JP7454088B2 - Information processing device, information processing method, and computer program - Google Patents

Description

Embodiments of the present invention relate to an information processing device, an information processing method, and a computer program.

In order to continue the safe and stable operation of a railway system, it is essential to confirm the soundness of the railway system through inspections and make repairs as necessary. In particular, it is important to inspect parts that wear out with each use, such as brake shoes or pantograph sliders. If excessive wear is found during inspection, the safety of the railway system can be ensured by replacing the worn parts.

Conventionally, as a method for estimating the amount of wear of brake shoes or sliders, there is a method of estimating the amount of wear of brake pads based on the distance traveled. Furthermore, there is also a method that estimates the amount of wear based on the cumulative travel distance and the temperature during braking. There is also a method that uses a video camera to detect the amount of wear on a pantograph.

However, when these methods are applied to a vehicle that is not equipped with a sensor necessary for estimation (for example, a temperature sensor at a specific location, a video camera, etc.), it is necessary to newly add the sensor.

Japanese Patent Application Publication No. 2016-117357 International Publication No. 2017/188282 International Publication No. 2016/181280

Embodiments of the present invention predict the wear state of wear parts using data measured for operation control in a vehicle.

The information processing device according to the present embodiment records the history of the measured value of the pressing force on the wear part by a mechanism that presses the wear part provided on the vehicle against an object and exerts an effect on the vehicle, and the wear part of the wear part. constructing an estimation model of the wear state based on a history of inspection values representing the state, and obtaining a predicted value of the pressing force against the wear parts by the mechanism of the vehicle based on operation plan data of the vehicle; A processing unit is provided that predicts a wear state of wear parts provided in the vehicle based on the predicted value of the pressing force and the estimation model.

FIG. 1 is a block diagram of a wear prediction device according to a first embodiment. The figure which shows the example of inspection data. The figure which shows the 1st example of vehicle data. An explanatory diagram of brake shoes, brake cylinder pressure, and air spring pressure. An explanatory diagram of a pantograph slider. The figure which shows the 2nd example of vehicle data. The figure which shows the example of learning data and teacher data. The figure which shows other examples of learning data and teacher data. A diagram showing an example of operation data (operation results and operation plan). The figure which shows the example of a data display based on a prediction result. The figure which shows the other data display example based on a prediction result. The figure which shows the example of still another data display based on a prediction result. Flowchart of learning phase processing. Flowchart of prediction phase processing. FIG. 2 is a block diagram of a wear prediction device according to a second embodiment. FIG. 2 is a diagram showing a hardware configuration of a wear prediction device according to any of the first to third embodiments.

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a wear prediction device according to an embodiment of the present invention. The wear prediction device of FIG. 1 includes an inspection data storage section 11, a vehicle data storage section 12, a model construction section 13, a model storage section 14, an operation data storage section 15, a wear prediction section (hereinafter referred to as a prediction section) 16, and a prediction result storage section. 17, and a prediction result output unit 18. The wear prediction device of FIG. 1 may further include an input device for inputting instructions or data by a user who is an operator or administrator of the device. In this case, the input device is, for example, a keyboard, mouse, touch pad, or smartphone.

The wear prediction device shown in Figure 1 collects and analyzes data related to vehicles such as railway vehicles, and creates a model that estimates the wear state (amount of wear, whether or not replacement is required, etc.) of wear parts used in vehicles. To construct. The model is then used to predict the future wear status of the wear parts of the target vehicle. This allows inventory management of wear parts and optimization of inspection timing. In this embodiment, a railway vehicle is used as the vehicle, but a similar embodiment can be realized with other types of vehicles such as automobiles. The railway vehicle may be a single vehicle or may be a vehicle formation in which a plurality of vehicles are connected.

[Learning phase]
The inspection data storage unit 11 stores a history of inspection values (inspection data) of wear parts to be monitored among the wear parts of each vehicle in each vehicle formation. Wear parts are parts that can be replaced due to wear. The inspection data includes a measurement history of wear parts or a history of replacement of wear parts. An example of a wear part is a brake shoe for braking the wheels of a vehicle. There are also pantograph sliders that collect current from the trolley wire (overhead wire). Other parts may also be used.

During operation of the vehicle, the wear parts are pressed against an object by a mechanism provided in the vehicle and exert an effect on the vehicle. If the wear part is a brake shoe, the object is a vehicle wheel, and the mechanism is a brake cylinder (BC: brake cylinder), the brake cylinder presses the brake shoe against the wheel and decelerates the vehicle (brake). ). Further, if the wear part is a pantograph slider, the object is an overhead wire, and the mechanism is a pantograph, the pantograph presses the slider against the overhead wire, and current is collected from the overhead wire to the vehicle.

Each wear part is given a wear part ID. For example, assume that the number of vehicles in a certain vehicle formation is 10, and brake shoes are installed at four locations as wear parts for each vehicle. If all brake shoes are to be monitored, inspection data for a total of 40 wear parts for the vehicle configuration is stored. Inspection data is similarly stored for other vehicle formations. Information on which position of each wear part in which vehicle of which vehicle formation is installed may be stored in the inspection data storage unit 11 or another storage unit (not shown). Note that when a wear part installed at a certain location is replaced with a new wear part, the wear part ID of the replaced wear part is assumed to be the same as that of the wear part before replacement. However, the ID may be changed before and after the exchange.

FIG. 2A shows an example of the measurement history of the wear amount of each wear part (brake shoes in this case) as inspection data. FIG. 2B shows an example of the replacement history of each wear part (brake shoes in this case) as inspection data. In FIG. 2(A) or FIG. 2(B), the measurement history or replacement history of the wear part with ID001 is displayed.

The measurement history in FIG. 2A includes the time of inspection (inspection time) and the measured value of the amount of wear. In this example, inspections (measurements) are performed approximately every three months. The amount of remaining thickness may be measured instead of the amount of wear.

