JPH01278865A - Trouble diagnoser for vehicle - Google Patents

Abstract

PURPOSE:To make a portion subject to a trouble cause narrowable down at fairly good accuracy by retrieving each of first and second trouble information on the basis of trouble symptoms inputted, while inferring damaged parts and a parts aggregate and the trouble probability, and installing a means for informing this inferred result. CONSTITUTION:On the basis of trouble symptoms inputted by an inputting means 1, first trouble information stored in a first memory means 5 is retrieved and parts being caused to the trouble symptoms and the trouble probability are inferred by an inferring means 7. In addition, second trouble information stored in a second memory means 6 is retrieved and a parts aggregate made up of gathering specified parts themselves and the probability are inferred by the inferring means 7. This inferred result is informed to a machinist by an informing means 8. In consequence, a user narrows the trouble part down at fairly good accuracy and thus he can find it out.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Field of Application) The present invention relates to a fault diagnosis system for 4J cars that is capable of determining the cause of a failure, that is, the location of a failure, based on the failure symptoms of a 1U car.

(Prior Art) Modern vehicles tend to have increasingly complex structures. For this reason, when both the jp and the iF are in poor condition, that is, when a failure symptom occurs in the iF, it tends to be difficult to identify the part that is causing the failure symptom, even for a fairly experienced chiropractor. is getting stronger. This plays a major role in quickly repairing breakdowns.

For this reason, recently, as shown in Japanese Unexamined Patent Publication No. 62-6856, a system has been proposed that uses what is called an expert system to identify the cause of a failure based on its symptoms. ing. This is done by storing the causal relationships between failure symptoms and component failures that cause them as a failure tree, and by inputting the failure symptoms one by one, the system finally searches for the failed part and then searches for this. is displayed. By also storing the probability of a causal relationship between a failure symptom and a component failure that causes it, it is possible to search for a malfunctioning component with higher accuracy.

(Problems to be Solved by the Invention) However, with the conventional light rain failure diagnosis device described in 1, not only does it take a great deal of time to finally locate the malfunctioning component, but also the fault diagnosis system that requires five inputs. If quality and clarity are not evenly matched, an extremely large number of parts with the same probability of failure will be displayed.7 This will greatly reduce the initial value of this type of device. .

Therefore, the object of the present invention is to
An object of the present invention is to provide a heavy-duty failure diagnosis device η that can more accurately provide information to a person who maintains a part that causes a failure even when the failure is not possible.

(r=stage 1 action for solving the problem) In order to achieve the above-mentioned object, the present invention has the following configuration. That is, as shown in Fig. 1, failures caused by parts and failures of the parts 1. A first storage stage that stores the causal relationship with the F condition as the first failure information, a parts assembly made up of a specific quantity of parts, and the cause of the failure of the parts assembly. a second storage stage for storing a causal relationship with a failure symptom that occurs as second failure information together with its probability; and an input stage for inputting the failure symptom of Oyama.

Based on the failure symptoms input by the person card stage, the first failure information is searched to infer the failure parts and their probabilities, and the second failure information is searched to determine the failure parts. An inference means for inferring aggregates and their probabilities; and a Hachicho stage for notifying the inference results inferred by the first inference stage.

This is a configuration with the following.

In the present invention configured in this way, the part causing the failure symptom is not only notified as a middle part, but also notified when a specific part is gathered together, such as a failure in the ignition system or a failure in the fuel system. It will also be notified at the IN position of the parts assembly. In each of these notifications, the probability of failure is also notified, so you can use your own knowledge of what to do during maintenance and ultimately know which parts have a high probability of failing. 3, especially the quality of failure symptoms: c
Even if 1 is not 1 minute, it will be notified with a fairly high probability for a collection of 5 parts, so even if the part is about 1i%, it will be notified.
Even if the number of failures is large, by comparing and examining the two, maintenance personnel will be able to narrow down the failure part to a fairly detailed range.

(Embodiments) Hereinafter, embodiments of the present invention will be explained based on the attached drawings (+i).

