A Dynamic Prediction Method and System for Elevator Overrun Accidents Based on Fault Tree

Abstract

The present invention discloses a dynamic prediction method and system for elevator overshoot accidents based on a fault tree, belonging to the technical field of elevator fault prediction. Collect information about elevator overshoot accident faults, and use the collected information to construct a fault tree for elevator overshoot accidents; obtain multiple common bottom events of the fault tree; perform frequency-domain feature extraction on the original vibration signal of the traction machine collected in real time by setting sensors on the outer bearing seat of the traction machine, convert the frequency-domain indicators reflecting the fault characteristics of the traction machine into failure probabilities, calculate the fuzzy failure probabilities for other bottom events, extract the time-domain index information of the traction machine, convert the failure probabilities, and conduct quantitative analysis of the fault tree. Then upload the results to the intelligent operation and maintenance cloud platform for dynamic risk assessment and display. According to the quantitative analysis of the fault tree, output the probability index of elevator overshoot accidents, automatically evaluate the risk, and solve the technical problems of fault prediction and diagnosis of elevator overshoot accidents and dynamic risk assessment.

Description

Technical Field

[0001] The present invention relates to a dynamic prediction method and system for elevator over - running accidents based on a fault tree, belonging to the technical field of elevator fault prediction. Background Art

[0002] Over - running accidents are elevator accidents with high frequency and high harmfulness. They are one of the five common elevator accidents and are extremely harmful. The occurrence of elevator over - running accidents not only has a huge impact on the economy, reputation, and development of enterprises; at the same time, it poses a threat to the lives and safety of passengers and waiting people, bringing a heavy blow to families; and even affects social stability.

[0003] In the prior art, the methods for predicting the risk of elevator over - running accidents mainly transform the health assessment of the entire elevator based on static fault trees, FMEA analysis, and the analytic hierarchy process. The technical problems existing in the prior art solutions are: there are many static methods for elevator risk assessment, and they cannot reflect the current elevator risk status in real - time. Therefore, the role of dynamic risk prediction for elevator over - running accidents is very limited. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, a dynamic prediction method and system for elevator over - running accidents based on a fault tree are provided. Its steps are simple, it can feedback the current elevator risk status in real - time, and can make a dynamic risk prediction for elevator over - running accidents.

[0005] To achieve the above - mentioned technical purpose, a dynamic prediction method for elevator over - running accidents based on a fault tree of the present invention is as follows:

[0006] S1. Collect data on elevator over - running accident faults, and use the collected data to construct a fault tree with the elevator over - running accident as the top event; it is obtained that the fault tree of the elevator over - running accident includes a total of 23 common fault tree bottom events.

[0007] S2. Through the structural importance ranking, it is confirmed that the importance of the elevator traction machine failure among the bottom events is the greatest. Therefore, the original vibration information of the traction machine is selected for monitoring.

[0008] S3. Perform frequency - domain feature extraction on the original vibration signal of the traction machine collected in real - time by setting sensors on the outer bearing seat of the elevator traction machine, convert the frequency - domain indicators reflecting the fault characteristics of the traction machine into failure probabilities, and calculate the fuzzy failure probabilities of the other 22 common fault tree bottom events through the fuzzy comprehensive evaluation theory.

[0009] S4. Use the edge - segment data acquisition device set at the elevator to extract the time - domain index information of the traction machine, convert the failure probability, and perform quantitative analysis of the fault tree. Then upload the results to the intelligent operation and maintenance cloud platform for dynamic risk assessment and display.

[0010] Furthermore, the calculation method of the failure probability of the elevator traction machine is as follows:

[0011] Use the following formula to calculate the time-domain feature extraction of the original vibration signal of the traction machine collected, including four time-domain index values of the traction machine: the mean value x of the vibration signal mean , the maximum value x of the vibration signal max , the root mean square value x of the vibration signal rms and the kurtosis value x of the vibration signal kurt : The average value and the root mean square value of the vibration signal can reflect the vibration amplitude of the traction machine; the maximum value and the kurtosis value can reflect the vibration impact of the traction machine;

[0012]

[0013] x max = Max(X),

[0014]

[0015]

[0016] In the formula: x i represents the vibration value, N represents the number of points sampled each time, and σ represents the standard deviation.

