] Optionally, the computer device determines, for each second historical network voltage, a plurality of second historical network voltage signals, a plurality of second fundamental wave network voltage signals and a second fundamental wave amplitude within the second historical period of the second historical network voltage. , extract the second historical network pressure A plurality of second historical grid voltage signals (denoted as , wherein 1≤i≤N); the second historical network pressure Perform bandpass filtering to obtain the second historical network voltage A plurality of second fundamental wave grid voltage signals (denoted as , where 1≤i≤N), and based on multiple second fundamental wave grid voltage signals, determine the second fundamental wave amplitude (denoted as ). The frequency of the second historical grid voltage signal is 50 Hz; the second historical period N=0.02s.

[0086] Optionally, when the computer device determines the second harmonic content corresponding to the second historical grid voltage in the second historical period based on multiple second historical grid voltage signals, multiple second fundamental wave grid voltage signals and second fundamental wave amplitudes, the following formula (2) may be used.

[0087] (2)

[0088] In formula (2), It represents the harmonic content of the source locomotive in one second historical period N; It represents the i-th second historical grid voltage signal within the second historical period N of the second historical grid voltage; It represents the second fundamental wave grid voltage signal of the second historical grid voltage in the second historical period N; It represents the second fundamental wave amplitude of the second historical network pressure within the second historical period N.

[0089] Step 3: For each first historical grid voltage, add a first label to the first historical grid voltage based on the first grid voltage characteristic parameter corresponding to the first historical grid voltage; and, for each second historical grid voltage, add a second label to the second historical grid voltage based on the second grid voltage characteristic parameter corresponding to the second historical grid voltage.

[0090] The first tag is used to represent the actual power quality corresponding to the first historical grid voltage, and the second tag is used to represent the actual power quality corresponding to the second historical grid voltage.

[0091] In some embodiments, the computer device adds a label to each first historical grid voltage based on the first grid voltage characteristic parameter corresponding to the first historical grid voltage, and the label may be added to the first historical grid voltage corresponding to the first historical period based on the harmonic content of the target locomotive in a first historical period N. Optionally, the computer device may use the rule shown in the following expression (3) to add a label to the first historical network voltage corresponding to the first historical period: .

[0092] (3)

[0093] In formula (3), It represents the harmonic content of the target locomotive in the first historical period N, which can be obtained by the above formula (1).

[0094] From formula (3), we can see that ≤15%, it indicates that the first historical grid voltage corresponding to the first historical period is normal; 15%≤ ≤50%, it indicates that the first historical network voltage corresponding to the first historical period is weakened; 50%≤ ≤75%, it indicates that the first historical grid voltage corresponding to the first historical period is abnormal; ≥75%, it indicates that the first historical grid voltage fault corresponding to the first historical period occurred.

[0095] Among them, when the first historical network voltage is normal, the computer device can add a label to the first historical network voltage corresponding to the first historical period. =1; when the first historical network voltage is weakened, the computer device may add a label to the first historical network voltage corresponding to the first historical period. =2; in the case of an abnormal first historical network voltage, the computer device may add a label to the first historical network voltage corresponding to the first historical period. =3; In the event of a first historical network voltage failure, the computer device may add a label to the first historical network voltage corresponding to the first historical period. =4.

[0096] In some embodiments, the computer device adds a label to each second historical grid voltage based on the second grid voltage characteristic parameter corresponding to the second historical grid voltage, and the label can be added to the second historical grid voltage corresponding to the second historical period based on the harmonic content of the source locomotive in a second historical period N. Optionally, the computer device may use the rule shown in the following expression (4) to add a label to the second historical network voltage corresponding to the second historical period: .

[0097] (4)

[0098] In formula (4), It represents the harmonic content of the source locomotive in a second historical period N, which can be obtained by the above formula (2).

[0099] From formula (4), we can see that ≤15%, it indicates that the second historical grid voltage corresponding to the second historical period is normal; 15%≤ ≤50%, it indicates that the second historical network voltage corresponding to the second historical period is weakened; 50%≤ ≤75%, it indicates that the second historical grid voltage corresponding to the second historical period is abnormal; ≥75%, it indicates that the second historical grid voltage corresponding to the second historical period is faulty.

