Automatic stereoscopic warehouse key fault screening method and device

Abstract

The invention discloses an automatic stereoscopic warehouse key fault screening method and device, and relates to the technical field of automatic stereoscopic warehouse management, and the method comprises the steps: constructing an interaction strength matrix based on fault data; building a Bayesian network model based on the fault data and the interaction strength matrix; based on the Bayesian network model, determining the fault risk degree of each fault mode in the automatic stereoscopic warehouse by using a potential failure mode and consequence analysis (FMEA); all fault modes of the automatic stereoscopic warehouse are arranged in a descending order based on the fault risk degrees; and determining that the plurality of fault modes in the previous preset proportion are key fault modes of the automatic stereoscopic warehouse. According to the method, the Bayesian network model and the FMEA are combined to carry out fault risk assessment, and objectivity and accuracy of automatic stereoscopic warehouse fault assessment can be improved.

Description

Technical Field

[0001] This application relates to the technical field of automated stereoscopic warehouse management, and particularly to a method and device for screening key faults in an automated stereoscopic warehouse. Background Art

[0002] An automated stereoscopic warehouse is an important part of an enterprise's production system, and its reliability has an important impact on the continuity and stability of the enterprise's production activities. By conducting a fault risk assessment on it to screen key fault modes and formulate targeted maintenance strategies, potential fault modes can be effectively prevented, and the reliable operation of the automated stereoscopic warehouse can be ensured.

[0003] Currently, there are still the following three problems in the fault research of automated stereoscopic warehouses: 1) Existing research mainly focuses on single component equipment such as stackers, and there is little research on the whole automated stereoscopic warehouse; 2) Existing research rarely considers the influence of the interaction between risk factors, fault modes, and the interaction between risk factors and fault modes, resulting in some potential key fault modes being ignored and reducing the accuracy of the fault risk assessment results; 3) When existing research uses Failure Mode and Effects Analysis (FMEA) for fault risk assessment, subjective methods such as the Analytic Hierarchy Process or the Fuzzy Analytic Hierarchy Process are mostly used to evaluate the occurrence probability, consequence severity, and detectability of fault modes. These methods are easily affected by subjective factors such as experience or preference, resulting in great randomness of the fault risk assessment results. Summary of the Invention

[0004] The purpose of this application is to provide a method and device for screening key faults in an automated stereoscopic warehouse, which can improve the objectivity and accuracy of the fault assessment of the automated stereoscopic warehouse.

[0005] To achieve the above purpose, this application provides the following solutions:

[0006] In the first aspect, this application provides a method for screening key faults in an automated stereoscopic warehouse, including:

[0007] Obtain the fault data of the automated stereoscopic warehouse;

[0008] Based on the fault data, construct an interaction intensity matrix;

[0009] Based on the fault data and the interaction intensity matrix, build a Bayesian network model;

[0010] Based on the Bayesian network model, use FMEA to respectively determine the fault risk degree of each fault mode in the automated stereoscopic warehouse;

[0011] Rank all the fault modes of the automated stereoscopic warehouse in descending order based on the fault risk degree;

[0012] Determine that multiple failure modes with a preset ratio before are the key failure modes of the automated stereoscopic warehouse.

[0013] Optionally, after respectively determining the failure risk degree of each failure mode in the automated stereoscopic warehouse by using FMEA based on the Bayesian network model, it further includes:

[0014] Obtain the maximum value of the failure risk degree and the minimum value of the failure risk degree;

[0015] Determine that the difference between the maximum value of the failure risk degree and the minimum value of the failure risk degree is the first intermediate value;

[0016] Determine that the product of the first intermediate value and the first coefficient is the second intermediate value; the first coefficient is a positive number less than 1;

[0017] Determine that the product of the first intermediate value and the second coefficient is the third intermediate value; the second coefficient is a positive number less than 1; the second coefficient is greater than the first coefficient;

[0018] Determine that the sum of the second intermediate value and the minimum value of the failure risk degree is the first demarcation value;

[0019] Determine that the sum of the third intermediate value and the minimum value of the failure risk degree is the second demarcation value;

[0020] Construct a closed interval with the minimum value of the failure risk degree as the left endpoint and the first demarcation value as the right endpoint as the corrective maintenance interval;

[0021] Determine that the corrective maintenance strategy is the maintenance strategy corresponding to the corrective maintenance interval;

[0022] Construct a left-open and right-closed interval with the first demarcation value as the left demarcation point and the second demarcation value as the right endpoint as the preventive maintenance interval;

[0023] Determine that the preventive maintenance strategy is the maintenance strategy corresponding to the preventive maintenance interval;

[0024] Construct a left-open and right-closed interval with the second demarcation value as the left demarcation point and the maximum value of the failure risk degree as the right endpoint as the condition-based maintenance interval;

[0025] Determine that the condition-based maintenance strategy is the maintenance strategy corresponding to the condition-based maintenance interval;

[0026] Determine any failure mode as the current failure mode;

[0027] Determine the maintenance strategy of the current failure mode according to the strategy interval where the failure risk degree of the current failure mode is located; the strategy interval is the corrective maintenance interval, the preventive maintenance interval or the condition-based maintenance interval.

[0028] Optionally, the failure data includes: the failure mode of the automated stereoscopic warehouse, risk factors, the severity level of the consequences of the failure mode, and the difficulty level of detecting the failure mode.

[0029] Optionally, based on the failure data, an interaction intensity matrix is constructed, including:

[0030] Determine all failure modes and risk factors in the automated stereoscopic warehouse as interaction objects;

[0031] Based on the interaction relationships between all interaction objects, construct a binary risk structure matrix; the element w in the i-th row and j-th column of the binary risk structure matrix ij is used to describe whether the j-th interaction object affects the i-th interaction object; when w ij = 1, it means the j-th interaction object affects the i-th interaction object, j = 1, 2, …, I), i = 1, 2, …, I, where I represents the number of interaction objects, and when w ij = 0, it means the j-th interaction object does not affect the i-th interaction object;

[0032] Determine the row vector where the main diagonal element w ii is located as the cause vector of the i-th interaction object;

[0033] Based on the cause vector of the i-th interaction object, construct the cause interaction comparison matrix of the i-th interaction object;

[0034] Determine the maximum eigenvector of the cause interaction comparison matrix of the i-th interaction object as the cause eigenvector of the i-th interaction object;

[0035] Take the cause eigenvector of the i-th interaction object as the i-th row vector, and construct a cause matrix based on the cause eigenvectors of I interaction objects;

[0036] Determine the column vector where the main diagonal element w ii is located as the influence vector of the i-th interaction object;

[0037] Based on the influence vector of the i-th interaction object, construct the influence interaction comparison matrix of the i-th interaction object;

[0038] Determine the maximum eigenvector of the influence interaction comparison matrix of the i-th interaction object as the influence eigenvector of the i-th interaction object;

[0039] Take the influence eigenvector of the i-th interaction object as the i-th column vector, and construct an influence matrix based on the influence eigenvectors of I interaction objects;

[0040] Construct an interaction intensity matrix based on the cause matrix and the impact matrix; the element rnm in the i-th row and j-th column of the interaction intensity matrix ij is used to describe the interaction intensity of the j-th interaction object on the i-th interaction object, ncv ij is the element in the i-th row and j-th column of the cause matrix, nev ji is the element in the i-th row and j-th column of the impact matrix, 0 ≤ rnm ij ≤ 1.

