JP2023062950A - Vehicle information control device, vehicle information control system, and abnormality determination method of air conditioner - Google Patents

Abstract

To accurately detect a failure sign or abnormality of an air conditioner for a railway vehicle.SOLUTION: A vehicle information control device has: a sensor group which acquires onboard data including at least air conditioner sensor information about an air conditioner installed in a vehicle and door opening information about door opening frequencies of a door of the vehicle; a feature extraction part which extracts, from the onboard data, a feature amount group including a temperature feature amount about the vehicle and a door opening frequency feature amount about the door opening frequencies for a prescribed inspection section, through analysis of the onboard data; and an abnormality determination part which generates an abnormality determination result indicating an abnormality or abnormality sign about the air conditioner based on the feature amount group and a normal model indicating an operation characteristic during a normal operation of the air conditioner, and outputs the abnormality determination result.SELECTED DRAWING: Figure 1

Description

The present disclosure relates to a vehicle information management device, a vehicle information management system, and an abnormality determination method for an air conditioner.

In order to keep air conditioners mounted on railcars in an efficient and cost-effective operating state, it is important to perform maintenance work, and various methods have been proposed for this maintenance work.
As an example, there is known a TBM (Time Based Maintenance) method in which maintenance of an air conditioner is performed at regular intervals regardless of the presence or absence of a failure. Also, in recent years, a CBM (Condition Based Maintenance) method has been proposed in which maintenance work is performed when there is a sign of failure of an air conditioner without waiting for the maintenance period.

With CBM, it is possible to suppress the occurrence of the process of performing maintenance even though there is no failure, which occurs in TBM, so it is possible to reduce the frequency of maintenance and reduce the maintenance cost of railway vehicle air conditioning equipment. There is expected.

In order to realize CBM in railway vehicle air conditioners, it is necessary to use data from temperature sensors that monitor the state of the air conditioners to determine signs of failure of the air conditioners. Determination of a sign of failure is performed by converting sensor information such as the temperature inside the vehicle and the outside temperature, which is a load factor for cooling, into an appropriate feature amount and inputting it into a method of determining a sign of failure. For example, Patent Document 1 proposes a method of extracting a feature quantity and a method of determining a sign of failure.

In Patent Document 1, there is a problem that "Even if the operating state is wasteful in terms of energy, it cannot be grasped at an early stage, and it is not notified until the degree of the state is extremely deteriorated. Output from the air conditioner 5 is a problem. A data acquisition unit 6 that acquires operation data at a predetermined sampling time, and an operation state estimation unit 7 that estimates whether the operation state of the air conditioner 5 is normal based on the operation data acquired by the data acquisition unit 6. , the driving state estimating unit 7 targets a plurality of types of driving data, uses data obtained by processing a plurality of sampling data included in a predetermined time period for each driving data as feature amount data, and uses these feature amount data as feature amount data. It is characterized by making one data set together, constructing a reference space from a plurality of data sets acquired sequentially, and determining whether the acquired data set is normal with respect to the reference space using the Mahalanobis distance. and "technology is disclosed.

JP-A-2004-232968

In Patent Document 1, data output from an air conditioner is used to divide the data into a section immediately after the start of air conditioning and a stable operation section. Means used as feature quantities are disclosed. In addition, regarding the failure sign judgment method, a reference space is constructed by accumulating multiple data sets of feature values obtained by processing air conditioner data for a certain period of time after maintenance of the air conditioner. A method is disclosed for determining whether a dataset is normal with respect to a reference space by the Mahalanobis distance.

In general, when a railway vehicle arrives at a station, the doors are opened and the outside air flows in, so the temperature inside the vehicle approaches the outside temperature. The longer the door is open in a certain time interval, that is, the more frequently the door is opened, the more the amount of outside air that flows in, and the more the in-vehicle temperature follows the target temperature. Therefore, the door opening frequency is a temperature variation factor that deviates the vehicle interior temperature from the target temperature in railway air conditioning.

However, when the follow-up performance of the air conditioner deteriorates mainly due to the high frequency of opening the door, the air conditioner itself is normal, so the failure determination method must also determine that the air conditioner is normal. Therefore, it is important for the failure determination method for railroad air conditioners to correctly determine whether the air conditioner is normal or abnormal, regardless of the influence of door opening frequency on the air conditioning performance.

In the abnormality detection means described in Patent Document 1, the indoor temperature or the like is used as a feature amount used for abnormality determination, but the temperature fluctuation factor such as the frequency of door opening is not included in the feature amount, and the door opening frequency is not included in the feature amount. Temperature fluctuation due to frequency etc. is not considered. Therefore, during railway operation, when the temperature inside the car is high and the Mahalanobis distance to the reference space is large, it is difficult to determine whether the cause is an air conditioning failure sign or the door opening frequency is high. If it is the cause, it may lead to an erroneous determination. For this reason, the abnormality detection means described in Patent Document 1 has a limited abnormality determination system.

As described above, when applying a failure sign determination method to a railroad air conditioner, it is important to consider various temperature fluctuation factors of railcars, and door opening frequency is one example. Therefore, there is a demand for means for detecting failure signs and abnormalities in air conditioners of railroad vehicles with high accuracy using temperature fluctuation factors such as door opening frequency as feature quantities.

Therefore, an object of the present disclosure is to provide a vehicle information management means that detects failure signs and abnormalities of air conditioners of railroad vehicles with high accuracy by using temperature fluctuation factors such as door opening frequency as feature quantities. .

In order to solve the above-mentioned problems, one representative in-vehicle information device of the present invention provides air conditioning sensor information regarding an air conditioner installed in a vehicle and door opening information regarding the door opening frequency of the vehicle door. A sensor group that acquires at least on-vehicle data, and a feature value group that includes a temperature feature value related to the vehicle and a door opening frequency feature value related to the door opening frequency for a predetermined inspection section by analyzing the on-vehicle data. Abnormality determination indicating presence/absence of an abnormality or a sign of abnormality regarding the air conditioner based on the feature extracting unit extracted from the on-vehicle data, the group of feature values, and a normal model indicating operating characteristics of the air conditioner when the air conditioner is in a normal state. and an anomaly determination unit that generates and outputs a result.

According to the present disclosure, it is possible to provide vehicle information management means for detecting failure signs and abnormalities in air conditioners of railroad vehicles with high accuracy by using temperature fluctuation factors such as door opening frequency as feature quantities.
Problems, configurations, and effects other than the above will be clarified by the description in the following modes for carrying out the invention.

FIG. 1 is a diagram illustrating an example configuration of a vehicle information management system according to a first embodiment of the present disclosure. FIG. 2 is a diagram illustrating an example of data used for feature quantity extraction according to the first embodiment of the present disclosure. FIG. 3 is a diagram illustrating an example of a follow section and a stable section according to the first embodiment of the present disclosure; FIG. 4 is a flow chart showing the flow of the feature quantity extraction method according to the first embodiment of the present disclosure. FIG. 5 is a diagram illustrating an example of a configuration of a tracking interval temporary storage database according to the first embodiment of the present disclosure; FIG. 6 is a diagram illustrating an example of a configuration of a stable interval temporary storage database according to the first embodiment of the present disclosure; FIG. 7 is a diagram illustrating an example of a configuration of a tracking segment storage database according to the first embodiment of the present disclosure; FIG. 8 is a diagram illustrating an example of a configuration of a stable interval storage database according to the first embodiment of the present disclosure; FIG. 9 is a diagram illustrating an example configuration of a vehicle information management system according to a second embodiment of the present disclosure. FIG. 10 is a diagram illustrating an example of data used for feature quantity extraction according to the second embodiment of the present disclosure. FIG. 11 is a diagram illustrating an example configuration of a vehicle information management system according to a third embodiment of the present disclosure; FIG. 12 is a diagram illustrating an example of data used for feature amount extraction according to the third embodiment of the present disclosure; FIG. 13 is a diagram illustrating an example configuration of a vehicle information management system according to a fourth embodiment of the present disclosure; FIG. 14 is a diagram illustrating an example of a travel section type correspondence table according to the fourth embodiment of the present disclosure. FIG. 15 is a diagram illustrating an example of data used for feature amount extraction according to the fourth embodiment of the present disclosure; FIG. 16 is a diagram illustrating an example of data used for feature amount extraction according to the fourth embodiment of the present disclosure;

Hereinafter, embodiments of the present invention will be described with reference to the drawings. It should be noted that the present invention is not limited by these examples. Moreover, in the description of the drawings, the same parts are denoted by the same reference numerals.