In the case of the replacement history shown in FIG. 2(B), it includes information on the time of inspection (inspection time) and whether or not there has been replacement. In this example, information on whether or not replacement is required is stored for each inspection period. However, the format of the exchange history is not limited to this. For example, only the time when worn parts were replaced may be stored.

The vehicle data storage unit 12 stores, as vehicle data, the history of measured values of sensors (on-vehicle sensors) mounted on the vehicles in the vehicle composition and the history of driving command values in the vehicle composition.

FIG. 3 shows an example (first example) of vehicle data stored in the vehicle data storage section 12. In this example, data such as a power running command, a brake command, air spring (AS) pressure, and brake cylinder (BC) pressure for each vehicle are stored in chronological order for a certain vehicle configuration. There is. In this example, the data is time series data every second. Similar vehicle data is stored for other vehicle formations as well. A vehicle formation ID may be assigned to each vehicle formation, and the vehicle data of these vehicle formations may be stored in the same database. The brake command value is a brake control command value corresponding to a plurality of brake notches displayed on a brake lever provided in the driver's cab. The larger the value of the brake command, the greater the braking force. The value of the power running command is a control command value for performing acceleration/deceleration corresponding to a plurality of power running notches displayed on a power running lever provided in the driver's cab. It means that the larger the value of the power running command, the faster the speed will be increased. Note that the brake lever and the power running lever are examples of means for giving a brake command and a power running command, and these commands may be given by other means, such as a handle.

Here, the brake shoes, BC pressure, and AS pressure will be explained using FIG. 4.
FIG. 4 schematically shows the wheels of a vehicle running on the rails 21 and their peripheral configuration. Wheels 22 are mounted on rails 21. The vehicle is provided with a tread brake 23, which is a type of air brake. The wheels 22 are braked by the tread brake 23. Although only one wheel 22 is shown here, in reality, one vehicle is provided with a plurality of pairs of left and right wheels. For example, four air brakes are provided for each vehicle. However, there may be vehicles that are not equipped with air brakes. Also, one vehicle may be provided with a plurality of air brakes.

The tread brake 23 is powered by an air cylinder 24. By increasing the brake cylinder pressure (BC pressure), which is the pressure inside the air cylinder 24, the brake shoes 25 are pressed against the tread surface of the wheel 22 (the surface in contact with the rail 21). The BC pressure is the force that presses the brake shoe 25 against the tread of the wheel, and the force with which the brake shoe 25 is pressed against the wheel 22 is proportional to the BC pressure. The frictional force between the wheel 22 and the brake shoe 25 becomes the braking force of the tread brake 23.

Due to the frictional force between the wheels 22 and the brake shoes 25, the brake shoes 25 wear out with use. When the brake shoes 25 wear out, the braking force may decrease, so it is necessary to replace the brake shoes 25 depending on the amount of wear. Therefore, during inspection, the amount of wear on the brake shoes 25 is measured, and if the amount of wear is greater than a specified value, work such as replacing them with new ones is performed. The measured amount of wear is stored in the inspection data storage unit 11 as an inspection value together with the inspection date and time.

In FIG. 4, the cylinder that applies pressure to the brake shoes 25 is an air cylinder whose working fluid is air, but it may also be a cylinder that uses a working fluid other than air, such as a hydraulic cylinder.

The braking force of the tread brake 23 varies not only due to wear of the brake shoes 25 but also due to the load on the vehicle. As shown in FIG. 4, a variable load device 26 is mounted on the vehicle. The variable load device 26 includes an air spring 27, and by detecting the air spring pressure (AS pressure) of the air spring 27, the load applied to the vehicle can be measured. AS (Air spring) pressure depends on the number of passengers in the vehicle. As the number of passengers increases, the AS pressure increases. The braking force of the brake is adjusted according to the AS pressure detected by the load adjustment device 26. For example, the greater the AS pressure, the stronger the braking force of the brake. The stronger the braking force of the brake, the greater the amount of wear on the brake shoes 25, so the AS pressure affects the amount of wear on the brake shoes 25.

As mentioned above, the wear parts may be other parts, such as pantograph sliders, in addition to brake shoes.

FIG. 5 is a perspective view of the pantograph. The pantograph 31 is a device that is installed on the roof of the vehicle and receives electricity from a trolley wire (overhead wire) 32. The pantograph sliding plate 33 is a friction member installed on the top of the pantograph 31. The slider plate 33 is a member that receives electricity while moving in contact with the trolley wire (while being rubbed against the trolley wire). Therefore, the slider plate 33 wears out due to friction with the trolley wire. If the amount of wear exceeds the specified value, perform work such as replacing it with a new one.

When the wear part is a pantograph slider, the vehicle data may include at least one of the pressure to press the slider, vertical displacement of the pantograph, speed, current flowing through the overhead wire, voltage, etc. For example, in the case of a configuration in which the slide plate is pressed against the overhead wire by a spring, spring pressure can be used as the pressure for pressing the slide plate. All or part of the items shown in FIG. 3 may be included in the vehicle data.

The vehicle data may not be time-series data as shown in FIG. 3, but may be data obtained by integrating the measured values of on-board sensors and operation commands for each operation of the vehicle formation. An example of vehicle data in this case will be explained using FIG. 6.

FIG. 6 shows another example (second example) of vehicle data. For a certain vehicle formation, a service ID is set for each service. For each operation, the departure time, end time, cumulative time for each power running command value, cumulative time for each brake command value, and the AS pressure cumulative value and BC pressure cumulative value for each car are stored. . Similar data is stored for other vehicle formations as well. Here, operation means that a vehicle formation runs from a departure point to an end point according to a predetermined timetable (operation plan). Here, the integrated value is stored for each service, but the integrated value may be stored in other units, such as for each station section.

