Vehicle cloud cooperation-based vehicle real-time fault prediction method and device

By adopting the vehicle-cloud collaboration method in intelligent connected vehicles, using training set data and naive Bayes algorithm combined with fault tree theory, the prior and posterior probability of vehicle failure is calculated, real-time prediction of vehicle failure is achieved, and the solution is to solve the problem of Problems that are difficult to early warning in the prior art have improved vehicle safety.

CN118915702APending Publication Date: 2024-11-08Z-ONE TECH CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411112779.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing technology is difficult to achieve real-time vehicle failure prediction, especially in the context of the complexity of intelligent connected vehicles. There are shortcomings in fault prediction and diagnosis technology, and it is difficult to conduct fault warning in advance to improve vehicle safety.

Method used

Using a method based on vehicle-cloud collaboration, the prediction model is trained using training set data, combined with Naive Bayes algorithm and fault tree theory, the prior and posterior probability of atomic failures are calculated, and the probability of vehicle failures is predicted, and the prior calculation module is used. , data association module and prediction module realize real-time fault prediction.

Benefits of technology

It realizes early warning of vehicle failures, timely avoids risks, improves vehicle safety, provides strong data support, and provides users with a higher safe car experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118915702A_ABST
    Figure CN118915702A_ABST
Patent Text Reader

Abstract

The invention discloses a vehicle real-time fault prediction method and device based on vehicle cloud collaboration, and the method comprises the steps: training a prediction model through training set data, calculating the prior probability distribution of an atomic fault according to the prediction model, the training set data is sample data of vehicle signals and faults; acquiring signal data reported by a vehicle, determining the signal data associated with a bottom event corresponding to an atomic fault based on a fault tree, and determining target prior probability distribution corresponding to the signal data according to the bottom event corresponding to the atomic fault; and calculating a posterior probability of an atomic fault corresponding to occurrence of the bottom event based on a naive Bayes algorithm and the target prior probability distribution, predicting a probability of occurrence of the fault of the bottom event based on the posterior probability, and predicting a probability of occurrence of a vehicle fault according to the probability of occurrence of the fault of the bottom event. According to the invention, the problem of real-time fault prediction according to the fault signal is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of software development, and in particular to a vehicle real-time fault prediction method and device based on vehicle-cloud collaboration. Background Art

[0002] With the continuous development of intelligent connected vehicles, the complexity and intelligence of vehicles are constantly improving, and vehicle fault prediction and diagnosis technology is becoming more and more important. The real-time acquisition of vehicle signals, signal transmission to the cloud and real-time computing technology have been further developed, providing infrastructure conditions for the establishment of a real-time fault prediction system for vehicle-cloud collaboration; at the same time, the vigorous development of prediction theory and fault analysis theory based on statistical learning provides theoretical support for the establishment of a real-time fault prediction system for vehicle-cloud collaboration. Summary of the invention

[0003] In response to the above technical problems, the present invention provides a vehicle real-time fault prediction method and device based on vehicle-cloud collaboration, which can provide early fault warning.

[0004] In a first aspect of the present invention, a vehicle real-time fault prediction method based on vehicle-cloud collaboration is provided, comprising the following steps: training a prediction model using training set data, calculating a priori probability distribution of atomic faults according to the prediction model, wherein the atomic fault is a category of a fault signal, and the training set data is sample data of vehicle signals and faults; Acquire signal data reported by the vehicle, determine the signal data associated with a bottom event corresponding to the atomic fault based on a fault tree, and determine a target priori probability distribution corresponding to the signal data according to the bottom event corresponding to the atomic fault; Based on the Naive Bayes algorithm and the target prior probability distribution, the posterior probability of the atomic fault corresponding to the bottom event is calculated, the probability of the bottom event failure is predicted based on the posterior probability, and the probability of the vehicle failure is predicted based on the probability of the bottom event failure.

[0005] In an optional implementation manner, the method of training a prediction model using training set data and calculating a priori probability distribution of atomic failures according to the prediction model includes: Select atomic fault sets and signal sets. Each atomic fault set includes multiple sample data. Each atomic fault corresponds to multiple discrete signals, and the multiple signals are independent of each other. Calculate the prior probability distribution of atomic faults by the following formula: Marginal probability distribution of atomic faults in is the fault label f ​​in the sample dataset i Take the number of 0 or 1, |D| is the number of sample sets; Conditional probability distribution of atomic failures in is the fault f in the sample data set i When 0 or 1 respectively, the signal x j Take the number of a certain value, is the fault label f i Takes the number of 0s or 1s.