[0100] Wherein, when the second historical network voltage is normal, the computer device can add a label to the second historical network voltage corresponding to the second historical period. =1; when the second historical network voltage is weakened, the computer device may add a label to the second historical network voltage corresponding to the second historical period. =2; in the case of an abnormality in the second historical network voltage, the computer device may add a label to the second historical network voltage corresponding to the second historical period. =3; in the event of a second historical network voltage failure, the computer device may add a label to the second historical network voltage corresponding to the second historical period =4.

[0101] Step 4: Construct a first historical data set based on multiple first historical grid voltages, multiple first historical primary currents, and a first label for each first historical grid voltage; and construct a second historical data set based on multiple second historical grid voltages, multiple second historical primary currents, and a second label for each second historical grid voltage.

[0102] The first historical data set can be expressed as the following formula (5).

[0103] (5)

[0104] The second historical data set can be expressed as the following formula (6).

[0105] (6)

[0106] Step 5: Based on the first historical data set and the second historical data set, the transfer learning model is trained to obtain a trained transfer learning model.

[0107] In some embodiments, a computer device trains a transfer learning model based on a first historical data set and a second historical data set to obtain a trained transfer learning model, which may include: normalizing the first historical data set to obtain a first training set, and normalizing the second historical data set to obtain a second training set; inputting the first training set and the second training set into the transfer learning model to obtain a first predicted power quality corresponding to each normalized first historical grid voltage in the first training set and a second predicted power quality corresponding to each normalized second historical grid voltage in the second training set; performing denormalization processing on multiple first predicted power qualities to obtain multiple processed first predicted power qualities, and performing denormalization processing on multiple second predicted power qualities to obtain multiple processed second predicted power qualities; training the transfer learning model in a direction of reducing the difference between each processed first predicted power quality and the corresponding first label, and in a direction of reducing the difference between each processed second predicted power quality and the corresponding second label to obtain a trained transfer learning model.

[0108] The first training set can be expressed as the following formula (7).

[0109] (7)

[0110] The second training set can be expressed as the following formula (8).

[0111] (8)

[0112] The following transfer learning model is Figure 3 Taking the transfer learning model shown as an example, the process of inputting the first training set and the second training set into the transfer learning model by a computer device to obtain the first predicted power quality corresponding to each normalized first historical grid voltage in the first training set and the second predicted power quality corresponding to each normalized second historical grid voltage in the second training set is explained.

[0113] First, the computer device can input X of the first training set l Input to the first input layer and get the output of the input layer corresponding to the target locomotive ;in, The expression of can be shown as the following formula (9).

[0114] (9)

[0115] In formula (9), represents the output of the first input layer; It means X l The i-th value in ; It represents the weight of the i-th neuron; represents the threshold of the i-th neuron; f() represents the activation function, such as the sigmoid function, that is, .

[0116] The computer device can also input X of the second training set s Input to the second input layer to get the output of the input layer corresponding to the source locomotive ;in, The expression of can be shown as the following formula (10).

[0117] (10)

[0118] In formula (10), represents the output of the second input layer; It means X s The i-th value in ; It represents the weight of the i-th neuron; represents the threshold of the i-th neuron; f() represents the activation function, such as the sigmoid function, that is, .

[0119] Secondly, the computer device can output the input layer corresponding to the target locomotive The output of the input layer corresponding to the source locomotive All are input into the first L-1 layer of latent layer, and the first L-1 layer of latent layer is trained at the same time to obtain the output of the first L-1 layer of latent layer corresponding to the target locomotive. The output of the first L-1 hidden layers corresponding to the source locomotive .

[0120] Then, the computer device can convert the output of the first L-1 hidden layer corresponding to the target locomotive into Input into the first latent layer of the Lth latent layer, and get the output of the Lth latent layer corresponding to the target locomotive , and the output of the first L-1 hidden layers corresponding to the source locomotive Input into the second latent layer of the Lth latent layer, and get the output of the Lth latent layer corresponding to the source locomotive .

[0121] Afterwards, the computer device can output the Lth hidden layer corresponding to the target locomotive Input to the first output layer and get the output of the output layer corresponding to the target locomotive (i.e. the first predicted power quality), and the output of the Lth latent layer corresponding to the source locomotive Input to the second output layer to get the output of the output layer corresponding to the source locomotive (i.e. the second predicted power quality).

[0122] Optionally, the computer device trains the transfer learning model in a direction of reducing the difference between each processed first predicted power quality and the corresponding first label, and in a direction of reducing the difference between each processed second predicted power quality and the corresponding second label to obtain a trained transfer learning model. The computer device may train the transfer learning model with the goal of determining the minimum value of the error function shown in the following formulas (11) and (12) to obtain a trained transfer learning model.

