Design support device and design support method
The design support device and method address the challenge of insufficient operation data by using simulation and Bayesian networks to predict and mitigate fire risks in battery systems and other equipment, ensuring effective fire risk reduction through causal relationship analysis.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- HITACHI LTD
- Filing Date
- 2022-11-25
- Publication Date
- 2026-06-16
AI Technical Summary
Existing design support systems for equipment, such as battery systems, are inadequate when operation data is insufficient, particularly for systems with a large number of connected batteries, as they cannot effectively predict and mitigate fire risks.
A design support device and method that utilizes an input unit, event identification, causal relationship storage and identification, and specification calculation to estimate fire probabilities and minimize accident occurrence by integrating simulation and Bayesian networks to analyze causal relationships between design specifications and fire risks.
Enables accurate estimation of fire probabilities and identification of design specifications that reduce fire risks even in data-insufficient conditions, providing a robust design support system for battery systems and general equipment.
Smart Images
Figure 0007874528000001 
Figure 0007874528000002 
Figure 0007874528000003
Abstract
Description
[Technical Field]
[0001] The present invention relates to a design support device and a design support method for various types of equipment. [Background technology]
[0002] Design support devices are being considered and put into practical use to assist in the design of various types of equipment. Specifically, for example, there is a demand for reducing carbon dioxide emissions in order to aim for a sustainable society. In response to this, renewable energy power plants such as wind and solar power, and electric vehicles are increasing. Renewable energy generation has the problem of voltage and frequency being prone to instability, and power sources that can compensate for these are needed. Electric vehicles also require power sources that can achieve both long-distance driving and high voltage.
[0003] One such power source is the small, high-energy-density lithium-ion battery. However, small, high-energy-density batteries like lithium-ion batteries can cause fires, and reports of such accidents are increasing.
[0004] Given the above background, inventions have been proposed that aim to prevent fire accidents or to propose operations that anticipate fire accidents. For example, Patent Document 1 proposes estimating the degradation state of electric vehicle batteries and raising insurance premiums if the degradation exceeds a certain level, as this increases the possibility of ignition. [Prior art documents] [Patent Documents]
[0005] [Patent Document 1] Japanese Patent Publication No. 2022-21895 [Overview of the project] [Problems that the invention aims to solve]
[0006] The above invention utilizes data during the operation of a battery system. However, when there is no data during the operation with a large number of batteries connected, there is a problem that the above prior art cannot be applied. This problem of lacking such operation data can also occur in the case of general target equipment other than the battery system.
[0007] Therefore, an object of the present invention is to provide a design support device and a design support method that can be applied even when operation data is insufficient.
Means for Solving the Problems
[0008] Based on the above, in the present invention, "a design support device for a battery system in which the target equipment for design support combines a plurality of unit cells be, comprises an input unit for inputting the design specifications of the battery system, an event identification unit for identifying events that cause accidents occurring in the target equipment of the design specifications, a causal relationship storage unit for storing causal relationship information between accidents and accident causes in the target equipment, a causal relationship identification unit for identifying causal relationship information from the accident related to the target equipment to the events that cause the accident using the events that cause the accident identified by the event identification unit and the causal relationship information stored in the causal relationship storage unit, a design specification calculation unit for calculating a design specification with the minimum occurrence probability of the accident in the target equipment using the causal relationship information identified by the causal relationship identification unit, and an output unit for outputting the design specification calculated by the design specification calculation unit. The causal relationship identification unit calculates the probability of the design specification that causes the accident at the time of accident occurrence and identifies the causal relationship information."
