Steel rail condensation prediction method and device based on time sequence data, equipment and medium

Abstract

The invention provides a steel rail condensation prediction method and device based on time sequence data, equipment and a medium. The method comprises the steps that the rail surface temperature in a future time period is predicted according to rail surface parameters of a time sequence in a past time period; according to the trackside parameters of the time sequence in the past time period, the trackside temperature, the trackside humidity and the trackside wind speed in the future time period are predicted; on the basis of the predicted trackside temperature and the predicted trackside humidity, the dew-point temperature in the future time period is predicted in combination with the relation between the dew-point temperature and the trackside temperature and the relation between the dew-point temperature and the trackside humidity; and predicting rail surface condensation based on the predicted rail surface temperature, the predicted trackside wind speed and the predicted dew point temperature. According to the method provided by the invention, the rail surface condensation is predicted according to the rail surface parameters and the trackside parameters of the time sequence in the past time period in combination with the relationship between the dew point temperature and the trackside temperature and the relationship between the dew point temperature and the trackside humidity, so that the pre-judgment of the condensation of the steel rail is realized, a powerful guarantee is provided for the operation safety of rail traffic, and potential risks and losses caused by condensation are reduced.

Description

Technical Field

[0001] The present application relates to the field of rail transit technology, and in particular to a method, device, equipment, and medium for predicting rail condensation based on time series data. Background Art

[0002] Air under natural conditions is composed of a small amount of dust, water vapor and absolutely dry air. The amount of water vapor that the air can hold is proportional to the ambient temperature, that is, the higher the ambient temperature, the more water vapor the air can hold. The so-called dew point temperature refers to the highest temperature at which condensation occurs in air with a specific humidity.

[0003] Water vapor contained in the air at a higher temperature will no longer be contained in the air due to the drop in temperature, and will precipitate in the form of liquid water. If the air with high humidity and relatively high temperature hits the track surface with a relatively low temperature (lower than the dew point temperature of the air under this condition), condensation will occur, and a certain amount of liquid water will be produced on the track surface.

[0004] A certain amount of condensation adheres to the rail surface, which may destroy the adhesion coefficient between the wheel and rail (<0.06) and cause the train to slip. In addition, on open-air lines, fallen leaves fall onto the rail surface. As the train passes, the fallen leaves stuck to the rail surface by condensation will be crushed, causing the organic matter and water in the fallen leaves to mix to form an oxide layer, destroying the adhesion coefficient between the wheel and rail (<0.03) and increasing the risk of train braking slippage. Summary of the invention

[0005] In order to solve one of the above-mentioned technical defects, the present application provides a rail condensation prediction method, device, equipment and medium based on time series data.

[0006] In a first aspect, the present application provides a rail condensation prediction method based on time series data, the method comprising:

[0007] Obtain track surface parameters and trackside parameters in the past time period; track surface parameters include: track surface temperature, air temperature, air humidity; trackside parameters include: trackside temperature, trackside humidity, trackside wind speed;

[0008] Predict the track surface temperature in the future period based on the track surface parameters in the past period; predict the trackside temperature, trackside humidity, and trackside wind speed in the future period based on the trackside parameters in the past period;

[0009] Based on the predicted trackside temperature and predicted trackside humidity, combined with the relationship between the dew point temperature and the trackside temperature and trackside humidity, the dew point temperature in the future period is predicted;

[0010] Predict rail surface condensation based on predicted rail surface temperature, predicted trackside wind speed, and predicted dew point temperature.

[0011] Optionally, the rail surface parameters and railside parameters are collected by patch temperature sensors and multi-factor small weather stations;

[0012] The past period includes multiple seasons and multiple climate conditions within its time span.

[0013] Optionally, predicting the rail surface temperature in a future time period based on the rail surface parameters in the past time period includes:

[0014] Through the recurrent neural network in the rail surface temperature prediction model, the change pattern and trend of the rail surface temperature over time are determined according to the rail surface temperature in the past period;

[0015] Through the multi-layer perceptron in the rail surface temperature prediction model, the temperature prediction features are extracted according to the time series of air temperature and air humidity in the past period;

[0016] The rail surface temperature prediction model is used to predict the rail surface temperature in the future period based on the change pattern and trend of the rail surface temperature over time and the temperature prediction characteristics.

[0017] Optionally, the rail surface temperature prediction model is a Decoder-only architecture.

[0018] Optionally, the trackside temperature, trackside humidity, and trackside wind speed in the future period are predicted based on the trackside parameters in the past period, including:

[0019] The trackside environment prediction model is used to predict the trackside temperature, trackside humidity and trackside wind speed in the future period based on the trackside parameters in the past period.