The model construction unit 13 uses vehicle data (history of measured values and history of operation commands) as learning data, inspection data (history of measurement of wear amount or history of replacement of worn parts) as training data, and uses the wear state (amount of wear or Build a model to estimate whether or not replacement is necessary.

An example of constructing an estimation model will be shown for each case of using the measurement history of wear parts (see FIG. 2(A)) and the case of using the replacement history (see FIG. 2(B)) as training data.

[When using measurement history of wear parts as training data]
Learning data corresponding to a plurality of explanatory variables and teacher data corresponding to a target variable are created for each measurement period (inspection period) of inspection data, using the point immediately after replacement (the point in time when the amount of wear is 0) as a reference. Learning data and teacher data are created for each wear part.

FIG. 7 shows an example of the created learning data and teacher data. In this example, the wear amount measurement history shown in FIG. 2(A) is used as training data.

For example, the amount of wear during the period (inspection interval) from the inspection time 2018/3/10 10:00 to the next inspection time 2018/6/8 17:00 in FIG. 2(A) increases by 0.1. Corresponding to the inspection time of 2018/6/8 17:00, in the first line of FIG. 7, a wear amount of 0.1 is set as teacher data.

Further, the values of the explanatory variables during the above period are calculated from the vehicle data, and the calculated values of the explanatory variables are cumulatively added and set as learning data. As explanatory variables, it is desirable to use data that affects the wear of wear parts. In the example in the figure, the cumulative value of power running command value 1, the cumulative value of power running command value 2, the cumulative value of power running command value 3...the cumulative value of power running command value 6, the cumulative value of AS pressure, the cumulative value of BC pressure etc. are set as learning data. A cumulative value of the power running command value 0 may be further set.

The cumulative value of BC pressure in FIG. 7 is the cumulative value of BC pressure measured for a vehicle in which a wear part is provided. However, it may be the average of the cumulative values of the BC pressures measured for a plurality of vehicles (for example, one vehicle in front and the rear) including the vehicle in question, or the average of the cumulative values of the BC pressures measured for all the vehicles. What has been described here applies to AS pressure as well.

The items of explanatory variables are not limited to the example shown in FIG. For example, it may be the cumulative number of times of sudden braking or the cumulative value of speed (speed is a value related to the rotational speed of the wheels). The speed may be corrected based on wheel diameter. Also, different explanatory variables may be used depending on whether the vehicle is stopped or in motion. Alternatively, the value of the explanatory variable may be calculated based only on data during braking.

Note that the format of the vehicle data used to create the learning data does not matter, and any of the vehicle data shown in FIGS. 3 and 6 may be used.

Similarly, teacher data and learning data are created for other inspection times in the inspection data.

As described above, once the learning data and teacher data as shown in FIG. 7 are prepared, an estimation model is constructed by machine learning to calculate the objective variable (in the example of FIG. 7, the amount of wear) from the explanatory variables. Any algorithm can be used for learning, including linear regression, logistic regression, decision trees, random forests, and neural networks. It is also possible to construct multiple types of models and select the model with the least model error (described later). Alternatively, a model that integrates the results of multiple models to produce a single solution, such as ensemble learning, may be used.

The method for constructing the estimation model will be explained below. Assume that a linear regression model is used. An example of an estimation model function (model function) is shown in equation (1).

y is an objective variable, x ₁ to x _n are explanatory variables, and b ₀ to b _n are regression coefficients. Note that in order to absorb differences in the measurement units of each explanatory variable, the objective variable and all explanatory variables may be normalized to have a mean value of 0 and a variance of 1.

Calculation of explanatory variables and regression coefficients (parameter calculation) is performed, for example, by minimizing an objective function that defines the difference between the output value of the function and the training data (ie, model error). For example, explanatory variables and regression coefficients are calculated by solving the following problem of minimizing the squared error. i represents the number of learning data (the number of teacher data).

[When using the replacement history of worn parts as training data]
When using the replacement history of worn parts as inspection data, a binary objective variable indicating whether or not it was replaced is used as training data (objective variable value) for each inspection period based on the point immediately after replacement. Set. Furthermore, the values of the explanatory variables for the period from the previous inspection time to the current inspection time are calculated from the vehicle data and set as learning data. The method for creating learning data is the same as the method described above.

FIG. 8 shows an example of the created learning data and teacher data. It is the same as FIG. 7 except that the value of the target variable (teacher data) has been changed to binary information indicating whether or not exchange is performed. Regarding the value of the objective variable, 1 means exchange is present and 0 means no exchange.

In this case as well, any algorithm used for learning may be used, such as SVM, logistic regression, decision tree, random forest, or neural network. For example, if the teacher data is whether exchange is required, the objective variable is, for example, whether exchange is necessary.

For example, in the case of logistic regression, a logistic function (sigmoid function) can be used as the model function. An example of the logistic function is shown in equation (3). The definition of each symbol in the formula is the same as in formula (1). Calculation of the regression coefficients b ₀ to b _n can be performed by a general method by optimizing (minimizing or maximizing) a predetermined objective function. The output value of the logistic function is in the range greater than 0 and less than 1. The output value of the logistic function corresponds to the probability that a worn part will need to be replaced.

It is also possible to construct a plurality of types of models and select a model whose statistical values, such as the average, minimum value, and maximum value of model errors, are the minimum or below a threshold value. Model error is the difference between the model output (value of objective variable) and the value of inspection data for testing. When using teacher data used during learning as test inspection data, explanatory variables calculated from vehicle data used for learning are used as model inputs. When inspection data not used for learning is used as inspection data for testing, explanatory variables calculated from vehicle data not used for learning are used as model inputs.