First, with reference to FIGS. 1 to 8, a vehicle information management system 1 according to a first embodiment of the present disclosure that performs abnormality determination of an air conditioner by using the door opening frequency of a railway vehicle as a feature quantity will be described.
(Configuration of Vehicle Information Management System 1)

FIG. 1 is a diagram showing an example of a configuration of a vehicle information management system 1 related to Example 1 of the present disclosure. As shown in FIG. 1, a vehicle information management system 1 according to the first embodiment of the present disclosure includes a vehicle 10, an air conditioner 101 installed in the vehicle 10, an air conditioning sensor 102 in the air conditioner 101, and various data management systems. and an abnormality determination device 30 for determining an abnormality.

The vehicle 10 is a housing in which an air conditioner 101, an air conditioning sensor 102, a vehicle information management device 103, and the like are mounted, and may be, for example, a railroad vehicle included in a train formation that travels on a railroad. However, the vehicle 10 related to the present disclosure is not limited to railway vehicles, and may be, for example, construction machines, automobiles, elevators, monorails, and the like.

The air conditioner 101 is an air conditioner mounted on the vehicle 10, and may have functions such as cooling, dehumidification, heating, and/or ventilation.

The air-conditioning sensor 102 is a sensor installed in the air-conditioning device 101, and acquires time-series air-conditioning sensor information related to, for example, the internal temperature of the air-conditioning device 101 (return temperature time history and outlet temperature time history), humidity, atmospheric pressure, and the like. be able to. The air conditioning sensor 102 may be installed, for example, at a return port, a discharge port, a refrigerant pipe, or the like of the air conditioner 101 .

The vehicle information management device 103 is a device that collects information on various devices installed in the vehicle 10 and sensor information from various sensors installed in the vehicle 10 and manages them as on-board data 20 . When the abnormality determination device 30, which will be described later, is implemented as a remote server or a cloud server that can be accessed from the vehicle information management device 103 via a communication network, the vehicle information management device 103 transmits the collected on-board data 20 by wireless communication. It may be transmitted to the abnormality determination device 30 via. On the other hand, when the abnormality determination device 30 is mounted in the vehicle 10, the vehicle information management device 103 may transfer the on-vehicle data 20 to the abnormality determination device 30 via local communication such as a LAN.

On-board data 20 managed by the vehicle information management device 103 includes air-conditioning sensor information acquired by the air-conditioning sensor 102 described above, and information such as vehicle occupancy rate, door opening/closing, cooling set temperature, heating set temperature, outside temperature, etc. and additional information obtained by other sensors installed on the vehicle 10 .
As will be described later, this on-board data 20 is acquired during a period in which the air conditioner 101 is operating normally, and is used to create a normal model. It may be either one of the on-vehicle data 202 during actual operation, which is acquired in a period in which whether or not it is unknown, and which is subject to abnormality determination.

The abnormality determination device 30 is a device that uses the on-board data 20 to create a normal model and determines an abnormality of the air conditioner. The abnormality determination device 30 may be implemented, for example, as ground equipment accessible from the vehicle 10 via a communication network, or may be configured integrally with the vehicle information management device 103 in the vehicle 10 . Moreover, in an embodiment, the functions of the abnormality determination device 30 may be implemented by a cloud that is communicably connected to the vehicle 10 via a network such as the Internet.
As shown in FIG. 1, the abnormality determination device 30 includes a feature amount extraction unit 301 for extracting feature amounts from the on-board data 20, a model creation unit 303 for creating a normal model, and a normal model. and an abnormality determination unit 305 for performing abnormality determination.

The vehicle information management related to this embodiment consists of two processes of normal model creation and abnormality determination during operation. Normal model creation is a process of accumulating data during a period in which the air conditioner 101 is operating normally and creating a normal model to be used for abnormality determination. This is a process of performing abnormality determination of the air conditioner 101 using the normal model described above.
(Normal model creation)

Hereinafter, creation of a normal model used for abnormality determination according to the embodiment of the present disclosure will be described with reference to functional units included in the abnormality determination device 30 shown in FIG. As described above, the normal model is a statistical model that is created from the normal on-board data 201 acquired while the air conditioner 101 is operating normally and that indicates the operating characteristics of the air conditioner 101 during normal operation. , the on-board data 202 during actual operation are compared to determine whether or not there is an abnormality.

First, the vehicle information management device 103 in the vehicle 10 acquires air conditioning sensor information from the air conditioning sensor 102 while the air conditioning device 101 is operating normally. After that, the vehicle information management device 103 transmits normal vehicle on-board data 201 including the acquired air conditioning sensor information to the abnormality determination device 30 .

Next, the normal vehicle on-board data 201 acquired while the air conditioner 101 is operating normally is input to the feature quantity extraction unit 301 of the abnormality determination device 30 . The feature amount extraction unit 301 is a functional unit that extracts a feature amount group from the on-board data 20 (the on-board data 201 during normal operation and the on-board data 202 during actual operation).
A feature amount group here is a data set containing a plurality of types of feature amounts corresponding to a certain time or time zone. Further, the feature value here is a representative value such as an average value or a median value calculated from data such as the vehicle interior temperature or the outside air temperature. As an example, if the vehicle interior temperature average value and the outside air temperature average value corresponding to a certain time period are used as feature amounts for abnormality determination, the feature amount group is a set of the vehicle interior temperature average value and the outside air temperature average value.

Next, the feature amount extraction unit 301 processes the normal on-board data 201 , extracts an appropriate combination of feature amounts as a normal feature amount group, and accumulates them in the normal accumulated feature amount group 302 . The normal accumulated feature amount group 302 here is obtained by accumulating the normal time feature amount group extracted from the normal on-vehicle data 201 by the feature amount extraction unit 301 over a plurality of times and environmental conditions.

Next, after a predetermined amount of normal feature amount groups are accumulated in the normal time accumulated feature amount group 302, the model creation unit 303 uses the normal time accumulated feature amount group 302 to generate a normal air conditioner Create a normal model that represents the driving characteristics of The form of the normal model here depends on the abnormality determination method used in the abnormality determination unit, which will be described later. For example, when abnormality determination is performed by a clustering method, the normal model may be a representative point of a cluster in which the normal accumulated feature amount group 302 exists in the feature amount space. Further, when abnormality determination is performed by deep learning, the normal model may be the structure and weighting of a learning model for estimating the vehicle interior representative temperature from information such as the outside air temperature. The vehicle interior representative temperature here is the representative temperature of the vehicle 10, and may be, for example, the time average value of the return temperature.

Next, the model creation unit 303 saves the created normal model in the model storage unit 304 . This model storage unit 304 is a storage unit for storing the normal model created by the model creation unit 303 .
(Abnormal judgment during operation)

Operational abnormality determination according to an embodiment of the present disclosure will be described below with reference to functional units included in the abnormality determination device 30 illustrated in FIG. 1 . As described above, the abnormality determination during operation is a process of determining abnormality of the air conditioner 101 using the air conditioning sensor data acquired during operation and the normal model described above.