[0020] Among them, the trackside environment prediction model is an Encoder-Decoder architecture;

[0021] The encoder in the encoder-decoder architecture is used to obtain the characteristics and patterns of the trackside environment according to the trackside parameters in the past time period, and output the encoding vector of the characteristics and patterns of the trackside environment;

[0022] The decoder in the encoder-decoder architecture is used to predict the trackside temperature, trackside humidity, and trackside wind speed in the future period based on the encoding vector.

[0023] Optionally, the relationship between the dew point temperature and the trackside temperature and trackside humidity is T d =T-((100-RH) / 5);

[0024] Among them, T d is the dew point temperature, T is the trackside temperature, and RH is the trackside humidity.

[0025] Optionally, based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature, the rail surface condensation is predicted, including:

[0026] If the predicted trackside wind speed is less than the wind speed threshold and the predicted rail surface temperature is less than the predicted dew point temperature, condensation is predicted to occur on the rail surface.

[0027] In a second aspect of the present application, a rail condensation prediction device based on time series data is provided, the device comprising:

[0028] The acquisition module is used to acquire the track surface parameters and trackside parameters of the past time period; the track surface parameters include: track surface temperature, air temperature, air humidity; the trackside parameters include: trackside temperature, trackside humidity, trackside wind speed;

[0029] The first prediction module is used to predict the rail surface temperature in the future period according to the rail surface parameters of the past period acquired by the acquisition module; and predict the railside temperature, railside humidity and railside wind speed in the future period according to the railside parameters of the past period acquired by the acquisition module;

[0030] A second prediction module is used to predict the dew point temperature in a future period based on the trackside temperature and the predicted trackside humidity predicted by the first prediction module and in combination with the relationship between the dew point temperature and the trackside temperature and the trackside humidity;

[0031] The third prediction module is used to predict rail surface condensation based on the rail surface temperature predicted by the first prediction module, the predicted trackside wind speed, and the dew point temperature predicted by the second prediction module.

[0032] In a third aspect of the present application, an electronic device is provided, including:

[0033] Memory;

[0034] Processor; and

[0035] Computer programs;

[0036] The computer program is stored in the memory and is configured to be executed by the processor to implement the method as described in the first aspect above.

[0037] In a fourth aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored; the computer program is executed by a processor to implement the method described in the first aspect above.

[0038] The present application provides a method, device, equipment and medium for predicting rail condensation based on time series data, the method comprising: obtaining rail surface parameters and trackside parameters in a time series in a past time period; wherein the rail surface parameters include: rail surface temperature, air temperature, air humidity; the trackside parameters include: trackside temperature, trackside humidity, and trackside wind speed; predicting the rail surface temperature in a future time period based on the rail surface parameters in a time series in the past time period; predicting the trackside temperature, trackside humidity, and trackside wind speed in a future time period based on the trackside parameters in a time series in the past time period; predicting the dew point temperature in a future time period based on the predicted trackside temperature and the predicted trackside humidity, combined with the relationship between the dew point temperature and the trackside temperature and the trackside humidity; predicting rail surface condensation based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature. The method provided in the present application predicts rail surface condensation based on the rail surface parameters and railside parameters in the past time period, combined with the relationship between dew point temperature and railside temperature and railside humidity, thereby realizing the pre-judgment of condensation on the rails, providing strong protection for the safety of rail transit operation, and reducing potential risks and losses caused by condensation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0040] Figure 1 A schematic flow chart of a rail condensation prediction method based on time series data provided in an embodiment of the present application;

[0041] Figure 2 A schematic diagram of the architecture of a rail surface temperature prediction model provided in an embodiment of the present application;

[0042] Figure 3 A schematic diagram of the architecture of a trackside environment prediction model provided in an embodiment of the present application;

[0043] Figure 4 A schematic diagram of the architecture of a rail condensation prediction method based on time series data provided in an embodiment of the present application;

[0044] Figure 5 A schematic diagram of the structure of a rail condensation prediction device based on time series data provided in an embodiment of the present application;

[0045] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to make the technical solutions and advantages in the embodiments of the present application more clearly understood, the exemplary embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than an exhaustive list of all the embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0047] In the process of realizing the present application, the inventors found that a certain amount of condensation attached to the rail surface may destroy the adhesion coefficient between the wheel and rail (<0.06), causing the train to slip; in addition, on open-air lines, fallen leaves fall onto the rail surface. As the train passes, the fallen leaves stuck to the rail surface by the condensation will be crushed, causing the organic matter and water in the fallen leaves to mix to form an oxide layer, destroying the adhesion coefficient between the wheel and rail (<0.03) and increasing the risk of train braking slippage.