[0006] In an optional implementation manner, the using of training set data to train the prediction model, and calculating the prior probability distribution of atomic failures according to the prediction model, further includes: Correct the marginal probability distribution of atomic failures: Correct the conditional probability distribution of atomic failures: in is the fault f in the sample data set i When 0 or 1 respectively, the signal x j Take the number of a certain value, is the fault label f i Take the number of 0 or 1; |D| is the number of sample sets, n is x j The number of possible values.

[0007] In an optional implementation, the vehicle real-time fault prediction method based on vehicle-cloud collaboration further includes constructing a fault tree: Matching the bottom events of the fault tree to the atomic faults one by one, and assigning the prior probability distribution of the atomic faults to the bottom events; A tree structure relationship between intermediate events and top events is constructed based on the fault rules and the bottom events, wherein the intermediate events are necessary and sufficient conditions for the top events.

[0008] In an optional implementation manner, the calculating the posterior probability of the atomic fault corresponding to the occurrence of the bottom event based on the naive Bayes algorithm and the target prior probability distribution includes: Calculating the probability of occurrence of atomic faults and the probability of non-occurrence of atomic faults based on the naive Bayes algorithm and the target prior probability distribution; The probability of atomic failure occurring and the probability of atomic failure not occurring are normalized to obtain the posterior probability of atomic failure.

[0009] In an optional implementation, the predicting the probability of the bottom event failure based on the posterior probability, and predicting the probability of the vehicle failure according to the probability of the bottom event failure, includes: The probability of failure of the intermediate event is calculated based on the probability of failure of the bottom event and the failure rule, the probability of failure of the top event is calculated based on the probability of failure of the intermediate event and the failure rule, and then the probability of vehicle failure is predicted based on the probability of failure of the top event.

[0010] In an optional embodiment, the signal data associated with the bottom event corresponding to the atomic fault is determined based on a fault tree, and the signal data corresponding to the bottom event corresponding to the atomic fault is determined to have a target prior probability distribution, including: searching for the atomic fault corresponding to the bottom event according to the fault tree, searching for the corresponding signal data according to the atomic fault, and calculating the conditional probability of the signal data when the atomic fault is known, and the conditional probability is the target prior probability.

[0011] A second aspect of the present invention provides a vehicle real-time fault prediction device based on vehicle-cloud collaboration, comprising: A priori calculation module is used to train a prediction model using training set data, and calculate the priori probability distribution of atomic faults according to the prediction model, wherein the atomic fault is a category of fault signal, and the training set data is sample data of vehicle signals and faults; a data association module is used to obtain signal data reported by the vehicle, determine the signal data associated with the bottom event corresponding to the atomic fault based on the fault tree, and determine that the signal data corresponds to a target priori probability distribution according to the bottom event corresponding to the atomic fault; A prediction module is used to calculate the posterior probability of the atomic fault corresponding to the occurrence of the base event based on the naive Bayes algorithm and the target prior probability distribution, predict the probability of the base event failure based on the posterior probability, and predict the probability of vehicle failure based on the probability of the base event failure.

[0012] According to a third aspect of the present invention, there is provided an electronic device, comprising: At least one processor; and at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the vehicle real-time fault prediction method based on vehicle-cloud collaboration as described in the first aspect of an embodiment of the present invention.

[0013] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a computer, the vehicle real-time fault prediction method based on vehicle-cloud collaboration as described in the first aspect of an embodiment of the present invention is executed.