[0041] Optionally, based on the cause vector of the i-th interaction object, construct a cause interaction comparison matrix for the i-th interaction object, including:

[0042] Based on the cause vector of the i-th interaction object, construct a first cause undetermined matrix for the i-th interaction object; the element W in the j-th row and l-th column of the first cause undetermined matrix jl is used for the paired comparison result of the j-th influence degree and the l-th influence degree; the j-th influence degree is the influence degree of the j-th interaction object on the i-th interaction object; the l-th influence degree is the influence degree of the l-th interaction object on the i-th interaction object, l(l = 1, 2,..., I);

[0043] Set the values of the main diagonal elements in the first cause undetermined matrix to 1 to obtain a second cause undetermined matrix;

[0044] Replace the values of all elements below the main diagonal in the second cause undetermined matrix with the reciprocals of the elements themselves to obtain the cause interaction comparison matrix of the i-th interaction object.

[0045] Optionally, based on the impact vector of the i-th interaction object, construct an impact interaction comparison matrix for the i-th interaction object, including:

[0046] Based on the impact vector of the i-th interaction object, construct a first impact undetermined matrix for the i-th interaction object;

[0047] Set the values of the main diagonal elements in the first impact undetermined matrix to 1 to obtain a second impact undetermined matrix;

[0048] Replace the values of all elements below the main diagonal in the second impact undetermined matrix with the reciprocals of the elements themselves to obtain the impact interaction comparison matrix of the i-th interaction object.

[0049] Optionally, based on the Bayesian network model, use FMEA to respectively determine the failure risk degree of each failure mode in the automated stereoscopic warehouse, including:

[0050] All risk factors of the automated stereoscopic warehouse are divided into affected factors and influencing factors according to whether they are affected by non-self risk factors;

[0051] Taking the interaction intensity matrix as the conditional probability, determine the first occurrence probability, the second occurrence probability, the third occurrence probability, the fourth occurrence probability, and the fifth occurrence probability; The first occurrence probability P(A k |R n ) is the occurrence probability of the k-th affected factor A n when the n-th influencing factor R k occurs, n = 1, 2, …, N), N is the number of influencing factors, k = 1, 2, …, K), K is the number of affected factors; The second occurrence probability P(A k |A k' ) is the occurrence probability of the k-th affected factor A k' when the k'-th affected factor A k occurs, k' = 1, 2, …, K; The third occurrence probability P(F m |R n ) is the occurrence probability of the m-th failure mode F n when the n-th influencing factor R m occurs, m = 1, 2, …, M, M is the number of failure modes; The fourth occurrence probability P(F m |A k ) is the occurrence probability of the m-th failure mode F k when the k-th affected factor A m occurs; The fifth occurrence probability P(F m |F m' ) is the occurrence probability of the m-th failure mode F m' when the m'-th failure mode F m occurs, m' = 1, 2, …, M;

[0052] Obtain the sixth occurrence probability, the first joint probability, and the second joint probability from the failure data of the automated stereoscopic warehouse; The sixth occurrence probability P(R n ) is the occurrence probability of the n-th influencing factor R n ; The first joint probability P(F m , S u ) is the joint probability of the m-th failure mode F m and the u-th level of consequence severity S u , u = 1, 2, …, U, U is the difficulty level of consequence severity; The second joint probability P(F m , D v ) is the joint probability of the m-th failure mode F m and the v-th level of detectability D v The joint probability P(F m , Dv ), (v = 1, 2, …, V, where V is the difficulty of detection;

[0053] Input the first occurrence probability, the second occurrence probability, the third occurrence probability, the fourth occurrence probability, and the fifth occurrence probability into the Bayesian network model, and use the noise model or the gate model to determine the seventh occurrence probability and the eighth occurrence probability; The seventh occurrence probability P(A k |X(A k )) is the occurrence probability of the kth affected factor A k under the combination of the states of its parent nodes; The eighth occurrence probability P(F m |X(F m )) = is the occurrence probability of the mth failure mode F m under the combination of the states of its parent nodes;

[0054] Determine the ninth occurrence probability according to the sixth occurrence probability, the seventh occurrence probability, and the eighth occurrence probability; The ninth occurrence probability P(F m ) is the occurrence probability of the mth failure mode F m ;

[0055] Determine the first-level probability and the second-level probability according to the ninth occurrence probability, the first joint probability, and the second joint probability; The first-level probability P(S u |F m ) is the probability that the failure consequence belongs to the u-level consequence severity level S m when the mth failure mode F u occurs; The second-level probability P(D v |F m ) is the probability that it belongs to the v-level difficulty of detection level D m when the mth failure mode F v occurs;

[0056] Input the first-level probability and the second-level probability into the Bayesian network model to obtain the tenth occurrence probability and the eleventh occurrence probability; The tenth occurrence probability P(S u |X(S u )) is the occurrence probability of the u-level consequence severity level S u under the combination of the states of its parent nodes; The eleventh occurrence probability P(D v |X(D v )) is the occurrence probability of the v-level difficulty of detection level D v under the combination of the states of its parent nodes;

[0057] Input the tenth occurrence probability and the eleventh occurrence probability into the Bayesian network model to obtain the Bayesian network probability model;

[0058] Based on the Bayesian network probability model, determine the twelfth occurrence probability; the twelfth occurrence probability P(F m |S u ,D v ) is the occurrence probability of the m-th failure mode F under the u-th consequence severity level S u and the v-th detectability level D v ; m

[0059] Determine that the product of the twelfth occurrence probability, the consequence severity level score, and the detectability level score corresponding to the same failure mode is the failure risk degree of the corresponding failure mode.

[0060] Optionally, the seventh occurrence probability is:

[0061]

[0062] where

[0063]

[0064] In the formula, is the first variable, is the second variable, X(A k ) is the parent node state combination of the k-th influencing factor A k ;

[0065] The eighth occurrence probability is:

[0066]

[0067] where

[0068]

[0069] In the formula, is the third variable, is the fourth variable, is the fifth variable, X(F m ) is the parent node state combination of the m-th failure mode F m ;

[0070] The ninth occurrence probability is:

[0071]

[0072] where

[0073] In the formula, is​Sixth variable, For Seventh variable, For Eighth variable, P(X(F m )) is the joint probability of the parent node state combinations of the fault mode F m ;

[0074] The first-level probability is:

[0075]

[0076] The second-level probability is:

[0077]

[0078] The tenth occurrence probability is:

[0079]

[0080] Among them,

[0081] In the formula, is the ninth variable, X(S u ) is the parent node state combination of the u-th level consequence severity level S u ;

[0082] The eleventh occurrence probability is:

[0083]

[0084] Among them,

[0085] In the formula, is the tenth variable, X(D v ) is the parent node state combination of the v-th level detectability level D v ;

[0086] The twelfth occurrence probability is:

[0087]

[0088] In the formula, P(F m , S u , D v ) is the joint probability of the m-th fault mode F m , the u-th level consequence severity S u and the v-th level detectability level D v , P(S u , D v ) is the joint probability of the u-th level consequence severity S u and the v-th level detectability level D vJoint probability.