[0048] In response to the above problems, the present application provides a method, device, equipment, and medium for predicting rail condensation based on time series data, the method comprising: obtaining rail surface parameters and trackside parameters in a time series in a past time period; wherein the rail surface parameters include: rail surface temperature, air temperature, and air humidity; the trackside parameters include: trackside temperature, trackside humidity, and trackside wind speed; predicting the rail surface temperature in a future time period based on the rail surface parameters in a time series in the past time period; predicting the trackside temperature, trackside humidity, and trackside wind speed in a future time period based on the trackside parameters in a time series in the past time period; based on the predicted trackside temperature and predicted trackside humidity, combined with the relationship between the dew point temperature and the trackside temperature and the trackside humidity, predicting the dew point temperature in a future time period; predicting the rail surface condensation based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature. The method provided in the present application predicts rail surface condensation based on the rail surface parameters and railside parameters in the past time period, combined with the relationship between dew point temperature and railside temperature and railside humidity, thereby realizing the pre-judgment of condensation on the rails, providing strong protection for the safety of rail transit operation, and reducing potential risks and losses caused by condensation.

[0049] See also Figure 1 This embodiment provides a rail condensation prediction method based on time series data, and the implementation process of the method is as follows:

[0050] 101, obtaining track surface parameters and trackside parameters in a time series in the past period.

[0051] Among them, the rail surface parameters include: rail surface temperature, air temperature, and air humidity.

[0052] Trackside parameters include: trackside temperature, trackside humidity, and trackside wind speed.

[0053] Track surface parameters and trackside parameters are collected through patch temperature sensors and multi-factor small weather stations. For example, track surface temperature is collected through patch temperature sensors, and air temperature, air humidity, trackside humidity, and trackside wind speed are collected through multi-factor small weather stations.

[0054] In addition, the time span of the past period includes multiple seasons and multiple climate conditions, that is, the time span of the past period is long enough, and the track surface parameters and trackside parameters obtained in step 101 need to cover the parameters of environmental changes under different seasons and climate conditions.

[0055] In addition, the track surface parameters and trackside parameters obtained in this step are preprocessed, such as data cleaning, using linear interpolation to fill missing data points; removing obvious outliers through the threshold method; extracting time features such as year, month, day, hour, etc., so that seasonal and cyclical changes can be learned in the subsequent prediction process.

[0056] 102. Predict the track surface temperature in a future time period based on the track surface parameters in the past time period, and predict the trackside temperature, trackside humidity, and trackside wind speed in a future time period based on the trackside parameters in the past time period.

[0057] For example, the track surface temperature prediction model can be used to predict the track surface temperature in the future period based on the track surface parameters in the past period. The trackside environment prediction model can be used to predict the trackside temperature, trackside humidity, and trackside wind speed in the future period based on the trackside parameters in the past period.

[0058] 1. The rail surface temperature prediction model is a decoder-only architecture.

[0059] like Figure 2 As shown in FIG. 1 , the rail surface temperature prediction model uses a recurrent neural network (RNN) and a multilayer perceptron (MLP) to predict the rail surface temperature.

[0060] Through the rail surface temperature prediction model, the process of predicting the rail surface temperature in the future period based on the rail surface parameters in the past period is as follows:

[0061] Through the RNN in the rail surface temperature prediction model, the changing pattern and trend of the rail surface temperature over time are determined based on the rail surface temperature in the past period.

[0062] Through the MLP in the rail surface temperature prediction model, the temperature prediction features are extracted according to the time series of air temperature and air humidity in the past period.

[0063] The rail surface temperature prediction model is used to predict the rail surface temperature in the future period based on the change pattern and trend of the rail surface temperature over time and the temperature prediction characteristics.

[0064] RNN can record past information and use it for current predictions. RNN can take the time series of rail surface temperature in the past period as input and process the rail surface temperature value at each time point in turn. By continuously updating its internal state, it can capture time series information such as seasonality, periodicity and trend of rail surface temperature, and learn the changing patterns and trends of rail surface temperature over time.

[0065] MLP takes the time series of air temperature, air humidity, and rail surface temperature in the past period as input, learns the complex relationship between air temperature, air humidity and rail surface temperature, and the impact of these factors on rail surface temperature, combines and transforms these input data, and extracts features related to rail surface temperature prediction.

[0066] The final rail surface temperature prediction model will predict the rail surface temperature in the future period based on the changing pattern and trend of the rail surface temperature over time and the temperature prediction characteristics.

[0067] In the specific implementation, the rail surface temperature prediction model is pre-trained. The training process can be to obtain a temperature prediction training data set, divide the temperature prediction training data set according to the ratio of 7:2:1, obtain a training set, input the training set into the rail surface temperature prediction model, and use the optimizer to continuously adjust the model parameters to minimize the loss function. The rail surface temperature prediction model predicts future rail surface temperature values ​​by learning the patterns and rules of the input sequence.