[0017] Furthermore, the performance evaluation method of the elevator traction machine is as follows:

[0018] According to the ISO-10816 vibration monitoring and evaluation standard, the safe state of the traction machine is divided into four intervals. The time-domain index values reflecting the performance of the traction machine are converted into failure probabilities. After combining expert experience and individual fuzzy opinion aggregation, the failure probability threshold area for the newly put into use traction machine is less than 0.03. The failure threshold area for the traction machine running without restriction for a long time is [0.03 - 0.1). The failure probability threshold area for the traction machine whose condition is not suitable for long-term continuous operation is [0.1 - 0.3]. When the failure probability threshold area is greater than 0.3, it means that the traction machine may be damaged at any time;

[0019] Calculate the failure probability P of the traction machine reflected by each index according to the following formula n :

[0020]

[0021] P n,average = average(P mean , P max , P rms , P kurt ),

[0022] where x nis the time-domain index value of the vibration signal of the traction machine. r0 - r4 are vibration thresholds selected according to the ISO-10816 vibration monitoring and evaluation standard. P n is the failure probability of the traction machine reflected by each time-domain index. P n,average is the overall failure probability of the traction machine reflected by the time-domain index. P mean , P max , P rms , P kurt are the failure probabilities calculated respectively according to the mean value, maximum value, root mean square value, and kurtosis value of the vibration signal.

[0023] Further, the failure probability of the traction machine within a period of time is diagnosed through the interval where the failure probability value is located, and the overall failure probability of the traction machine can intuitively reflect the overall health condition of the traction machine;

[0024] The safety status of the traction machine corresponding to the failure probability is shown in the following table:

[0025]

[0026]

[0027] Further, use the fault tree to quantitatively analyze the failure probability of the traction machine:

[0028] Since the occurrence probability of the bottom event is relatively low, and the occurrence probabilities of higher-order multi-bottom events are all below the minimum value, the minimum cut set independent approximation method is used to calculate the top event to reduce the calculation amount.

[0029] Use the minimum cut set independent approximation method to calculate the probability of the top event of the elevator overshooting accident:

[0030]

[0031] In the formula, P(T) is the probability value of the top event T, and K i is the i-th minimum cut set.

[0032] A dynamic prediction system for elevator overshooting accidents based on a fault tree includes a vibration sensor installed on the traction machine. The vibration sensor is connected to a fault tree analysis module through a data acquisition module and a data conversion module. The fault tree analysis module is connected to an intelligent operation and maintenance cloud platform through a wireless transmission module;

[0033] The data acquisition module is used to control the operation of the vibration sensor and send the vibration sensor data to the data conversion module;

[0034] The data conversion module is used to convert the received vibration sensor data of the analog signal into a digital signal and send it to the fault tree analysis module;

[0035] The fault tree analysis module classifies the bottom events of the fault tree into traction machine failures and other bottom events according to the fault tree. For traction machine failures, the failure probability is converted based on the vibration sensor data, and quantitative analysis is performed synchronously according to the failure probability. The failure probability of the overshoot accident and the key indicators of probability importance are output and uploaded to the intelligent operation and maintenance cloud platform. For other bottom events, through the fuzzy comprehensive evaluation theory, expert judgment language is collected, expert opinions are aggregated, and the fuzzy failure probability is calculated. Finally, through the characteristics of the overshoot fault tree, quantitative analysis is performed synchronously to dynamically reflect the probability of the top event occurring and some importance risk indicators.

[0036] The intelligent operation and maintenance cloud platform is used to perform dynamic risk assessment and display on the data sent by the fault tree analysis module.

[0037] An electronic device, characterized in that it includes:

[0038] A processor;

[0039] A memory for storing executable instructions of the processor;

[0040] Wherein, the processor is configured to execute the instructions to implement a dynamic prediction method for elevator overshoot accidents based on a fault tree.