Alternatively, an estimation model may be constructed using survival time analysis. In this case, an estimation model can be constructed by selecting one explanatory variable, treating the value of the selected explanatory variable as time, and performing survival time analysis. Any algorithm such as Kaplan-Meier method or Weibull method may be used. As an example, when BC pressure is used as an explanatory variable, a graph in which the horizontal axis is the cumulative value of BC pressure and the vertical axis is the survival probability is created as an estimation model based on vehicle data and inspection data. The training data uses binary information indicating whether or not the data has been exchanged. The cumulative value of BC pressure is calculated based on the test vehicle data, and the survival probability corresponding to the calculated value is identified from the graph. If the identified survival probability is less than or equal to or greater than, it is determined that the worn part needs to be replaced.

The model storage unit 14 stores the estimated model constructed by the model construction unit 13. Since an estimated model is generated for each wear part to be monitored, the model storage unit 14 stores an estimated model for each wear part to be monitored. Further, the type of model to be constructed may be changed depending on the type of knitting, and the materials and types of sliders and brake shoes.

[Prediction phase]
The operation data storage unit 15 stores operation results and operation plans (diagram data) of each vehicle formation as operation data.

Examples of operation data are shown in FIGS. 9(A) and 9(B). FIG. 9(A) shows operation performance data including departure time, fixed time, departure station, fixed station, etc. regarding the past operation of each vehicle composition. FIG. 9B shows operation plan data including departure time, fixed time, departure station, fixed station, etc. for the planned operation of each vehicle formation. A service ID is assigned to each service. Data associating the operation ID with the ID of the vehicle formation may be stored in the operation data storage unit 15 or a storage unit (not shown). Information on other items may be further added to the operation record data in FIG. 9(A) and the operation plan data in FIG. 9(B). For example, information about stations stopped or passed along the way may be added.

The prediction unit 16 uses the estimated model stored in the model storage unit 14, operation plan data of the target vehicle, and operation performance data of at least one vehicle (which may or may not include the target vehicle). and vehicle data (history of measured values and history of driving commands) in the vehicle data storage unit 12 to predict the wear state of the wear parts used in the target vehicle. Specifically, the prediction unit 16 predicts the values of the explanatory variables when the operation plan indicated by the operation plan data of the target vehicle is executed, and assigns the predicted values of the explanatory variables to the explanatory variables of the estimation model (estimated (used as input variables of the model), and calculate the output value (objective variable) of the estimated model. The output value of the estimation model is a predicted value of the wear state of the wear part.

In order to calculate predicted values of explanatory variables, cases similar to the operation plan of the target vehicle (target operation record data) are extracted from past operation record data. For example, assume that there is a plan (operation plan for the target vehicle) for a certain vehicle formation to travel from 〇〇 station to △△ station over a two-hour period as an operation plan one day after the current day. In this case, the past operation results are searched for the most similar case to the operation plan. For example, cases that operate in the same section from 〇〇 station to △△ station are identified, and from among the identified cases, a case is selected based on the difference in operating time (2 hours) of the target vehicle. . For example, a case with the smallest difference in operating times is searched as a similar case. Calculate the value of the explanatory variable from the vehicle data corresponding to the found case. The calculated value of the explanatory variable is used as a predicted value of the explanatory variable for the operation plan of the target vehicle. Although one case is selected here as a similar case, a plurality of cases may be selected. In this case, for example, the values of the vehicle data corresponding to the selected cases may be averaged, and the value of the explanatory variable may be calculated based on the averaged vehicle data.

When searching for past similar cases, for example, the similarity of vehicles or equipment used in vehicles (wear parts or other parts, etc.) may be considered. For example, the search may be narrowed down to data on vehicle formations with the same type of wear parts (for example, the same type of brake shoes used in each vehicle). Thereby, it is possible to improve the calculation accuracy of the predicted value of the explanatory variable.

After calculating the predicted value of the explanatory variable, the prediction unit 16 inputs the calculated predicted value to the estimation model and calculates the output value of the estimation model (value of the objective variable). The calculated value is a predicted value of the wear condition. The predicted value is the amount of wear when the teacher data used when constructing the model is the measured value of the amount of wear. If the training data used at the time of model construction indicates whether or not replacement is required, the predicted value is the presence or absence or probability that replacement is necessary. Whether the predicted value is whether or not replacement is necessary or its probability depends on the algorithm used to build the model. For example, in the case of logistic regression, the predicted value is the probability that replacement is necessary, and in the case of a decision tree, the predicted value is the probability that replacement is necessary. However, it is also possible to construct a model that predicts the probability using a decision tree. Note that instead of the probability that replacement is necessary, the probability that replacement is not necessary may be used.

The prediction unit 16 can obtain time-series data of the predicted value of the wear state of the wear part of the target vehicle by advancing the prediction target day one business day at a time and repeatedly performing the same prediction process.

The prediction result storage unit 17 stores the predicted value of the wear state calculated by the prediction unit 16. As the prediction unit 16 repeatedly performs the prediction process, the prediction result storage unit 17 stores time-series data of predicted values of wear information. By similarly performing prediction processing on other wear parts of the target vehicle, predicted values of the wear states of other wear parts or their time-series data are stored in the prediction result storage unit 17.

The prediction result output unit 18 includes a display device that displays data or information on a screen. The prediction result output unit 18 generates display data based on the predicted values stored in the prediction result storage unit 17, and displays the generated data on the screen. For example, based on the predicted values stored in the prediction result storage unit 17, time-series data of the predicted values is displayed on the screen.

FIG. 10 shows an example of data display of the prediction result output unit 18. The trends in the predicted wear amount are displayed. The horizontal axis represents time, and the vertical axis represents the predicted value of wear amount. The specified value (threshold value) of wear amount, which is a guideline for replacement, and the inspection date are also displayed. The specified value is a standard for replacing worn parts, and when the amount of wear exceeds the specified value, it becomes a standard for replacing worn parts. In the illustrated example, the predicted value of the amount of wear is less than the specified value on the next inspection date, but the predicted value of the amount of wear is greater than or equal to the specified value on the next inspection date. Therefore, it can be seen at a glance that the worn parts will be replaced or are likely to be replaced on the following inspection dates.