[0068] 2. The trackside environment prediction model is an Encoder-Decoder architecture.

[0069] like Figure 3 As shown, the trackside environment prediction model uses an Encoder encoder and a Decoder decoder to predict the trackside environment.

[0070] Through the trackside environment prediction model, the trackside temperature, trackside humidity and trackside wind speed in the future period are predicted based on the trackside parameters in the past period:

[0071] The Encoder in the Encoder-decoder architecture is used to obtain the characteristics and patterns of the trackside environment based on the trackside parameters in the past time period, and output the encoding vector of the characteristics and patterns of the trackside environment.

[0072] The decoder in the encoder-decoder architecture is used to predict the trackside temperature, trackside humidity, and trackside wind speed in the future period based on the encoding vector.

[0073] In specific implementation, the encoder can take the trackside parameters of the time series in the past period as input and convert them into an intermediate representation, usually called a coding vector or hidden state, which mainly extracts the key features and patterns of the input data. The encoder can include an input layer, an RNN layer, and a coding vector generation layer.

[0074] Input layer: receives trackside parameters of the time series in the past period.

[0075] The trackside parameters of the past time period can be univariate or multivariate time series.

[0076] RNN layer: processes input data (i.e., trackside parameters of time series in the past period) time step by time step, updates its internal state, and remembers long-term dependencies.

[0077] Encoding vector generation layer: Generates an encoding vector by aggregating the output of the last time step of the RNN layer or the output of all time steps. This vector contains a compressed representation of the input time series, capturing the characteristics and patterns of the trackside environment.

[0078] The decoder decoder gradually predicts the trackside temperature, trackside humidity, and trackside wind speed in the future period based on the encoding vector generated by the encoder encoder. The decoder decoder can include an initial state setting layer, an RNN layer, and an output layer.

[0079] Initial state setting layer: The encoding vector generated by the Encoder encoder is used as the initial state of the Decoder decoder.

[0080] RNN layer: receives the output of the previous time step and the current input at each time step, and updates its internal state. The current input is composed of the historical temperature at the corresponding moment of the prediction and the horizontal historical data of the past n days and m hours. Adding past references for temperature prediction at a certain moment in the future improves the accuracy of the prediction.

[0081] Output layer: Based on the output of the Decoder RNN layer, the trackside temperature, trackside humidity, and trackside wind speed in the future period are predicted. The output layer can be a linear layer or other appropriate function to convert the output of the Decoder decoder into the required output format.

[0082] In the specific implementation, the trackside environment prediction model is pre-trained. The training process can be to obtain an environmental prediction training data set, divide the environmental prediction training data set according to the ratio of 7:2:1, obtain a training set, and input the historical environment sequence and its related features in the training set into the Encoder encoder. The Encoder encoder outputs the encoding vector of the characteristics and patterns of the trackside environment. The Decoder decoder uses the optimizer to continuously adjust the model parameters according to the output of the Encoder encoder and the historical information of the target sequence to minimize the loss function, and gradually predict the trackside temperature, trackside humidity, and trackside wind speed in the future period.

[0083] 103, based on the predicted trackside temperature and the predicted trackside humidity, combined with the relationship between the dew point temperature and the trackside temperature and the trackside humidity, the dew point temperature in the future period is predicted.

[0084] Dew point temperature refers to the temperature at which the water vapor in the air reaches saturation under a certain air pressure. When the air temperature drops to or below the dew point temperature, the water vapor in the air will condense into water droplets. The relationship between the dew point temperature and the trackside temperature and trackside humidity is T d =T-((100-RH) / 5).

[0085] Among them, T d is the dew point temperature, T is the trackside temperature (unit: Celsius), and RH is the trackside humidity (unit: percentage).

[0086] In specific implementation, a corresponding table of the relationship between the dew point temperature, the trackside temperature and the trackside humidity can be pre-stored according to the relationship between the dew point temperature, the trackside temperature and the trackside humidity, as shown in Table 1. In step 103, the table can be looked up to obtain the dew point temperature in the future period.

[0087] Table 1

[0088]

[0089]

[0090] 104. Predict rail surface condensation based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature.

[0091] After the predicted dew point temperature is obtained, the rail surface condensation can be predicted based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature.

[0092] For example, if the predicted trackside wind speed is less than a wind speed threshold (such as 2 m / s), and the predicted rail surface temperature is less than the predicted dew point temperature, it is predicted that condensation will occur on the rail surface.

[0093] If at a certain time t1 in the future, the rail surface temperature at time t1 is predicted by the rail surface temperature prediction model in step 102 as According to the trackside environment prediction model, the trackside temperature at time t1 is predicted as Trackside humidity is The trackside wind speed is Step 103 According to and The dew point temperature predicted for the future period is In wind speed Under the premise of It is predicted that condensation will form on the rail surface, that is, condensation will occur.