Figure 1 is a block diagram showing the entire system of the present invention, and this system is constructed using a computer. In this Figure 1, the first stage of failure symptom input is
1 (< by four blocks 1, 2, 3, 4).

That is, the first stage of failure symptom input includes failure symptom input section 1,
No.:S+iL! It is composed of a storage section 2, a question display section 3, and a one-input emotion storage section 4. The failure symptom input unit 1 is used to input primary failure symptoms for machines using this system.The third storage unit 2 is used to input the components that cause the failure symptoms The 6-question display section 3, which is stored together with the probabilities of causal relationships and is stored in the form of a fault tree as described in the official uniform of IFJ, is stored in the third storage section of -. A question limited to the range of the stored fault tree is displayed to the operator. The input information storage section 4 stores input information, and this stored information is used as input information of failure symptoms for parts constituting the present invention, which will be described later.

In the case of utilizing the present invention, each of the elements 1, 2, 3, and 4 as described above is configured in order to limit the B+ symptoms to some extent.

In FIG. 1, 5 is a first storage section, 6 is a second storage section, 7 is an inference section, and 8 is an inference result notification section. The first storage unit 5 is a part-sized medium, and stores the causal relationship between this part and the failure symptoms caused by the failure of this part, together with its probability, as the first failure information. This is done in the form of an inference section, which will be described later. The second storage unit 6 is a medium-sized part assembly made up of a plurality of specific parts, and the second storage unit 6 stores the causal relationship between this parts assembly and a failure symptom caused by a failure of this part assembly, along with its probability. This is stored as failure information. The storage in this second storage section 6 is also performed in the form of an inference section, which will be described later. For example, such a parts assembly is referred to as an ignition system, which is an assembly of parts that play an important role in ignition, such as a spark plug, a destroyer, an igniter, and a ignition system. The inference unit 7 selects 1-memory 1 and each of the second storage units 5 according to the information stored in the input information storage unit 4.
．． 6 to infer the parts that are likely to be faulty and their probabilities, and 1 to search the second storage section to infer a collection of parts that are likely to be faulty along with their probabilities; , this inference result is notified to the user by means of the equipment 8. Note that the above-mentioned failure symptom input section 1 is constituted by, for example, a keyboard. For each storage section 2, 5, and 6, an external storage device such as a floppy disk or a hard disk is used since a large storage capacity is required. Input information storage section 4
It is also possible to use the same storage unit as each storage unit 2.5.6, but it may also be a small internal storage device of the storage unit-I. For example, a CRT is used for the question display section 3 and the inference result display section 8, and this 1t! ! Of course, the inference section 7 can also be used with a printer.
, 2. 3. 4) 1st nL! The stored contents of storage unit I are in the form of a fault tree as shown in FIG. 2, if shown schematically. This is based on 1, for example, not cranking as a failure symptom, and the causes become more specific as you move toward F, making a causal relationship. It is now possible to reach . Then, the degree of causality of the tree-to-subordinate relationship is expressed by its probability. FIG. 3 schematically shows what is shown in FIG. 2. Specifically, such a fault tree is stored as a large number of rules as shown in FIG. In other words, X is a failure symptom (or a cause if you go to the bottom of the tree), Y is a cause of confidence in the relationship). As shown in FIG. 5, the reliability of this failure is, for example, 1. OVl, MIDDLE, III
A roughly three-level classification such as GII may be used together, or only this rough classification may be used.

FIG. 12 is a flowchart showing the procedure until information is stored in the failure information storage 1164 in FIG. 1, and FIG. 12 will be described below. First, P(
Step) In step 1, import the first rule (the rule with 8 in Figure 3 as the highest position). Then in [)2,
Look at the X in the rule and search to see if there is a rule with this X in the pond. , and among the rules that satisfy X, the rule with the highest confidence level N is selected. .

1) After 2, in P; 3, it is determined whether Y in the selected rule matches the failure symptom or not. If they match, in P4, the current X, Y, The relationship of N is stored in the input storage section 4.

Then, in P5, after adding Y to X (Y is updated as X in the rule), it is determined in P6 whether there is a rule that causes X. If the determination in [)6 is NO, the process is immediately terminated. Further, if the determination at P6 is YES, the processes from P2 onwards are repeated.