[0014] The present invention constructs an atomic fault prediction model based on the naive Bayesian theory and integrates the fault tree analysis theory into the model, solving the problem of real-time fault prediction based on fault signals. The present invention can provide users and vehicle manufacturers with early fault warnings and timely avoid high-risk faults, thereby avoiding accidents, providing strong data support and providing users with a safer car experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 A flowchart of a vehicle real-time fault prediction method based on vehicle-cloud collaboration in an embodiment of the present invention; Figure 2 This is an architecture diagram of a vehicle real-time fault prediction method based on vehicle-cloud collaboration in an embodiment of the present invention; Figure 3 is a schematic diagram of a fault tree in an embodiment of the present invention; Figure 4 This is a module schematic diagram of a vehicle real-time fault prediction device based on vehicle-cloud collaboration in an embodiment of the present invention; Figure 5 FIG. 1 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0018] It should be understood that the terms "first", "second", "third", etc. in the claims, specifications and drawings of the present invention are used to distinguish different objects rather than to describe a specific order. The terms "include" and "comprise" used in the specification and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0019] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the claims, the singular forms of "a", "an" and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in the specification of the present invention and the claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0020] See also Figure 1 The vehicle real-time fault prediction method based on vehicle-cloud collaboration provided by the present invention includes the following steps.

[0021] S100: training a prediction model using training set data, and calculating a priori probability distribution of atomic faults according to the prediction model, wherein the atomic faults are categories of fault signals, and the training set data are sample data of vehicle signals and faults.

[0022] In some embodiments, a large amount of sample data of vehicle signals and faults is collected as training set data, and continuous signal values ​​are discretized according to signal value features to set hyperparameters for model training.

[0023] Select an atomic fault set F = {f1, f2, ..., f m}, for any atomic fault f i , each has a training sample data set That is, for any atomic fault f i , there are n samples; where the signal set: X = {x1, x2, ..., x d}, that is, for any atomic fault f i , d signals are selected and it is assumed that the d signals are independent of each other; the class variable set Y = {0, 1}, where 0 represents that the fault has not occurred and 1 represents that the fault has occurred.

[0024] For a signal set X = {x1, x2, ..., x d}, according to the value characteristics of the signal, use data binning methods to discretize it, such as equal-width binning, equal-frequency binning, chi-square binning and other methods.

[0025] Learning atomic faults f based on sample data statistics i The marginal probability distribution of atomic failures f i The prediction model is trained based on the conditional probability distribution of . For any known atomic fault, the prior probability of the atomic fault can be calculated based on the prediction model.

[0026] S200: Acquire signal data reported by a vehicle, determine the signal data associated with a bottom event corresponding to an atomic fault based on a fault tree, and determine a target priori probability distribution corresponding to the signal data according to the bottom event corresponding to the atomic fault.

[0027] The fault tree constructed by known fault analysis experience or expert experience can be used to obtain the target prior probability distribution by mapping the probability of atomic fault occurrence to the probability of occurrence of the bottom event of the fault tree through qualitative analysis of the fault tree. For example, if the signal data corresponds to the bottom event of brake failure (such as brake pad aging), the prior probability can be obtained according to the mapping relationship.

[0028] S300: Calculate the posterior probability of the atomic fault corresponding to the occurrence of the base event based on the naive Bayes algorithm and the target prior probability distribution, predict the probability of the base event failure based on the posterior probability, and predict the probability of vehicle failure based on the probability of the base event failure.

[0029] The statistical Naive Bayesian theory ensures the robustness of the system. The Naive Bayesian theory has strong mathematical interpretability and high tolerance to signal loss. Under the condition of a given atomic fault, the posterior probability of the atomic fault can be calculated based on the Naive Bayesian algorithm.

[0030] Specifically, calculate the atomic fault f i When calculating the prior probability of atomic faults, we select atomic fault sets and signal sets. Each atomic fault set includes multiple sample data. Each atomic fault corresponds to multiple discrete signals, and the multiple signals are independent of each other. The prior probability distribution of atomic faults is calculated by the following formula: Marginal probability distribution of atomic failures in is the fault label f ​​in the sample dataset i Take the number of 0 or 1, |D| is the number of sample sets; Conditional probability distribution of atomic failures in is the fault f in the sample data set i When 0 or 1 respectively, the signal x j Take the number of a certain value, is the fault label f i Takes the number of 0s or 1s.

[0031] Then, the posterior probability of atomic fault is calculated. Based on the above atomic fault prior probability distribution, the newly reported signal data X of the vehicle is used as the new data, x j For data signals, using the Bayesian formula The d signals are independent of each other. Under the condition of a given atomic fault,

[0032] Furthermore, if the data signal x j The value of has never appeared in the sample data set, or appears but P(x j ∣f i )=0, the probability of subsequent fault prediction is 0, which erases the information of other signals, which is obviously not in line with reality. Therefore, the present invention believes that this situation is caused by the lack of training samples, and thus makes corrections for this situation.