[0123] (11)

[0124] (12)

[0125] in, It represents the error between the processed first predicted power quality corresponding to the target locomotive and the corresponding first label; It represents the error between the processed second predicted power quality corresponding to the source locomotive and the corresponding second label.

[0126] Step 6: Use the trained transfer learning model as a power quality assessment model.

[0127] By adopting this implementation mode, the computer device can obtain a trained transfer learning model by training the pre-built transfer learning model, and use the trained transfer learning model as a power quality assessment model, which can be beneficial to the subsequent assessment of the power quality of the target locomotive and improve the accuracy of the power quality assessment.

[0128] The following is an overall description of the power quality assessment method for the high-voltage side of a locomotive provided in the embodiment of the present application. Figure 4 , Figure 4 This is a flow chart of another method for evaluating the power quality of a locomotive high-voltage side provided by an embodiment of the present application. Figure 4 As shown, the power quality assessment method for the high-voltage side of the locomotive may include but is not limited to the following steps:

[0129] S401. Obtain a first historical grid voltage and a first historical primary current of a target locomotive in each first historical period, and obtain a second historical grid voltage and a second historical primary current of a source locomotive in each second historical period.

[0130] Among them, the number of the first historical period is much smaller than that of the second historical period.

[0131] S402: Determine a first grid voltage characteristic parameter corresponding to each first historical grid voltage, and determine a second grid voltage characteristic parameter corresponding to each second historical grid voltage.

[0132] S403. For each first historical grid voltage, add a first label to the first historical grid voltage based on a first grid voltage characteristic parameter corresponding to the first historical grid voltage; and, for each second historical grid voltage, add a second label to the second historical grid voltage based on a second grid voltage characteristic parameter corresponding to the second historical grid voltage.

[0133] The first tag is used to represent the actual power quality corresponding to the first historical grid voltage, and the second tag is used to represent the actual power quality corresponding to the second historical grid voltage.

[0134] S404. Construct a first historical data set based on multiple first historical grid voltages, multiple first historical primary currents, and a first label for each first historical grid voltage; and construct a second historical data set based on multiple second historical grid voltages, multiple second historical primary currents, and a second label for each second historical grid voltage.

[0135] In an optional implementation manner, the relevant descriptions of steps S401 to S404 can be referred to the descriptions of steps 1 to 4 in the aforementioned determination of the power quality assessment model, and will not be repeated here.

[0136] S405 : performing normalization processing on the first historical data set to obtain a first training set, and performing normalization processing on the second historical data set to obtain a second training set.

[0137] S406: Construct a transfer learning model, and train the transfer learning model using the first training set and the second training set to obtain a trained transfer learning model.

[0138] In an optional embodiment, the structural diagram of the transfer learning model can be as follows Figure 3 The computer device trains the transfer learning model using the first training set and the second training set. For the relevant description of the trained transfer learning model, please refer to the relevant description in the above step 5, which will not be repeated here.

[0139] S407: Using the trained transfer learning model as a power quality assessment model.

[0140] S408. Obtain the grid voltage and primary current corresponding to the target locomotive in real time.

[0141] S409: Construct a data set corresponding to the target locomotive based on the grid voltage and the primary current.

[0142] S410: Calling a power quality assessment model, and determining a target power quality assessment result of the target locomotive based on a data set corresponding to the target locomotive.

[0143] In the embodiment of the present application, since the power quality assessment model is obtained by training a pre-constructed transfer learning model using a second historical data set with a large amount of operating data (i.e., the historical data set corresponding to the source locomotive) and a first historical data set with a small amount of operating data (i.e., the historical data set corresponding to the target locomotive), the pre-trained power quality assessment model has learned the data characteristics of each data in the second historical data set and the data characteristics of each data in the first historical data set. Therefore, for electric locomotives with a small amount of operating data, the accuracy of the power quality assessment results can be improved by using the pre-trained power quality assessment model.

[0144] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0145] Based on the same inventive concept, embodiments of the present application also provide a locomotive high-voltage-side power quality assessment device for implementing the locomotive high-voltage-side power quality assessment method described above. The solution provided by this device is similar to the solution described in the method described above. Therefore, the specific limitations of one or more locomotive high-voltage-side power quality assessment device embodiments provided below can be found in the above-described limitations of the locomotive high-voltage-side power quality assessment method and will not be further elaborated here.