[0009] Also, in the present invention, "a design support method in a design support device comprising an input unit, an event identification unit, a causal relationship storage unit, a causal relationship identification unit, a design specification calculation unit, and an output unit. The input unit is for a battery system in which the target equipment for design support combines a plurality of unit cells be,Input the design specifications of the battery system. The event identification unit identifies events that cause accidents occurring in the target equipment of the design specifications. The causal relationship storage unit stores causal relationship information between accidents and accident causes in the target equipment. The causal relationship identification unit calculates the probability of the design specifications that cause the accidents when the accidents occur, using the events that cause the accidents identified by the event identification unit and the causal relationship information stored in the causal relationship storage unit, and identifies the causal relationship information from the accidents related to the target equipment to the events that cause the accidents. The design specification calculation unit calculates the design specifications that minimize the occurrence probability of the accidents in the target equipment, using the causal relationship information identified by the causal relationship identification unit. The output unit outputs the design specifications calculated by the design specification calculation unit. This is a design support method characterized by the above.
Effect of the Invention
[0010] According to the present invention, it is possible to provide a design support system applicable even when the data during operation is insufficient. In the case of a specific battery system, even when there is no data during the operation of a large number of connected storage batteries, the fire probability and the design specifications that can reduce the fire probability can be estimated.
Brief Description of the Drawings
[0011] [Figure 1] A diagram showing a configuration example of a design support system according to an embodiment of the present invention configured using a computer device. [Figure 2] A diagram showing a screen example of the display unit 40 at the time of input. [Figure 3] A diagram showing the stored content of the circuit specification storage unit DB2. [Figure 4] A diagram showing a configuration example of a battery circuit. [Figure 5] A diagram showing a configuration example of the Bayesian network structure storage unit DB4. [Figure 6] A diagram showing a configuration example of the Bayesian network probability storage unit DB3. [Figure 7] A diagram showing a screen example of a specification proposal selection display. [Figure 8] A diagram showing a schematic processing flow example of the design support system 100. [Figure 9] A diagram showing subroutine SUB01 within the main routine. [Figure 10] A diagram showing subroutine SUB02 within the main routine. [Modes for carrying out the invention]
[0012] The embodiments of the present invention will be described below with reference to the drawings. While the design support system according to the embodiments of the present invention can support the design of any equipment, the specific examples shown below will use the design of a battery system in Embodiment 1 as an example, and a universal structure that does not limit the design support target in Embodiment 2. [Examples]
[0013] Figure 1 shows a design support system for a battery system as an example of the configuration of a design support system 100 according to an embodiment of the present invention, which is configured using a computer device. In a battery system, multiple sets of unit battery cells are connected in series or in parallel, and the desired characteristics are obtained by these combinations. The design support system is required to perform a design that can reduce the probability of fires and other incidents.
[0014] The design support system 100 in Figure 1 consists of an input unit 30, a display unit 40, a storage unit DB, and a calculation unit 10, all connected via a bus 20. The storage unit DB includes a temporary storage unit DB1, a circuit specification storage unit DB2, a Bayesian network probability storage unit DB3, and a Bayesian structure storage unit DB4. Functionally, the processing in the calculation unit 10 can be described as having the processing functions of a battery circuit simulator 11, a battery degradation simulator 12, a heat quantity calculation unit 13, a Bayesian network generation unit 15, a fire probability calculation unit 16, and a specification search unit 17. The temporary storage unit DB1 stores data as appropriate during the calculation process, and a detailed explanation is omitted.
[0015] The design support system 100 in Figure 1 first prompts the user (designer) to input the necessary specifications from the input unit 30 according to the content displayed on the display unit 40. Figure 2 shows an example of the screen of the display unit 40 at this time, and the item values D23 of the necessary specification items D21 are input using the input units (802, D23) formed on the screen. The specific specification items D21 are item values D23 (conditions (e.g., series / parallel) and numerical values (number, etc.)) for circuit connection patterns D21a, number of series cells N (D21b), number of parallel cells M (D21c), cell capacitance D21d, cell resistance D21e, etc., and the user inputs the corresponding item values D23 on the screen where these are already displayed. Note that some of the item values D23 may be pre-initialized in the form of reference values or default values.