[0041] A computer-readable storage medium, when the instructions in the computer-readable storage medium are executed by the processor of the electronic device, enables the electronic device to execute a dynamic prediction method for elevator overshoot accidents based on a fault tree.

[0042] Beneficial effects: The present invention constructs a fault tree for elevator overshoot accidents, collects the original signals of key components in real time for frequency domain feature extraction, converts the frequency domain indicators reflecting the fault characteristics of key components into failure probabilities, and constructs a bridge for the traction machine spectrum data acquisition and fault tree analysis system. Using the edge segment data acquisition device to set up a fault tree analysis module, quantitative analysis is performed synchronously according to the failure probability, which is real-time converted into a fault probability. At the same time, according to the quantification of the fault tree analysis, important indicators such as the probability of elevator overshoot accidents occurring are output, and the risk is evaluated automatically, solving the technical problems of fault prediction and diagnosis and dynamic risk assessment of elevator overshoot accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a flow block diagram of the dynamic prediction system for elevator overshoot accidents based on a fault tree of the present invention.

[0044] Figure 2 It is a structural block diagram of the dynamic prediction system for elevator overshoot accidents based on a fault tree of the present invention. DETAILED DESCRIPTION

[0045] The following further describes the present invention with reference to the accompanying drawings:

[0046] As Figure 1 shown, a dynamic prediction method for elevator overshoot accidents based on a fault tree of the present invention includes the following steps:

[0047] 1. Taking the elevator overshoot event as the top event, construct an elevator overshoot accident fault tree from top to bottom. A total of 23 key basic event causes are excavated. Among them, the basic event of traction machine failure has the highest importance. It is necessary to monitor the original vibration information of the traction machine and then perform failure probability conversion;

[0048] 2. Real-time collect the original signal of the traction machine for frequency-domain feature extraction, and perform failure probability conversion on the frequency-domain indicators reflecting the fault characteristics of the traction machine; for the remaining 22 basic events, directly through the fuzzy comprehensive evaluation theory, collect expert judgment language, aggregate expert opinions, and calculate the fuzzy failure probability;

[0049] 3. Construct a dynamic prediction system for elevator overshoot accidents based on the fault tree. According to the extraction, failure probability conversion and quantitative analysis of the fault tree of the time-domain indicators of the traction machine, upload the results to the intelligent operation and maintenance cloud platform, and combine the quantitative analysis of the fault tree to output the probability index of elevator overshoot accidents, and conduct dynamic risk assessment and display.

[0050] Construction of the elevator overshoot accident fault tree:

[0051] Taking the elevator overshoot event as the top event, construct an elevator overshoot accident fault tree from top to bottom. A total of 23 key basic event causes are excavated, which are specifically as follows:

[0052] The elevator overshoot accident includes two major categories: traction system failure and safety protection system failure.

[0053] Among them, the traction system failure includes insufficient traction force and brake failure; the reasons for insufficient traction force include: serious wear and diameter reduction of the traction steel wire rope, serious wear of the traction wheel groove, oil contamination between the traction wheel groove and the steel wire rope, excessive tension of the traction ropes on both sides of the traction wheel, and traction machine failure; the reasons for brake failure include brake circuit failure, brake arm jamming, and too little friction between the brake wheel and the brake shoe. The reasons for too little friction between the brake wheel and the brake shoe include serious wear of the brake wheel and the brake shoe, too large gap between the brake wheel and the brake shoe, oil contamination on the surface of the brake wheel, and too loose adjustment of the brake spring;

[0054] Among them, the failures of the safety protection system include the failures of the speed limiter and the safety gear. The factors leading to the speed limiter failure include the short circuit of the speed limiter safety circuit, the breakage of the speed limiter wire rope, and the insufficient formation of the speed limiter wire rope lift. The reasons for the insufficient formation of the speed limiter wire rope lift include the wear of the tension pulley groove, the loosening of the nut in the speed adjustment part, the damage of the speed limiter spring due to long-term expansion and contraction, and the existence of errors on the part surface; the factors of the safety gear failure include the short circuit of the safety gear safety circuit, incorrect installation and improper adjustment, and the inability of the two-side safety gears to work properly. The reasons for the inability of the two-side safety gears to work properly include the excessive clearance of the safety wedge block, the grinding flat of the safety wedge opening, the inability of the two-side wedge blocks to act synchronously, and the presence of oil stains on the surface of the slider / wedge block.