[0089] In a second aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the above-mentioned key fault screening method for an automated storage and retrieval system.

[0090] According to the specific embodiments provided by the present application, the following technical effects are disclosed:

[0091] The present application provides a method and device for screening key faults in an automated storage and retrieval system. First, the design structure matrix is used to calculate the interaction strengths among risk factors, fault modes, and between risk factors and fault modes in the automated storage and retrieval system. Then, based on this, combined with the causal relationships among each fault mode, its consequence severity, detectability, and risk factors, a Bayesian network model is constructed to calculate the occurrence probabilities of each fault mode under different consequence severities and detectabilities. Furthermore, the risk priority number method of FMEA is used to evaluate the risk levels of each fault mode, so as to more objectively and accurately screen key fault modes and formulate corresponding maintenance strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0093] Figure 1 It is a flowchart of a method for screening key faults in an automated storage and retrieval system according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0094] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present application.

[0095] To make the above objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0096] In an exemplary embodiment, as Figure 1 shown, a method for screening key faults in an automated storage and retrieval system is provided, including:

[0097] Step 101: Obtain the fault data of the automated storage and retrieval system. The fault data includes: the fault mode, risk factors, severity level of the consequences of the fault mode, and difficulty level of detecting the fault mode of the automated storage and retrieval system.

[0098] Based on the fault data of the automated storage and retrieval system, summarize the types, risk factors, and consequences of the fault modes, and classify the severity of the consequences and difficulty of detection of each fault mode according to Table 1.

[0099] Table 1 Table of Severity and Difficulty Levels of Fault Mode Consequences

[0100]

[0101]

[0102] Step 102: Based on the fault data, construct an interaction intensity matrix. Use the design structure matrix to calculate the interaction intensity between risk factors, between fault modes, and between risk factors and fault modes in the automated storage and retrieval system.

[0103] Step 102 includes:

[0104] Step 102-1: Determine all fault modes and risk factors in the automated storage and retrieval system as interaction objects.

[0105] Step 102-2: Based on the interaction relationships between all interaction objects, construct a binary risk structure matrix. The element w in the i-th row and j-th column of the binary risk structure matrix ij is used to describe whether the j-th interaction object affects the i-th interaction object. When w ij = 1, it means that the j-th interaction object affects the i-th interaction object, j = 1, 2,..., I), i = 1, 2,..., I, where I represents the number of interaction objects, and when w ij = 0, it means that the j-th interaction object does not affect the i-th interaction object.

[0106] Step 1 Take the risk factors and fault modes of the automated storage and retrieval system as interaction objects. Define the binary risk structure matrix according to whether there is an interaction relationship between a pair of interaction objects where, and i ≠ j.

[0107] Step 102-3: Determine that the row vector where the main diagonal element w in the binary risk structure matrix ii is located is the cause vector of the i-th interaction object.

[0108] Step 102-4: Based on the cause vector of the i-th interaction object, construct the cause interaction comparison matrix of the i-th interaction object.

[0109] Step 102-5: Determine that the maximum eigenvector of the cause interaction comparison matrix of the i-th interaction object is the cause eigenvector of the i-th interaction object.

[0110] Step 102-5 includes:

[0111] Step 102-5-1: Based on the cause vector of the i-th interaction object, construct the first cause undetermined matrix of the i-th interaction object. The element W in the j-th row and l-th column of the first cause undetermined matrix jl is used for the paired comparison result between the j-th influence degree and the l-th influence degree. The j-th influence degree is the influence degree of the j-th interaction object on the i-th interaction object. The l-th influence degree is the influence degree of the l-th interaction object on the i-th interaction object, where l (l = 1, 2,..., I).

[0112] Step 102-5-2: Set the values of the main diagonal elements in the first cause undetermined matrix to 1 to obtain the second cause undetermined matrix.

[0113] Step 102-5-3: Replace the values of all elements below the main diagonal in the second cause undetermined matrix with the reciprocals of the elements themselves to obtain the cause interaction comparison matrix of the i-th interaction object.

[0114] Step 102-6: Use the cause eigenvector of the i-th interaction object as the i-th row vector, and construct a cause matrix based on the cause eigenvectors of I interaction objects.

[0115] Step 102-7: Determine that the column vector where the main diagonal element w ii is located in the binary risk structure matrix is the influence vector of the i-th interaction object.

[0116] Step 102-8: Based on the influence vector of the i-th interaction object, construct the influence interaction comparison matrix of the i-th interaction object.

[0117] Step 102-9: Determine that the maximum eigenvector of the influence interaction comparison matrix of the i-th interaction object is the influence eigenvector of the i-th interaction object.

[0118] Step 102-9 includes:

[0119] Step 102-9-1: Based on the influence vector of the i-th interaction object, construct the first influence undetermined matrix of the i-th interaction object.

[0120] Step 102-9-2: Set the values of the main diagonal elements in the first influence undetermined matrix to 1 to obtain the second influence undetermined matrix.

[0121] Step 102-9-3: Replace the values of all elements below the main diagonal in the second influence to-be-determined matrix with the reciprocals of the elements themselves, to obtain the influence interaction comparison matrix of the i-th interaction object.

[0122] Step 102-10: Use the influence eigenvector of the i-th interaction object as the i-th column vector, and construct an influence matrix based on the influence eigenvectors of I interaction objects.

[0123] Step 102-11: Based on the cause matrix and the influence matrix, construct an interaction strength matrix. The element rnm in the i-th row and j-th column of the interaction strength matrix ij is used to describe the interaction strength of the j-th interaction object on the i-th interaction object. ncv ij is the element in the i-th row and j-th column of the cause matrix, and nev ji is the element in the i-th row and j-th column of the influence matrix, where 0 ≤ rnm ij ≤ 1.

[0124] For the i-th interaction object, decompose the interaction objects related to it into two parts. One part forms a cause vector BCV i =(w i1 w i2 …w iI ), representing the interaction objects that may affect it, and the other part forms an influence vector representing the interaction objects that it may affect.