[0016] These inputs (item values D23) are stored in the circuit specification storage unit DB2 in Figure 1 in the data format shown in the upper part of Figure 3. The circuit specification storage unit DB2 records the design specifications that will be used as input values for the battery degradation simulator 12 and the battery circuit simulator 11, which will be described later. The design specifications consist of D21, which represents the specification item name in Figure 3, D22, which indicates how the specification item D21 will be handled in the simulation, and D23, which records the specific value of the specification item D21.
[0017] Specification item D21 is the name of the item that indicates the specifications of the equivalent circuit of the battery cells that make up the circuit, as shown in Figure 3. Specification item D21 takes the value of either a constant or a random variable, as shown in item type D22. When it is set to a random variable, such as cell capacitance D21d and cell resistance D21e, the simulator performs a simulation that takes errors into account. Cell capacitance D21d and cell resistance D21e are treated in this way because they contain errors depending on the state of degradation. Item type D22 and specification item D21 are defined when creating a system to which the present invention is applied, but the item value D23 is entered from the screen in Figure 2.
[0018] According to the data group shown in the upper part of the circuit specification storage unit DB2 in FIG. 3, it means that a set of design specification plans has been defined. The circuit specification storage unit DB2 can store a plurality of sets of design specification plans given by the designer. With this data input, the circuit design of the battery system as shown in FIG. 4 is completed. For example, the left side of FIG. 4 represents that a series circuit composed of three cells has been designed, and the right side of FIG. 4 represents that a parallel circuit composed of three cells has been designed. Each individual unit cell is represented as an equivalent circuit in which a resistor (cell resistor D21e) and a capacitor (cell capacitance D21d) are connected in parallel.
[0019] Using these design specification plans, the battery circuit simulator 11 performs circuit simulation of the battery circuit. The circuit to be simulated is a large number of equivalent circuits of battery cells, which are the minimum units of the storage battery, connected in large quantities as shown in FIG. 4. The information on the circuit connection and the specifications of the battery cells are recorded in the circuit specification storage unit DB2 as pattern D21a, series / parallel cell numbers D21b, D21c, cell capacitance D21d, and cell resistor D21e.
[0020] That equivalent circuit is composed of N resistors R1~R N and capacitances C1~C N when there are N cells. The circuit specification storage unit DB2 can calculate the current values I1~I N flowing through the resistors R1~R N during the charge and discharge of each battery cell, and the voltage values V1~V N of the equivalent circuit. Thus, it can be confirmed that the battery circuit of the design specification plan satisfies the desired characteristics.
[0021] On the other hand, the storage battery cell deteriorates after going through charge and discharge many times, and the N resistors R1~R N and the N capacitances C1~C N gradually change. It is the battery deterioration simulator 12 that simulates such deterioration. In the battery deterioration simulator 12, the current values I1(k)~I N (k) and the voltage values V1(k)~V NR1(k+1) ~ R from (k) to the k+1th charge / discharge cycle N (k+1) and C1(k+1)~C N (k+1) is calculated. Note that simulators that simulate battery degradation, such as the battery degradation simulator 12, are well-known technologies, and these can be applied as appropriate in the present invention.
[0022] In this embodiment, the battery degradation simulator 12 and the battery circuit simulator 11 are operated in combination as follows. In the battery degradation simulator 12, R1(k)~R at the kth charge / discharge cycle N (k) and C1(k)~C N (k) First, using the battery circuit simulator 11, we obtain the cell current value I1(k)~I N (k) and cell voltage value V1(k)~V N We estimate (k). Next, I1(k)~I N (k) and V1(k)~V N (k) From the battery circuit simulator 11, R1(k+1)~R N (k+1) and C1(k+1)~C N (k+1) is determined. By using the battery circuit simulator 11 and the battery degradation simulator 12 alternately in this way, the cell current values I1~I that take battery degradation into account are calculated. N and voltage values V1~V N It can be calculated.