First, it is acquired in a period different from the normal time on-board data 201 used for normal time model creation (for example, a period corresponding to the operation of the vehicle 10 to verify whether the air conditioner 101 is operating normally). On-board data 202 during actual operation is input to feature quantity extraction unit 301 of abnormality determination device 30 . The feature quantity extraction unit 301 extracts a feature quantity group 312 during operation from the on-board data 202 during actual operation.
The operational feature amount group 312 here is a feature amount group extracted from the actual operation on-board data 202 by the feature amount extraction unit 301, and is of substantially the same type as the normal accumulated feature amount group 302 described above. contains features of

Next, the abnormality determination unit 305 performs abnormality determination of the air conditioner 101 using the normal model created by the model creation unit 303 described above and the operating feature amount group 312, and outputs an abnormality determination result 306. do.
More specifically, the abnormality determination unit 305 receives as input the normal model and the in-operation feature quantity group 312 stored in the model storage unit 304, and uses a predetermined abnormality determination method to determine the presence or absence of an anomaly sign or the presence of an anomaly. is generated. The abnormality determination method used by the abnormality determination unit 305 is not particularly limited, and may be a clustering method, deep learning, or the like.

As an example, when the abnormality determination is performed by a clustering method, the abnormality determination unit 305 calculates the distance between the representative point of the normal cluster and the point representing the feature amount group during operation in the feature amount space, and calculates the abnormality according to the distance. may be determined. When the calculated distance is equal to or greater than a predetermined threshold, the abnormality determination unit 305 determines that an abnormality predictor has occurred in the air conditioner 101, and outputs an abnormality determination result 306 indicating that an abnormality predictor has occurred. do. On the other hand, when the calculated distance is less than the predetermined threshold, the abnormality determination unit 305 determines that no abnormality sign has occurred in the air conditioner 101, and outputs an abnormality determination result 306 indicating that no abnormality sign has occurred. do.

As another example, when the abnormality determination is performed by deep learning, the abnormality determination unit 305 inputs, for example, the feature value group 312 during operation other than the vehicle interior representative temperature to the learning model to estimate the first vehicle interior representative temperature, and the operation Abnormality determination may be performed according to the value of the difference between the second vehicle interior representative temperature and the first vehicle interior temperature in the time feature quantity group 312 . When this difference is equal to or greater than a predetermined threshold, the abnormality determination unit 305 determines that an abnormality sign has occurred in the air conditioner 101, and outputs an abnormality determination result 306 indicating that an abnormality sign has occurred. . On the other hand, if the difference is less than the predetermined threshold, the abnormality determination unit 305 determines that no abnormality sign has occurred in the air conditioner 101, and outputs an abnormality determination result 306 indicating that no abnormality sign has occurred. .

As described above, the abnormality determination result 306 is information indicating whether or not there is an abnormality sign output from the abnormality determination unit 305, or whether or not an abnormality has occurred. The abnormality determination result 306 may be represented by integers such as "0: no abnormality sign", "1: abnormality sign occurred", and "2: abnormality occurred".

According to the vehicle information management system 1 described above, by using temperature fluctuation factors such as the frequency of door opening as feature quantities, it is possible to determine with high accuracy signs of failure and abnormalities in air conditioners of railroad vehicles.

Next, with reference to FIG. 2, data used for feature amount extraction and a feature amount extraction method according to the first embodiment of the present disclosure will be described.
FIG. 2 is a schematic diagram of data extracted from the on-board data 20 described above and used for extracting feature amounts in this embodiment. As shown in FIG. 2, the time-series data of the vehicle interior temperature G102 and the time-series data of the outside air temperature G103, the set temperature G104, the passenger ratio G105, and the door open signal G106 may be included as fluctuation factors of the vehicle interior temperature.
Feature amounts related to the embodiment of the present disclosure are extracted from the above data shown in FIG. 2, and the extracted feature amounts are collected as a feature amount group.

These feature amounts are extracted for a predetermined inspection section G101. This inspection interval G101 is a time interval corresponding to the data from which the feature amount is extracted. As an example, a period from a predetermined first time to a predetermined second time may be set as the inspection section G101. The length of the inspection section G101 may be determined according to the cycle of abnormality determination and the characteristics of the on-board data 20 .

For example, if the abnormality determination is performed every day, the length of the inspection section G101 is one day, and if the operation start time is 5:00 and the operation end time is 23:00, the inspection section G101 is 5 hours long. :00 to 23:00. Also, if there is knowledge that the temperature inside the vehicle varies depending on the time period, for example, due to an increase in the temperature inside the vehicle due to sunshine, the inspection section G101 may be divided into time periods, and the length of the inspection section G101 may be set to one hour.
(Feature extraction method)

For the inspection section G101 set in this manner, a representative value (for example, average value, median value, etc.) of the on-board data 20 corresponding to the inspection section G101 is calculated and extracted as a feature amount. For example, when the on-board data 20 acquired for the inspection section G101 includes the outside air temperature, the inside temperature, the set temperature (temperature feature value), and the passenger ratio, the average inside temperature and the average outside temperature in the inspection section G101 , the set temperature average value, and the passenger load average value may be extracted as feature amounts.

Furthermore, in addition to the above features, the door open frequency corresponding to the inspection section G101 is extracted as a feature amount (door open frequency feature amount) from the data of the door open signal G106. The door open signal G106 here is binary data consisting of 0 and 1, 0 indicating door closed and 1 indicating door open. Further, the door opening frequency here is a value indicating the frequency with which the doors of the vehicle 10 are opened in the inspection section G101. may be calculated.
As a result, feature values such as the average vehicle interior temperature, the average outside temperature, the average set temperature, the average passenger ratio, and the door opening frequency in the inspection section are extracted and used as a feature value group (first feature value group). are gathered.

According to the feature amount extraction method described above, it is possible to extract the door opening frequency as a feature amount in addition to the information such as the vehicle interior temperature, the set temperature, the outside air temperature, and the passenger ratio. In this way, by performing the abnormality determination using the feature amount group (first feature amount group) including the door opening frequency, it is possible to realize highly accurate air conditioning system abnormality determination that takes into consideration the fluctuation in the vehicle interior temperature due to the door opening. can be made

As an example, consider a case where abnormality determination is performed using a feature amount group that includes only the vehicle interior temperature average value, the preset temperature average value, the outside air temperature average value, and the passenger load average value in the inspection section G101. Here, if the average set temperature is around 24°C, the average outside air temperature is around 35°C, and the average passenger load is around 50%, the difference between the average cabin temperature and 26°C in the feature value during operation is 3°C. In the following cases, abnormality determination means for determining that the air conditioner is normal is used.
At this time, if the average set temperature is 24°C, the average outside temperature is 35°C, the average occupancy rate is 50%, and the average temperature inside the car is 30°C, it is judged to be abnormal. Become. However, in reality, the air conditioning system is functioning normally, the doors are frequently opened, and the amount of air flowing in is large, which can lead to an increase in the average temperature inside the vehicle. be.

On the other hand, as described in the present disclosure, by using a feature amount group (first feature amount group) that further includes the door opening frequency, which is a temperature variation factor, as a feature amount, erroneous determination of abnormality can be reduced. .
For example, when the average set temperature is around 24°C, the average outside air temperature is around 35°C, the average passenger load is around 50%, and the door opening frequency is around 0.10, the average temperature inside the car is within 26±3°C. If the air conditioner is normal, the average set temperature is around 24°C, the average outside temperature is around 35°C, the average passenger load is around 50%, and the door opening frequency is around 0.20, then the average interior temperature is Abnormality determination means is used to determine that the air conditioner is normal if the temperature falls within 28±3°C.
In this case, if the average set temperature is 24°C, the average outside temperature is 35°C, the average passenger load is 50%, and the average interior temperature is 30°C, if the door opening frequency is 0.10, the air conditioner will It is determined that an abnormality has occurred, and if the value is 0.20, no abnormality has occurred in the air conditioner. Therefore, it is possible to eliminate erroneous determination of abnormality when the door opening frequency is high (for example, 0.20).

In the above, with reference to FIG. 2, the data used for the feature-value extraction and the feature-value extraction method in connection with Example 1 of this indication were demonstrated. However, in practice, the doors of the vehicle may be open for a long time, for example, when the vehicle stops at a station for a long time. In such a case, in order to realize highly accurate abnormality determination, it is desirable to set the above-described inspection interval based on the door open time and perform feature extraction.
Therefore, next, with reference to FIG. 3, a method of extracting a feature amount for further improving the accuracy of abnormality determination will be described.