FIG. 11 is a table storing IDs of wear parts whose predicted wear amount is equal to or greater than a specified value among the wear parts to be monitored. The data is sorted in descending order of the date when the value exceeded the specified value. It is predicted that the wear amount of the wear part with ID 122 will exceed the specified value before the next scheduled inspection date. For this reason, the background of the data of ID122 is painted in a color to call attention to it.

FIG. 12 is a graph showing the predicted number of replacements of worn parts in chronological order. The horizontal axis represents time, and the vertical axis represents the predicted cumulative replacement number of worn parts. The predicted cumulative number of replacements is calculated for all vehicle formations (or specific formations) that use wear parts of the same model (for example, wear parts of the same model number), and the wear parts whose predicted amount of wear exceeds the specified value. is the total number of The cumulative number of exchanges up to the previous month is plotted on each day of each month. In the illustrated example, brake shoes and pantograph sliders are shown as wear parts.

By displaying a graph like that shown in Figure 12, it is easy to know when and how many wear parts are needed. The stock of wear parts may be managed in a database. In this case, the time when stock will run short may be identified and a warning mark displayed. In Figure 12, the cumulative predicted number of brake shoes to be replaced in May will be 12, exceeding the stock of 10, so a warning mark M is displayed.

In FIGS. 10 to 12, the output value of the estimation model is the estimated value of the amount of wear, but the data can be displayed in the same way if it is the probability that a worn part needs to be replaced or whether or not it needs to be replaced. In the case of the probability that a worn part needs to be replaced, if the probability is greater than or equal to a threshold value, it may be determined that replacement is necessary, and if it is less than the threshold value, it may be determined that replacement is not necessary.

FIG. 13 is a flowchart of the learning phase according to this embodiment. This process is started by a learning phase start trigger (S100). The start trigger may be, for example, receiving an instruction from the user of the device from the input device, a predetermined time, or the establishment of any other event.

When the learning phase starts, the model construction unit 13 reads inspection data from the inspection data storage unit 11 and reads vehicle data from the vehicle data storage unit 12 (S101).

Based on the read inspection data, teacher data corresponding to the target variable is generated, and based on the vehicle data, learning data corresponding to one or more explanatory variables is generated (S102).

Based on the generated teaching data and learning data, machine learning using any regression algorithm (linear regression, decision tree, random forest, neural network, etc.) is performed to determine the objective variable (e.g. wear amount) from one or more explanatory variables. ) is constructed (S103).

The model storage unit 14 internally stores the estimated model constructed by the model construction unit 13 in association with the model ID (S104). The model ID differs depending on the estimated model.

Such an estimation model is constructed for each wear part to be monitored (S105). The wear parts to be monitored are, for example, all brake shoes provided in each vehicle of each vehicle formation. However, one brake shoe may be selected as a representative of the group for each of a plurality of vehicles, and the selected brake shoe may be monitored. In this case, when replacing the brake shoe to be monitored, it may be determined that other brake shoes belonging to the same group are also replaced at the same time. The wear parts to be monitored may be determined using methods other than those described herein.

FIG. 14 is a flowchart of the prediction phase according to this embodiment. This process is started by a prediction phase start trigger (S200). The start trigger may be, for example, receiving an instruction from the user of the device, a predetermined time, or the establishment of any other event. This process is performed for each wear part to be monitored for the target vehicle.

The prediction unit 16 reads the operation plan data of the target vehicle from the operation data storage unit 15 (S201). As an example, the operation plan data of the target vehicle for the business day following the day when a wear part to be monitored in the target vehicle is replaced is read.

Cases similar to the operation plan data of the target vehicle are identified from the operation performance data of one or more vehicles in the operation data storage unit 15 (S202).

Vehicle data corresponding to the identified case is read from the vehicle data storage unit 12, and predicted values of explanatory variables are calculated for the wear parts to be monitored (wear parts mounted on the target vehicle) (S203).

The wear state (for example, the amount of wear) is predicted based on the calculated predicted value of the explanatory variable and the estimated model corresponding to the wear part (S204).

The prediction result storage unit 17 stores the predicted value of the wear state in association with the wear part ID (S205).

The prediction result output unit 18 generates display data (for example, a transition graph of predicted values) based on the predicted values stored in the prediction result storage unit 17, and displays the generated data (S206).

The above processing is performed, for example, while advancing the operating days (business days) one day at a time. When calculating the value of an explanatory variable (for example, cumulative value of BC pressure) for each day after the second day (prediction target day), the value of the explanatory variable (for example, the cumulative value of BC pressure) is calculated based on the vehicle data for the case specified for each day up to the prediction target date It is sufficient to calculate the value of the explanatory variable for the day. Note that instead of proceeding with this process one day at a time immediately after replacing the worn parts, it is naturally possible to predict the date several days later. In that case, similarly, similar past cases are identified based on the operation plan for each day from immediately after the replacement of worn parts until the prediction target date, and an explanation of the prediction target date is given based on vehicle data corresponding to the identified cases. All you have to do is calculate the value of the variable.

As described above, according to the present embodiment, it is possible to predict with high accuracy the wear state (amount of wear, whether or not replacement is required, etc.) of wear parts at a future point in time. The related technology methods only estimate the current amount of wear, and cannot estimate the amount of wear at a future point in time. In contrast, in this embodiment, the wear state at each point in the future can be predicted with high accuracy at the present time. This makes it possible to optimize the timing of inspection of wear parts, and also to optimize the inventory management of wear parts. Furthermore, according to the present embodiment, since learning is performed using data measured by the vehicle for operation control (BC pressure, AS pressure, power running command, brake command, etc.), it is not necessary to add new sensors or the like to the vehicle. is not necessarily necessary.