[0033] Correct the marginal probability distribution of atomic failures: Correct the conditional probability distribution of atomic failures: in is the fault f in the sample data set i When 0 or 1 respectively, the signal x j Take the number of a certain value, is the fault label f i Take the number of 0 or 1; |D| is the number of sample sets, n is x j The number of possible values.

[0034] The posterior probability distribution of atomic failure can be further obtained:

[0035] In some embodiments, the probability of occurrence of an atomic failure and the probability of non-occurrence of an atomic failure are first calculated based on the naive Bayes algorithm and the target prior probability distribution; then the probability of occurrence of an atomic failure and the probability of non-occurrence of an atomic failure are normalized to obtain a posterior probability of an atomic failure.

[0036] Since the new data sample X is fixed, the occurrence probability P(X) of the data X is a fixed value and is eliminated after normalization, so it can be ignored.

[0037] Atomic Fault i Occurrence (f i =1) is:

[0038] Atomic Fault i Not occurred (f i =0) is:

[0039] Then normalize it:

[0040] After normalization, p(f i =1) is the atomic fault predicted by the cloud model f i The probability of occurrence of atomic failure f i The probability of not happening p(f i =0)Same.

[0041] The present invention integrates prior expert experience knowledge and prior analysis experience into the fault tree, combines Bayesian theory with fault tree theory, and achieves the effect of enhancing the accuracy of fault prediction, which greatly facilitates the application of data in the cloud.

[0042] In some embodiments of the present invention, when constructing a fault tree, the bottom events of the fault tree are matched one-to-one with the atomic faults, and the prior probability distribution of the atomic faults is assigned to the bottom events. A tree structure relationship between intermediate events and top events is constructed based on the fault rules and the bottom events, wherein the intermediate events are necessary and sufficient conditions for the top events.

[0043] For example, based on the actual situation and expert experience, whether a fault occurs is matched to whether a bottom event occurs. Therefore, the bottom event set E = {e1, e2, ..., e m} and the atomic fault set F = {f1, f2, ..., f m}One to one correspondence.

[0044] If we only consider the occurrence of faults, that is, the occurrence of bottom events (we can also consider the occurrence of bottom events when faults do not occur), then the bottom event can be defined as:

[0045] Defining intermediate events That is, the intermediate event is an intermediate fault formed by the bottom event or other intermediate events that satisfies the fault rule θ, where the fault rule θ = {and, or}. Define the top event That is, the top event is a fault formed by the intermediate event or the bottom event that satisfies the fault rule θ.

[0046] For details, please refer to Figure 2 As shown, Figure 2 Schematic diagram of the vehicle real-time fault prediction method based on vehicle-cloud collaboration in the present invention. Multiple cars Car1, Car2, Car3, Car4... generate signals X1, X2, X2...X d , the signal corresponds to the atomic fault f i , atomic fault f i With the bottom event (e 1、 e2…e m ) phase mapping, the bottom event generates the intermediate event according to the AND or OR relationship of the fault rule θ Intermediate Events According to the AND or OR relationship of the fault rule θ, a top event is generated, and the top event is a specific vehicle fault.

[0047] Therefore, the three types of events, namely bottom events, intermediate events and top events, can be qualitatively analyzed based on the fault tree. Based on the expert experience and knowledge, an undesirable product failure event (for preventive safety analysis) or an already occurred failure event (for troubleshooting) is taken as the top event Φ. By strictly analyzing the fault rules θ from top to bottom, the necessary and sufficient conditional independent direct causes of the top event are found layer by layer as the intermediate events. And the fault rule θ; based on the intermediate event Similarly, follow the causal logic until the bottom event set E = {e1, e2, ..., e m} in the bottom event e i , and qualitatively analyze the atomic fault rule θ of the bottom event.

[0048] Of course, in other embodiments, two types of events, bottom events and top events, may be constructed based on the fault tree; in other embodiments, two types of events, bottom events and intermediate events, may also be constructed based on the fault tree.

[0049] Then, quantitative analysis can be performed to calculate the probability of occurrence of bottom events, intermediate events, and top events.

[0050] In some embodiments, the probability of failure of the intermediate event is calculated based on the probability of failure of the bottom event and the failure rule, the probability of failure of the top event is calculated based on the probability of failure of the intermediate event and the failure rule, and then the probability of vehicle failure is predicted based on the probability of failure of the top event.