[0146] See Figure 5 , Figure 5 This is a schematic diagram of the structure of a power quality assessment device for the high voltage side of a locomotive provided in an embodiment of the present application. Figure 5 As shown, the power quality assessment device on the high-voltage side of the locomotive may include but is not limited to:

[0147] An acquisition module 501 is used to obtain the grid voltage and primary current corresponding to the target locomotive in real time;

[0148] A construction module 502 is used to construct a data set corresponding to the target locomotive based on the grid voltage and the primary current;

[0149] An evaluation module 503 is configured to call a pre-trained power quality evaluation model and determine a target power quality evaluation result of a target locomotive based on a data set;

[0150] The power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set;

[0151] The first historical data set includes a first historical grid voltage and a first historical primary current corresponding to the target locomotive; the second historical data set includes a second locomotive grid voltage and a second historical primary current corresponding to the source locomotive.

[0152] In one embodiment, when the evaluation module 503 is used to call a pre-trained power quality evaluation model and determine the target power quality evaluation result of the target locomotive based on a data set, it is specifically used to: normalize the data set to obtain a normalized data set; input the normalized data set into the pre-trained power quality evaluation model to obtain an initial power quality evaluation result of the target locomotive; and denormalize the initial power quality evaluation result to obtain the target power quality evaluation result of the target locomotive.

[0153] In one embodiment, the construction module 502 is further used to: construct a first input layer for inputting data corresponding to the target locomotive, and construct a second input layer for inputting data corresponding to the source locomotive; construct L latent layers, wherein the first L-1 latent layers are shared latent layers, and the Lth latent layer includes a first latent layer and a second latent layer, the first latent layer is associated with the data corresponding to the target locomotive, and the second latent layer is associated with the data corresponding to the source locomotive; construct a first output layer for outputting the power quality of the target locomotive, and construct a second output layer for outputting the power quality of the source locomotive; and construct a transfer learning model based on the first input layer, the second input layer, the L latent layers, the first output layer, and the second output layer.

[0154] In one embodiment, the device may also include a training module. The training module is used to obtain the first historical grid voltage and the first historical primary current of the target locomotive in each first historical period, and to obtain the second historical grid voltage and the second historical primary current of the source locomotive in each second historical period; wherein the number of first historical periods is much smaller than the number of second historical periods; a determination module is used to determine the first grid voltage characteristic parameter corresponding to each first historical grid voltage, and to determine the second grid voltage characteristic parameter corresponding to each second historical grid voltage; for each first historical grid voltage, based on the first grid voltage characteristic parameter corresponding to the first historical grid voltage, a first label is added to the first historical grid voltage, and, for each second historical grid voltage, based on the second grid voltage characteristic parameter corresponding to the second historical grid voltage, a first label is added to the first historical grid voltage. characteristic parameters, adding a second label to the second historical grid voltage; the first label is used to characterize the actual power quality corresponding to the first historical grid voltage, and the second label is used to characterize the actual power quality corresponding to the second historical grid voltage; based on multiple first historical grid voltages, multiple first historical primary currents and the first label of each first historical grid voltage, a first historical data set is constructed, and, based on multiple second historical grid voltages, multiple second historical primary currents and the second label of each second historical grid voltage, a second historical data set is constructed; based on the first historical data set and the second historical data set, a transfer learning model is trained to obtain a trained transfer learning model; and the trained transfer learning model is used as a power quality assessment model.

[0155] In one embodiment, when the training module is used to determine the first grid voltage characteristic parameter corresponding to each first historical grid voltage, and to determine the second grid voltage characteristic parameter corresponding to each second historical grid voltage, it is specifically used to: for each first historical grid voltage, determine multiple first historical grid voltage signals, multiple first fundamental grid voltage signals and first fundamental amplitude of the first historical grid voltage within the first historical period; based on the multiple first historical grid voltage signals, multiple first fundamental grid voltage signals and first fundamental amplitude, determine the first harmonic content corresponding to the first historical grid voltage within the first historical period, and use the first harmonic content as the first grid voltage characteristic parameter corresponding to the first historical grid voltage; for each second historical grid voltage, determine multiple second historical grid voltage signals, multiple second fundamental grid voltage signals and second fundamental amplitude of the second historical grid voltage within the second historical period; based on the multiple second historical grid voltage signals, multiple second fundamental grid voltage signals and second fundamental amplitude, determine the second harmonic content corresponding to the second historical grid voltage within the second historical period, and use the second harmonic content as the second grid voltage characteristic parameter corresponding to the second historical grid voltage.