(Modified example)
In the first embodiment described above, the prediction unit 16 predicts the future wear state based on the operation plan data of the target vehicle, but the prediction unit 16 estimates the current or past wear state based on the operation record data of the target vehicle. Good too. For example, the state of wear on an arbitrary day after the previous inspection time may be estimated from actual operation data.

(Second embodiment)
FIG. 15 is a block diagram of a wear prediction device according to a second embodiment of the present invention. An environmental data storage section 19 is added to the wear prediction device of FIG. The environmental data storage unit 19 stores data (environmental data) related to the operating environment during operation of the vehicle formation.

An example of the environmental data is route information such as route gradient and route curve. Route information is associated with position information on the map. Furthermore, in this embodiment, position information is also associated with vehicle data (see FIG. 3). Specific examples of location information include location information in kilometers (distance from the starting point) or GPS (Global Positioning System).

Other examples of environmental data include weather information such as temperature and precipitation. Weather information is associated with location information and time information. Public data from the Japan Meteorological Agency or the like may be used as the weather information. When using weather information, use the weather information of the observation position closest to the position information of the vehicle data.

In this embodiment, the model construction unit 13 constructs an estimation model using the history of environmental data. For example, when the model construction unit 13 creates learning data, the BC pressure cumulative value of the slope section, the BC pressure cumulative value of the non-gradient section, the BC pressure cumulative value of the curve section, the BC pressure cumulative value of the non-curve section, the BC pressure cumulative value of the non-curve section, and Add explanatory variables according to environmental conditions, such as the cumulative BC pressure value. Here, a slope section is an area where the slope value is above a certain value, a non-slope section is an area where the slope value is less than a certain value, a curve section is an area where the radius of curvature is below a certain value, and a non-curve section is an area where the radius of curvature is a constant value. Assume that the interval is as follows. Here, the gradient is divided into two sections and the curve is divided into two sections, but each section may be divided into three or more sections, and explanatory variables may be added according to each section. Since the amount of wear may change depending on the degree of slope and curve, or weather conditions such as rain or snow, estimation accuracy can be improved by increasing the number of explanatory variables for each environmental condition. .

When performing estimation based on the estimation model, it is sufficient to calculate the operating environment during the operation period of the operation plan data of the target vehicle, and calculate the value of the added explanatory variable based on the environmental data representing the calculated operating environment.

Furthermore, in this embodiment, when searching for cases similar to the operation plan data of the target vehicle from the operation record data, the similarity of the environmental data may be used. For example, the operation environment (season in this case) during the operation period of the operation plan data of the target vehicle is calculated, and the search range in the operation performance data is narrowed down to the same season as the calculated season. As another example, values such as the amount of precipitation are identified for each case from past weather information, and cases with values close to the weather forecast values are selected. Thereby, it is possible to improve the calculation accuracy of the predicted value of the explanatory variable. The selection of cases by this method may or may not be combined with the creation of learning data of this embodiment described above.

(Third embodiment)
In the first or second embodiment described above, a correction term (explanatory variable) depending on the configuration of the vehicle or equipment mounted on the vehicle may be added to the model function. For example, there is a correction factor depending on what car number the brake shoes belong to, whether it is an M car (vehicle with power such as a motor) or a T car (vehicle without power such as a motor, i.e. an accompanying vehicle). A correction item depending on whether the brake shoe is on the front wheel or the rear wheel, a correction item depending on whether the brake shoe is on the front wheel or the rear wheel, and whether the brake shoe is provided on the left wheel or the right wheel in the direction of travel. There are correction terms depending on the type of brake shoe and correction term depending on the type of brake shoe. Further, there are correction terms depending on the type of vehicle formation or correction terms depending on the type of brake structure. When the wear part is a pantograph, there are also correction terms depending on the type of pantograph structure or the type of slider.

For example, if it is a correction term based on car number, prepare a variable for each car, and for the variable G1 representing car No. 1, set the value g1 (g1 is any real number other than 0) for the brake shoe located in car No. 1. , for other car numbers, give 0. Similarly, in the case of the brake shoe located in the second car, g2 (g2 is any real number other than 0) is given, and in the case of other cars, 0 is given. Correction terms are defined in the same manner for the third and subsequent cars. An example of the model function in this case is shown in equation (4) below. In this example, there are cars number 1 to 10. Other types of correction terms mentioned above can be added in a similar manner.

Different estimation models may be constructed by learning separately according to each condition without adding a correction term to the model function. For example, separate models may be constructed, such as a model for car No. 1 and a model for car No. 2.

The model construction unit 13 may analyze vehicle data and select explanatory variables to be used in the model function through learning. For example, the correlation with the target variable may be calculated for all variables of the vehicle data, and a predetermined number of variables may be selected in order of the highest correlation. For example, calculate each statistic (maximum value, minimum value, cumulative value, etc.) of all variables in vehicle data, calculate the correlation of each variable with the objective function for the statistic, and then select a predetermined number of variables in order of the highest correlation. may be selected. At this time, this process may be performed for each of the conditions described above to construct a model for each condition.

When the wear part is a pantograph slider, explanatory variables may be defined separately depending on whether the vehicle is stopped or moving, or only one of the explanatory variables may be employed. Similarly, explanatory variables may be defined separately depending on whether the pantograph is raised or not, or only one of the explanatory variables may be employed. Also, a correction term for whether or not the vehicle configuration is within the air section section (for example, value a if it is within the air section section, value b if outside the section, etc., where a and b are arbitrary real numbers), or the vehicle configuration Add a correction term for whether the distance from the substation is short (for example, the value c if the distance is below a certain value, the value d if the distance is above a certain value, etc., where a and b are arbitrary real numbers) to the model function. You can.