[0094] The rail condensation prediction method based on time series data provided in this embodiment is based on a small weather station and a patch temperature sensor, collects a large amount of historical environmental temperature and humidity and rail surface temperature data, and uses a time series prediction method combining the decoder-only and encoder-decoder architectures to effectively predict the rail surface temperature and trackside temperature and humidity for a period of time in the future. Combined with the corrected dew point temperature table, the condensation generation conditions are judged, and a prediction result of whether condensation will occur in the future is given.

[0095] The rail condensation prediction method based on time series data provided in this embodiment can be implemented based on Figure 4 The architecture shown is implemented. Using patch temperature sensors and multi-factor small weather stations, the track surface temperature, air temperature, air humidity, trackside temperature, trackside humidity, trackside wind speed, etc. are collected. After obtaining the track surface parameters and trackside parameters of the time series in the past period in step 101, these track surface parameters and trackside parameters are input into the track surface temperature prediction model and the trackside environment prediction model to execute step 102, and the track surface temperature, trackside temperature, trackside humidity, and trackside wind speed in the next n hours are predicted respectively. The track surface temperature, trackside temperature, trackside humidity, and trackside wind speed in the next n hours are input into the track surface condensation prediction module to execute steps 103 and 104, and the relationship between the rail transit dew point temperature and the trackside temperature and trackside humidity is combined to predict the occurrence of track surface condensation.

[0096] The rail condensation prediction method based on time series data provided in this embodiment collects a large number of rail surface parameters and trackside parameters through patch temperature sensors and multi-factor small weather stations, effectively predicts the rail surface temperature, trackside temperature, trackside humidity, and trackside wind speed at a certain time in the future, and gives a prediction result of whether condensation will occur on the rail surface at a certain time in the future, and issues an early warning in time to remind operators to adjust the control strategy to avoid idling and slipping of trains due to condensation on the rail surface. It provides strong protection for the safety of rail transit operation and reduces potential risks and losses caused by condensation.

[0097] The rail condensation prediction method based on time series data provided in this embodiment collects the historical temperature and humidity data of the rails, and uses the time series prediction algorithm to accurately predict the temperature and humidity at a certain moment in the future. At the same time, the dew point temperature and the rail surface temperature are combined for judgment. When specific conditions are met, it is determined whether condensation occurs, and an early warning is issued in time. The rail condensation prediction method based on time series data provided in this embodiment can effectively predict the condensation phenomenon that may occur on the rails in advance, provide strong protection for the safety of rail transit operation, and reduce the potential risks and losses caused by condensation.

[0098] The rail condensation prediction method based on time series data provided in this embodiment can provide timely warnings to operating personnel, adjust driving strategies, and reduce the impact of condensation on open-air sections of rail transit, which may cause train idling and slipping.

[0099] The rail condensation prediction method based on time series data provided in this embodiment can combine the horizontal historical data of the past n days and the vertical historical data at the time of prediction as input, thereby effectively improving the prediction accuracy.

[0100] This embodiment provides a method for predicting rail condensation based on time series data, and obtains the rail surface parameters and trackside parameters of the time series in the past time period; wherein the rail surface parameters include: rail surface temperature, air temperature, air humidity; the trackside parameters include: trackside temperature, trackside humidity, and trackside wind speed; predict the rail surface temperature in the future time period based on the rail surface parameters of the time series in the past time period; predict the trackside temperature, trackside humidity, and trackside wind speed in the future time period based on the trackside parameters of the time series in the past time period; based on the predicted trackside temperature and the predicted trackside humidity, combined with the relationship between the dew point temperature and the trackside temperature and the trackside humidity, predict the dew point temperature in the future time period; based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature, predict the rail surface condensation. The method provided in this embodiment predicts the rail surface condensation based on the rail surface parameters and trackside parameters of the time series in the past time period, combined with the relationship between the dew point temperature and the trackside temperature and the trackside humidity, thereby realizing the prejudgment of condensation on the rails, providing a strong guarantee for the safety of rail transit operation, and reducing the potential risks and losses caused by condensation.

[0101] The same inventive concept as the rail condensation prediction method based on time series data, this embodiment provides a rail condensation prediction device based on time series data, see Figure 5 The rail condensation prediction device based on time series data includes:

[0102] The acquisition module 501 is used to acquire the track surface parameters and trackside parameters of the past time period.

[0103] Among them, the track surface parameters include: track surface temperature, air temperature, air humidity. Trackside parameters include: trackside temperature, trackside humidity, trackside wind speed.

