System and method for estimating durability of a fuel cell system

Abstract

To estimate durability according to a method of machine learning even regarding a fuel cell system of which sufficient pieces of data for learning do not exit.SOLUTION: A system comprises: a first storage device storing actuation data of a fuel cell system of a first type as data for learning; a first arithmetic unit including a machine learning model which for estimating durability of the fuel cell system of the first type by performing machine learning using the data for learning; a second storage device storing results of durability tests of fuel cell systems of the first type and a second type; and a second arithmetic unit which determines a conversion formula or a conversion model for leading the result of the durability test of the fuel cell system of the second type from the result of the durability test of the fuel cell system of the first type using the results of the durability tests. The second arithmetic unit estimates the durability of the fuel cell system of the second type by applying the conversion formula or the conversion model determined by the second arithmetic unit to the durability of the fuel cell system of the first type estimated by the machine learning model.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The technology disclosed in this specification relates to a system and method for estimating the durability of a fuel cell system.

Background Art

[0002] Patent Document 1 describes a technology for reusing a battery mounted on a vehicle for other products. In this technology, both the operation data of the battery before reuse and the operation data of the battery after reuse are collected, and a machine learning model for estimating the durability (life) of the battery is created by machine learning using the collected operation data as learning data.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By using machine learning techniques as described above, it is possible to estimate the durability of industrial products. In this regard, since a fuel cell system is, for example, mounted on a vehicle or reused in other products later, it is desirable to accurately estimate its durability. When using machine learning techniques, a large amount of learning data is required. For example, in the case of an existing fuel cell system that has already been provided to the public, operating data when it is actually operating can be obtained as learning data from a large number of fuel cell systems used by the public. On the other hand, for a new fuel cell system obtained by improving the existing fuel cell system, such learning data cannot be obtained sufficiently, and it is difficult to estimate the durability by machine learning techniques only from the results of limited durability tests conducted before being provided to the public. This specification provides a technique capable of estimating the durability by machine learning techniques even for a fuel cell system lacking sufficient learning data.

Means for Solving the Problems

[0005] The technology disclosed in this specification is embodied in a system for estimating the durability of a fuel cell system. In a first aspect, the system includes a first storage device that stores, as learning data, operation data when a first type of fuel cell system actually operates, for one or more first type of fuel cell systems; a first arithmetic device that is a machine learning model that performs machine learning using the learning data stored in the first storage device and estimates the durability of the first type of fuel cell system; a second storage device that stores the results of a durability test of the first type of fuel cell system and the results of a durability test of a second type of fuel cell system different from the first type of fuel cell system, for the same usage conditions; and a second arithmetic device that determines a conversion formula or conversion model for deriving the results of the durability test of the second type of fuel cell system from the results of the durability test of the first type of fuel cell system, using the results of the durability test stored in the second storage device. The second arithmetic device may apply the conversion formula or conversion model determined by the second arithmetic device to the durability of the first type of fuel cell system estimated by the machine learning model of the first arithmetic device to estimate the durability of the second type of fuel cell system.

[0006] In the above-described system, operation data from one or more first-type fuel cell systems is stored as learning data. Then, a machine learning model for estimating the durability of the first-type fuel cell system is created by performing machine learning using the learning data. Further, using the results of the durability tests of the first-type and second-type fuel cell systems, a conversion formula or conversion model for deriving the results of the durability test of the second-type fuel cell system from the results of the durability test of the first-type fuel cell system is created. Then, the conversion formula or conversion model is applied to the estimated durability result of the first-type fuel cell system described above. According to such a configuration, by applying the conversion formula or conversion model based on the results of the durability tests of the two types of fuel cell systems to the estimated durability result of the first-type fuel cell system by the machine learning model, the durability of the second-type fuel cell system can be derived. Thereby, the durability of the second-type fuel cell system can be estimated depending not only on the results of the durability test of the second-type fuel cell system but also on the operation data of the known first-type fuel cell system.

[0007] In a second aspect, in the above first aspect, the system may further include a receiving device that receives operation data from a first-type fuel cell system provided to the public or an industrial product equipped with the same. The first storage device may be connected to the receiving device and store the operation data received by the receiving device. According to such a configuration, a large amount of operation data can be collected from a large number of first-type fuel cell systems provided to the public or industrial products equipped with the same.

[0008] In a third aspect, in the above first or second aspect, the second arithmetic unit may determine a conversion formula or conversion model by machine learning using the results of the durability test stored in the second storage device. According to such a configuration, the conversion formula or conversion model can be determined with high accuracy. However, as another embodiment, the conversion formula or conversion model may have a simple structure, such as being represented by a linear function, for example.