FIG. 3 is a diagram illustrating examples of a follow-up interval and a stable interval according to the first embodiment of the present disclosure; More specifically, FIG. 3 is a diagram showing changes in the on-board data 20 such as the train interior temperature G205 and the outside air temperature G204 when there is a station where the doors are open for a long time, and how the inspection section is set.

When the train stops at the first station G211 for a long time due to the reason of passing through a terminal station or other operation type, connection, etc., the phenomenon that the inside temperature G205 deviates greatly from the set temperature G206 occurs due to the doors of the train being open while the train is stopped. do. For example, when the door of the car is opened for about 10 minutes, the temperature inside the car G205 may rise by about 5°C from immediately after stopping at the first station G211 to just before departing, depending on the outside air temperature. In such a case, after the train departs from the first station G211, the temperature inside the train G205 gradually converges to the set temperature G206 until it arrives at the second station G212, where the doors are open for a long time. It tends to exhibit a behavior of stabilizing in the vicinity.

In order to suppress fluctuations in the temperature inside the vehicle due to the door of the vehicle being open for a long time and the resulting influence on the abnormality determination of the air conditioner, the present disclosure does not include the period in which the door of the vehicle is open for a long time. It is desirable to set the inspection interval as follows. For example, as shown in FIG. 3, the inspection section G201 may be between a first station G211 and a second station G212 where the doors are open for a long time. As a result, the vehicle stops at the first station G211 and the second station G212, and the period during which the doors are open for a long time is not included in the inspection section G201. The influence on abnormality determination can be suppressed.

Furthermore, in the inspection section G201, there are two types of phenomena: the operation in which the vehicle interior temperature G205 follows the set temperature G206, and the phenomenon in which it stabilizes near the set temperature G206. Suppose that it divides|segments into the stable area G203. Here, the follow-up section G202 is a section in which the vehicle interior temperature G205 approaches the set temperature G206, and the stable section G203 is a section in which it stabilizes around the set temperature G206.

Next, feature amounts are extracted for each of the following section G202 and the stable section G203.
In the follow-up section G202, the vehicle interior temperature G205 decreases or increases toward the set temperature G206. It is preferable to use the rate G207A as a feature amount. The section temperature change rate G207A is the absolute value of the difference between the vehicle interior temperature G205 immediately after the door of the first station G211 is closed and the vehicle interior temperature G205 at the end of the follow-up section G202 divided by the time width of the follow-up section G202. It can be obtained by calculation.

In addition, in the stable section G203, the vehicle interior temperature average value is used as a feature quantity because the influence of the increase in the vehicle interior temperature G205 due to the door being open for a long period of time is excluded.

On-vehicle data 20 other than the in-vehicle temperature G205, that is, the outside air temperature G204, the set temperature G206, the passenger rate G209, and the door open signal G210 are not affected by the door opening for a long period of time. For each section, the average outside air temperature, average set temperature, average passenger ratio, and door open frequency are calculated and used as feature values.

Based on the above, when the inspection section G201 is separated into the follow-up section G202 and the stable section G203, for the follow-up section G202, the average outside air temperature, the average set temperature, the average passenger ratio, the door open frequency, and the interval temperature change rate G207A are defined as a feature quantity group (second feature quantity group). In addition, for the stable section G203, the feature amounts of the vehicle interior temperature average value, the outside air temperature average value, the set temperature average value, the occupancy average value, and the door opening frequency in the stable section G203 are added to a feature amount group (third feature amount group ).

Further, when separating the inspection section G201 into the following section G202 and the stable section G203, the abnormality determination unit 305 performs model creation and abnormality determination using the second feature amount group and Model creation and abnormality determination are performed separately, and an abnormality determination result A based on the second feature amount group and an abnormality determination result B based on the third feature amount group are obtained. After that, the abnormality determination unit 305 analyzes both the abnormality determination result A and the abnormality determination result B, and outputs an abnormality determination result 306 .
Here, when both the abnormality determination result A and the abnormality determination result B indicate that there is a sign of abnormality, the abnormality determination unit 305 may output the abnormality determination result 306 indicating that there is a sign of abnormality. Moreover, when both the abnormality determination result A and the abnormality determination result B indicate that there is no abnormality, the abnormality determination unit 305 may output the abnormality determination result 306 indicating that there is no abnormality. On the other hand, if one of the abnormality determination result A and the abnormality determination result B indicates that there is a sign of abnormality and the other indicates that there is no abnormality, the abnormality determination unit 305 outputs the abnormality determination result 306 indicating that there is a sign of abnormality. may Thereby, overlooking of an abnormality can be suppressed.

As described above, in the follow-up section G202, the vehicle interior temperature G205 decreases or increases toward the set temperature G206. Although it has been described that it is preferable to use the interval temperature change rate G207A as the feature amount, the effect of using the interval temperature change rate G207A as the feature amount will be described here. When an abnormality occurs in the air conditioning and the vehicle interior temperature G205 at the start and end of the follow-up section G202 does not change, but the time required for the vehicle interior temperature G205 to converge to the set temperature G206 increases, the vehicle interior temperature average value in the follow-up section G202 does not change, but the temperature change rate G207A decreases. Therefore, the interval temperature change rate G207A can express performance degradation in which the time required for the vehicle interior temperature G205 to converge to the set temperature G206 is extended. can be improved.

In addition, the effect of the method of determining the inspection section in consideration of the phenomenon that the door is open for a long time will be described. By taking into account the phenomenon of the door being open for a long time, it is possible to extract more appropriate feature values and improve the abnormality detection performance compared to extracting feature values without considering the phenomenon of the door being open for a long time. .

For example, without considering the phenomenon that the door is open for a long time, if the inspection section is set every hour and the feature value such as the average temperature inside the train is extracted, there is a station where the door is open for a long time in the inspection section every hour. In this case, since the vehicle interior temperature fluctuates, the vehicle interior temperature average value may fluctuate and deviate from the set temperature average value. In such a case, there is a risk that the vehicle interior temperature average value deviates from the set temperature average value, and thus it is determined to be abnormal. On the other hand, if the vehicle interior temperature average value is limited to the stable interval as in the present embodiment, the temperature fluctuation effect due to the door being open for a long period of time can be eliminated. It is possible to reduce the judgment that there is an abnormality because the temperature average value deviates from the set temperature average value.

Next, with reference to FIG. 4, a feature amount extraction method according to the first embodiment of the present disclosure will be described.

FIG. 4 is a flow chart showing the flow of the feature quantity extraction method according to the first embodiment of the present disclosure. The feature quantity extraction method according to the first embodiment of the present disclosure, shown in FIG. G202 and G203 are processing for extracting feature amounts in the stable interval G203, which are performed by the feature amount extraction unit 301 of the abnormality determination device 30 shown in FIG.
Below, the feature quantity extraction method according to the first embodiment will be described along the step numbers shown in FIG.

First, in S101, the feature quantity extraction unit 301 reads the on-board data received from the vehicle information management device 103 in chronological order, and calculates the door opening time at the time of arrival at the station. A station G211 and then a second station G212 whose doors are open for a long time are specified. After reading the on-board data is completed up to the second station, the process proceeds to S102.
Regarding the door open time calculation method, specifically, the sensor detects when the door open signal changes from 0 to 1, and then calculates the time until the door open signal changes from 1 to 0. Door open time can be calculated. Further, when the door open time calculated in this way is equal to or greater than the predetermined threshold value Tth, it is considered that the door is open for a long time.

Next, in S102, the feature amount extraction unit 301 sets the inspection section from the time immediately after the door of the first station G211 to the time immediately before the door of the second station G212 opens, and from the on-board data 20, The data is extracted as inspection section on-board data, and the process proceeds to S103.

S103 to S106 are a loop process that is repeatedly performed on inspection section on-board data at each time.