[0156] In one embodiment, when the training module is used to train the transfer learning model based on the first historical data set and the second historical data set to obtain the trained transfer learning model, it is specifically used to: normalize the first historical data set to obtain a first training set, and normalize the second historical data set to obtain a second training set; input the first training set and the second training set into the transfer learning model to obtain the first predicted power quality corresponding to each normalized first historical grid voltage in the first training set and the second predicted power quality corresponding to each normalized second historical grid voltage in the second training set; perform denormalization on multiple first predicted power qualities to obtain multiple processed first predicted power qualities, and perform denormalization on multiple second predicted power qualities to obtain multiple processed second predicted power qualities; train the transfer learning model in the direction of reducing the difference between each processed first predicted power quality and the corresponding first label, and in the direction of reducing the difference between each processed second predicted power quality and the corresponding second label to obtain the trained transfer learning model.

[0157] Each module in the locomotive high-voltage-side power quality assessment device described above can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a terminal device in hardware form, or can be stored in a memory in the terminal device in software form, so that the processor can call and execute the corresponding operations of each module.

[0158] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 6As shown. The computer device includes a processor, memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are connected to the system bus via the input / output interface. 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 internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals via wired or wireless means. The wireless means can be implemented via Wi-Fi, mobile cellular networks, near-field communication (NFC), or other technologies. When executed by the processor, the computer program implements a method for assessing the power quality of the high-voltage side of a locomotive. The display unit of the computer device is used to produce a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse.

[0159] 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 shown in the figure, or combine certain components, or have a different component arrangement.

[0160] In an exemplary embodiment, the present application provides a computer device including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps in the above-mentioned method for evaluating the power quality of the high-voltage side of a locomotive are implemented.

[0161] In an exemplary embodiment, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the steps in the above-mentioned method for evaluating the power quality of the high-voltage side of a locomotive when the computer program is executed by a processor.

[0162] In an exemplary embodiment, the present application provides a computer program product, including a computer program, which implements the steps in the above-mentioned method for evaluating the power quality of the high-voltage side of a locomotive when executed by a processor.

[0163] It should be noted that the data involved in this application (including but not limited to the grid voltage and primary current corresponding to the target locomotive, the data set corresponding to the target locomotive, the power quality assessment results, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0164] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. 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. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile memory 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), magnetic 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. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), quantum computing-based data processing logic devices, artificial intelligence (AI) processors, and the like.

[0165] The technical features of the above embodiments can be combined arbitrarily. In order 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 application.

[0166] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for evaluating the power quality of a locomotive high-voltage side, characterized in that: The method comprises: Obtain the grid voltage and primary current corresponding to the target locomotive in real time; Constructing a data set corresponding to the target locomotive based on the grid voltage and primary current; calling a pre-trained power quality assessment model and determining a target power quality assessment result of the target locomotive based on the data set; The power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set; The first historical data set includes a first historical grid voltage and a first historical primary current corresponding to the target locomotive; the second historical data set includes a second locomotive grid voltage and a second historical primary current corresponding to the source locomotive.

2. The method according to claim 1, characterized in that The calling of a pre-trained power quality assessment model and determining a target power quality assessment result of the target locomotive based on the data set includes: Normalizing the data set to obtain a normalized data set; Inputting the normalized data set into a pre-trained power quality assessment model to obtain an initial power quality assessment result of the target locomotive; The initial power quality evaluation result is subjected to denormalization processing to obtain a target power quality evaluation result of the target locomotive.

3. The method according to claim 1, characterized in that The transfer learning model is constructed in the following way: constructing a first input layer for inputting data corresponding to the target locomotive, and constructing a second input layer for inputting data corresponding to the source locomotive; Constructing L latent layers, wherein the first L-1 latent layers are shared latent layers, and the Lth latent layer includes a first latent layer and a second latent layer, wherein the first latent layer is associated with the data corresponding to the target locomotive, and the second latent layer is associated with the data corresponding to the source locomotive; constructing a first output layer for outputting the power quality of the target locomotive, and constructing a second output layer for outputting the power quality of the source locomotive; A transfer learning model is constructed based on the first input layer, the second input layer, the L-layer latent layer, the first output layer, and the second output layer.