[0051] Let atomic fault F = {f1, f2, ..., f m}, and the probability of the event occurring is mapped one by one to the probability of the event occurring P(E) = {p(e1), p(e2), ..., p(e m )}.

[0052] Furthermore, when focusing on the intermediate events, the bottom events are also independent of each other. By calculating the relationship with, the probability of the intermediate event is By calculating the OR relationship, the probability of the intermediate event is

[0053] Similarly, when focusing on the top event, the intermediate events are also independent of each other. Rules based on three types of events can be Figure 2 From the middle bottom to the upper part until the probability distribution P(Φ) of the top event is calculated.

[0054] The present invention will be further described below with reference to specific examples.

[0055] See also Figure 3 , Figure 3 The fault tree is constructed. The following 10 signals (Table 1) and fault records are used as training sample data, and the following fault tree is constructed based on expert knowledge. (Note: The more actual training samples there are, the closer it is to the actual situation, and the vehicle offline time and the total mileage of the vehicle, weather and humidity may have a certain correlation, but the present invention assumes that these signals are independent of each other).

[0056] Table 1 Serial number <![CDATA[x1]]> <![CDATA[x2]]> <![CDATA[x3]]> <![CDATA[x4]]> <![CDATA[x5]]> <![CDATA[x6]]> <![CDATA[f1]]> <![CDATA[f2]]> <![CDATA[f3]]> <![CDATA[f4]]> <![CDATA[f5]]> <![CDATA[f6]]> 1 4 12321 434 sunny 8 high 0 0 0 0 0 0| 2 2 28233 760 cloudy 22 low 0 0 0 0 0 0| 3 3 34333 1245 rainy 28 high 0 0 0 1 0 1 4 3 72986 1200 rainy 23 high 1 0 1 0 1 1 5 7 55655 412 rainy 12 middle 0 0 0 0 1 0 6 5 130234 1230 cloudy 9 middle 1 1 0 0 0 1 7 8 110223 904 sunny 15 low 0 0 0 1 0 0 8 11 122248 801 cloudy 8 high 1 0 1 0 1 0 9 10 112334 1102 sunny 18 high 1 1 0 1 0 1 10 7 94456 385 cloudy 27 high 0 0 1 0 0 0

[0057] Signal x includes x1: vehicle offline time, unit: year; x2: vehicle total mileage, unit: km; x3: vehicle load, unit: kg; x4: weather; x5: temperature, unit: °C; x6: humidity. Atomic fault f includes f1: whether the brake pad is aging; f2: whether the tire is seriously worn; f3: whether the brake oil level is too low; f4: whether the brake oil is deteriorated; f5: whether the road is slippery; f6: whether the vehicle is overloaded.

[0058] The above 10 signals are preprocessed, and the signals x1, x2, x3, and x5 are discretized using the equal-width binning method.

[0059] Right now

[0060] The training data after preprocessing is shown in Table 2.

[0061] Table 2 Serial number <![CDATA[x1]]> <![CDATA[x2]]> <![CDATA[x3]]> <![CDATA[x4]]> <![CDATA[x5]]> <![CDATA[x6]]> <![CDATA[f1]]> <![CDATA[f2]]> <![CDATA[f3]]> <![CDATA[f4]]> <![CDATA[f5]]> <![CDATA[f6]]> 1 1 1 1 sunny 1 high 0 0 0 0 0 0 2 1 1 2 cloudy 3 low 0 0 0 0 0 0 3 1 1 3 rainy 3 high 0 0 0 1 0 1 4 1 2 3 rainy 3 high 1 0 1 0 1 1 5 2 2 1 rainy 2 middle 0 0 0 0 1 0 6 2 3 3 cloudy 1 middle 1 1 0 0 0 1 7 2 3 2 sunny 2 low 0 0 0 1 0 0 8 3 3 2 cloudy 1 high 1 0 1 0 1 0 9 3 3 3 sunny 2 high 1 1 0 1 0 1 10 2 2 1 cloudy 3 high 0 0 1 0 0 0

[0062] Can be combined Figure 3 As shown, the prior probability of atomic failure is first calculated.

[0063] Depend on Compute the prior probability distribution of atomic fault f1.