[0023] Using these simulation results, the heat quantity calculation unit 13 calculates the current values I1 to I of the cell. N and voltage values V1~V N From the resistance values R1~R N The amount of electrical energy converted into heat (heat generation) is calculated. This can be calculated according to Ohm's law by multiplying the resistance value by the square of the current. The calculated heat generation Q is added and stored as heat generation D21f in the circuit specification memory DB2, as shown in the lower part of Figure 3.
[0024] The calculated heat output Q is then incorporated into the Bayesian network as the cause of the cell fire. Here, a Bayesian network is a network constructed from causal relationships, such as physical events, and is configured within the Bayesian network probability storage unit DB3 and the Bayesian structure storage unit DB4 by the Bayesian network generation unit 15 in Figure 1. The former represents causal relationships as probabilities, and the latter represents causal relationships as structures, and both methods make it possible to probabilistically grasp the occurrence of causal relationships.
[0025] Of these, the Bayesian network structure memory unit DB4 is a database that stores a network structure showing causal relationships, as shown in Figure 5. The network structure showing causal relationships is configured in two stages, for example. The lower layer 350 represents the causal relationship from the design specifications to the heat generation Q of the battery cell, and the upper layer 300 represents the causal relationship from battery cell ignition to a fire in the entire battery system.
[0026] More specifically, the lower layer 350 represents the causal relationship between the design specifications (in the examples in Figures 2 and 3, such as the circuit connection pattern D21a, the number of series cells N (D21b), the number of parallel cells M (D21c), the cell capacity D21d, and the cell resistance D21e) and the heat generated by the battery cells Q. The lower layer 350 corresponds to the results simulated by the battery degradation simulator 12 and the battery circuit simulator 11, while the Bayesian network structure memory unit DB4 stores only the structure information.
[0027] The upper layer 300 represents the direct causes (X2: failure of the fire detection device, X3: failure of the fire suppression system, X4: battery module fire, X5: damage to the inter-module firewall) or indirect causes of the final result, a fire event (X1: battery system fire), as a network. As an example of an indirect cause, Figure 5 shows a series of events in which, due to the high heat generation of the battery cells, X7: one of the battery cells ignites, then X6: damage to the inter-cell firewall progresses, and at this time, X8: the amount of heat removed by the cooling fan is considered, but it leads to X4: a battery module fire.
[0028] Here, each event X(X1-X8) can be expressed as a probability of occurrence. If the probability of occurrence of an individual event is low, the progression of the event is unlikely, and conversely, if it is high, it is easy to reach a fire accident X1. According to this, one of the causes of the final result, the fire event (X1: battery system fire occurs), is X2: failure of the fire detection device. For simplicity, these events and design specifications are shown in Figure 5 as events X1-X9, number of series battery cells N, number of parallel battery cells M, and cell capacities C1-C N , cell resistance R1~R N This is represented by the variable X1 = Yes. When an event occurs, it is written here as X1 = Yes. The Bayesian network structure memory unit DB4 is to be constructed using the Bayesian network generation unit 15 in Figure 1 when designing the system to which the present invention is applied.
[0029] Next, the Bayesian network probability memory unit DB3 stores the probability parameters between the variables representing the result and cause in the Bayesian network structure memory unit DB4, as shown in Figure 6. However, while some variables have a defined cause, such as X4: battery module fire for the fire event (X1: battery system fire) which is the final result in Figure 5, others do not have a defined cause, such as X2: fire detection device failure and X3: fire extinguishing system failure.
[0030] Therefore, the Bayesian network structure memory unit DB4 is configured as shown in Figure 6, with a cause-enabled variable table TB31 and a cause-non-variable table TB32, depending on whether or not a cause is set. Variables with a cause, such as X1, are stored in the cause-enabled variable table TB31, while variables without a cause are stored in the cause-non-variable table TB32.
[0031] The causal variable table TB31 stores the values of the causal variable D31, the value of the outcome variable D32, and the conditional probability D33 for each of those values. For example, looking at the first row of D31, D32, and D33, the causal variable D31 is X2=Yes, X3=Yes, X4=Yes, X5=Yes, which indicates the case where all of the failure phenomena (X2, X3, X4, X5) in Figure 5 have occurred. In other words, when all of these failure phenomena (X2, X3, X4, X5) have occurred, the probability of X1=Yes (a fire occurs), as shown in conditional probability D33, is 99%.