Next, in S104, the feature amount extraction unit 301 determines whether the inspection section on-board data at each time exists in the follow-up section G202 of the inspection section G201 or in the stable section G203. Determined using a threshold ath. If the rate of change in temperature at a specific time is equal to or greater than the threshold ath, the feature quantity extraction unit 301 determines that the temperature in the vehicle G205 is approaching the set temperature G206, and determines that the time is in the follow-up interval G202, and proceeds to S105. Transition. On the other hand, if the temperature change rate at the time is smaller than the threshold ath, the feature amount extraction unit 301 determines that the change in the vehicle interior temperature G205 is small and is stable around the set temperature G206, and that the time is in the stable interval G203. Then, the process proceeds to S115.

Here, the rate of change in temperature at a specific time is information indicating the magnitude of temperature change in the vehicle at a certain time t1. The temperature change rate at this specific time may be obtained, for example, from the moving median value of the temperature change rate instantaneous value G207B (see FIG. 3) of the vehicle interior temperature at a certain time. Further, the temperature change rate instantaneous value G207B is obtained by using the vehicle interior temperature T1 at a certain time t1 and the vehicle interior temperature T0 at the time t0 before that time in the on-vehicle data, |(T1-T0)/(t1- t0)|

The moving median value of the instantaneous temperature change rate value G207B is a value obtained by calculating the median value from the instantaneous temperature change rate instantaneous values G207B calculated at a certain time and at a plurality of times before and after that time.

By using the moving median value of the instantaneous temperature change rate G207B, fluctuations in the instantaneous temperature change rate G207B caused by opening the door for a short period of time can be removed as noise, and the temperature change rate at a specific time temporarily exceeds the threshold ath As a result, it is possible to prevent an erroneous determination that the on-board data has entered a stable section. Therefore, by using the moving median of the instantaneous temperature change rate value, it is possible to prevent the data of the follow-up interval from being mixed in the stable interval. It is possible to suppress deterioration in determination performance.

In this embodiment, as a specific example, the instantaneous temperature change rate G207B at a certain time and the median of the temperature change rate instantaneous values G207B corresponding to six points before and after the time are calculated, and used as the temperature change rate at that time. Also, here, the above threshold ath is assumed to be 0.25° C./min. However, the present disclosure is not limited to this, and the data points and threshold ath used for calculating the temperature change rate may be set as appropriate.

Next, in S105, the feature amount extraction unit 301 processes the inspection section on-board data determined to be in the following section, records it in the following section temporary storage database, and proceeds to S106. The following segment temporary storage database here is a database for processing and storing the data necessary for feature quantity extraction using the inspection segment on-board data at the time determined to be the following segment. may be stored within

Next, with reference to FIG. 5, a tracking interval temporary storage database according to the first embodiment of the present disclosure will be described. A table T10 shown in FIG. 5 is a table showing an example of the configuration of the following segment temporary storage database.

As shown in FIG. 5, in table T10, the moving median value T108 of the instantaneous temperature change rate value T107 is 0.25° C./min or less, and the times (13:00.00, 13:00.05, . . . 13:00.05, . . . 13 : 15.25), the vehicle interior temperature T102, the set temperature T103, the outside temperature T104, the occupancy rate T105, and the door open signal T106 are stored as information used for extracting feature values from the inspection section on-board data of 15.25). Table T10 further includes the result of calculating an instantaneous temperature change rate value T107 based on the vehicle interior temperature T102 and time T101, and the result of calculating an instantaneous temperature change rate moving median value T108 based on the instantaneous temperature change rate value T107. and are stored

Returning to the description of FIG. 4, in S115, the feature amount extraction unit 301 processes the inspection section on-board data determined to be in the stable section, records it in the stable section temporary storage database, and proceeds to S106. The stable section temporary storage database here is a database that processes and stores the data necessary for feature amount extraction using the inspection section on-board data at the time determined to be a stable section. 301 may be stored.

Next, with reference to FIG. 6, the stable section temporary storage database according to the first embodiment of the present disclosure will be described. A table T20 shown in FIG. 6 is a table showing an example of the configuration of the stable interval temporary storage database.

As shown in FIG. 6, in Table T20, the moving median value T208 of the instantaneous temperature change rate T207 is smaller than 0.25° C./min, and the time (13:15.30, 13:15.35, . The vehicle interior temperature T202, the set temperature T203, the outside air temperature T204, the passenger ratio T205, and the door open signal T206 are stored from the inspection section on-board data as information used for extracting the feature quantity. Table T20 further includes the result of calculating an instantaneous temperature change rate value T207 based on the vehicle interior temperature T202 and time T201, and the result of calculating an instantaneous temperature change rate moving median value T208 based on the instantaneous temperature change rate value T207. and are stored.

Returning to the description of FIG. 4, in S107, the feature amount extraction unit 301 stores the feature amount extracted using the information in the following section temporary storage database in the following section storage database, and also stores the information in the stable section temporary storage database. The feature values extracted using the method are stored in the stable interval storage database.

Next, with reference to FIG. 7, the following section storage database according to the first embodiment of the present disclosure will be described. A table T30 shown in FIG. 7 is a table showing an example of the configuration of the following segment storage database.

As shown in FIG. 7, for example, the row corresponding to "1" in No T301 stores on-board data from 13:00.00 to 13:15.25 shown in table T10 in FIG. This is a table that stores the results of calculating a second feature quantity group based on the information in the following segment temporary storage database.

More specifically, in Table T30, set temperature average value T303 is the average value of set temperature T103, outside air temperature average value T304 is the average value of outside temperature T104, and passenger load average value T305 is the average value of passenger load T105. . As shown in FIG. 7, these pieces of information are recorded every time T302.
As described above, the set temperature average value T303, the outside air temperature average value T304, and the passenger load average value T305 may be feature amounts included in the feature amount group (second feature amount group).

Furthermore, the section temperature change rate T306 is obtained by subtracting the vehicle interior temperature T102 corresponding to the end of time T101 from the vehicle interior temperature T102 corresponding to the start of time T101 shown in the table T10 of FIG. is the value obtained by dividing by the value obtained by subtracting the value of

Further, the door open frequency accumulated value T307 is the total value of the time when the door open signal T106 shown in the table T10 of FIG. 5 is 1. The intra-interval time T308 is the time width of the follow-up interval. In No T301-1 shown in FIG. 7, the door open frequency of 0.10 is calculated by dividing the time in the interval of 925.0 from the door open time cumulative value of 92.8 sec.

In addition, among the columns described in Table T30, the second feature amount group used in the abnormality sign determination in this embodiment is the average set temperature T303, the average outside air temperature T304, the average passenger ratio T305, and the temperature decrease It includes feature amounts of rate average value T306 and door open frequency T309.

Next, with reference to FIG. 8, a stable interval storage database according to the first embodiment of the present disclosure will be described. A table T40 shown in FIG. 8 is a table showing an example of a configuration of the stable interval storage database.

As shown in FIG. 8, for example, the row corresponding to "1" in No T401 stores on-board data from 13:15.30 to 13:40.00 shown in table T20 in FIG. It is a table containing the results of extracting feature amounts based on the information in the stable interval temporary storage database. In addition, in the stable section storage database, instead of the median value of the instantaneous temperature decrease rate, the result of calculating the average value of the vehicle interior temperature T202 shown in Table T20 of FIG. 6 is recorded.

As shown in Table T40, for No T401, average set temperature T403, average outside air temperature T404, average passenger load T405, average interior temperature T406, cumulative door open time T407 (seconds), in-section time T408, And the door opening frequency T409 is recorded every time T402.
Among the columns described in Table T40, the feature amount group used in the abnormality determination in this embodiment is the set temperature average value T403, the outside air temperature average value T404, the passenger load average value T405, the vehicle interior temperature average value T406, the door It includes the feature quantity of the open frequency T409.

Returning to the description of FIG. 4, in S108, the feature quantity extraction unit 301 deletes the information in the following segment temporary storage database and the stable segment temporary storage database in order to calculate the information of the next following segment and stable segment, Go to S109.