[0064] Probability Probability of not happening

[0065] In some embodiments, the atomic fault corresponding to the bottom event is found according to the fault tree, the corresponding signal data is found according to the atomic fault, and the conditional probability of the signal data when the atomic fault is known is calculated, and the conditional probability is the target prior probability.

[0066] according to Figure 3 From the fault tree shown, we can see that the atomic fault f1 is only related to the signals x1, x2, and x3, and has nothing to do with other signals. Therefore, the conditional probability distribution Compute the conditional probability of signal x1 given the atomic fault f1.

[0067] Since the sample data in this example is small and some information is missing, the conditional probability is 0. To simplify the calculation, no probability correction is made for the conditional probability of 0 in this example.

[0068] Calculate the conditional probability of signal x1 when fault f1 is known:

[0069] Calculate the conditional probability of signal x2 when fault f1 is known:

[0070] Calculate the conditional probability of signal x3 when fault f1 is known:

[0071] Similarly, the conditional probability of each signal when the remaining atomic faults f2, f3, f4, f5, and f6 occur is statistically calculated, and then the posterior probability is calculated.

[0072] There is a set of signals X={x1=2, x2=3, x3=3, x4=cloudy, x5=1, x6=high}, i.e., the vehicle is 5-10 years old, the total mileage is between 100,000 and 150,000 kilometers, the load is between 1,000 and 1,500 kg, the weather is cloudy, the temperature is between 0 and 10°C, and the humidity is high, and the probability of brake failure is determined.

[0073] Depend on Calculate the posterior probability of atomic fault f1 occurring, Under the condition of signal X, the probability of atomic fault f1 occurring, namely brake pad aging, is 86.7%.

[0074] Similarly, the conditional probability of atomic fault f2 is calculated. According to the fault diagram, atomic fault f2 is only related to signals x1, x2, and x3, and has nothing to do with other signals. Therefore, the prior probability of atomic fault f2 is

[0075] The conditional probability of atomic fault f2 under the condition of signal X = {x1 = 2, x2 = 3, x3 = 3, x4 = cloudy, x5 = 1, x6 = high} is calculated:

[0076] Calculate the posterior probability of atomic fault f2 occurring: Under the condition of signal X, the probability of atomic fault f2 occurring, that is, severe tire wear, is 84.2%.

[0077] Similarly, the conditional probability of the occurrence of atomic fault f3 is calculated. According to the fault diagram, it can be seen that the atomic fault f3 is only related to signals x1 and x2, and has nothing to do with other signals.

[0078] Therefore, the prior probability of atomic fault f3 is:

[0079] Calculate the posterior probability of atomic fault f3 occurring: Under the condition of signal X, the probability that atomic fault f3 occurs, ie, the brake oil level is too low, is 20.2%.

[0080] Similarly, the conditional probability of atomic fault f4 occurring is calculated. According to the fault diagram, it can be seen that atomic fault f4 is only related to signals x1, x2, and x6, and has nothing to do with other signals.

[0081] Therefore, the prior probability of atomic fault f4 is:

[0082] The conditional probability of atomic fault f4 under the condition of signal X = {x1 = 2, x2 = 3, x3 = 3, x4 = cloudy, x5 = 1, x6 = high} is calculated:

[0083] Calculate the posterior probability of atomic fault f4 occurring: Under the condition of signal X, the probability of atomic fault f4 occurring, namely the deterioration of brake fluid, is 46.3%.

[0084] Similarly, the conditional probability of the atomic fault f5 is calculated. According to the fault diagram, the atomic fault f5 is only related to signals x4, x5, and x6, and has nothing to do with other signals.

[0085] Therefore, the prior probability of atomic fault f5 is:

[0086] The conditional probability of atomic fault f5 under the condition of signal X = {x1 = 2, x2 = 3, x3 = 3, x4 = cloudy, x5 = 1, x6 = high} is calculated:

[0087] Calculate the posterior probability of atomic fault f5 occurring: Under the condition of signal X, the probability of atomic fault f5 occurring, namely, the road surface being slippery, is 29.8%.

[0088] Similarly, the conditional probability of atomic fault f6 is calculated. According to the fault diagram, it can be seen that atomic fault f6 is only related to signal x3 and has nothing to do with other signals.

[0089] Therefore, the prior probability of atomic fault f6 is:

[0090] Calculate the posterior probability of atomic fault f6 occurring: Under the condition of signal X, the atomic fault f6 occurs, that is, the probability of vehicle overloading is 100%.