[0009] The technology disclosed in this specification is embodied in a method for estimating the results of a durability test on a fuel cell system. The method includes, for one or more first-type fuel cell systems, storing the operating data when the first-type fuel cell system actually operates as learning data in a first storage device; using the learning data stored in the first storage device to perform machine learning on a machine learning model that the first computing device has, and estimating the durability of the first-type fuel cell system; storing in a second storage device the results of the durability test of the first-type fuel cell system and the results of the durability test of a second-type fuel cell system different from the first-type fuel cell system under the same usage conditions; determining, in a second computing device, a conversion formula or conversion model for deriving the results of the durability test of the second-type fuel cell system from the results of the durability test of the first-type fuel cell system using the results of the durability test stored in the second storage device; and applying, in the second computing device, the conversion formula or conversion model determined in the second computing device to the durability of the first-type fuel cell system estimated by the machine learning model of the first computing device to estimate the durability of the second-type fuel cell system. According to such a method, by applying a conversion formula or conversion model based on the results of the durability tests of two types of fuel cell systems to the estimated durability results of the first-type fuel cell system by the machine learning model, the durability of the second-type fuel cell system can be derived. Thereby, the durability of the second-type fuel cell system can be estimated depending not only on the results of the durability test of the second-type fuel cell system but also on the operating data of the known first-type fuel cell system.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Embodiments for Carrying Out the Invention

[0011] (Embodiment) Referring to FIGS. 1 to 3, the estimation system 10 of this embodiment will be described. In this estimation system 10, the durability of a fuel cell system is estimated. Although it is an example, the fuel cell system is mounted on a vehicle. However, it is not limited to a vehicle, and the fuel cell system may be mounted on other industrial products other than vehicles.

[0012] As shown in FIG. 1, the estimation system 10 includes a receiving device 12, a first storage device 14, a second storage device 16, a first arithmetic device 18, a second arithmetic device 20, and a communication device 22. The receiving device 12 is a device that receives operation data of a fuel cell system (hereinafter referred to as an existing type FC system 2a) mounted on each of a plurality of vehicles 2 (hereinafter simply referred to as vehicles 2) provided to the public. Although it is an example, part or all of the receiving device 12 can be composed of a mobile communication network provided by a mobile phone carrier. The receiving device 12 wirelessly receives the operation data of the existing type FC system 2a transmitted from the plurality of vehicles 2 by any communication protocol. The receiving device 12 is communicably connected to the first storage device 14 using a wired and / or wireless line. The receiving device 12 transmits the stored operation data to the first storage device 14. Note that the operation data transmitted by the receiving device 12 to the first storage device 14 is not limited to the data of the existing type FC system 2a of the plurality of vehicles 2, and may be the data of the existing type FC system 2a of a single vehicle 2. However, the larger the number of vehicles 2 providing the operation data, the better.

[0013] In addition to the aforementioned existing type FC system 2a, the vehicle 2 is provided with a plurality of sensors, a memory, and a communication device. Each sensor measures values of various parameters (voltage, current, voltage rate, temperature (temperature inside the tank, outside air temperature, etc.), gas flow rate, gas pressure, etc.) related to the existing type FC system 2a of the vehicle 2 at predetermined time intervals. The operation data is time-series data obtained by associating the measured values of various parameters with time information together with a control signal (such as a target current, etc.) for controlling the fuel cell system. The memory stores this operation data. The communication device reads the operation data from the memory and periodically transmits the operation data to the receiving device 12 together with the unique vehicle identification information for each vehicle 2. Although not particularly limited, the communication device may be, for example, a wireless communication-capable mobile device such as a mobile phone owned by the user of the vehicle 2.

[0014] The first storage device 14 is communicably connected to the first arithmetic device 18. The first storage device 14 has a memory. The first storage device 14 receives the operation data transmitted from the receiving device 12 and stores the received operation data as learning data. The first storage device 14 transmits the stored learning data to the first arithmetic device 18. The first arithmetic device 18 executes estimation processing. The first arithmetic device 18 has a memory and a CPU. The first arithmetic device 18 has a learning model 19 stored in the memory. The first arithmetic device 18 causes the CPU to execute various processes based on the learning model 19 and / or other programs stored in the memory. The first arithmetic device 18 causes the learning model 19 to perform machine learning on the learning data received from the first storage device 14. The first arithmetic device 18 estimates the durability result of the existing type FC system 2a using the machine-learned learning model 19. The communication device 22 is a communication network that connects the first arithmetic device 18 and the second arithmetic device 20 to each other. The first arithmetic device 18 is communicably connected to the second arithmetic device 20 via the communication device 22. The first arithmetic device 18 transmits the durability result estimated by the learning model 19 to the second arithmetic device 20. Note that the communication device 22 may be constituted by, for example, a communication instruction program incorporated in each of the first arithmetic device 18 and the second arithmetic device 20, or a primary storage device provided in one or both of the first arithmetic device 18 and the second arithmetic device 20.