[0104] The first prediction module 502 is used to predict the rail surface temperature in the future period according to the rail surface parameters in the past period obtained by the acquisition module 501. The rail side temperature, rail side humidity and rail side wind speed in the future period are predicted according to the rail side parameters in the past period obtained by the acquisition module 501.

[0105] The second prediction module 503 is used to predict the dew point temperature in a future period based on the trackside temperature and the predicted trackside humidity predicted by the first prediction module 502 and the relationship between the dew point temperature and the trackside temperature and the trackside humidity.

[0106] The third prediction module 504 is used to predict rail surface condensation based on the rail surface temperature predicted by the first prediction module 502 , the predicted trackside wind speed, and the dew point temperature predicted by the second prediction module 503 .

[0107] Among them, the track surface parameters and trackside parameters are collected through patch temperature sensors and multi-factor small meteorological stations.

[0108] The past period includes multiple seasons and multiple climate conditions within its time span.

[0109] The second prediction module 503 is used to determine the change pattern and trend of the rail surface temperature over time according to the rail surface temperature in the past time period through the recurrent neural network in the rail surface temperature prediction model.

[0110] Through the multi-layer perceptron in the rail surface temperature prediction model, the temperature prediction features are extracted according to the air temperature and air humidity in the past time period.

[0111] The rail surface temperature prediction model is used to predict the rail surface temperature in the future period based on the change pattern and trend of the rail surface temperature over time and the temperature prediction characteristics.

[0112] Among them, the rail surface temperature prediction model is a decoder-only architecture.

[0113] The second prediction module 503 is used to predict the trackside temperature, trackside humidity and trackside wind speed in the future time period according to the trackside parameters in the past time period through the trackside environment prediction model.

[0114] Among them, the trackside environment prediction model is an Encoder-Decoder architecture.

[0115] The Encoder in the Encoder-decoder architecture is used to obtain the characteristics and patterns of the trackside environment based on the trackside parameters in the past time period, and output the encoding vector of the characteristics and patterns of the trackside environment.

[0116] The decoder in the encoder-decoder architecture is used to predict the trackside temperature, trackside humidity, and trackside wind speed in the future period based on the encoding vector.

[0117] The relationship between dew point temperature and trackside temperature and trackside humidity is T d =T-((100-RH) / 5).

[0118] Among them, T d is the dew point temperature, T is the trackside temperature, and RH is the trackside humidity.

[0119] The third prediction module 504 is used to predict the generation of condensation on the rail surface if the predicted trackside wind speed is less than the wind speed threshold and the predicted rail surface temperature is less than the predicted dew point temperature.

[0120] The device provided in this embodiment predicts rail surface condensation based on the rail surface parameters and railside parameters in the past time period, combined with the relationship between dew point temperature and railside temperature and railside humidity, thereby realizing the pre-judgment of condensation on the rails, providing strong protection for the safety of rail transit operation, and reducing potential risks and losses caused by condensation.

[0121] The same inventive concept as the rail condensation prediction method based on time series data, this embodiment provides an electronic device, the electronic device is as follows Figure 6 As shown, it includes: a memory 601, a processor 602, and a computer program.

[0122] The computer program is stored in the memory 601 and is configured to be executed by the processor 602 to implement the above-mentioned rail condensation prediction method based on time series data.

[0123] Specifically,

[0124] Get the track surface parameters and trackside parameters of the past time period. The track surface parameters include: track surface temperature, air temperature, air humidity. The trackside parameters include: trackside temperature, trackside humidity, trackside wind speed.

[0125] The track surface temperature in the future period is predicted based on the track surface parameters in the past period. The track side temperature, track side humidity and track side wind speed in the future period are predicted based on the track side parameters in the past period.

[0126] Based on the predicted trackside temperature and the predicted trackside humidity, combined with the relationship between the dew point temperature and the trackside temperature and the trackside humidity, the dew point temperature in the future period is predicted.

[0127] Predict rail surface condensation based on predicted rail surface temperature, predicted trackside wind speed, and predicted dew point temperature.

[0128] Optionally, track surface parameters and trackside parameters are collected through patch temperature sensors and multi-factor small weather stations.

[0129] The past period includes multiple seasons and multiple climate conditions within its time span.

[0130] Optionally, predicting the rail surface temperature in a future time period based on the rail surface parameters in the past time period includes:

[0131] Through the recurrent neural network in the rail surface temperature prediction model, the changing pattern and trend of the rail surface temperature over time are determined based on the rail surface temperature in the past period.

[0132] Through the multi-layer perceptron in the rail surface temperature prediction model, the temperature prediction features are extracted according to the air temperature and air humidity in the past time period.

[0133] The rail surface temperature prediction model is used to predict the rail surface temperature in the future period based on the change pattern and trend of the rail surface temperature over time and the temperature prediction characteristics.