Next, in S109, the feature quantity extraction unit 301 determines whether or not vehicle operation for a certain day has been completed. If the operation has not been completed, the process proceeds to S110. On the other hand, when the operation is completed, the operation of this flowchart for that day is completed.

Next, in S110, the feature amount extraction unit 301 updates the second station G212 to the first station G211 in order to perform feature amount extraction for the next inspection section, and Next, the station where the door is open for a long time is updated to the second station G212, and the process returns to S102.

As described above, in the feature amount extraction method according to the first embodiment of the present disclosure, in addition to the feature amounts related to the temperature and the occupancy rate, the door opening frequency is included in the feature amount group. Air conditioner abnormality detection that takes into account is possible. In addition, by setting the interval between the first station where the doors are open for a long time and the second station where the doors are open for a long time as the inspection section, the temperature change behavior between the stations where the doors are open for a long time can be expressed. A feature amount can be extracted, and the abnormality determination performance of an air conditioner can be improved.

Furthermore, by using the moving median of the instantaneous value of the temperature change rate as the method for determining when to switch between the follow-up interval and the stable interval, it prevents erroneous determination of the switching time between the follow-up interval and the stable interval, and suppresses performance degradation in abnormality determination. can be made

Next, with reference to FIGS. 9 and 10, a vehicle information management system according to a second embodiment of the present disclosure, which performs abnormality determination of an air conditioner further using the window opening ratio of the windows of a railway vehicle as a feature quantity, will be described.

In the above-described first embodiment, the configuration for determining whether the air conditioner is abnormal has been described using the door opening frequency as the feature amount in addition to the feature amount relating to the temperature and the occupancy rate. However, in reality, the temperature fluctuation factor that fluctuates the temperature inside the vehicle and affects the abnormality determination of the air conditioner is not limited to the frequency of opening the door, and for example, the inflow of air through the opening of the window can be considered. Therefore, the second embodiment of the present disclosure adopts the window opening rate as an evaluation index of window opening in addition to the temperature, the occupancy rate, and/or the door opening frequency. It is a configuration used as a quantity.

Next, a vehicle information management system according to a second embodiment of the present disclosure will be described with reference to FIG.

FIG. 9 is a diagram showing an example of the configuration of the vehicle information management system 2 according to the second embodiment of the present disclosure. As shown in FIG. 9 , the vehicle information management system 2 according to the second embodiment of the present disclosure includes, in addition to the air conditioning sensor 102, the vehicle information management device 103, and the abnormality determination device 30 in the first embodiment described above, the window opening ratio Further includes an estimator 104 .
Note that the vehicle information management system 2 according to the second embodiment is the same as the vehicle information management system 1 according to the above-described second embodiment except that the window opening ratio estimation unit 104 is further included. A detailed description is omitted here.

The window opening ratio estimating unit 104 is a functional unit for estimating the opening ratio of windows present in the vehicle 10 .
The operation of the vehicle information management system 2 according to the second embodiment of the present disclosure will be described below.

First, the window opening ratio estimating unit 104 estimates the opening ratio of the windows present in the vehicle 10 and transmits the window opening ratio to the vehicle information management device 103 . The vehicle information management device 103 generates second on-vehicle data including at least the air conditioning sensor information acquired by the air conditioning sensor 102 and the window opening ratio estimated by the window opening ratio estimation unit 104, and calculates the second on-board data. The on-board data is transmitted to the abnormality determination device 30 .

After that, the abnormality determination device 30 extracts the feature amount using the vehicle interior temperature and the like included in the received second on-board data, and also uses the window opening ratio at each time to determine the window opening rate in the inspection section. A feature quantity corresponding to the aperture ratio (window aperture ratio feature quantity) is extracted.
It should be noted that the window opening rate here is a value indicating the rate at which the windows of the vehicle 10 are open. Further, the window opening ratio is obtained, for example, by dividing the opening area of the windows present in the vehicle 10 by the window area when all the windows are open.

The window opening ratio estimating unit 104 acquires first video data indicating the state of the vehicle interior using an imaging unit (first imaging unit) such as a camera installed in the vehicle 10, and calculates the first image data. The window opening area and the window opening ratio of the vehicle 10 may be estimated by applying the image processing method to the image data of . In an embodiment, the window opening ratio estimating unit 104 may be configured to predetermine the window opening area of the vehicle 10 and output the window opening ratio input as a constant to the vehicle information management device. . However, the means for estimating the window aperture ratio in the present disclosure is not particularly limited, and may be changed as appropriate.

Next, with reference to FIG. 10, data used for feature amount extraction according to the second embodiment of the present disclosure will be described.

FIG. 10 shows the vehicle interior temperature G302, the outside temperature G303, the set temperature G304, the passenger ratio G305, and the door open signal G306 included in the second on-board data, as well as the window opening rate estimation unit 104. It is a figure which shows the window aperture ratio G307 in each time.

The feature quantity extraction unit 301 in the abnormality determination device 30 uses the vehicle interior temperature G302, the outside air temperature G303, the passenger ratio G305, and the door open signal G306 to extract the feature quantity in the same manner as in the first embodiment, and in the inspection section, The average value of the window aperture ratio G307 is further extracted as a feature amount.

After the feature amount extraction is completed, the abnormality determination unit 305 in the abnormality determination device 30 performs abnormality determination on the fourth feature amount group obtained by adding the average value of the window opening ratio G307 in the inspection section G301 to the first feature amount group. Alternatively, a fifth feature amount group in which the average value of the window opening ratio G307 in the following interval is added to the second feature amount group, and the average value of the window opening ratio G307 in the stable interval is added to the third feature amount group Abnormality determination is performed with respect to the sixth feature amount group.

As described above, in the feature amount extraction method according to the second embodiment of the present disclosure, in addition to the temperature, the boarding rate, and the door opening frequency, the window opening rate is included in the feature amount group, so that the temperature fluctuation in the vehicle is Since it is possible to distinguish whether it is caused by a high window opening rate or by an air conditioning failure, it is possible to further reduce erroneous determination of the air conditioner.

Next, with reference to FIGS. 11 and 12, a description will be given of a vehicle information management system according to a third embodiment of the present disclosure, which further uses the sunshine intensity as a feature amount to determine whether an air conditioner is abnormal.

In the first embodiment and the second embodiment described above, the configuration for performing the abnormality determination of the air conditioner using the door opening frequency and the window opening ratio as the feature values has been described. However, in reality, temperature fluctuation factors that affect the temperature inside the vehicle and affect the judgment of an air conditioning system abnormality are not limited to the door opening frequency or the window opening rate, but can be heat radiation from the sun to the vehicle body, for example. be done. Also, the amount of thermal radiation from the sun varies depending on the weather. Therefore, in Example 3 of the present disclosure, in addition to temperature, occupancy rate, and/or door opening frequency and window opening rate, solar irradiance is adopted as an evaluation index for thermal radiation, and features in Example 1 or Example 2 This is a configuration in which the intensity of sunlight is added as a feature amount to the amount group.

Next, a vehicle information management system according to a third embodiment of the present disclosure will be described with reference to FIG.

FIG. 11 is a diagram showing an example configuration of a vehicle information management system 3 according to the third embodiment of the present disclosure. As shown in FIG. 11 , the vehicle information management system 3 according to the third embodiment of the present disclosure includes the air conditioning sensor 102, the vehicle information management device 103, and the abnormality determination device 30 in the above-described first embodiment, as well as the sunshine intensity estimation. It further includes a unit 106 and a camera 105 (second imaging unit).
Note that the vehicle information management system 3 according to the third embodiment is the same as the vehicle information management system 1 according to the first embodiment described above except that it further includes the sunshine intensity estimation unit 106 and the camera 105. Therefore, for convenience of explanation, , repetitive descriptions are omitted here.