[0125] After vector decomposition, perform row comparison and column comparison on the i-th interaction object respectively. Among them, row comparison is to make pairwise comparisons of the influence degrees between the i-th interaction object and the interaction objects that affect it, and column comparison is to make pairwise comparisons of the influence degrees between the i-th interaction object and the interaction objects that it affects. Thus, a cause interaction comparison matrix CCM i and an influence interaction comparison matrix ECM i are formed respectively as follows:

[0126]

[0127]

[0128] In the cause interaction comparison matrix, W jl represents the pairwise comparison result of the influence degree of the j-th interaction object on the i-th interaction object and the influence degree of the l (l = 1, 2, …, I) -th interaction object on the i-th interaction object. In the influence interaction comparison matrix, W jlIndicates the pairwise comparison result of the influence degree of the i-th interaction object on the j-th interaction object and the influence degree of the i-th interaction object on the l-th interaction object, and i, j, and l are not equal to each other.

[0129] Calculate the cause interaction comparison matrix CCM of the i-th interaction object i and the effect interaction comparison matrix ECM i of the maximum eigenvector NCV i =(ncv i1 ncv i2 … ncv iI ) and to form the cause matrix and the effect matrix and calculate the interaction intensity matrix accordingly as:

[0130]

[0131] Combined with the causal relationships among the various failure modes of the automated storage and retrieval system, their consequence severity, detectability, and risk factors, construct a Bayesian network model, and calculate the occurrence probabilities of each failure mode under different consequence severities and detectabilities using the noisy model or the OR gate model.

[0132] Step 103: Based on the failure data and the interaction intensity matrix, build a Bayesian network model. Establish a Bayesian network according to the causal relationships among the various failure modes of the automated storage and retrieval system, their consequence severity, detectability, and risk factors.

[0133] Step 104: Based on the Bayesian network model, use FMEA to determine the failure risk degree of each failure mode in the automated storage and retrieval system respectively.

[0134] Step 104 includes:

[0135] Step 104-1: Divide all the risk factors of the automated storage and retrieval system into affected factors and influencing factors according to whether they are affected by non-self risk factors.

[0136] Step 104-2: Use the interaction intensity matrix as the conditional probability to determine the first occurrence probability, the second occurrence probability, the third occurrence probability, the fourth occurrence probability, and the fifth occurrence probability. The first occurrence probability P(A k |R n ) is the occurrence probability of the k-th affected factor A n when the n-th influencing factor R k occurs, n = 1, 2, …, N), N is the number of influencing factors, k = 1, 2, …, K), K is the number of affected factors. The second occurrence probability P(A k |A k' ) is the k'-th affected factor Ak' The occurrence probability of the k-th affected factor A when it occurs k , where k' = 1, 2, …, K. The third occurrence probability P(F m |R n ) is the occurrence probability of the m-th failure mode F when the n-th influencing factor R n occurs, where m = 1, 2, …, M and M is the number of failure modes. The fourth occurrence probability P(F m |A m ) is the occurrence probability of the m-th failure mode F when the k-th affected factor A k occurs. The fifth occurrence probability P(F k |F m ) is the occurrence probability of the m-th failure mode F when the m'-th failure mode F m occurs, where m' = 1, 2, …, M. m' (m' = 1, 2, …, M). m' occurs, where m' = 1, 2, …, M. m occurs, where m' = 1, 2, …, M.

[0137] According to whether it is affected by other risk factors, risk factors are divided into affected factors and influencing factors. Among them, an affected factor is a risk factor affected by other risk factors, and an influencing factor is a risk factor not affected by other risk factors. Considering the interaction strength obtained from the design structure matrix as a conditional probability, the occurrence probability P(A n (n = 1, 2, …, N) when the n-th influencing factor R k (k = 1, 2, …, K) of the k-th affected factor A k |R n ), the occurrence probability P(A k' |A k ) of the k-th affected factor A when the k'-th (k' = 1, 2, …, K) affected factor A k |A k' ) occurs, the occurrence probability P(F n |R m (m = 1, 2, …, M) of the m-th failure mode F when the n-th influencing factor R m |R n ) occurs, the occurrence probability P(F k |A m ) of the m-th failure mode F when the k-th affected factor A m |A k ) occurs, and the occurrence probability P(F m' |F m ) of the m-th failure mode F when the m'-th (m' = 1, 2, …, M) failure mode F m |F m' ) occurs. Among them, k' ≠ k and m' ≠ m.

[0138] Step 104-3: Obtain the sixth occurrence probability, the first joint probability, and the second joint probability from the failure data of the automated storage and retrieval system. The sixth occurrence probability P(R n ) is the occurrence probability of the nth influencing factor R n . The first joint probability P(F m , S u ) is the joint probability of the mth failure mode F m and the u-level consequence severity S u , where u = 1, 2, …, U, and U is the difficulty of detecting the consequence severity level. The second joint probability P(F m , D v ) is the joint probability of the mth failure mode F m and the v-level detectability D v , and P(F m , D v ), (v = 1, 2, …, V), and V is the detectability of the detectability.

[0139] According to the failure data of the automated storage and retrieval system, count the occurrence probability P(R n ) of the nth influencing factor R n , the joint probability P(F m , S u ) of the mth failure mode F m and the u (u = 1, 2, …, U)-level consequence severity S u , and the joint probability P(F m , D v ) of the mth failure mode F m and the v (v = 1, 2, …, V)-level detectability D v .

[0140] Step 104-4: Input the first occurrence probability, the second occurrence probability, the third occurrence probability, the fourth occurrence probability, and the fifth occurrence probability into the Bayesian network model, and use the noise model or the gate model to determine the seventh occurrence probability and the eighth occurrence probability. The seventh occurrence probability P(A k |X(A k )) is the occurrence probability of the kth affected factor A k under the state combination of its own parent nodes. The eighth occurrence probability P(F m |X(F m )) = is the occurrence probability of the mth failure mode F m under the state combination of its own parent nodes.

[0141] The seventh occurrence probability is:

[0142]

[0143] Among them,

[0144]

[0145] In the formula, is the first variable, is the second variable, X(A k ) is the state combination of the parent nodes of the k-th affected factor A k .

[0146] The probability of the eighth occurrence is:

[0147]

[0148] Among them,

[0149]

[0150] In the formula, is the third variable, is the fourth variable, is the fifth variable, X(F m ) is the state combination of the parent nodes of the m-th failure mode F m .

[0151] When the n-th influencing factor R n occurs, the occurrence probability P(A k |R k ), the occurrence probability P(A n |A k' ) when the k'-th affected factor A k occurs, the occurrence probability P(F k |R k' ) when the n-th influencing factor R n occurs, the occurrence probability P(F m |A m ) when the k-th affected factor A n occurs, the occurrence probability P(F k |A m ), the occurrence probability P(F m |A k ) when the m'-th failure mode F m' occurs, and the occurrence probability P(F m |F m ) are input into the Bayesian network model, and the noise model or the OR gate model is used to calculate the occurrence probability of the k-th affected factor A m' under the state combination of its own parent nodes and the m-th failure mode F k m ​The occurrence probability under the state combination of its own parent node.

[0152] Step 104-5: Determine the ninth occurrence probability according to the sixth occurrence probability, the seventh occurrence probability, and the eighth occurrence probability. The ninth occurrence probability P(F m ) is the occurrence probability of the mth failure mode F m .