[0091] The probability of atomic failure is mapped to the probability of the bottom event, calculated according to the relationship formula:

[0092] The probability of the intermediate event occurring: p(e7)=1-(1-p(e1))*(1-p(e2))*(1-p(e3))*(1-p(e4)) =1-(1-0.867)*(1-0.842)*(1-0.202)*(1-0.463) =1-0.009005045364 ≈99.1%

[0093] Basis and relationship calculation formula: The probability of the intermediate event occurring: p(e8)=p(e5)*p(e6) =0.298*1 =0.298 =29.8%.

[0094] The probability of the top event occurring: p(e9)=1-(1-p(e7))*(1-p(e8)) =1-0.009*0.702 =1-0.006318 =0.993682≈99.4%.

[0095] That is, under a set of signals X={x1=2, x2=3, x3=3, x4=cloudy, x5=1, x6=high}, i.e., the vehicle is 5-10 years old, the total mileage is between 100,000 and 150,000 kilometers, the vehicle load is between 1,000 and 1,500 kg, the weather is cloudy, the temperature is between 0 and 10°C, and the humidity is high, the probability of brake failure is determined to be 99.4%.

[0096] See also Figure 4 The present invention also provides a vehicle real-time fault prediction device based on vehicle-cloud collaboration, including: a priori calculation module 41, a data association module 42, and a prediction module 43.

[0097] The prior calculation module 41 is used to train the prediction model using the training set data, and calculate the prior probability distribution of the atomic fault according to the prediction model. The atomic fault is the category of the fault signal, and the training set data is the sample data of the vehicle signal and the fault.

[0098] The data association module 42 is used to obtain signal data reported by the vehicle, determine the signal data associated with the bottom event corresponding to the atomic fault based on the fault tree, and determine that the signal data corresponds to a target prior probability distribution according to the bottom event corresponding to the atomic fault.

[0099] The prediction module 43 is used to calculate the posterior probability of the atomic fault corresponding to the occurrence of the base event based on the naive Bayes algorithm and the target prior probability distribution, predict the probability of the base event failure based on the posterior probability, and predict the probability of vehicle failure based on the probability of the base event failure.

[0100] For the description of the vehicle real-time fault prediction device based on vehicle-cloud collaboration, please refer to the description of the vehicle real-time fault prediction method based on vehicle-cloud collaboration, which will not be repeated here.

[0101] like Figure 5 As shown, the present invention also provides an electronic device, including: At least one processor; and at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the above-mentioned vehicle real-time fault prediction method based on vehicle-cloud collaboration.

[0102] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned vehicle real-time fault prediction method based on vehicle-cloud collaboration is implemented.

[0103] It can be understood that computer-readable storage media may include: any entity or device capable of carrying a computer program, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc. A computer program includes computer program code. The computer program code may be in source code form, object code form, executable file, or some intermediate form, etc. A computer-readable storage medium may include: any entity or device capable of carrying a computer program code, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc.

[0104] In certain embodiments of the present invention, the electronic device may include a controller or a processor, and the controller is a single-chip microcomputer chip that integrates a processor, a memory, a communication module, etc. The processor may refer to the processor contained in the controller. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.

[0105] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0106] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0107] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vehicle real-time fault prediction method based on vehicle-cloud collaboration, characterized in that: include: Using training set data to train a prediction model, and calculating a priori probability distribution of atomic faults according to the prediction model, wherein the atomic fault is a category of a fault signal, and the training set data is sample data of vehicle signals and faults; Acquire signal data reported by the vehicle, determine the signal data associated with a bottom event corresponding to the atomic fault based on a fault tree, and determine a target priori probability distribution corresponding to the signal data according to the bottom event corresponding to the atomic fault; Based on the Naive Bayes algorithm and the target prior probability distribution, the posterior probability of the atomic fault corresponding to the bottom event is calculated, the probability of the bottom event failure is predicted based on the posterior probability, and the probability of the vehicle failure is predicted based on the probability of the bottom event failure.