[0055] Among them, the structure importance degree of the traction machine failure in this fault tree is 0.0615, ranking first among all bottom events. That is, in terms of structure, the traction machine failure has the greatest impact on the occurrence of the top event.

[0056] Once the traction machine fails, the probability of the elevator running over the top will increase greatly. Therefore, its monitoring and diagnosis are particularly important.

[0057] Calculation method for the failure probability and performance evaluation method of elevator traction machine:

[0058] Extract the time-domain characteristics of the collected original vibration signal, including but not limited to the average value of the vibration signal (x mean ), maximum value (x max ), root mean square value (x rms ) and kurtosis value (x kurt ):

[0059]

[0060] x max =Max(X),

[0061]

[0062]

[0063] In the formula: x i represents the vibration value, N represents the number of points for each sampling, and σ represents the standard deviation;

[0064] According to the ISO-10816 vibration monitoring and evaluation standard, it is known that the safe state of the traction machine is divided into four intervals.

[0065] Combined with expert experience and individual fuzzy opinion aggregation, the steps are as follows: Invite 3 experts in the industry to evaluate the basic events of elevator overshoot accidents using 5 linguistic values. The evaluation languages include: large, relatively large, medium, relatively small, and small, which are used to represent the influence degree of accident-causing factors on the occurrence of elevator overshoot accidents respectively. When using natural language for uncertain description, it is necessary to quantitatively express it using a fuzzy membership function. Since the trapezoidal fuzzy function has the property of a relatively wide distance and the calculation process is simple and efficient, the trapezoidal fuzzy membership function is used to map the expert judgment language. The expression is:

[0066]

[0067] f(x) = 1, b < x ≤ c

[0068]

[0069] f(x) = 0, otherwise

[0070] In the formula, a and b respectively represent the upper and lower bounds of the trapezoidal fuzzy number, and the region [b, c] represents the median value of the trapezoidal fuzzy number.

[0071] Regarding the views on the same thing, due to the similarities or differences in their experiences or backgrounds, different experts often have conflicting or consistent judgments. However, due to the fuzzy and difficult-to-quantify nature of the elevator overshoot accident itself, it is impossible to prove whether the expert judgments are correct or not. The algorithm proposed by Hus and Chen is a method to aggregate individual fuzzy opinions into a group of fuzzy consistent opinions, which can eliminate the above-mentioned contradictions in expert judgments and reduce subjectivity. Therefore, this algorithm will be adopted to aggregate the expert opinions, and the steps are as follows:

[0072] 1) Determine the similarity

[0073]

[0074] In the formula, Z k and Z y are the evaluation languages of the kth and yth experts respectively; q is a fuzzy number; S(Z k , Z y ) is the similarity between the evaluation languages of the two experts, and its value range is [0, 1].

[0075] 2) Determine the average degree of agreement

[0076]

[0077] In the formula, n is the total number of experts; B k is the average degree of agreement among the experts

[0078] 3) Determine the relative degree of agreement

[0079]

[0080] 4) Determine the aggregation weight

[0081] Since the experiences and backgrounds of the experts making the judgments vary, the professionalism and authority of the judgments on the elevator overshoot events also differ. Different weight values need to be assigned to each expert. By assigning different values to the experts in terms of professional title, education level, and years of service, the weight values of different experts are given.

[0082] 5) The calculation formula for the aggregation weight is

[0083]

[0084] In the formula, β is the relaxation factor, which reflects whether to emphasize individual opinions or group opinions, and β ∈ [0, 1].

[0085] 6) Summarize the expert opinions and determine the fuzzy numbers of the basic events

[0086] To facilitate the synthesis operation of the fuzzy numbers, the α - cut set theory in fuzzy sets is applied to convert the fuzzy numbers into interval numbers for calculation. The α - cut set interval number calculation formula for the trapezoidal fuzzy number F α is

[0087] F α = [a + (b - a)α, d - (d - c)α]

[0088] The specific cases of the fuzzy number form and the α - cut set are shown in the following table.