[0153] The ninth occurrence probability is:

[0154]

[0155] Among them,

[0156] In the formula, is Sixth variable, is Seventh variable, is Eighth variable, P(X(F m )) is the joint probability of the state combination of the parent node of the failure mode F m .

[0157] Combined with the occurrence probability P(A k |X(A k )) of the kth influencing factor A k under the state combination of its own parent node, the occurrence probability P(F m |X(F m )) of the mth failure mode F m under the state combination of its own parent node, and the occurrence probability P(R n ) of the nth influencing factor R n ), calculate the occurrence probability of the mth failure mode F m .

[0158] Step 104-6: Determine the first-level probability and the second-level probability according to the ninth occurrence probability, the first joint probability, and the second joint probability. The first-level probability P(S u |F m ) is the probability that the failure consequence belongs to the u-level consequence severity level S m when the mth failure mode F u occurs. The second-level probability P(D v |F m ) is the probability that it belongs to the v-level difficulty detection level D m when the mth failure mode F v occurs.

[0159] The first-level probability is:

[0160]

[0161] The probability of the second level is:

[0162]

[0163] Based on the m-th failure mode F m and the severity level S of the u-th level u of the combined probability P(F m , S u ), the m-th failure mode F m and the undetected level D of the v-th level v of the combined probability P(F m , D v ), and the occurrence probability P(F m ) of the m-th failure mode F m ), calculate the probability that the failure consequence belongs to the severity level S of the u-th level m when the m-th failure mode F u occurs and its probability of belonging to the undetected level D of the v-th level v .

[0164] Step 104-7: Input both the probability of the first level and the probability of the second level into the Bayesian network model to obtain the tenth occurrence probability and the eleventh occurrence probability. The tenth occurrence probability P(S u |X(S u )) is the occurrence probability of the severity level S of the u-th level u under the combined state of its parent nodes. The eleventh occurrence probability P(D v |X(D v )) is the occurrence probability of the undetected level D of the v-th level v under the combined state of its parent nodes.

[0165] The tenth occurrence probability is:

[0166]

[0167] Among them,

[0168] In the formula, is the ninth variable, and X(S u ) is the combined state of the parent nodes of the severity level S of the u-th level u .

[0169] The eleventh occurrence probability is:

[0170]

[0171] Among them,

[0172] In the formula, is the tenth variable, X(D v ) is the combined state of the parent nodes of the difficulty level D of the v-th level v .

[0173] Input the probability P(S m |F u ) that the failure consequence belongs to the u-th level of severity S when the m-th failure mode F u occurs and the probability P(D m |F v ) that it belongs to the v-th level of difficulty D v |F m ) into the Bayesian network model to calculate the occurrence probability of the u-th level of severity S u under its own combined state of parent nodes and the occurrence probability of the v-th level of difficulty D v under its own combined state of parent nodes.

[0174] Step 104-8: Input both the tenth occurrence probability and the eleventh occurrence probability into the Bayesian network model to obtain the Bayesian network probability model.

[0175] Step 104-9: Based on the Bayesian network probability model, determine the twelfth occurrence probability. The twelfth occurrence probability P(F m |S u ,D v ) is the occurrence probability of the m-th failure mode F u under the u-th level of severity S v and the v-th level of difficulty D m .

[0176] The twelfth occurrence probability is:

[0177]

[0178] In the formula, P(F m ,S u ,D v ) is the joint probability of the m-th failure mode F m , the u-th level of severity S u and the v-th level of difficulty D v , and P(S u ,D v ) is the joint probability of the u-th level of severity S u and the v-th level of difficulty D v .

[0179] Input the u-th level of severity S uThe occurrence probability P(S u |X(S u )) at the combined state of its own parent nodes and the difficulty detection level D of the v-th level v The occurrence probability P(D v |X(D v )) at the combined state of its own parent nodes are input into the Bayesian network to obtain a complete Bayesian network probability model. Then, combined with the occurrence probability P(R n ) of the n-th influencing factor R n , the occurrence probability P(A k |X(A k )) of the k-th affected factor A k at the combined state of its own parent nodes, the occurrence probability P(F m |X(F m )) of the m-th failure mode F m at the combined state of its own parent nodes, the occurrence probability P(S u |X(S u )) of the severity level S of the u-th level u and the occurrence probability P(D v |X(D v )) of the difficulty detection level D of the v-th level v at the combined state of its own parent nodes, calculate the occurrence probability of the m-th failure mode F u under the severity level S of the u-th level v and the difficulty detection level D of the v-th level m .

[0180] Step 104 - 10: Determine that the product of the twelfth occurrence probability, the severity level score, and the difficulty detection level score corresponding to the same failure mode is the failure risk degree of the corresponding failure mode.

[0181] Combined with the occurrence probability P(F u |S v , D m ) of the m-th failure mode F m under the severity level S of the u-th level u and the difficulty detection level D of the v-th level v and the grade scores of severity and difficulty detection in Table 1, calculate that the failure risk degree of the m-th failure mode F m is

[0182] RPN m = P(F m |S u , D v ) × S(u) × D(v).

[0183] Wherein, S(u) is the grade score of the severity of the u-th level consequence, D(v) is the grade score of the difficulty of detection of the v-th level, and 0 ≤ RPN m ≤ 49.

[0184] Step 105: Arrange all the failure modes of the automated stereoscopic warehouse in descending order based on the degree of failure risk.

[0185] Step 106: Determine that a plurality of failure modes with a preset proportion in the front are the key failure modes of the automated stereoscopic warehouse.

[0186] According to the degree of failure risk of each failure mode of the automated stereoscopic warehouse, sort each failure mode, and screen out the failure modes ranked in the top 20% as the key failure modes of the automated stereoscopic warehouse.

[0187] After step 104, it further includes:

[0188] Step 107: Obtain the maximum value of the degree of failure risk and the minimum value of the degree of failure risk.

[0189] Step 108: Determine that the difference between the maximum value of the degree of failure risk and the minimum value of the degree of failure risk is the first intermediate value.

[0190] Step 109: Determine that the product of the first intermediate value and the first coefficient is the second intermediate value. The first coefficient is a positive number less than 1.

[0191] Step 1010: Determine that the product of the first intermediate value and the second coefficient is the third intermediate value. The second coefficient is a positive number less than 1. The second coefficient is greater than the first coefficient.

[0192] Step 1011: Determine that the sum of the second intermediate value and the minimum value of the degree of failure risk is the first demarcation value.

[0193] Step 1012: Determine that the sum of the third intermediate value and the minimum value of the degree of failure risk is the second demarcation value.

[0194] Step 1013: Construct a closed interval with the minimum value of the degree of failure risk as the left endpoint and the first demarcation value as the right endpoint as the corrective maintenance interval.

[0195] Step 1014: Determine that the corrective maintenance strategy is the maintenance strategy corresponding to the corrective maintenance interval.

[0196] Step 1015: Construct a left-open and right-closed interval with the first demarcation value as the left demarcation point and the second demarcation value as the right endpoint as the preventive maintenance interval.