[0015] The second storage device 16 is communicably connected to the durability test device 4 and the second arithmetic device 20. The second storage device 16 has a memory. The second storage device 16 stores the results of durability tests of various types of fuel cell systems received from the durability test device 4. The results of the durability tests of various types of fuel cell systems include the results of the durability test of the existing FC system 2a and the results of the durability test of the new FC system under the same usage conditions. The new FC system is different from the existing FC system 2a and is a fuel cell system obtained by improving the existing FC system 2a. Although it is an example, the output (kW) of the new FC system is about ±20% of the output (kW) of the existing FC system 2a. The second storage device 16 transmits the stored results of the durability tests of various types of fuel cell systems to the second arithmetic device 20. Here, in a fuel cell system, when the usage conditions experienced by the fuel cell system are different, the possible values of the operating parameters (for example, the change rates of current and voltage, outside air temperature, etc.) that affect the deterioration of the fuel cell system can be affected. Therefore, the results of the durability test of the existing FC system 2a and the results of the durability test of the new FC system, which are carried out under the same usage conditions as each other, are transmitted to the second arithmetic device 20. The same usage conditions in this specification refer to those in which the ranges and occurrence frequencies of the possible values of each operating parameter are about the same. Here, the existing FC system 2a is an example of the "first type of fuel cell system" in the technology disclosed in this specification, and the new FC system is an example of the "second type of fuel cell system" in the technology disclosed in this specification.

[0016] The durability test device 4 is a device for conducting a durability test on a fuel cell system. The durability test device 4 includes, for example, a chassis dynamometer, and can conduct a durability test using a test vehicle equipped with the fuel cell system. Similar to the vehicle 2 described above, the test vehicle measures values of various parameters (voltage, current, voltage rate, temperature (temperature inside the tank, outside air temperature, etc.), gas flow rate, gas pressure, etc.) related to the fuel cell system at predetermined time intervals by means of various sensors provided on the actual vehicle. The durability test device 4 records, as a result of the durability test, time-series data in which time information is associated with those measurement data, together with control signals (such as target current) for controlling the fuel cell system. The durability test device 4 transmits the result of the durability test to the second storage device 16. Note that the specific configuration of the durability test device 4 is not particularly limited. Instead of using the test vehicle, the durability test device 4 may use a simulator device capable of reproducing the operation of the fuel cell system mounted on the actual vehicle.

[0017] The second arithmetic device 20 has a memory and a CPU. The CPU of the second arithmetic device 20 determines a conversion model for deriving the result of the durability test of the new FC system from the result of the durability test of the existing FC system 2a using the results of the durability tests of the existing and new FC systems received from the second storage device 16. The second arithmetic device 20 determines the conversion model by having the machine learning model stored in the memory perform machine learning. Thereby, the conversion model can be determined with high accuracy. However, the conversion model does not necessarily have to be determined by machine learning, and may have a simple structure such as being represented by a linear function. For example, if the conversion formula is y = Ax. Here, x is the index value of the existing FC system 2a, which is the difference between the value after the durability test and the initial value, y is the index value of the new FC system, which is the difference between the value after the durability test and the initial value, A is a constant, and is determined from the ratio of x and y. When the durability of the new FC system is improved compared to the existing FC system 2a, A becomes a value smaller than 1.

[0018] The CPU of the second arithmetic unit 20 applies the above-described conversion formula or conversion model to the durability of the existing FC system 2a estimated by the learning model 19 of the first arithmetic unit 18. Thereby, the second arithmetic unit 20 can estimate the durability of the new FC system.