[0089] Fuzzy language Fuzzy number form α-cut set Small (0.1，0.2，0.2，0.3) [0.1α+0.1，-0.1α+0.3] Relatively small (0.2，0.3，0.4，0.5) [0.1α+0.2，-0.1α+0.5] Medium (0.4，0.5，0.5，0.6) [0.1α+0.4，-0.1α+0.6] Relatively large (0.5，0.6，0.7，0.8) [0.1α+0.5，-0.1α+0.8] Large (0.7，0.8，0.8，0.9) [0.1α+0.7，-0.1α+0.9]

[0090] Then the average fuzzy number of the expert group evaluation under the α - cut set is

[0091]

[0092] 7) Calculate the fuzzy possibility value

[0093] After quantifying the judgment language made by the experts, it is still a fuzzy number and still has uncertainty. Therefore, the left - right fuzzy ranking method is used to convert it into the fuzzy possibility value F ps . First, the fuzzy maximum set and the minimum set need to be obtained, which are respectively

[0094] f max (x) = x, 0 ≤ x ≤ 1

[0095] f max (x) = 0, otherwise

[0096] f minf(x) = 1 - x, 0 ≤ x ≤ 1

[0097] f min f(x) = 0, otherwise

[0098] Then the left and right fuzzy possibility degrees are respectively

[0099] F PS，R (D) = sup[f D (x) ∧ f max (x)]

[0100] F PS，L (D) = sup[f D (x) ∧ f min (x)]

[0101] The fuzzy possible value is

[0102]

[0103] 8) Calculate the fuzzy failure probability.

[0104] The finally calculated fuzzy failure probability is an exact probability value, which is calculated through the Onisawa formula, that is

[0105] F PS，T ≠0

[0106] F FR = 0, F PS,T = 0

[0107]

[0108] Risk analysis of elevator overshoot accident based on fuzzy fault tree:

[0109] Industry experts' opinions are used to identify the risk factors that may cause elevator overshoot accidents, and the final determined risk factors. The bottom events of elevator overshoot accidents are independent of each other and have only two states: occurrence and non-occurrence. Then, according to the established fault tree, the fuzzy set theory is introduced to analyze the elevator overshoot accident as follows.

[0110] 1) Qualitative analysis

[0111] Structure importance analysis related only to the fault tree structure:

[0112]

[0113] In the formula, k represents the total number of minimum cut sets, K j represents the jth minimum cut set, and N j represents the number of basic events in the jth minimum cut set.

[0114] Let the structural importance of the basic event \(X_1\), \(I_{\varphi}(1)=0.123\), and so on. The order of structural importance is as follows:

[0115] \(I_{\varphi}(1)=I_{\varphi}(2)=I_{\varphi}(3)=I_{\varphi}(4)=0.123\gt\)

[0116] \(I_{\varphi}(5)=I_{\varphi}(6)=I_{\varphi}(7)=I_{\varphi}(8)=I_{\varphi}(9)=I_{\varphi}(20)=0.0293\gt\)

[0117] \(I_{\varphi}(10)=I_{\varphi}(11)=I_{\varphi}(12)=I_{\varphi}(13)=I_{\varphi}(14)=I_{\varphi}(15)=I_{\varphi}(16)=I_{\varphi}(17)=I_{\varphi}(18)=I_{\varphi}(19)\)

[0118] \(=I_{\varphi}(21)=I_{\varphi}(22)=0.000451\). From the above sorting, it can be seen that the basic events \(X_1 - X_4\) have the greatest impact on the occurrence of the elevator overshoot accident, followed by \(X_5 - X_9\), \(X_{20}\). In order to reduce the risk of elevator overshoot, countermeasures should be proposed specifically for the above factors.