[0032] Similarly, the second row shows that when all of the failure phenomena in Figure 5 (X2, X3, X4, X5) occur, the probability that X1 = No (no fire occurs) is 1%, and the third row shows that when X2 = No, X3 = No, X4 = Yes, and X5 = Yes, the probability that X1 = Yes (a fire occurs) is 10%.
[0033] The probabilities presented here are those calculated by the fire probability calculation unit 16 in Figure 1 using a simulator for each individual case, and the calculation results are reflected in the probability D33 of the Bayesian network probability storage unit DB3. The calculation of probabilities will be explained in detail separately.
[0034] Unlike the cause-variable table TB31, the cause-variable table TB32 is a variable that does not have a causal variable, such as X2: Fire detection device in Figure 5. Therefore, the cause-variable table TB32 stores the value of variable D34 and the probability D35 that it occurs.
[0035] The fire probability calculation for each event is performed by the fire probability calculation unit 16 in Figure 1, and a usage search unit 17 is provided to extract an appropriate one from a set of pre-stored specifications and verify its contents.
[0036] The calculation of the probability of fire occurrence for each event will be explained in detail below, but the calculated probability will be displayed as the fire occurrence probability D21f on the screen in Figure 2, along with the personal item D21 and item D23. In addition, the screen in Figure 2 should also display the fire occurrence probability D21f, the upper limit of the probability D21g, and alternative specification proposals 825 in case the upper limit is exceeded.
[0037] Figure 7 shows a screen for sequentially comparing multiple specification proposals when they are stored in the circuit specification storage unit DB2. For example, the usage search unit 17 in Figure 1 displays the specification proposal screen in Figure 7, and the specification proposal selection buttons (previous specification proposal selection 905, next specification proposal selection 910) enable sequential display or comparative display of multiple proposals.
[0038] The design support system 100 shown in Figure 1 executes the processing flows shown in Figures 8, 9, and 10 during its use. First, Figure 8 shows the general processing flow of the design support system 100.
[0039] Figure 8 shows the main routine, illustrating the processing flow from start to finish of the present invention. Subroutines SUB01 and SUB02 are called in intermediate processing steps S505 and S504, but the detailed processing of these subroutines SUB will be explained in Figures 9 and 10, and only the main routine will be explained here for now.
[0040] In the main routine's processing step S502 shown in Figure 8, the user is prompted to input information from the screen 40 in Figure 2, and the input content is recorded in the circuit specification storage unit DB2. In Figure 2, specification item D21 displays the names D21a-D21e, and the user is prompted to input the item value D23. After the user inputs the item value D23, they press the input button 802, which is part of the input unit 30, and the input content is recorded in the circuit specification storage unit DB2. Multiple items can be entered, and these are recorded as an array, as shown in the contents of the circuit specification storage unit DB2. The objective of this invention is to search for candidates that can reduce the probability of fire occurrence from the design specifications entered here.
[0041] In processing step S505, the probability distribution of the heat generated by the battery cells, which is the main cause of fire, is calculated. This probability distribution is calculated for each combination of design specifications entered in processing step S502, for example, 10W of heat generation is 10%, 20% of heat generation is 5%, and so on, and is added and stored as D21f in the bottom row of the lower section of Figure 3.
[0042] In processing step S520, using the heat generation probability from processing step S505, the contents of the circuit specification storage unit DB2, and the contents of the Bayesian structure storage unit DB4, a Bayesian network algorithm is used to calculate the probability of fire occurrence in the battery system P(X1=yes|X9, N, M, C1~C) N R1~R N The fire probability P is calculated when a battery system is configured based on the combination of design specifications. This process is performed by the fire probability calculation unit 16. Since this algorithm is publicly known, the details are omitted. This fire probability corresponds to the fire probability entered in processing step S502 and is displayed on the screen in Figure 2 as D21f.