Camera 105 is an imaging device configured to acquire second video data including the sky around the train.
The sunshine estimation unit 106 is a functional unit that estimates the current sunshine intensity based on the second image data of the camera 105 .
The operation of the vehicle information management system 3 according to the third embodiment of the present disclosure will be described below.

First, the camera 105 photographs the sky around the train, acquires second video data, and transmits the acquired second video data to the sunshine estimation unit 106 .

The sunshine estimation unit 106 estimates the current sunshine intensity based on the received second video data, and transmits the estimated sunshine intensity to the vehicle information management device 103 . More specifically, the sunshine estimation unit 106 classifies the sky and the clouds in the sky by using, for example, a semantic segmentation technique for the video data, and identifies the pixels occupied by the sky and the clouds. Based on the percentage of pixels occupied, the current weather and irradiance may be estimated. As an example, the sunshine estimation unit 106 sets the sunshine to "1" when the percentage of clouds in an image of the sky is 100%, and sets the sunshine to "0.5" when the percentage of clouds is 50%. ”, and if the cloud ratio is 0%, the sunshine intensity may be set to “0”. Moreover, when the image of the sky cannot be acquired, the sunshine intensity estimation unit 106 may estimate that the vehicle is traveling in a tunnel or an underground section, and set the sunshine intensity to "0".

Next, the vehicle information management device 103 generates third on-vehicle data including the air-conditioning sensor information acquired by the air-conditioning sensor 102 and the sunshine intensity estimated by the sunshine intensity estimation unit 106. on-board data is transmitted to the abnormality determination device 30 .
The abnormality determination device 30 extracts a feature amount using the vehicle interior temperature and the like included in the third on-vehicle data, and uses the sunshine intensity at each time to obtain a feature amount (day illuminance feature quantity).

Next, data used for feature amount extraction according to the third embodiment of the present disclosure will be described with reference to FIG. 12 .

FIG. 12 is a diagram schematically showing the third on-board data. The lowermost graph in FIG. 12 shows the time history of sunshine intensity in the third on-vehicle data related to Example 3 of the present disclosure.

As described above, in the third embodiment of the present disclosure, in addition to the temperature, the occupancy rate, and/or the door opening frequency and the window opening rate, the sunshine intensity is further used as a feature amount to determine the abnormality of the air conditioner. be. The feature amount extraction unit 301 in the abnormality determination device 30 uses the vehicle interior temperature G302, the outside air temperature G303, the passenger ratio G305, and the door open signal G306 to extract the feature amount in the same manner as in the first embodiment, and A time average value of the sunshine intensity G307 is further extracted as a feature amount.

After the extraction of the feature amount is completed, the abnormality determination unit 305 in the abnormality determination device 30 performs abnormality determination on the seventh feature amount group obtained by adding the time average value of the illuminance G307 in the inspection section G301 to the first feature amount group. Alternatively, an eighth feature amount group in which the average value of the sunshine intensity G307 in the following interval is added to the second feature amount group, and an average value of the sunshine intensity G307 in the stable interval is added to the third feature amount group Abnormality determination is performed for the 9 feature amount groups.

As described above, in the feature amount extraction method according to the third embodiment of the present disclosure, in addition to the temperature, the boarding rate, and the door opening frequency, the sunshine intensity is included in the feature amount group. Since it is possible to distinguish whether it is caused by high illuminance or by an air conditioning failure, it is possible to further reduce erroneous determination of the air conditioner.

Next, with reference to FIGS. 13 to 16, a description will be given of a vehicle information management system according to a fourth embodiment of the present disclosure, which further uses the degree of sunshine as a feature quantity to determine whether an air conditioner is abnormal.

In the above-described third embodiment, a configuration has been described in which the abnormality determination of the air conditioner is performed using the sunlight intensity obtained by analyzing the image data acquired by the camera installed in the vehicle as the feature value. However, from the viewpoint of suppressing equipment costs for installing a camera in a vehicle, it may be desirable to determine whether the air conditioner is abnormal using the intensity of sunlight as a feature value without installing the camera. Therefore, the fourth embodiment of the present disclosure is configured to perform an abnormality sign of an air conditioner by further using the intensity of sunlight as a feature amount without using a camera.

FIG. 13 is a diagram for explaining the vehicle information management system 4 according to the fourth embodiment of the present disclosure. As shown in FIG. 13 , the vehicle information management system 4 according to the fourth embodiment of the present disclosure includes above-ground and underground determination in addition to the air conditioning sensor 102, the vehicle information management device 103, and the abnormality determination device 30 in the first embodiment described above. A unit 307 is further included in the abnormality determination device 30 .
Note that the vehicle information management system 4 according to the fourth embodiment is the same as the vehicle information management system 1 according to the first embodiment described above, except that the above-ground/underground determination unit 307 is further included. Description is omitted here.

The above-ground/underground determination unit 307 determines whether the vehicle 10 is traveling above ground or underground based on the section information of the vehicle 10 and the travel section type correspondence table created in advance, and estimates the sunlight intensity. It is a functional part that
The operation of the vehicle information management system 4 according to the fourth embodiment of the present disclosure will be described below.

First, the vehicle information management device 103 creates fourth on-board data including air-conditioning sensor information acquired by the air-conditioning sensor 102 and on-rail section information of the vehicle 10 , and transmits the fourth on-board data to the abnormality determination device 30 .
The on-rail section information here is information indicating the section in which the vehicle 10 is currently traveling. In the on-rail section information, the travel route of the vehicle 10 may be divided into a plurality of travel sections based on the distance, surrounding environment such as stations, tunnels, underground, etc., and each travel section may be associated with a different symbol. As an example, when a certain station is divided into a plurality of travel sections 1-1, 1-2, 1-3, etc., if the vehicle 10 exists in the travel section 1-2, the on-rail section information is "1-2". can be indicated as

After the abnormality determination device 30 receives the fourth on-board data transmitted from the vehicle information management device 103, the above-ground/underground determination unit 307 in the abnormality determination device 30 detects the on-rail section included in the fourth on-board data. Based on the information, it is determined whether the vehicle 10 is running on the ground or underground at the current time.

In addition, the above-ground/underground determination unit 307 estimates the time-series data of the sunshine intensity using the above-ground/underground determination result, and stores the estimated time-series data of the sunshine intensity and the above-described fourth on-vehicle data. On-board data is created and transferred to the feature value extraction unit.
After that, the feature amount extraction unit 301 extracts a feature amount using the vehicle interior temperature and the like included in the fifth on-board data, and uses the time-series data of the sunshine intensity to determine the sunshine intensity in the inspection section. Extract features.

Next, with reference to FIG. 14, a travel section type correspondence table according to the fourth embodiment of the present disclosure will be described.

FIG. 14 is a diagram showing an example of a travel section type correspondence table T50 according to the fourth embodiment of the present disclosure. The travel section type correspondence table T50 is a table indicating the section type T502 (tunnel, aboveground, underground) for each section of the above-described travel section T501. For example, as shown in FIG. 14, in the travel section type correspondence table T50, the travel section 1-5 is a tunnel, the travel section 11-6 is above ground, and the travel sections 11-7 and 11-8 are underground. It shows that

Next, feature quantity extraction according to the fourth embodiment of the present disclosure will be described with reference to FIG. 15 .

FIG. 15 is a diagram for explaining data used for feature quantity extraction according to the fourth embodiment of the present disclosure. A table T60 shown in FIG. 15 is a table in which the above-ground/underground determination unit 307 records the result of obtaining the illuminance using the traveling section type correspondence table T50 together with the vehicle interior temperature and the like included in the fourth on-board data. is.
As shown in FIG. 15, in a table T60, an on-rail section T602, a traveling section type T603, a sunshine T604, a vehicle interior temperature (°C) T605, a set temperature (°C) T606, an outside temperature (°C) T607, a passenger load T608, and Information on door opening/closing T609 is recorded every time T601.
The information of Table T60 is transferred to the feature amount extraction unit 301 as the fifth on-board data, and the feature amount is extracted.