[0197] Step 1016: Determine that the preventive maintenance strategy is the maintenance strategy corresponding to the preventive maintenance interval.

[0198] Step 1017: Construct an open - left and closed - right interval, i.e., the condition - based maintenance interval, with the second demarcation value as the left demarcation point and the maximum value of the failure risk level as the right endpoint.

[0199] Step 1018: Determine the condition - based maintenance strategy as the maintenance strategy corresponding to the condition - based maintenance interval.

[0200] Step 1019: Determine any failure mode as the current failure mode.

[0201] Step 1020: Determine the maintenance strategy of the current failure mode according to the strategy interval in which the failure risk level of the current failure mode is located. The strategy intervals are the corrective maintenance interval, the preventive maintenance interval, or the condition - based maintenance interval.

[0202] The failure maintenance strategies of the automated storage and retrieval system are divided into three types: condition - based maintenance, preventive maintenance, and corrective maintenance. And based on the RPN values of each failure mode, the three - level demarcation values are calculated as

[0203]

[0204] where c is the maximum RPN value of each failure mode and b is the minimum RPN value of each failure mode. Then, according to the three - level demarcation values G 1 and G 2 and the characteristics of each maintenance strategy, the maintenance strategy decision rules for the failure modes of the automated storage and retrieval system are established as shown in Table 2.

[0205] Table 2 Maintenance strategy judgment table for failure modes

[0206]

[0207] This application takes the whole automated storage and retrieval system as the research object. On the basis of considering the influence of the interaction between risk factors, between failure modes, and between risk factors and failure modes on the failure risk assessment results, it uses the more objective Bayesian FMEA to screen key failure modes, which can effectively screen out potential failure modes that are easily triggered by risk factors or other failure modes or have strong failure propagation effects, so as to achieve a more accurate screening of key failure modes of the automated storage and retrieval system. This application establishes the maintenance strategy decision rules for the failure modes of the automated storage and retrieval system based on the risk assessment results of each failure mode and the adaptation characteristics between the failure mode and the maintenance strategy, so as to formulate targeted maintenance strategies for different failure modes to ensure the safe and reliable operation of the automated storage and retrieval system. Next, taking the automated storage and retrieval system of Company A as an example, the content provided by this application will be specifically described.

[0208] Company A is mainly engaged in the manufacturing and sales of household appliances and the business of goods import and export. Its automated stereoscopic warehouse covers an area of 1,718 square meters and is mainly used for the conveying, warehousing storage, picking, and outbound distribution of materials such as finished products. The warehouse consists of five areas: the vertical storage area, the second-floor inbound area, the second-floor re-packaging inbound and outbound area, the first-floor outbound shipping area, and the first-floor transfer inbound area, and is composed of a rack system, a stacker system, and an inbound and outbound transportation system.

[0209] Based on the failure data and relevant literature of Company A's automated stereoscopic warehouse, 34 failure modes and 12 risk factors are summarized as shown in Table 3 and Table 4. Then, Table 1 is used to classify the consequence severity and detectability of these 34 failure modes. Subsequently, according to the design structure matrix, an evaluation is carried out on whether there are interaction relationships among the 34 failure modes, among the 12 risk factors, and between the 34 failure modes and the 12 risk factors, and the risk structure matrix L is obtained. 1 Among them, there are 5 pairs of risk factors with interaction relationships, 57 pairs of failure modes, and 164 pairs of failure modes and risk factors.

[0210] Table 3 Failure Mode Table of Automated Stereoscopic Warehouse

[0211]

[0212]

[0213] Table 4 Influence Factor Table of Automated Stereoscopic Warehouse

[0214]

[0215] Risk Structure Matrix Then, the risk structure matrix L 1 is decomposed and transformed to obtain the interaction intensity matrix among risk factors, among failure modes, and between risk factors and failure modes as follows:

[0216]

[0217] Construct a Bayesian network for each failure mode, its risk factors, consequence severity levels, and difficulty of detection levels of the automated storage and retrieval system based on the interaction relationships among 34 failure modes, 12 risk factors, and between 34 failure modes and 12 risk factors; secondly, calculate the occurrence probabilities of each affected factor when each influencing factor occurs, the occurrence probabilities of each affected factor when other affected factors occur, the occurrence probabilities of each failure mode when each risk factor occurs, and the occurrence probabilities of each failure mode when other failure modes occur obtained from the interaction intensity matrix, calculate the occurrence probabilities of each affected factor and failure mode under the combination of the states of their own parent nodes, and calculate the occurrence probabilities of each failure mode by combining the occurrence probabilities of each affected factor and failure mode under the combination of the states of their own parent nodes with the occurrence probabilities of each influencing factor obtained statistically; thirdly, calculate the probabilities that the failure consequences of each failure mode belong to each consequence severity level and the probabilities that the difficulty of detecting the failure belongs to each difficulty of detection level when each failure mode occurs by combining the occurrence probabilities of each failure mode with the joint probabilities of each failure mode, its consequence severity level, and difficulty of detection level obtained statistically, and calculate the occurrence probabilities of each consequence severity level and difficulty of detection level under the combination of the states of their own parent nodes based on the probabilities that the failure consequences of each failure mode belong to each consequence severity level and the probabilities that the difficulty of detecting the failure belongs to each difficulty of detection level when each failure mode occurs; finally, calculate the occurrence probabilities of each failure mode at each consequence severity level and difficulty of detection level by combining the occurrence probabilities of each influencing factor, each affected factor, failure mode, consequence severity level, and difficulty of detection level under the combination of the states of their own parent nodes as shown in Table 5.

[0218] Table 5 Occurrence Probability Table of Each Failure Mode F at Each Consequence Severity Level S and Difficulty of Detection Level D

[0219]

[0220]

[0221] Combine the occurrence probabilities of each failure mode at each consequence severity level and difficulty of detection level in Table 5 and the grade scores of each failure mode consequence severity level and difficulty of detection level in Table 1 to calculate the RPN of each failure mode, and sort the 34 failure modes of the automated storage and retrieval system accordingly. The sorting results are shown in Table 6; secondly, according to the sorting results of the RPN values of the failure modes in Table 6, screen the failure modes ranked in the top 20% as the key failure modes of Company A's automated storage and retrieval system, namely F6, F17, F20, F12, F14, F8; finally, combine the maximum and minimum RPN values in Table 6 to calculate the third-level demarcation values G 1 = 16.823, G 2 = 32.916, and according to the decision rules in Table 2, obtain the maintenance strategies for the failure modes of Company A's automated storage and retrieval system as shown in Table 7.