[0043] Processing step S530 determines whether the calculated fire occurrence probability is at an acceptable level. When the user inputs one of the design specification proposals into the design support system 100 in processing step S502, the design support system 100 displays the occurrence probability on the screen in Figure 2 as shown in D21f. At this time, in addition to simply displaying the probability, a cell D21g is provided for entering an upper limit to make the determination easier. The user enters an upper limit for the probability into D21g, and if it exceeds the limit, a button 825 is displayed to suggest an alternative specification proposal. For clarity, button 825 displays the message "The upper limit has been exceeded." Pressing button 825 moves to processing step S540. If the upper limit is not exceeded in the first place, this main routine terminates.
[0044] In processing step S540, if the probability of occurrence presented in processing step S530 exceeds the upper limit, the system searches for and presents a specification proposal that can lower the probability of fire occurrence. This process is performed by the usage search unit 17 in Figure 1. The screen being presented is shown in Figure 7. The specification proposal is displayed in specification item D21 and item value D23, and the reduced probability of fire occurrence in that case is displayed in D21f. The probability upper limit value D21g is the same value as the value entered in the probability upper limit value D21g in Figure 2. The contents of this subroutine SUB02 will be described later.
[0045] Figure 9 illustrates the subroutine SUB01 in the main routine shown in Figure 8. Subroutine SUB01 details the processing step S505 and is a process for calculating the heat generated by the battery cell using the battery circuit simulator 11 and the battery degradation simulator 12 alternately.
[0046] Processing steps S605 and S610 in Figure 9 are processes that generate all combinations of item values D23 where item type D22 is "constant" from the circuit specification storage unit DB2. A combination of item values D23 is, for example, (circuit connection pattern = series N, number of series battery cells N = 100). If there are 3 types of circuit connection patterns and 10 types of cell count N, then 3 x 10 = 30 combinations will be generated. The next loop from processing steps S615 to S660 is executed for each of these combinations, and the heat generation amount Q1 to Q N We estimate this.
[0047] In processing step S615, one combination of item values D23 to be calculated is selected. This is a process to manage the loop processing.
[0048] Processing steps S620 to S645 are processes for estimating the battery cell resistance value D21d and battery cell capacity D21e, which inevitably contain errors because item type D22 is a "random variable". First, in processing step S625, the battery cell resistance R1 to R N and battery cell capacity C1~C NOnly one set is sampled. Sampling is performed according to a probability distribution linked to the battery cell resistance and battery cell capacity of item value D23, which in this example follows a truncated normal distribution.
[0049] In processing step S630, the sampled battery cell resistances R1~R N and battery cell capacity C1~C N The battery circuit simulator 11 is operated using this to obtain cell voltage values V1~V N and current values I1~I N The following is calculated. Then, in processing step S630, the cell voltage values V1~V N and current values I1~I N The battery degradation simulator 12 is run using the input, and the cell resistance values R1~R N and cell capacity C1~C N The following is calculated. Processing steps S630 and S635 are repeated the number of charge-discharge cycles. The number of charge-discharge cycles is determined by the number of charge-discharge cycles and the period until the battery life is reached, as specified by the battery manufacturer.
[0050] By repeating the process a fixed number of times, the cell resistance values R1~R at the end of the battery's lifespan are determined. N and cell capacity C1~C N This can be calculated. This calculation process is looped in processing step S645 and repeated many times. The number of repetitions should preferably be until the histogram of the sampling result is sufficiently close to the original truncated normal distribution, but since the computation time increases with the number of repetitions, it is also acceptable to repeat it only up to a number of times that is acceptable in terms of computation time.