If the determination in [)3 is NO, a rule with a lower confidence level N is selected in l)7, and then it is determined in P8 whether a similar rule still exists. Then, if there are other rules, return to 1-)3,
If not, at P9 the level is lower than the current one (
The rules at the bottom of the fault tree are stored in the input storage section 4. Note that when reaching 1]9, it is the case that the part that will ultimately cause the failure symptom cannot be reached to the level of the top part of the failure tree.

When storing information in the storage unit 4 using the fault tree described above, it is necessary to summarize the question items along this fault tree to some extent.
Creating a sheet and inputting it will save you a lot of effort. However, in this case, please limit the questions to areas that can be easily determined, such as first inspection and basic inspection.

,+,・J2;''・L5.6 First, if we classify the structure of heavy vehicles from large concepts to small concepts, we get the result shown in Figure 6. That is, heavy vehicles are classified into major concepts (system level) such as engines and missions, depending on the person. In addition, each system level includes, for example, an ignition system and a fuel system for an engine.
It is classified as a middle concept (subsystem level). and,
Things at each subsystem level are classified into individual component levels, such as fuel pumps, injectors, etc. in the fuel system, for example.

The inference part is 1-1 ignition system from the relationship as described above.

Two types of systems are set: a system corresponding to a subsystem such as a fuel system, that is, an assembly of specific parts, and a system at the component level.

FIG. 7 shows an example of the inference section regarding a parts assembly, and FIG. 7 shows an ignition system as an example.

In other words, when the ignition system malfunctions, the failure symptom is associated with the broad concept of poor starting, and the next sub-concept (failure symptom or cause of failure) is that the spark plug is wet. It is associated that the igniter is generating a false signal. In this way, the inference unit calculates the failure symptoms caused by the failure of the system as a collection of parts by -L
From the top down, the concepts are related in order from large concepts to more specific concepts. Of course, the causal relationship under L in this inference section is given a prediction as a failure probability.

Figure 8 schematically shows the inference part as described in 1-1. Specifically, such an inference part uses rules as shown in Figure 9, just as in the case of fault trees. is remembered as. In other words, if we focus on rule l,
B1 at the F level for A+ in FIG. +32 is stored along with its confidence level and confidence factor.

Figure 10 shows the inference section for each part, and the 9th point is that the individual parts come first instead of the part assembly.
It is different from the one shown in the figure (Fig. 7). As shown in FIG. 11, the inference section for each component is also stored as a rule along with its confidence level and degree of confidence. Note that in the case of a reasoning section for one component, there are many cases where the same level of people in the reasoning section are correlated and exhibit one failure symptom, and in this case, a separate rule is formed as AND information ( (See Rule 3 in Figure 11)', 7. [,,
8 The inference unit 7 uses the above-mentioned inference unit for parts assemblies and the inference part (rules) for each part to determine which parts assemblies are likely to be at fault, depending on the input failure symptoms. and parts, along with their failure probabilities. In other words, if there is a rule that stores contents that match the input failure symptoms, the failure probability of the part assembly or part will fail based on the confidence level and confidence level stored in this existing rule. infer that it is.

Then, the inferred result is finally determined according to the correlation between one part assembly and the parts, the confidence level, and the degree of confidence.
An appropriate selection is made to obtain the final inference result. Of course, this final inference result will be notified to the inference notification section 8.

An example of the final inference result by the inference unit 7 is as follows.

■About the parts assembly Control system (IIIGI O, 8) Fuel system
(HIGII O, 3) ■About parts Water temperature sensor (tl I Gl (0,5) Control unit (l (I Gll 0.3) Injector
(MIDDL, E O18) [:, GR valve (+
-OW 0.7). Looking at these results, the maintenance person suspects that the water temperature sensor, control unit, and injector are malfunctioning based on the correlation between the parts and the parts, and the confidence level and confidence level. On the other hand, the EGR valve is considered to be excluded from suspicion of malfunction.