2. The vehicle real-time fault prediction method based on vehicle-cloud collaboration according to claim 1 is characterized in that: The method of training the prediction model using the training set data and calculating the prior probability distribution of atomic failures according to the prediction model includes: Atomic fault sets and signal sets are selected. Each atomic fault set includes multiple sample data. Each atomic fault corresponds to multiple discrete signals, and the multiple signals are independent of each other. The prior probability distribution of atomic faults is calculated using the following formula: Marginal probability distribution of atomic failures ;in is the fault label f ​​in the sample dataset i Take the number of 0 or 1, |D| is the number of sample sets; Conditional probability distribution of atomic failures ;in is the fault f in the sample data set i When 0 or 1 respectively, the signal x j Take the number of a certain value, is the fault label f i Takes the number of 0s or 1s.

3. The vehicle real-time fault prediction method based on vehicle-cloud collaboration according to claim 2 is characterized in that: The method of training the prediction model using the training set data and calculating the prior probability distribution of atomic failures according to the prediction model further includes: Correct the marginal probability distribution of atomic failures: ; Correct the conditional probability distribution of atomic failures: ;in is the fault f in the sample data set i When 0 or 1 respectively, the signal x j Take the number of a certain value, is the fault label f i Take the number of 0 or 1; |D| is the number of sample sets, n is x j The number of possible values.

4. The vehicle real-time fault prediction method based on vehicle-cloud collaboration according to claim 2 is characterized in that: Also includes building a fault tree: Matching the bottom events of the fault tree to the atomic faults one by one, and assigning the prior probability distribution of the atomic faults to the bottom events; A tree structure relationship between intermediate events and top events is constructed based on the fault rules and the bottom events, wherein the intermediate events are necessary and sufficient conditions for the top events.

5. The vehicle real-time fault prediction method based on vehicle-cloud collaboration according to claim 4 is characterized in that: The calculating the posterior probability of the atomic fault corresponding to the bottom event based on the naive Bayes algorithm and the target prior probability distribution includes: Calculating the probability of occurrence of atomic faults and the probability of non-occurrence of atomic faults based on the naive Bayes algorithm and the target prior probability distribution; The probability of atomic failure occurring and the probability of atomic failure not occurring are normalized to obtain the posterior probability of atomic failure.

6. The vehicle real-time fault prediction method based on vehicle-cloud collaboration according to claim 5 is characterized in that: The predicting the probability of the bottom event failure based on the posterior probability, and predicting the probability of the vehicle failure according to the probability of the bottom event failure, includes: The probability of failure of the intermediate event is calculated based on the probability of failure of the bottom event and the failure rule, the probability of failure of the top event is calculated based on the probability of failure of the intermediate event and the failure rule, and then the probability of vehicle failure is predicted based on the probability of failure of the top event.

7. The vehicle real-time fault prediction method based on vehicle-cloud collaboration according to claim 5 is characterized in that: Determining the signal data associated with the bottom event corresponding to the atomic fault based on the fault tree, and determining that the signal data corresponds to a target priori probability distribution according to the bottom event corresponding to the atomic fault, comprises: An atomic fault corresponding to the bottom event is searched according to the fault tree, corresponding signal data is searched according to the atomic fault, and a conditional probability of the signal data when the atomic fault is known is calculated, the conditional probability being the target prior probability.

8. A vehicle real-time fault prediction device based on vehicle-cloud collaboration, characterized in that: include: A priori calculation module, used to train a prediction model using training set data, and calculate a priori probability distribution of atomic faults according to the prediction model, wherein the atomic fault is a category of a fault signal, and the training set data is sample data of vehicle signals and faults; A data association module, used to obtain signal data reported by the vehicle, determine the signal data associated with the bottom event corresponding to the atomic fault based on the fault tree, and determine that the signal data corresponds to a target prior probability distribution according to the bottom event corresponding to the atomic fault; A prediction module is used to calculate the posterior probability of the atomic fault corresponding to the occurrence of the base event based on the naive Bayes algorithm and the target prior probability distribution, predict the probability of the base event failure based on the posterior probability, and predict the probability of vehicle failure based on the probability of the base event failure.

9. An electronic device, characterized in that: include: at least one processor; And at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the vehicle real-time fault prediction method based on vehicle-cloud collaboration as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is run by a computer, it executes the vehicle real-time fault prediction method based on vehicle-cloud collaboration as described in any one of claims 1 to 7.

Citation Information

Cited By

  • Distributed photovoltaic power station intelligent operation and maintenance method based on fault tree analysis

    CN120914982A