(Fourth embodiment)
FIG. 16 shows the hardware configuration of the wear prediction device according to this embodiment. As the wear prediction device according to this embodiment, any of the wear prediction devices according to the first to third embodiments can be used. The wear prediction device according to this embodiment is configured by a computer device 100. The computer device 100 includes a CPU 101, an input interface 102, a display device 103, a communication device 104, a main storage device 105, and an external storage device 106, which are interconnected by a bus 107.

A CPU (central processing unit) 101 executes a wear prediction program, which is a computer program, on the main storage device 105. The wear prediction program is a program that realizes each of the above-described functional configurations of the wear prediction device. Each functional configuration is realized by the CPU 101 executing the wear prediction program.

The input interface 102 is a circuit for inputting operation signals from input devices such as a keyboard, a mouse, and a touch panel to the wear prediction device.

The display device 103 displays data or information output from the wear prediction device. The display device 103 is, for example, an LCD (liquid crystal display), a CRT (cathode ray tube), or a PDP (plasma display), but is not limited thereto. The data or information generated by the prediction result output unit 18 can be displayed on this display device 103.

The communication device 104 is a circuit for the wear prediction device to communicate with an external device wirelessly or by wire. Inspection data, vehicle data, operation data, and environmental data can be input from an external device via the communication device 104. Inspection data, vehicle data, operation data, and environmental data input from an external device can be stored in the inspection data storage section 11, the vehicle data storage section 12, the operation data storage section 15, and the environmental data storage section 19. As an example, the communication device 104 may acquire vehicle data by communicating with a communication device mounted on the vehicle, and may store the acquired vehicle data in the vehicle data storage unit 12.

The main storage device 105 stores a wear prediction program, data necessary for execution of the wear prediction program, data generated by execution of the wear prediction program, and the like. The wear prediction program is developed on the main storage device 105 and executed. The main storage device 105 is, for example, RAM, DRAM, or SRAM, but is not limited thereto. The inspection data storage section 11, the vehicle data storage section 12, the operation data storage section 15, and the environmental data storage section 19 may be constructed on the main storage device 105.

The external storage device 106 stores a wear prediction program, data necessary for executing the wear prediction program, data generated by executing the wear prediction program, and the like. These programs and data are read into the main storage device 105 when the wear prediction program is executed. The external storage device 106 is, for example, a hard disk, an optical disk, a flash memory, or a magnetic tape, but is not limited thereto. The inspection data storage section 11, the vehicle data storage section 12, the operation data storage section 15, and the environmental data storage section 19 may be constructed on the external storage device 106.

Note that the wear prediction program may be installed in the computer device 100 in advance, or may be stored in a storage medium such as a CD-ROM. Further, the wear prediction program may be uploaded onto the Internet.

Further, the wear prediction device may be configured by a single computer device 100, or may be configured as a system consisting of a plurality of computer devices 100 interconnected via a network. The wear prediction device may be placed on the cloud and receive operation input from the user via the Internet.

It should be noted that the present invention is not limited to the above-described embodiments as they are, but can be implemented by modifying the constituent elements within the scope of the invention at the implementation stage. Moreover, various inventions can be formed by appropriately combining a plurality of components disclosed in each of the above embodiments. For example, a configuration in which some components are deleted from all the components shown in each embodiment is also conceivable. Furthermore, the components described in different embodiments may be combined as appropriate.

[Item 1]
Based on the history of measured values of the pressing force on the wear parts by a mechanism that presses the wear parts provided on the vehicle against the target object and exerts an effect on the vehicle, and the history of inspection values representing the wear state of the wear parts. a model construction unit that constructs an estimation model of the wear state;
Based on the operation plan data of the target vehicle, a predicted value of the pressing force on the wear parts by the mechanism of the target vehicle is obtained, and based on the predicted value of the pressing force and the estimation model, a prediction unit that predicts the wear state of wear parts;
An information processing device equipped with
[Item 2]
The prediction unit selects at least one target operation performance data from a plurality of operation performance data of at least one vehicle based on the operation plan data of the target vehicle, and selects at least one target operation performance data from among a plurality of operation performance data of at least one vehicle, and The information processing device according to item 1, wherein the measured value of the pressing force during the indicated operation period is used as the predicted value of the pressing force.
[Item 3]
The plurality of operation record data include environmental data indicating the operation environment of the vehicle,
The information processing device according to item 2, wherein the prediction unit calculates the operation environment of the target vehicle based on the operation plan data, and specifies the target operation performance data based on the calculated operation environment.
[Item 4]
The information processing device according to item 2 or 3, wherein the prediction unit sets operation record data having the same departure point and the same terminal point as the operation plan data as the target operation record data.
[Item 5]
The information processing device according to item 4, wherein the prediction unit sets, as the target operation performance data, operation performance data for which the difference from the operation time indicated by the operation plan data is the minimum or less than a threshold value.
[Item 6]
The model construction unit constructs the estimation model based on a history of environmental data representing the operating environment of the vehicle,
The information processing according to any one of items 1 to 5, wherein the prediction unit calculates the operating environment of the target vehicle based on the operation plan data, and predicts the wear state based on the calculated operating environment. Device.
[Item 7]
The estimation model cumulatively adds the predicted value of the pressing force during the operation period indicated by the operation plan data of the target vehicle, and calculates the output value of the estimation model by using the cumulative value as input to the estimation model. Item 1 ~ 6. The information processing device according to any one of 6.
[Item 8]
The wear part is a brake shoe,
The mechanism is a brake cylinder,
The target object is a wheel of the vehicle,
the action is deceleration of the vehicle;
The information processing device according to any one of items 1 to 7, wherein the inspection value represents the amount of wear of the brake shoe or whether or not it needs to be replaced.
[Item 9]
The information processing device according to item 8, wherein the pressing force is the pressure of the cylinder.
[Item 10]
The mechanism is a pantograph,
The wear part is a slide plate of the pantograph,
The target object is an overhead wire,
The action is to collect current to the vehicle,
The information processing device according to any one of items 1 to 7, wherein the inspection value represents the amount of wear of the pantograph or whether or not it needs to be replaced.
[Item 11]
Based on the history of measured values of the pressing force on the wear parts by a mechanism that presses the wear parts provided on the vehicle against the target object and exerts an effect on the vehicle, and the history of inspection values representing the wear state of the wear parts. a model construction step of constructing an estimation model of the wear state;
Based on the operation plan data of the target vehicle, a predicted value of the pressing force on the wear parts by the mechanism of the target vehicle is obtained, and based on the predicted value of the pressing force and the estimation model, a prediction step for predicting the wear state of the wear parts;
An information processing method with
[Item 12]
Based on the history of measured values of the pressing force on the wear parts by a mechanism that presses the wear parts provided on the vehicle against the target object and exerts an effect on the vehicle, and the history of inspection values representing the wear state of the wear parts. a model construction step of constructing an estimation model of the wear state;
Based on the operation plan data of the target vehicle, a predicted value of the pressing force on the wear parts by the mechanism of the target vehicle is obtained, and based on the predicted value of the pressing force and the estimation model, a prediction step for predicting the wear state of the wear parts;
A computer program that causes a computer to execute