[0119] 2) Fuzzy probability solution

[0120] First, calculate the failure probability of each basic event. Taking the serious wear of the traction steel wire rope as an example for introduction. Obtain the evaluation language from three industry experts involved in the project construction. The evaluation languages of the three experts for this event are respectively "smaller", "small", and "medium". Calculate the average consistency degree of the three experts' opinions \(B_1 = 0.7825\), \(B_2 = 0.675\), \(B_3 = 0.6875\), the relative consistency degree \(R_1 = 0.3663\), \(R_2 = 0.3140\), \(R_3 = 0.3198\); the ability weights of the three experts \(\varphi_1 = 0.348\), \(\varphi_2 = 0.370\), \(\varphi_3 = 0.283\). Aggregate the evaluation opinions. Take \(\beta = 0.5\) in the calculation formula, indicating that personal opinions and group opinions are equally valued. Then \(W_1 = 0.3572\), \(W_2 = 0.342\), \(W_3 = 0.301\). Then, according to the formula, obtain the left and right possibility values \(F\) PS，R(W) \(= 0.4179\), \(F\) PS，L(W) \(= 0.724\), \(F\) PS，T(W) \(= 0.347\). Finally, defuzzify to obtain the fuzzy number \(F\) FR \(= 0.0014\), that is, the probability of serious wear and diameter reduction of the traction steel wire rope is \(0.0014\).

[0121] Calculate the failure probabilities of all fuzzy events obtained successively through the above process

[0122]

[0123] Take the maximum value. The maximum failure probability of the newly put - into - use traction machine is 0.03. Therefore, the upper threshold of the newly put - into - use traction machine is set to be less than 0.03. Similarly, according to the failure probability calculated by expert evaluation, the failure threshold area where the traction machine can operate without restrictions for a long time is [0.03 - 0.1), the failure probability threshold area where the condition of the traction machine is not suitable for long - term continuous operation is [0.1 - 0.3], and when the failure probability threshold is greater than 0.3, it means that the traction machine may be damaged at any time;

[0124] The failure probability \(P\) of the traction machine reflected by each index is calculated according to the following formula n :

[0125]

[0126] \(P\) n,average = average(\(P\) mean , \(P\) max , \(P\) rms , \(P\) kurt ),

[0127] where \(x\) n is the time - domain index value of the vibration signal of the traction machine, \(r0 - r4\) are the vibration thresholds selected according to the ISO - 10816 vibration monitoring and evaluation standard, \(P\) n is the failure probability of the traction machine reflected by each time - domain index, \(P\) n,average is the overall failure probability of the traction machine reflected by the time - domain index, \(P\) mean , \(P\) max , \(P\) rms , \(P\) kurt are the failure probabilities calculated according to the mean value, maximum value, root mean square value, and kurtosis value of the vibration signal respectively.

[0128] Each time - domain index can reflect the safety state of the traction machine: the mean value and root mean square value of the vibration signal can reflect the vibration amplitude of the traction machine; the maximum value and kurtosis value can reflect the vibration impact of the traction machine. The failure probability of the traction machine within a certain period can be diagnosed through each time - domain index value, and the overall failure probability of the traction machine can directly reflect the overall health condition of the traction machine. The safety state of the traction machine corresponding to the failure probability is shown in the following table.

[0129]

[0130]

[0131] Dynamic Prediction System for Elevator Over - running Accidents

[0132] The architecture of the dynamic prediction and diagnosis system for elevator over - running accidents is as Figure 2As shown in the figure. The vibration sensor is connected to the edge - end data acquisition device. When the edge - end data acquisition device collects the vibration sensor signal, after calculating the failure probability through the built - in module, and then through the quantitative analysis algorithm of the fault tree for analysis and processing, the prediction result is uploaded to the cloud platform.

[0133] Quantitative analysis of the fault tree:

[0134] There are no repeated bottom events in the fault tree of the elevator over - running accident, which means that the minimum cut sets are independent of each other and do not contain the same bottom events.