[0222] Table 6 RPN values ​​and rankings of various fault modes

[0223]

[0224]

[0225]

[0226] Table 7 Maintenance strategy table for each failure mode

[0227]

[0228] To further verify the effectiveness and superiority of the Bayesian FMEA in the proposed method, it is compared with the traditional FMEA and fuzzy FMEA, and the specific results are shown in Table 8. It can be seen from Table 8 that: 1) In the RPN ranking of the above three methods, the RPN rankings of the eight fault modes F3, F6, F7, F8, F12, F14, F17, and F20 are all relatively high, while the RPN rankings of the six fault modes F18, F22, F29, F30, F31, and F32 are all relatively low, and the ranking deviations are all within [0,5], which is not large. Obviously, the rankings of 82.4% of the fault modes are basically consistent, which proves that the proposed Bayesian FMEA is effective; 2) In the RPN ranking of the traditional FMEA, the RPN rankings of F26 and F28 are both 17, while in the proposed Bayesian FMEA, the RPN rankings of the two are 21 and 13 respectively. In practice, when F26 occurs, it can be effectively stopped and the fault location is accurate, while F28 has no emergency stop measures and requires system detection to determine the cause of the fault, that is, the failure risk of F26 is less than that of F28, which is consistent with the conclusion of Bayesian FMEA. This shows that compared with traditional FMEA, the proposed Bayesian FMEA can effectively distinguish the ranking of fault modes; 3) In the proposed Bayesian FMEA, the RPN rankings of F1, F13, and F15 are 19, 7, and 12, respectively, which are significantly higher than 31, 11, and 18 of fuzzy FMEA. This is because the two fault modes of F1 and F15 are susceptible to multiple risk factors and other fault modes, resulting in a higher probability of occurrence. F13 will affect the four fault modes of F7, F8, F9, and F20, and has a strong fault propagation effect, which can easily cause a chain reaction and lead to more serious failure consequences. In addition, in the proposed Bayesian FMEA, the RPN rankings of F5, F23, and F33 are 23, 31, and 30, respectively, which are significantly lower than those of fuzzy FMEA, which are 17, 26, and 25. This is because F5, F23, and F33 do not affect other failure modes, that is, they do not have a failure propagation effect, and are rarely affected by other risk factors and failure modes.

[0229] Obviously, compared with traditional FMEA and fuzzy FMEA, the proposed Bayesian FMEA can more comprehensively reflect the influence of the interaction among risk factors, among failure modes, and between risk factors and failure modes on the degree of failure risk, thus more objectively and accurately reflecting the failure risk ranking among various failure modes, which fully demonstrates the superiority of the proposed method.

[0230] Table 8 Comparison table of RPN ranking results

[0231]

[0232]

[0233] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communicating with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for screening key failures of an automated storage and retrieval system.

[0234] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0235] In an exemplary embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0236] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0237] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0238] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0239] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0240] In this article, specific examples are used to elaborate on the principles and implementation manners of the present application. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A method for screening critical faults in an automated warehouse, characterized in that: include: Obtain fault data of automated warehouses; Based on the fault data, construct an interaction intensity matrix; Building a Bayesian network model based on the fault data and the interaction intensity matrix; Based on the Bayesian network model, FMEA is used to determine the failure risk degree of each failure mode in the automated warehouse; Arrange all failure modes of the automated warehouse in descending order based on the failure risk level; It is determined that multiple failure modes with preset proportions are key failure modes of the automated warehouse.

2. The method for screening critical faults in an automated warehouse according to claim 1, characterized in that: After determining the degree of failure risk of each failure mode in the automated warehouse using FMEA based on the Bayesian network model, the method further includes: Obtaining a maximum value of the fault risk level and a minimum value of the fault risk level; Determine a difference between a maximum value of the failure risk level and a minimum value of the failure risk level as a first intermediate value; Determine that the product of the first intermediate value and the first coefficient is a second intermediate value; the first coefficient is a positive number less than 1; Determine that the product of the first intermediate value and the second coefficient is a third intermediate value; the second coefficient is a positive number less than 1; the second coefficient is greater than the first coefficient; Determine the sum of the second intermediate value and the minimum value of the fault risk level as a first dividing value; Determine the sum of the third intermediate value and the minimum value of the fault risk level as the second dividing value; Taking the minimum value of the fault risk degree as the left endpoint and the first boundary value as the right endpoint, a closed interval is constructed as the post-maintenance interval; Determine the post-maintenance strategy as the maintenance strategy corresponding to the post-maintenance interval; Taking the first demarcation value as the left demarcation point and the second demarcation value as the right endpoint, a left-open and right-closed interval is constructed as a regular maintenance interval; Determining a regular maintenance strategy as a maintenance strategy corresponding to the regular maintenance interval; Taking the second demarcation value as the left demarcation point and the maximum value of the fault risk level as the right endpoint, a left-open and right-closed interval is constructed as a condition-based maintenance interval; Determine the condition-based maintenance strategy as the maintenance strategy corresponding to the condition-based maintenance interval; Determine any failure mode as the current failure mode; The maintenance strategy of the current fault mode is determined according to the strategy interval in which the fault risk level of the current fault mode is located; the strategy interval is a post-maintenance interval, a regular maintenance interval or a condition-based maintenance interval.

3. The method for screening critical faults in an automated warehouse according to claim 1, characterized in that: The fault data includes: the fault mode of the automated warehouse, risk factors, the severity level of the consequences of the fault mode, and the difficulty level of the fault mode.

4. The method for screening critical faults in an automated high-bay warehouse according to claim 1, characterized in that: Based on the fault data, an interaction intensity matrix is ​​constructed, including: Determine all failure modes and risk factors in the automated warehouse as interaction objects; Based on the interaction relationship between all the interaction objects, a binary risk structure matrix is ​​constructed; the element w in the i-th row and j-th column of the binary risk structure matrix ij It is used to describe whether the jth interaction object affects the ith interaction object; w ij =1 means that the jth interaction object affects the ith interaction object, j = 1, 2, ..., I, i = 1, 2, ..., I, I represents the number of interaction objects, w ij = 0 means that the jth interaction object does not affect the ith interaction object; Determine the main diagonal element w in the binary risk structure matrix ii The row vector is the cause vector of the i-th interaction object; Based on the cause vector of the i-th interaction object, construct the cause interaction comparison matrix of the i-th interaction object; Determine the maximum eigenvector of the cause interaction comparison matrix of the i-th interaction object as the cause eigenvector of the i-th interaction object; The cause feature vector of the i-th interaction object is taken as the i-th row vector, and a cause matrix is ​​constructed based on the cause feature vectors of I interaction objects; Determine the main diagonal element w in the binary risk structure matrix ii The column vector is the influence vector of the i-th interaction object; Based on the influence vector of the i-th interaction object, construct the influence interaction comparison matrix of the i-th interaction object; Determine the maximum eigenvector of the influence interaction comparison matrix of the i-th interaction object as the influence eigenvector of the i-th interaction object; The influence feature vector of the i-th interaction object is taken as the i-th column vector, and the influence matrix is ​​constructed based on the influence feature vectors of I interaction objects; Based on the cause matrix and the impact matrix, an interaction intensity matrix is ​​constructed; the element rnm in the i-th row and j-th column of the interaction intensity matrix ij It is used to describe the interaction strength of the jth interaction object on the ith interaction object. ncv ij is the element in the i-th row and j-th column of the cause matrix, nev ji is the element in the i-th row and j-th column of the matrix, 0≤rnm ij ≤1.