[0019] Referring to FIG. 2, the estimation by the learning model 19 and the application of the conversion model to the estimation will be described. As shown in FIG. 2, when the usage conditions of the existing FC system 2a to be estimated are input to the learning model 19 of the first arithmetic unit 18, the learning model 19 outputs an estimation result of the durability of the existing FC system 2a. The input parameters (usage conditions) include a current value, temperature, voltage change rate, operating time, and the like. Here, the output parameter (estimation result of durability) is, for example, the voltage value of the fuel cell system when the current in the fuel cell system is a predetermined value. The higher the voltage value, the higher the durability of the fuel cell system is determined to be, and the lower the voltage value, the lower the durability of the fuel cell system is determined to be. The learning model 19 may be optimized using some of the learning data read from the first storage device 14 so that the error between the estimated result of the output and the result of the operation data is minimized when an input of usage conditions is given. Also, in order to ensure the generalization performance of the learning model 19, some of the learning data may be secured for verification. In this case, the secured data may be used to avoid overfitting of the learning model 19.

[0020] When the estimation result of the existing FC system 2a from the first arithmetic unit 18 is input to the second arithmetic unit 20, the second arithmetic unit 20 applies a conversion model or conversion formula to the input estimation result and outputs a result of the durability of the new FC system. The output parameter here is, similar to the first arithmetic unit 18 (i.e., the learning model 19), for example, the voltage value of the fuel cell when the current in the fuel cell system is a predetermined value.

[0021] Note that the output parameters of the learning model 19 are not limited to the voltage values at specific current values, and may be voltage values under specific conditions such as voltage values at specific temperatures / gas supply amounts. The output parameter may be an index value representing the degree of degradation of the fuel cell system. In a modification, the output parameter may be, for example, the amount of gas leaking from the gas tank when hydrogen is pressurized under specific conditions.

[0022] In the estimation system 10 in this embodiment, ElasticNet is adopted as the learning model 19. However, the learning model 19 is not particularly limited, and other machine learning models such as regression different from ElasticNet (for example, Ridge regression, Lasso regression, etc.) and neural networks may be adopted.

[0023] Conventionally, by using machine learning techniques, the durability of industrial products can be estimated. In this regard, since the fuel cell system is, for example, mounted on a vehicle or reused in other products later, it is desirable to accurately estimate its durability. When using machine learning techniques, a large amount of learning data is required. For example, for an existing fuel cell system that has already been provided to the public, operating data when actually operating can be obtained as learning data from a large number of fuel cell systems used by the public. On the other hand, for a new fuel cell system obtained by improving the existing fuel cell system, such learning data cannot be obtained sufficiently, and it is difficult to estimate the durability by machine learning techniques only based on the results of limited durability tests conducted before being provided to the public.

[0024] In the above-described estimation system 10, operation data from one or more existing FC systems 2a is stored as learning data. Then, a learning model 19 for estimating the durability of the existing FC system 2a is created by performing machine learning using the learning data. Further, using the results of the durability tests of the existing and new FC systems, a conversion formula or conversion model for deriving the results of the durability test of the new FC system from the results of the durability test of the existing FC system 2a is created. Then, the conversion formula or conversion model is applied to the estimated durability results of the existing FC system 2a described above. According to such a configuration, by applying the conversion formula or conversion model based on the results of the durability tests of the two types of fuel cell systems (existing and new FC systems) to the estimated durability results of the existing FC system 2a by the learning model 19, the durability of the new FC system can be derived. Thereby, the durability of the new FC system can be estimated depending not only on the results of the durability test of the new FC system but also on the operation data of the known existing FC system 2a.

[0025] Next, with reference to FIG. 3, the procedure of the durability estimation process (estimation method) executed by the estimation system 10 will be described. As shown in FIG. 3, first, in step S12, the estimation system 10 performs a first storage step of storing learning data. Specifically, in the first storage step, the estimation system 10 stores the operation data of a plurality of existing FC systems 2a transmitted from the receiving device 12 in the first storage device 14 as learning data. Next, in step S14, the estimation system 10 executes a machine learning step of performing machine learning on the learning data. Specifically, in the machine learning step, the estimation system 10 performs machine learning on the learning model 19 of the first arithmetic device 18 using the learning data stored in the first storage device 14 in S12.

[0026] Next, in step S16, the estimation system 10 executes a first estimation step using the machine learning model. Specifically, in the first estimation step, the estimation system 10 estimates the durability results of the existing FC system 2a using the learning model 19 machine-learned by the first arithmetic device 18.

[0027] Next, in step S18, the estimation system 10 executes a second storage process of storing the results of the durability test. Specifically, in the second storage process, the estimation system 10 causes the second storage device 16 to store the results of the durability tests of the existing FC system 2a and the new FC system. Next, in step S20, the estimation system 10 executes a determination process of determining a conversion model using the results of the durability test. Specifically, in the determination process, the estimation system 10 causes the second arithmetic unit 20 to use the results of the durability tests of the existing FC system 2a and the new FC system stored in the first storage device 14 in S18 to determine a conversion formula or conversion model for deriving the results of the durability test of the new FC system from the results of the durability test of the existing FC system 2a.