[0135] When accurately calculating the occurrence probability of the top event of the elevator over - running accident, it needs to be expanded according to the probability formula of logical union in Boolean algebra (i.e., the inclusion - exclusion theorem). However, due to the large number of minimum cut sets generated by the fault tree of the elevator over - running accident, if the inclusion - exclusion theorem is used for accurate calculation, the number of calculation terms is prone to the "combinatorial explosion" phenomenon, and the calculation amount is huge;

[0136] Since the occurrence probability of the event is relatively low, and the occurrence probabilities of high - order multi - events are all below the minimum value, a good result can be obtained by using the independent approximation calculation of the minimum cut sets for the top event.

[0137] Using the independent approximation calculation of the minimum cut sets to calculate the probability of the top event of the elevator over - running accident:

[0138]

[0139] In the formula, P(T) is the probability value of the top event T, and Ki i is the i - th minimum cut set, and there are a total of k minimum cut sets.

[0140] As Figure 2 shown, a dynamic prediction system for elevator over - running accidents based on a fault tree includes a vibration sensor set on the traction machine. The vibration sensor is connected to a fault tree analysis module through a data acquisition module and a data conversion module, and the fault tree analysis module is connected to an intelligent operation and maintenance cloud platform through a wireless transmission module;

[0141] The data acquisition module is used to control the operation of the vibration sensor and send the vibration sensor data to the data conversion module;

[0142] The data conversion module is used to convert the received vibration sensor data of the analog signal into a digital signal and send it to the fault tree analysis module;

[0143] The fault tree analysis module divides the bottom events of the fault tree into traction machine failures and other bottom events according to the fault tree. For traction machine failures, the failure probability conversion is performed based on the vibration sensor data, and quantitative analysis is synchronously carried out according to the failure probability. The failure probability and probability importance key indicators of the overshoot accident are output and uploaded to the intelligent operation and maintenance cloud platform. For other bottom events, through the fuzzy comprehensive evaluation theory, expert judgment language is collected, expert opinions are aggregated, and the fuzzy failure probability is calculated. Finally, through the characteristics of the overshoot fault tree, quantitative analysis is synchronously carried out to dynamically reflect the probability of the top event occurring and some importance risk indicators.

[0144] The intelligent operation and maintenance cloud platform is used to perform dynamic risk assessment and display on the data sent by the fault tree analysis module.

Claims

1. A dynamic prediction method for elevator overshoot accidents based on fault tree, characterized in that, The specific steps are as follows: S1. Collect information on elevator overshoot accidents and faults, and use the collected information to construct a fault tree with the elevator overshoot accident as the top event; it is obtained that the fault tree of the elevator overshoot accident includes a total of 23 common fault tree bottom events; S2. Confirm through the structural importance ranking that the importance of the elevator traction machine failure among the bottom events is the greatest, so select to monitor the original vibration information of the traction machine; S3. Extract the frequency-domain characteristics of the original vibration signal of the traction machine collected in real time by setting sensors on the outer bearing seat of the elevator traction machine, convert the frequency-domain indicators reflecting the fault characteristics of the traction machine into failure probabilities, and calculate the fuzzy failure probabilities of the other 22 common fault tree bottom events through the fuzzy comprehensive evaluation theory; S4. Use the edge data acquisition device set at the elevator to extract the time-domain index information of the traction machine, convert the failure probability, and conduct quantitative analysis of the fault tree, and then upload the results to the intelligent operation and maintenance cloud platform for dynamic risk assessment and display.

2. The dynamic prediction method for elevator overshoot accidents based on a fault tree according to claim 1, characterized in that: The calculation method of the failure probability of the elevator traction machine is as follows: The original vibration signal of the traction machine collected is used to calculate the time-domain feature extraction by the following formula, including four time-domain index values of the traction machine: the mean value x of the vibration signal mean , the maximum value x of the vibration signal max , the root mean square value x of the vibration signal rms and the kurtosis value x of the vibration signal kurt : The magnitude of the vibration amplitude of the traction machine can be reflected by the average value and the root mean square value of the vibration signal; the magnitude of the vibration impact of the traction machine can be reflected by the maximum value and the kurtosis value; x max = Max(X), where: x i represents the vibration value, N represents the number of points for each sampling, and σ represents the standard deviation.