4. The method according to any one of claims 1 to 3, characterized in that The method further comprises: Obtaining a first historical grid voltage and a first historical primary current of the target locomotive in each first historical period, and obtaining a second historical grid voltage and a second historical primary current of the source locomotive in each second historical period; wherein the number of the first historical periods is much smaller than the number of the second historical periods; Determine a first grid voltage characteristic parameter corresponding to each of the first historical grid voltages, and determine a second grid voltage characteristic parameter corresponding to each of the second historical grid voltages; For each first historical grid voltage, a first label is added to the first historical grid voltage based on the first grid voltage characteristic parameter corresponding to the first historical grid voltage; and for each second historical grid voltage, a second label is added to the second historical grid voltage based on the second grid voltage characteristic parameter corresponding to the second historical grid voltage; the first label is used to characterize the actual power quality corresponding to the first historical grid voltage, and the second label is used to characterize the actual power quality corresponding to the second historical grid voltage; Constructing a first historical data set based on a plurality of first historical grid voltages, a plurality of first historical primary currents, and a first label of each of the first historical grid voltages; and constructing a second historical data set based on a plurality of second historical grid voltages, a plurality of second historical primary currents, and a second label of each of the second historical grid voltages; Training the transfer learning model based on the first historical data set and the second historical data set to obtain a trained transfer learning model; The trained transfer learning model is used as a power quality assessment model.

5. The method according to claim 4, characterized in that Determining a first grid voltage characteristic parameter corresponding to each of the first historical grid voltages, and determining a second grid voltage characteristic parameter corresponding to each of the second historical grid voltages, includes: For each first historical grid voltage, determining multiple first historical grid voltage signals, multiple first fundamental grid voltage signals, and first fundamental amplitude values ​​of the first historical grid voltage within the first historical period; determining a first harmonic content corresponding to the first historical grid voltage within the first historical period based on the multiple first historical grid voltage signals, the multiple first fundamental grid voltage signals, and the first fundamental amplitude value, and using the first harmonic content as a first grid voltage characteristic parameter corresponding to the first historical grid voltage; For each second historical grid voltage, determine multiple second historical grid voltage signals, multiple second fundamental grid voltage signals and second fundamental amplitude of the second historical grid voltage within the second historical period; based on the multiple second historical grid voltage signals, multiple second fundamental grid voltage signals and the second fundamental amplitude, determine the second harmonic content corresponding to the second historical grid voltage within the second historical period, and use the second harmonic content as the second grid voltage characteristic parameter corresponding to the second historical grid voltage.

6. The method according to claim 4, characterized in that The training of the transfer learning model based on the first historical data set and the second historical data set to obtain a trained transfer learning model includes: Normalizing the first historical data set to obtain a first training set, and normalizing the second historical data set to obtain a second training set; Inputting the first training set and the second training set into the transfer learning model to obtain a first predicted power quality corresponding to each normalized first historical grid voltage in the first training set and a second predicted power quality corresponding to each normalized second historical grid voltage in the second training set; performing denormalization processing on the plurality of first predicted power qualities to obtain a plurality of processed first predicted power qualities, and performing denormalization processing on the plurality of second predicted power qualities to obtain a plurality of processed second predicted power qualities; The transfer learning model is trained in a direction of reducing the difference between each of the processed first predicted power qualities and the corresponding first label, and in a direction of reducing the difference between each of the processed second predicted power qualities and the corresponding second label to obtain a trained transfer learning model.

7. A power quality assessment device for the high-voltage side of a locomotive, characterized in that: The device comprises: Acquisition module, used to obtain the grid voltage and primary current corresponding to the target locomotive in real time; A construction module, configured to construct a data set corresponding to the target locomotive based on the grid voltage and the primary current; An evaluation module, configured to call a pre-trained power quality evaluation model and determine a target power quality evaluation result of the target locomotive based on the data set; The power quality assessment model is obtained by training a pre-built transfer learning model using a first historical data set corresponding to the target locomotive and a second historical data set corresponding to the source locomotive; the amount of data in the first historical data set is much smaller than the amount of data in the second historical data set; The first historical data set includes a first historical grid voltage and a first historical primary current corresponding to the target locomotive; the second historical data set includes a second locomotive grid voltage and a second historical primary current corresponding to the source locomotive.

8. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images