[0134] Optionally, the rail surface temperature prediction model is a Decoder-only architecture.

[0135] Optionally, the trackside temperature, trackside humidity, and trackside wind speed in the future period are predicted based on the trackside parameters in the past period, including:

[0136] The trackside environment prediction model is used to predict the trackside temperature, trackside humidity and trackside wind speed in the future period based on the trackside parameters in the past period.

[0137] Among them, the trackside environment prediction model is an Encoder-Decoder architecture.

[0138] The Encoder in the Encoder-decoder architecture is used to obtain the characteristics and patterns of the trackside environment based on the trackside parameters in the past time period, and output the encoding vector of the characteristics and patterns of the trackside environment.

[0139] The decoder in the encoder-decoder architecture is used to predict the trackside temperature, trackside humidity, and trackside wind speed in the future period based on the encoding vector.

[0140] Optionally, the relationship between the dew point temperature and the trackside temperature and trackside humidity is T d =T-((100-RH) / 5).

[0141] Among them, T d is the dew point temperature, T is the trackside temperature, and RH is the trackside humidity.

[0142] Optionally, based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature, the rail surface condensation is predicted, including:

[0143] If the predicted trackside wind speed is less than the wind speed threshold and the predicted rail surface temperature is less than the predicted dew point temperature, condensation is predicted to occur on the rail surface.

[0144] The electronic device provided in this embodiment has a computer program executed by a processor to predict rail surface condensation based on rail surface parameters and railside parameters in a time series in a past period, combined with the relationship between dew point temperature and railside temperature and railside humidity, thereby realizing the pre-judgment of condensation on the rails, providing strong protection for the safety of rail transit operation, and reducing potential risks and losses caused by condensation.

[0145] With the same inventive concept as the rail condensation prediction method based on time series data, this embodiment provides a computer-readable storage medium, and a computer program is stored thereon. The computer program is executed by a processor to implement the rail condensation prediction method based on time series data.

[0146] Specifically,

[0147] Get the track surface parameters and trackside parameters of the past time period. The track surface parameters include: track surface temperature, air temperature, air humidity. The trackside parameters include: trackside temperature, trackside humidity, trackside wind speed.

[0148] The track surface temperature in the future period is predicted based on the track surface parameters in the past period. The track side temperature, track side humidity and track side wind speed in the future period are predicted based on the track side parameters in the past period.

[0149] Based on the predicted trackside temperature and the predicted trackside humidity, combined with the relationship between the dew point temperature and the trackside temperature and the trackside humidity, the dew point temperature in the future period is predicted.

[0150] Predict rail surface condensation based on predicted rail surface temperature, predicted trackside wind speed, and predicted dew point temperature.

[0151] Optionally, track surface parameters and trackside parameters are collected through patch temperature sensors and multi-factor small weather stations.

[0152] The past period includes multiple seasons and multiple climate conditions within its time span.

[0153] Optionally, predicting the rail surface temperature in a future time period based on the rail surface parameters in the past time period includes:

[0154] Through the recurrent neural network in the rail surface temperature prediction model, the changing pattern and trend of the rail surface temperature over time are determined based on the rail surface temperature in the past period.

[0155] Through the multi-layer perceptron in the rail surface temperature prediction model, the temperature prediction features are extracted according to the air temperature and air humidity in the past time period.

[0156] The rail surface temperature prediction model is used to predict the rail surface temperature in the future period based on the change pattern and trend of the rail surface temperature over time and the temperature prediction characteristics.

[0157] Optionally, the rail surface temperature prediction model is a Decoder-only architecture.

[0158] Optionally, the trackside temperature, trackside humidity, and trackside wind speed in the future period are predicted based on the trackside parameters in the past period, including:

[0159] The trackside environment prediction model is used to predict the trackside temperature, trackside humidity and trackside wind speed in the future period based on the trackside parameters in the past period.

[0160] Among them, the trackside environment prediction model is an Encoder-Decoder architecture.

[0161] The Encoder in the Encoder-decoder architecture is used to obtain the characteristics and patterns of the trackside environment based on the trackside parameters in the past time period, and output the encoding vector of the characteristics and patterns of the trackside environment.

[0162] The decoder in the encoder-decoder architecture is used to predict the trackside temperature, trackside humidity, and trackside wind speed in the future period based on the encoding vector.

[0163] Optionally, the relationship between the dew point temperature and the trackside temperature and trackside humidity is T d =T-((100-RH) / 5).

[0164] Among them, T d is the dew point temperature, T is the trackside temperature, and RH is the trackside humidity.

[0165] Optionally, based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature, the rail surface condensation is predicted, including:

[0166] If the predicted trackside wind speed is less than the wind speed threshold and the predicted rail surface temperature is less than the predicted dew point temperature, condensation is predicted to occur on the rail surface.