FIGS. 13 and 14 are flowcharts showing the sequence of steps for obtaining an inference result using the above-mentioned inference book. As mentioned above, the 1 input information storage unit 4 stores failure symptoms that have been input so far and parts that are considered to be at fault according to these, within a somewhat limited range. This means that it is only necessary to perform the calculation based on the information stored in the storage section 4.

First, at Ql in FIG. 13, the process described in box 1 and shown in FIG. 12 is performed.

Next, Q2. In Q3, it is inferred whether the parts assembly is at fault or not, along with its failure probability, according to the failure symptoms (search of the second +iL! storage unit 6). Also, Q4. Q
In step 5, it is inferred whether or not one part t11 has failed or not, together with its failure probability. In addition, L;
L! Regarding Q2, Q3, Q4, and Q5, after the inference of one of them is completed, the inference of the other is performed.

In Q6, based on the two types of inference results for one part assembly and the part, those that are less likely to be faulty are removed according to the comprehensive inference, that is, the correlation between the two, the confidence level, and the confidence level, and this process is performed. The final inference result after iii is Ql
The notification is notified to the notification section 8 at .

The details of Q2 (Q4) mentioned above are as shown in FIG. First, [? In step 1, a parts collection or a reasoning book corresponding to the parts is selected (selection in the first storage unit 5 or the second storage unit 6). With me.

1 (In step 2, select the rule located closest to -L.

At step 113, it is determined whether the information stored in the selected rule matches the input failure symptom. When the judgment in R3 is YES, in R4, the part assembly (parts/components) currently being inferred is stored together with the confidence level and confidence in the rule for which YES was determined in R3. After this, in R5, the rule is updated to the next lower level rule in the inference book, and the processing for 1 (3 or less) is repeated.

When the judgment in R3 is NO, another rule having the same level of hierarchy in the inference tree is selected in R6, and then it is judged in R7 whether such a rule exists. If the judgment of R7 is YES, return to R3,
Conversely, if NO, the search is terminated (search terminated).

(Effects of the Invention) As is clear from the above description, the present invention notifies parts assemblies and parts that are suspected of having failed, along with their respective failure probabilities. I) Even in cases where it is not sufficient, the knowledge of the person in charge of maintenance is used to identify suspected malfunctions based on the recorded correlation between the parts assembly and the parts and their failure probabilities. It is possible to narrow down a certain part to a considerable degree of 1+1 degrees, and it is extremely effective in E for specifying a faulty part.

[Brief explanation of the drawing]

FIG. 1 is an overall system diagram showing an embodiment of the present invention. FIG. 2 is a diagram showing an example of a fault tree. FIG. 3 is a diagram schematically showing FIG. 2. FIG. 4 is a diagram showing an example of storing the contents of FIGS. 2 and 3. FIG. 5 is a diagram showing an example of setting confidence levels and confidence degrees. FIG. 6 is a diagram showing how the structure of a heavy vehicle is classified from an E-level concept to a lower level concept. FIG. 7 is a diagram showing an example of an inference book about a parts assembly. FIG. 8 is a diagram schematically showing FIG. 7. FIG. 9 is a diagram showing an example in which the contents of FIGS. 7 and 8 are stored. FIG. 10 is a diagram schematically showing an example of an inference book about parts. FIG. 11 is a diagram showing an example of storing the contents of FIG. 10. 12 to 14 are flowcharts showing control examples of the present invention. Figure 4 Levelley Town Figure 7 Figure 8 Figure 9 Figure 12 Figure 13

Claims (1)

[Claims] (1) A first storage means that stores a causal relationship between a component and a failure symptom caused by a failure of the component together with its probability as first failure information; a second storage means for storing a causal relationship with a failure symptom caused by a failure of a parts assembly together with its probability as second failure information; an input means for inputting a failure symptom of the vehicle; and an input means for inputting a failure symptom of the vehicle; Based on the failure symptoms, the first failure information is searched to infer a failure component and its probability, and the second failure information is searched to infer a failure component assembly and its probability. A vehicle failure diagnosis device comprising: an inference unit; and a notification unit that notifies an inference result inferred by the inference unit.