[0028] Next, in step S22, the estimation system 10 executes an application process of applying the conversion model to the estimation by the learning model 19. Specifically, in the application process, the estimation system 10 causes the conversion formula or conversion model determined in S20 to be applied to the durability of the existing FC system 2a estimated by the learning model 19 of the first arithmetic unit 18. Through the series of steps S12 to S22 described above, the durability of the new FC system can be estimated depending not only on the results of the durability test of the new FC system but also on the operation data of the known existing FC system 2a.

[0029] The estimation system 10 in this embodiment includes a receiving device 12 that receives operation data from a vehicle 2 equipped with the existing FC system 2a provided to the public. The first storage device 14 is connected to the receiving device 12 and stores the operation data received by the receiving device 12. With such a configuration, a large amount of operation data can be collected from a large number of vehicles 2 equipped with the existing FC system 2a provided to the public.

[0030] In the estimation system 10 of this embodiment, the operation data from the receiving device 12 is transferred to the first arithmetic device 18 via the first storage device 14. However, it is not necessarily required to pass through the first storage device 14, and it may be directly transmitted from the receiving device 12 to the first arithmetic device 18. In this case, the memory of the first arithmetic device 18 may function as the first storage device 14. Similarly, the results of the endurance test from the endurance test device 4 do not necessarily have to pass through the second storage device 16, and may be directly transmitted from the endurance test device 4 to the second arithmetic device 20.

[0031] As described above, specific examples of the technology disclosed in this specification have been described in detail. However, these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and changes of the specific examples illustrated above. The technical elements described in this specification or the drawings exhibit technical utility alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing. The technology illustrated in this specification or the drawings can achieve multiple objectives simultaneously, and achieving one of these objectives itself has technical utility.

Description of Reference Numerals

[0032] 2a: Existing FC system 10: Estimation system 12: Receiving device 14, 16: Storage devices 18, 20: Arithmetic devices 19: Learning model

Claims

1. A system for estimating the durability of a fuel cell system, comprising: a first storage device that stores, as learning data, operation data when the first type of fuel cell system actually operates, for one or more first type of fuel cell systems; a first arithmetic unit having a machine learning model that performs machine learning using the learning data stored in the first storage device and estimates the durability of the first type of fuel cell system; a second storage device that stores the results of the durability test of the first type of fuel cell system and the results of the durability test of a second type of fuel cell system different from the first type of fuel cell system, under the same usage conditions; a second arithmetic unit that determines a conversion formula or conversion model for deriving the results of the durability test of the second type of fuel cell system from the results of the durability test of the first type of fuel cell system, using the results of the durability test stored in the second storage device; wherein the second arithmetic unit applies the conversion formula or conversion model determined by the second arithmetic unit to the durability of the first type of fuel cell system estimated by the machine learning model of the first arithmetic unit, to estimate the durability of the second type of fuel cell system. A system.

2. The system according to claim 1, further comprising a receiving device that receives the operation data from the first type of fuel cell system provided to the public or an industrial product equipped with the same, wherein the first storage device is connected to the receiving device and stores the operation data received by the receiving device.

3. The system according to claim 1 or 2, wherein the second arithmetic unit determines the conversion formula or conversion model by machine learning using the results of the durability test stored in the second storage device.

4. A method for estimating the durability of a fuel cell system, comprising: storing, as learning data in a first storage device, operation data when a first type of fuel cell system actually operates, for one or more first type of fuel cell systems; estimating the durability of the first type of fuel cell system using a machine learning model that performs machine learning using the learning data stored in the first storage device, by a first arithmetic unit having the machine learning model; A step of storing, in a second storage device, the result of the durability test of the first type of fuel cell system and the result of the durability test of a second type of fuel cell system different from the first type of fuel cell system, regarding the same usage conditions; A step of causing a second arithmetic device to determine a conversion formula or a conversion model for deriving the result of the durability test of the second type of fuel cell system from the result of the durability test of the first type of fuel cell system, using the result of the durability test stored in the second storage device; A step of causing the second arithmetic device to apply the conversion formula or the conversion model determined by the second arithmetic device to the durability of the first type of fuel cell system estimated by the machine learning model of the first arithmetic device, to estimate the durability of the second type of fuel cell system; A method comprising the above.

Images