[0167] The computer-readable storage medium provided in this embodiment has a computer program executed by a processor to predict rail surface condensation based on rail surface parameters and railside parameters in a time series within a past period, combined with the relationship between dew point temperature and railside temperature and railside humidity, thereby realizing the pre-judgment of condensation on the rails, providing strong protection for the safety of rail transit operation, and reducing potential risks and losses caused by condensation.

[0168] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt 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.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0169] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes 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 generate 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.

[0170] 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.

[0171] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions 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.

[0172] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0173] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A rail condensation prediction method based on time series data, characterized in that: The method comprises: Acquire track surface parameters and trackside parameters in the past time period; wherein the track surface parameters include: track surface temperature, air temperature, air humidity; the trackside parameters include: trackside temperature, trackside humidity, trackside wind speed; Predicting the rail surface temperature in a future period based on the rail surface parameters in the past period; predicting the railside temperature, railside humidity, and railside wind speed in the future period based on the railside parameters in the past period; Based on the predicted trackside temperature and the predicted trackside humidity, combined with the relationship between the dew point temperature and the trackside temperature and the trackside humidity, predict the dew point temperature in the future period; Predict rail surface condensation based on predicted rail surface temperature, predicted trackside wind speed, and predicted dew point temperature.

2. The method according to claim 1, characterized in that The rail surface parameters and railside parameters are collected by patch-type temperature sensors and multi-factor small weather stations; The past period includes a plurality of seasons and a plurality of climate conditions within its time span.

3. The method according to claim 1, characterized in that The method of predicting the rail surface temperature in a future time period according to the rail surface parameters in the past time period includes: Through the recurrent neural network in the rail surface temperature prediction model, the change pattern and trend of the rail surface temperature over time are determined according to the rail surface temperature in the past period; Through the multi-layer perceptron in the rail surface temperature prediction model, the temperature prediction features are extracted according to the time series of air temperature and air humidity in the past period; The rail surface temperature prediction model is used to predict the rail surface temperature in the future period based on the change pattern and trend of the rail surface temperature over time and the temperature prediction characteristics.

4. The method according to claim 3, characterized in that The rail surface temperature prediction model is a decoder-only architecture.

5. The method according to claim 1, characterized in that The predicting of the trackside temperature, trackside humidity and trackside wind speed in the future time period based on the trackside parameters in the past time period includes: Using the trackside environment prediction model, the trackside temperature, trackside humidity and trackside wind speed in the future period are predicted based on the trackside parameters in the past period; Wherein, the trackside environment prediction model is an Encoder-Decoder architecture; The encoder in the encoder-decoder architecture is used to obtain the characteristics and patterns of the trackside environment according to the trackside parameters in the past time period, and output the encoding vector of the characteristics and patterns of the trackside environment; The Decoder decoder in the Encoder-decoder architecture is used to predict the trackside temperature, trackside humidity, and trackside wind speed in the future period according to the encoding vector.

6. The method according to claim 1, characterized in that The relationship between the dew point temperature and the trackside temperature and trackside humidity is T d =T-((100-RH) / 5); Among them, T d is the dew point temperature, T is the trackside temperature, and RH is the trackside humidity.

7. The method according to claim 1, characterized in that The predicting of rail surface condensation based on the predicted rail surface temperature, the predicted trackside wind speed, and the predicted dew point temperature includes: If the predicted trackside wind speed is less than the wind speed threshold and the predicted rail surface temperature is less than the predicted dew point temperature, condensation is predicted to occur on the rail surface.

8. A rail condensation prediction device based on time series data, characterized in that: The device comprises: An acquisition module is used to acquire track surface parameters and trackside parameters in a past time period; wherein the track surface parameters include: track surface temperature, air temperature, air humidity; the trackside parameters include: trackside temperature, trackside humidity, trackside wind speed; A first prediction module is used to predict the rail surface temperature in a future period according to the rail surface parameters of the past period acquired by the acquisition module; and predict the railside temperature, railside humidity and railside wind speed in the future period according to the railside parameters of the past period acquired by the acquisition module; A second prediction module is used to predict the dew point temperature in the future period based on the trackside temperature and the predicted trackside humidity predicted by the first prediction module and in combination with the relationship between the dew point temperature and the trackside temperature and the trackside humidity; The third prediction module is used to predict rail surface condensation based on the rail surface temperature predicted by the first prediction module, the predicted trackside wind speed, and the dew point temperature predicted by the second prediction module.

9. An electronic device, characterized in that: include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and is configured to be executed by the processor to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon; the computer program is executed by a processor to implement the method according to any one of claims 1 to 7.

Images