11: Inspection data storage unit 12: Vehicle data storage unit 13: Model construction unit 14: Model storage unit 15: Operation data storage unit 16: Wear prediction unit (prediction unit)
17: Prediction result storage unit 18: Prediction result output unit 19: Environmental data storage unit

Claims (13)

Based on the history of measured values of the pressing force on the wear parts by a mechanism that presses the wear parts provided on the vehicle against the target object and exerts an effect on the vehicle, and the history of inspection values representing the wear state of the wear parts. and constructing an estimation model of the wear state,
Based on the operation plan data of the vehicle, a predicted value of the pressing force on the wear parts by the mechanism of the vehicle is obtained, and based on the predicted value of the pressing force and the estimation model, a wear part provided on the vehicle is acquired. An information processing device equipped with a processing unit that predicts the wear status of parts. The processing unit selects at least one target operation record data from a plurality of operation record data of at least one vehicle based on the operation plan data of the vehicle, and selects at least one target operation record data from among the plurality of operation record data of at least one vehicle, and selects at least one target operation record data indicated by the at least one target operation record data. The information processing device according to claim 1, wherein a measured value of the pressing force during an operation period is used as a predicted value of the pressing force. The plurality of operation record data include environmental data indicating the operation environment of the vehicle,
The information processing device according to claim 2, wherein the processing unit calculates the operating environment of the vehicle based on the operation plan data, and specifies the target operation performance data based on the calculated operating environment. The information processing device according to claim 2 or 3, wherein the processing unit sets operation record data having the same departure point and the same terminal point as the operation plan data as the target operation record data. The information processing device according to claim 4, wherein the processing unit sets, as the target operation performance data, operation performance data for which the difference from the operation time indicated by the operation plan data is the minimum or less than a threshold value. The processing unit constructs the estimation model based on a history of environmental data representing the operating environment of the vehicle,
The information processing according to any one of claims 1 to 5, wherein the processing unit calculates an operating environment of the vehicle based on the operation plan data, and predicts the wear state based on the calculated operating environment. Device. The estimation model cumulatively adds the predicted value of the pressing force during the operation period indicated by the operation plan data of the vehicle, and calculates the output value of the estimation model by using the cumulative value as input to the estimation model. 6. The information processing device according to any one of 6. The wear part is a brake shoe,
The mechanism is a brake cylinder,
The target object is a wheel of the vehicle,
the action is deceleration of the vehicle;
The information processing device according to any one of claims 1 to 7, wherein the inspection value represents the amount of wear of the brake shoe or whether or not it needs to be replaced. The information processing device according to claim 8, wherein the pressing force is pressure of the cylinder. The mechanism is a pantograph,
The wear part is a slide plate of the pantograph,
The target object is an overhead wire,
The action is to collect current to the vehicle,
The information processing device according to any one of claims 1 to 7, wherein the inspection value represents the amount of wear on the pantograph or whether or not it needs to be replaced. The processing unit includes a model construction unit that constructs an estimation model of the wear state based on the history of the measured values and the history of the inspection values;
Based on the operation plan data of the vehicle, a predicted value of the pressing force on the wear parts by the mechanism of the vehicle is obtained, and based on the predicted value of the pressing force and the estimation model, The information processing device according to any one of claims 1 to 10, further comprising a prediction unit that predicts a wear state of the wear part. Based on the history of measured values of the pressing force on the wear parts by a mechanism that presses the wear parts provided on the vehicle against the target object and exerts an effect on the vehicle, and the history of inspection values representing the wear state of the wear parts. and constructing an estimation model of the wear state;
Based on the operation plan data of the vehicle, a predicted value of the pressing force on the wear parts by the mechanism of the vehicle is obtained, and based on the predicted value of the pressing force and the estimation model, a wear part provided on the vehicle is acquired. predicting the wear state of the parts;
An information processing method with Based on the history of measured values of the pressing force on the wear parts by a mechanism that presses the wear parts provided on the vehicle against the target object and exerts an effect on the vehicle, and the history of inspection values representing the wear state of the wear parts. and constructing an estimation model of the wear state;
Based on the operation plan data of the vehicle, a predicted value of the pressing force on the wear parts by the mechanism of the vehicle is obtained, and based on the predicted value of the pressing force and the estimation model, a wear part provided on the vehicle is acquired. predicting the wear state of the parts;
A computer program that causes a computer to execute

Images