For example, as shown in FIG. 15, at time 13:00:00, the vehicle 10 is in section 1-6, and according to table T50, section 1-6 is on the ground. T603 is on the ground, and the illuminance T604 is 1.
At time 13:00:45, the vehicle 10 is in section 1-7, and according to table T50, section 1-7 is underground. T604 is 0.

Next, data used for feature amount extraction according to the fourth embodiment of the present disclosure will be described with reference to FIG. 16 .

FIG. 16 is a diagram schematically showing the fifth on-board data. FIG. 16 is a diagram showing the temporal history of the in-vehicle temperature G402, the outside temperature G403, the set temperature G404, the passenger ratio G405, and the door open signal G406, as well as the sunshine intensity G407.

As described above, in the fourth embodiment of the present disclosure, in addition to the temperature, the occupancy rate, and/or the door opening frequency and the window opening rate, the sunshine intensity is further used as a feature amount to determine the abnormality of the air conditioner. be. The feature amount extraction unit 301 in the abnormality determination device 30 uses the vehicle interior temperature G402, the outside air temperature G403, the passenger ratio G405, and the door open signal G406 to extract the feature amount in the same manner as in the first embodiment, A time average value of the illuminance G407 is further extracted as a feature amount.

After the feature amount extraction is completed, the abnormality determination unit 305 in the abnormality determination device 30 performs abnormality determination for the tenth feature amount group obtained by adding the time average value of the sunshine intensity G407 in the inspection section G401 to the first feature amount group. Alternatively, the eleventh feature amount group in which the average value of the sunshine intensity G407 in the following interval is added to the second feature amount group, and the average value of the sunshine intensity G407 in the stable interval is added to the third feature amount group. Abnormality determination is performed for the 12 feature quantity groups.

As described above, in the feature amount extraction method according to the fourth embodiment of the present disclosure, in addition to the temperature, the boarding ratio, and the door opening frequency, the sunshine intensity is included in the feature amount group. Since it is possible to distinguish whether it is caused by high illuminance or by an air conditioning failure, it is possible to further reduce erroneous determination of the air conditioner.
In addition, in the fourth embodiment of the present disclosure, since the sunshine intensity is represented by two integers 0 and 1, the estimation accuracy of the sunshine intensity is lowered compared to continuous values as in the third embodiment. However, on the other hand, in Example 4, it is possible to estimate the illuminance without installing additional equipment such as a camera. Therefore, according to the fourth embodiment of the present disclosure, it is possible to estimate the intensity of sunlight with the configuration of the existing vehicle information management system while suppressing the cost of additional equipment such as a camera.

Note that the present disclosure is not limited to the above-described embodiments. Variations of any of the components of the above-described embodiments are possible within the scope of the invention. Also, arbitrary components can be added or omitted in the above-described embodiments.

1, 2, 3, 4 Vehicle information management system 10 Vehicle 20 On-board data 30 Abnormality determination device 101 Air conditioner 102 Air conditioning sensor 103 Vehicle information management device 104 Window opening ratio estimation unit 105 Camera 106 Sunlight intensity estimation unit 307 Above ground and underground determination unit

Claims (10)

A vehicle information management device,
a group of sensors for acquiring on-vehicle data including at least air-conditioning sensor information relating to an air-conditioning device installed in the vehicle and door opening information relating to the frequency of opening doors of the vehicle;
a feature extraction unit for extracting from the on-vehicle data a feature amount group including a temperature feature amount for the vehicle and a door opening frequency feature amount for the door opening frequency for a predetermined inspection section by analyzing the on-vehicle data;
an abnormality determination unit that generates and outputs an abnormality determination result indicating the presence or absence of an abnormality or an abnormality sign regarding the air conditioner based on the feature amount group and a normal model that indicates the operating characteristics of the air conditioner in a normal state;
A vehicle information management device comprising: The door opening information is
Information indicating the ratio of the door open time of the vehicle to the length of time the vehicle travels in a specific section,
The vehicle information management device according to claim 1, characterized by: The feature extraction unit is
The inspection section is from the time when the door is closed at the first station where the door open time of the vehicle is equal to or longer than a predetermined threshold to the time immediately before the door is opened at the second station where the door open time is equal to or longer than the threshold. set,
The vehicle information management device according to claim 1, characterized by: The feature extraction unit is
the inspection section,
a follow-up section in which the rate of temperature change in the vehicle is greater than a predetermined value;
and a stable interval in which the rate of temperature change in the vehicle is equal to or less than a predetermined value;
The temperature change rate is
is the moving median of the instantaneous temperature change rate in the vehicle;
The vehicle information management device according to claim 3, characterized by: The vehicle information management device is
a first imaging unit installed in the vehicle for obtaining first video data showing an interior state of the vehicle;
a window opening rate estimation unit that determines window opening rate information indicating the window opening rate of the vehicle by using a video processing method for the first video data acquired from the first imaging unit;
The feature extraction unit is
extracting a window opening rate feature amount related to the vehicle from the window opening rate information and including it in the feature amount group;
The vehicle information management device according to claim 1, characterized by: The vehicle information management device is
a second imaging unit that acquires second video data showing the sky above the vehicle;
a first sunshine intensity estimating unit that determines a sunshine intensity time history by using a video processing method for the second image data acquired from the second imaging unit;
The feature extraction unit is
extracting a sunshine feature amount related to sunshine intensity from the sunshine time history and including it in the feature amount group;
The vehicle information management device according to claim 1, characterized by: The vehicle information management device is
for each section, based on a travel section type correspondence table indicating whether a specific section is an above-ground section or an underground section, and the section where the vehicle is on the line, and identifies the travel section type of the section where the vehicle is located,
a second sunshine intensity estimating unit that determines a sunshine intensity time history based on the travel section type and the weather information about the on-rail section;
The feature extraction unit is
extracting a sunshine feature amount related to sunshine intensity from the sunshine time history and including it in the feature amount group;
The vehicle information management device according to claim 1, characterized by: A vehicle information management system,
a group of sensors for acquiring on-vehicle data including at least air-conditioning sensor information regarding an air-conditioning device installed in the vehicle and temperature variation factor information regarding a temperature variation factor affecting an interior temperature of the vehicle;
A vehicle information management device that is installed in the vehicle and collects the on-vehicle data acquired by the sensor group and transmits it to the outside;
an abnormality determination device that receives the on-vehicle data from the vehicle information management device via a communication network and determines whether or not there is an abnormality in the air conditioner;
including
The abnormality determination device is
a feature extracting unit that analyzes the on-vehicle data to extract, from the on-vehicle data, a feature amount group including a temperature feature amount related to the vehicle and a temperature variation feature amount related to a temperature variation factor for a predetermined inspection section;
an abnormality determination unit that generates and outputs an abnormality determination result indicating the presence or absence of an abnormality or an abnormality sign regarding the air conditioner based on the feature amount group and a normal model that indicates the operating characteristics of the air conditioner in a normal state;
A vehicle information management system comprising: The temperature variation factor information is
Information on the door opening frequency of the vehicle door, information on the window opening rate of the window of the vehicle, and at least one of the on-rail section of the vehicle,
The vehicle information management system according to claim 8, characterized by: A method for determining abnormality of an air conditioner,
a step of acquiring on-vehicle data including at least air conditioning sensor information regarding an air conditioner installed in a vehicle and door opening information regarding a door opening frequency of a door of the vehicle;
An inspection section is set from the time when the door is closed at the first station where the door open time of the vehicle is equal to or longer than a predetermined threshold to the time immediately before the door is opened at the second station where the door open time is equal to or longer than the threshold. and
a step of extracting from the on-vehicle data a feature quantity group including a temperature feature quantity relating to the vehicle and a door opening frequency feature quantity relating to the door opening frequency for the inspection section by analyzing the on-vehicle data;
a step of generating and outputting an abnormality determination result indicating the presence or absence of an abnormality or a sign of abnormality regarding the air conditioner, based on the feature quantity group and a normal model indicating the operating characteristics of the air conditioner in a normal state;
An abnormality determination method comprising:

Images