[0051] In processing step S655, the cell resistance values R1~R at the end of the battery life are determined. N and cell capacity C1~C N The amount of heat generated by the battery cell is calculated from this. As mentioned above, this can be done by multiplying the resistance value by the square of the current. This allows us to calculate the amount of heat generated Q1~Q for each combination extracted in processing step S615. NThe set is calculated for the number of samples. The histogram of this set itself becomes the probability distribution calculated in processing step S655. This histogram of heat generation is added as an array, D21f, to the last row of the bottom row of Figure 2. The table after the addition is the table in the bottom row of Figure 2.
[0052] The above processing steps S615 to S660 are repeated for the number of combinations, generating heat Q1 to Q N The conditional probability is estimated. This completes the termination of this subroutine SUB01.
[0053] Figure 10 illustrates the subroutine SUB02 within the main routine shown in Figure 8. Subroutine SUB02 details the processing step S540, and searches for a specification that does not exceed the upper limit of the fire probability.
[0054] First, in processing step S705, the probability of fire occurrence P(X1=yes|X9, N, M, C1~C) for all combinations of item values of the specification draft created in processing step S615 is calculated. N R1~R N We estimate ).
[0055] In processing step S710, a specification proposal is selected from the estimated probabilities above that are less than or equal to the upper limit of the probability entered in D21g in Figure 2, and presented in specification item D21 and item value D23 in Figure 7. If multiple specification proposals are less than or equal to the upper limit, links to buttons 905 and 910 are presented to move to the previous or next specification proposal, and the user can switch the displayed specification proposal by pressing them. This concludes the subroutine SUB02.
[0056] In essence, the first example described above can be summarized as follows: First, there is little data available on the operation of a system with a large number of connected batteries. Therefore, simulations based on electrical circuits and physical models are used until one of the batteries ignites, and the effects after ignition are based on FTAs (Free Trade Analysis) derived from actual data and experience. A Bayesian network is then constructed that integrates both. This Bayesian network is used to search for design specifications that reduce the probability of fire to an acceptable level.
[0057] A Bayesian network is a method for estimating probabilities used in causal analysis, etc., and involves multiple causes X1~X n The probability of accident Y occurring is P(Y|X1~X) n ) can be estimated. However, this Bayesian network requires the following probability parameters. These are estimated from both simulation and FTA. These are the probability of occurrence of causes such as P(X1) and P(X2), and causes X1~X such as P(Y|X1). n The conditional probability of Y occurring is given by P(X1|X2, X3), and if the causes are related to each other, it is given by the conditional probability of the causes as P(X1|X2, X3). [Examples]
[0058] Example 1 showed an example of a design support system applied to a battery system, but the present invention can also be applied as a design support device for general equipment. In this case, the present invention is preferably provided with the following functions.
[0059] First, in Figure 1, the input unit 30 is used to input the design specifications of the target equipment.
[0060] Furthermore, in Figure 1, the component consisting of the circuit specification storage unit DB2, the battery degradation simulator 12, the battery circuit simulator 11, and the heat quantity calculation unit 13 can be described as an event identification unit that identifies events that cause accidents occurring in the equipment targeted by the design specifications. In this case, the accident is a fire, and the event causing the accident corresponds to the design specifications.
[0061] Furthermore, in Figure 1, the component consisting of the Bayesian network generation unit 15, the Bayesian network probability storage unit DB3, and the Bayesian structure storage unit DB4 can be described as a causal relationship storage unit that stores information on the causal relationship between accidents and their causes in the target equipment.
[0062] Furthermore, the processing in processing steps S505 (Figure 9) and S520 in Figure 8 can be described as a causal relationship identification unit that identifies the causal relationship information from the accident related to the target equipment to the event that caused the accident, using the event that caused the accident identified by the event identification unit and the causal relationship information stored in the causal relationship storage unit.
[0063] Furthermore, the processing in step S540 (Figure 10) in Figure 8, or the specification search unit 17 in Figure 1, can be described as a design specification calculation unit that uses the causal relationship information identified by the causal relationship identification unit to calculate the design specifications that minimize the probability of the aforementioned accident occurring in the battery system.
[0064] In Figure 1, the output unit 40 is an output unit that outputs the design specifications calculated by the design specification calculation unit.