5. The method for screening critical faults in an automated high-bay warehouse according to claim 4, characterized in that: Based on the cause vector of the i-th interaction object, the cause interaction comparison matrix of the i-th interaction object is constructed, including: Based on the cause vector of the i-th interaction object, a first cause-to-be-determined matrix of the i-th interaction object is constructed; the element W in the j-th row and l-th column of the first cause-to-be-determined matrix jl The paired comparison result of the j-th influence degree and the l-th influence degree; the j-th influence degree is the influence degree of the j-th interactive object on the i-th interactive object; the l-th influence degree is the influence degree of the l-th interactive object on the i-th interactive object, l (l = 1, 2, ..., I); The values ​​of the main diagonal elements in the first cause-to-be-determined matrix are all set to 1 to obtain a second cause-to-be-determined matrix; The values ​​of all elements on the lower left of the main diagonal in the second cause-to-be-determined matrix are replaced with the reciprocals of the elements themselves to obtain the cause interaction comparison matrix of the i-th interaction object.

6. The method for screening critical faults in an automated high-bay warehouse according to claim 4, characterized in that: Based on the influence vector of the ith interaction object, an influence interaction comparison matrix of the ith interaction object is constructed, including: Based on the influence vector of the i-th interaction object, construct the first influence matrix to be determined of the i-th interaction object; The values ​​of the main diagonal elements in the first influence matrix to be determined are all set to 1, so as to obtain a second influence matrix to be determined; The values ​​of all elements on the lower left of the main diagonal in the second influence matrix to be determined are replaced by the reciprocals of the elements themselves to obtain the influence interaction comparison matrix of the i-th interaction object.

7. The method for screening critical faults in an automated high-bay warehouse according to claim 1, characterized in that: Based on the Bayesian network model, FMEA is used to determine the failure risk level of each failure mode in the automated warehouse, including: According to whether they are affected by non-self risk factors, all risk factors of the automated warehouse are divided into affected factors and influencing factors; The interaction intensity matrix is ​​used as a conditional probability to determine the first occurrence probability, the second occurrence probability, the third occurrence probability, the fourth occurrence probability and the fifth occurrence probability; the first occurrence probability P(A k |R n ) is the nth influencing factor R n (When the kth affected factor A occurs k The probability of occurrence of the second occurrence P(A k |A k' ) is the k'(th affected factor A k' When the kth affected factor A occurs k The probability of occurrence, k'=1,2,…,K); the third probability of occurrence P(F m |R n ) is the nth influencing factor R n When the mth failure mode F occurs m The probability of occurrence, m = 1, 2, ..., M, M is the number of failure modes; the fourth probability of occurrence P (F m |A k ) is the kth affected factor A k When the mth failure mode F occurs m The probability of occurrence; the fifth probability of occurrence P(F m |F m' ) is the m'(th failure mode F m' When the mth failure mode F occurs m The probability of occurrence, m'=1,2,…,M; The sixth occurrence probability, the first joint probability and the second joint probability are obtained from the fault data of the automated warehouse; the sixth occurrence probability P(R n ) is the nth influencing factor R n The probability of occurrence; the first joint probability P(F m ,S u ) is the mth fault mode F m and the severity of the consequence at level u u The joint probability of u = 1, 2, ..., U, U is the difficulty of detection of the severity level of the consequence; the second joint probability P (F m ,D v ) is the mth fault mode F m With level v difficulty D v The joint probability P(F m ,D v ), (v=1,2,…,V, V is the difficulty of detection; The first occurrence probability, the second occurrence probability, the third occurrence probability, the fourth occurrence probability and the fifth occurrence probability are all input into the Bayesian network model, and the seventh occurrence probability and the eighth occurrence probability are determined by using a noise model or a gate model; the seventh occurrence probability P(A k |X(A k )) is the kth affected factor A k The probability of occurrence under the state combination of its own parent node; The eighth probability of occurrence P(F m |X(F m ))=is the mth fault mode F m The probability of occurrence under the combination of its own parent node status; According to the sixth occurrence probability, the seventh occurrence probability and the eighth occurrence probability, a ninth occurrence probability is determined; the ninth occurrence probability P(F m ) is the mth fault mode F m The probability of occurrence; Determine the first level probability and the second level probability according to the ninth occurrence probability, the first joint probability and the second joint probability; the first level probability P(S u |F m ) is the mth fault mode F m When the failure occurs, the consequences belong to the uth consequence severity level S u The probability of the second level P(D v |F m ) The mth failure mode F m When it occurs, it belongs to the level v difficult to detect level D v The probability of The first level probability and the second level probability are input into the Bayesian network model to obtain the tenth occurrence probability and the eleventh occurrence probability; the tenth occurrence probability P(S u |X(S u )) is the u-th level consequence severity level S u The probability of occurrence under the state combination of its own parent node; the eleventh probability of occurrence P(D v |X(D v )) is the vth level of difficulty to detect D v The probability of occurrence under the combination of its own parent node status; Inputting the tenth occurrence probability and the eleventh occurrence probability into a Bayesian network model to obtain a Bayesian network probability model; Based on the Bayesian network probability model, the twelfth occurrence probability is determined; the twelfth occurrence probability P(F m |S u ,D v ) is the u-th level consequence severity level S u Level V Difficulty Level D v Next mth failure mode F m The probability of occurrence; The product of the twelfth occurrence probability, consequence severity level score and difficulty of detection level score corresponding to the same failure mode is determined as the failure risk degree of the corresponding failure mode.

8. The method for screening critical faults in an automated high-bay warehouse according to claim 7, characterized in that: The seventh probability of occurrence is: in, In the formula, is the first variable, is the second variable, X(A k ) is the kth affected factor A k The parent node state combination; The eighth occurrence probability is: in, In the formula, is the third variable, is the fourth variable, is the fifth variable, X(F m ) is the mth fault mode F m The parent node state combination; The ninth probability of occurrence is: P(F m )=∑ X(Fm) P(F m |X(F m ))P(X(F m )); in, In the formula, is the sixth variable, is the seventh variable, is the eighth variable, P(X(F m )) is the failure mode F m The joint probability of the combination of parent node states; The first level probability is: The second level probability is: The tenth probability of occurrence is: in, In the formula, is the ninth variable, X(S u ) is the u-th level consequence severity level S u The parent node state combination; The eleventh probability of occurrence is: in, In the formula, is the tenth variable, X(D v ) is the vth level of difficulty to detect D v The parent node state combination; The twelfth probability of occurrence is: In the formula, P(F m ,S u ,D v ) is the mth fault mode F m , u-th level consequence severity S u Level V Difficulty Level D v The joint probability, P(S u ,D v ) is the severity of the u-th level consequence S u Level V Difficulty Level D v The joint probability of .

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement The present invention provides a method for screening critical faults in an automated high-bay warehouse as described in any one of claims 1 to 8.

Images