3. The dynamic prediction method for elevator overshoot accidents based on a fault tree according to claim 2, wherein, The performance evaluation method of the elevator traction machine is specifically as follows: According to the ISO-10816 vibration monitoring and evaluation standard, it is known that the safe state of the traction machine is divided into four intervals. Convert the time-domain index values reflecting the performance of the traction machine into failure probabilities. Combine expert experience and individual fuzzy opinion aggregation to obtain that the failure probability threshold area for the newly put into use of the traction machine is less than 0.03, the failure threshold area for the traction machine running without restriction for a long time is [0.03 - 0.1), the failure probability threshold area for the traction machine whose condition is not suitable for long-term continuous operation is [0.1 - 0.3), and when the failure probability threshold area is greater than 0.3, it means that the traction machine may be damaged at any time; The failure probability P of the traction machine is calculated according to the following formula based on various indicators n : P n,average = average(P mean , P max , P rms , P kurt ), where x n is the time-domain index value of the vibration signal of the traction machine, and r0 - r4 are the vibration thresholds selected according to the ISO-10816 vibration monitoring and evaluation standard, P n is the failure probability of the traction machine reflected by each time-domain index, P n,average is the overall failure probability of the traction machine reflected by the time-domain index, P mean , P max , P rms , P kurt are the failure probabilities calculated based on the mean value, maximum value, root mean square value, and kurtosis value of the vibration signal, respectively.

4. The dynamic prediction method for elevator overshoot accidents based on a fault tree according to claim 3, characterized in that: Diagnose the failure probability of the traction machine through the interval where the failure probability value is located, and the overall failure probability of the traction machine can intuitively reflect the overall health of the traction machine; The safe state of the traction machine corresponding to the failure probability is shown in the following table:

5. The dynamic prediction method for elevator overshoot accidents based on a fault tree according to claim 1, characterized in that: Quantitatively analyze the failure probability of the traction machine using the fault tree: Since the occurrence probability of the bottom event is relatively low, and the occurrence probabilities of higher-order multi-bottom events are all below the minimum value, the minimum cut set independent approximation is used to calculate the top event to reduce the calculation amount, Adopt the minimum cut set independent approximation to calculate the probability of the top event of the elevator overshoot accident: Where P(T) is the probability value of the top event T, and K i is the i-th minimal cut set.

6. A prediction system for the dynamic prediction method of elevator overshoot accidents based on a fault tree as described in any one of claims 1-5, characterized in that: It includes vibration sensors set on the traction machine. The vibration sensors are connected to a fault tree analysis module through a data acquisition module and a data conversion module. The fault tree analysis module is connected to an intelligent operation and maintenance cloud platform through a wireless transmission module; The data acquisition module is used to control the operation of the vibration sensors and send the vibration sensor data to the data conversion module; The data conversion module is used to convert the received vibration sensor data of the analog signal into a digital signal and send it to the fault tree analysis module; The fault tree analysis module classifies the bottom events of the fault tree into traction machine failures and other bottom events according to the fault tree. For traction machine failures, the failure probability is converted based on the vibration sensor data, and quantitative analysis is performed synchronously according to the failure probability. The failure probability of the overshoot accident and the key indicators of probability importance are output and uploaded to the intelligent operation and maintenance cloud platform. For other bottom events, through the fuzzy comprehensive evaluation theory, expert judgment languages are collected, expert opinions are aggregated, and the fuzzy failure probability is calculated. Finally, through the characteristics of the overshoot fault tree, quantitative analysis is performed synchronously to dynamically reflect the probability of the top event occurring and some importance risk indicators. The intelligent operation and maintenance cloud platform is used to perform dynamic risk assessment and display on the data sent by the fault tree analysis module.

7. An electronic device, characterized in that, It includes: A processor; A memory for storing executable instructions of the processor; Wherein, the processor is configured to execute the instructions to implement the dynamic prediction method for elevator overshoot accidents based on the fault tree as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device can execute the dynamic prediction method for elevator overshoot accidents based on the fault tree as described in any one of claims 1 to 5.

Images