[0065] Furthermore, in Figure 1, the fire probability calculation unit 16 calculates the probability of the design specifications causing the accident and identifies the causal relationship information. [Explanation of Symbols]
[0066] 10: Arithmetic section 11: Battery Circuit Simulator 12: Battery Degradation Simulator 13: Calorie calculation part 15: Bayesian network generation unit 16: Fire Probability Calculation Unit 17: Specification Search Unit 20: Bus 30: Input section 40: Display section 100: Design support system DB: Storage section DB1: Temporary storage DB2: Circuit specification storage unit DB3: Bayesian Network Probability Memory Unit DB4: Bayesian Structure Memory Unit D21: Specification Item D21a: Circuit connection pattern D21b: Number of serial cells N D21c: Number of parallel cells M D21d: Cell capacity D21e: Cell Resistor D22: Item type D23: Item value
Claims
1. A design support device comprising: an input unit for inputting the design specifications of a battery system, the target equipment for design support being a battery system combining multiple unit cells; an event identification unit for identifying events that cause accidents occurring in the target equipment; a causal relationship storage unit for storing causal relationship information between accidents and their causes in the target equipment; a causal relationship identification unit for identifying causal relationship information from an accident to the event causing the accident, using the event that causes the accident identified by the event identification unit and the causal relationship information stored in the causal relationship storage unit; a design specification calculation unit for calculating the design specifications that minimize the probability of the accident occurring in the target equipment, using the causal relationship information identified by the causal relationship identification unit; and an output unit for outputting the design specifications calculated by the design specification calculation unit, wherein the causal relationship identification unit calculates the probability of the design specifications causing the accident and identifies the causal relationship information.
2. A design support device according to claim 1, The design support device is characterized in that the event identification unit includes a simulator for identifying events that cause accidents occurring in the target equipment.
3. A design support device according to claim 1, The design support device is characterized in that the causal relationship storage unit includes causal relationship information between accidents and their causes in the target equipment as a Bayesian network.
4. A design support device according to claim 3, The Bayesian network is a design support device characterized by including a causal structure and accident probability.
5. A design support device according to claim 1, The aforementioned event identification unit is a design support device characterized by comprising a battery circuit simulator for the configuration of the battery system as defined in the design specifications and a battery degradation simulator for simulating battery degradation, as a simulator for identifying events that cause fires, which are accidents that occur in the battery system, in order to determine the amount of heat generated by the battery.
6. A design support device according to claim 1, The design support device is characterized in that the causal relationship storage unit is composed of a Bayesian network and comprises a Bayesian network structure storage unit relating to the structure of the causal relationship between accidents and their causes in the target equipment, and a Bayesian network probability storage unit relating to the probability of accident occurrence.
7. A design support device according to claim 1, The design support device is characterized in that the event identification unit is composed of simulations based on electrical circuits and physical models, and the causal relationship storage unit is composed of a Bayesian network based on FTAs that are based on actual data and experience.
8. A design support method in a design support device comprising an input unit, an event identification unit, a causal relationship storage unit, a causal relationship identification unit, a design specification calculation unit, and an output unit, The input unit receives the design specifications of a battery system in which the equipment to be designed is a battery system made up of multiple unit cells, and The aforementioned event identification unit identifies the events that cause accidents occurring in the equipment subject to the design specifications, The aforementioned causal relationship storage unit stores information on the causal relationship between an accident and its cause in the target equipment. The causal relationship identification unit uses the event that causes the accident identified by the event identification unit and the causal relationship information stored in the causal relationship storage unit to calculate the probability of the design specification causing the accident, and identifies the causal relationship information from the accident related to the target equipment to the event that caused the accident. The design specification calculation unit uses the causal relationship information identified by the causal relationship identification unit to calculate the design specifications that minimize the probability of the accident occurring in the target equipment. The output unit outputs the design specifications calculated by the design specification calculation unit. A design support method characterized by the following features.