Evaluation method for failure mode of tank system based on interval intuitionistic fuzzy rough number
By using the interval intuitionistic fuzzy rough number method to process the failure modes of liquefied natural gas storage tank systems, this method solves the problems of insufficient objective reflection of risk factor weights and insufficient consistency of expert evaluation in traditional FMEA methods, and achieves more accurate risk ranking and identification.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- LIAONING UNIVERSITY OF PETROLEUM AND CHEMICAL TECHNOLOGY
- Filing Date
- 2026-04-21
- Publication Date
- 2026-07-14
AI Technical Summary
Existing Failure Mode and Effects Analysis (FMEA) methods for liquefied natural gas storage tank systems suffer from problems such as insufficient objective reflection of risk factor weights, difficulty in expressing ambiguity and uncertainty, and insufficient consistency in expert evaluations, leading to unstable and inaccurate risk ranking.
We adopt a method based on interval intuitionistic fuzzy rough numbers. We convert expert language evaluations into interval intuitionistic fuzzy numbers through interval intuitionistic fuzzy sets IVIFNS, determine expert weights by combining an improved consensus mechanism, calculate objective weights using the CRITIC method, construct a combined weighting model by the minimum deviation principle of game theory, and rank risk priorities by combining the ExpTODIM-PROMETHEEII method.
It effectively characterizes the ambiguity and uncertainty in expert evaluation, enhances the consistency of group decision-making, improves the coordination and robustness of risk factor weights, enhances the ability to distinguish failure mode risks, and provides a reliable risk priority ranking.
Smart Images
Figure CN122389604A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of storage and transportation safety risk assessment technology, and in particular to a failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy rough numbers. Background Technology
[0002] Liquefied natural gas (LNG), as a clean, efficient, low-carbon, and environmentally friendly energy source, has been widely adopted under the optimization and adjustment of the energy structure. As LNG receiving terminals, storage tanks, and their transmission and distribution systems continue to develop towards larger scale, integration, and complexity, the systems exhibit high-risk, highly coupled, and highly uncertain safety characteristics during operation. Once a storage tank system experiences leakage, seal failure, valve malfunction, or related auxiliary facility failure, it will not only affect the safe and stable operation of the facility but may also lead to serious consequences such as fires, explosions, and environmental pollution. Therefore, conducting scientific and accurate risk assessments of LNG storage tank systems is of great significance.
[0003] Existing Failure Mode and Effects Analysis (FMEA) methods typically evaluate and rank failure modes using risk priority numbers, which has a certain application basis in engineering risk identification. However, traditional FMEA methods still have several shortcomings in practical applications: First, risk factor weights are usually weighted equally, making it difficult to objectively reflect the importance of different evaluation indicators; second, the product form of risk priority numbers can easily lead to different failure modes having the same ranking value, thus weakening the ranking's ability to distinguish between them; third, expert evaluation information is often ambiguous, hesitant, and uncertain, and traditional precise numerical methods cannot fully express expert cognition; fourth, in group decision-making scenarios, existing methods do not adequately consider the level of consistency and consensus among experts, which can easily affect the reliability and robustness of the evaluation results.
[0004] While methods such as fuzzy FMEA, multi-attribute decision-making, and combined weighting have improved risk assessment capabilities to some extent, they still suffer from limitations in characterizing fuzzy and hesitant information, poor coordination between subjective and objective weights, and limited interpretability of ranking. Particularly in high-risk and complex projects like LNG storage tank systems, how to simultaneously handle the fuzzy uncertainties in expert evaluations, balance subjective and objective information, and achieve stable identification of failure mode risk priorities remains a key technical challenge. Summary of the Invention
[0005] The purpose of this invention is to provide a failure mode evaluation method for tank systems based on interval intuitionistic fuzzy roughness numbers, aiming to solve or improve at least one of the above-mentioned technical problems.
[0006] To achieve the above objectives, the present invention provides the following solution: A failure mode evaluation method for tank systems based on interval intuitionistic fuzzy rough numbers includes: Step 1: Using the interval intuitionistic fuzzy set IVIFNS, the linguistic evaluation information of experts on each failure mode and risk factor of the liquefied natural gas storage tank system is converted into the corresponding interval intuitionistic fuzzy representation. The expert weights are determined through an improved consensus mechanism, and the evaluation results of each expert are aggregated to obtain the failure mode evaluation matrix represented in the form of interval intuitionistic fuzzy rough number IVIFRN. Step 2: Based on the failure mode evaluation matrix, the subjective weights of risk factors are calculated using the interval intuitionistic fuzzy rough weighted geometric operator, the objective weights of risk factors are calculated using the CRITIC method, and a combined weighting model is constructed by combining the game theory minimum deviation principle to obtain the comprehensive weights of each risk evaluation index. Step 3: Based on the combined weights of the failure mode evaluation matrix and risk factors, the ExpTODIM-PROMETHEEII method is used to rank the risk priorities of each failure mode and output the risk priority sequence of the failure modes.
[0007] Further, step 1 includes: Construct a language scale based on interval intuitionistic fuzzy sets IVIFNS; Obtain expert language evaluations, convert the language evaluation results into corresponding interval intuition fuzzy numbers based on the language scale, and generate an initial evaluation matrix; The expert weights are calculated based on the similarity and consistency levels among the language evaluation results of each expert, and an expert weight vector is generated. The similarity aggregation method SAM is used to achieve consistency fusion of expert evaluations, generate a failure mode evaluation matrix, and represent it in the form of interval intuitionistic fuzzy rough number IVIFRN.
[0008] Furthermore, expert language evaluations are obtained, and the language evaluation results are converted into corresponding interval intuitionistic fuzzy numbers based on the language scale to generate an initial evaluation matrix, including: The set of expert language evaluations is represented in the form of an IVIFN set as follows: In the formula, For language evaluation Interval intuitive fuzzy numbers; This is the lower bound of the membership degree; This is the upper bound of the membership degree; This is the lower bound of the non-membership degree; This is the upper bound of the non-membership degree; but The lower approximation Approximate with the above It can be determined as follows: In the formula, The following approximation is used; It is the set of all objects in the language evaluation set; , is the lower union of the upper approximation; It is the lower union of the upper approximation; Improved IVIFRN lower limit and upper limit The expression is: In the formula, For the improved IVIFRN lower limit; To be downward and concentrated, satisfying the condition The total number of language assessments; For the evaluation of language involved in the calculation; For the improved IVIFRN upper limit; To be upward and concentrated, satisfying the condition The total number of language assessments; For language evaluation IVIFRN value; Then the interval value intuition fuzzy class Improved IVIFRN for: In the formula, The resulting improved interval is intuitively fuzzy and coarse. , , , These are the arithmetic mean of the components corresponding to the lower and upper limits, respectively.
[0009] Furthermore, the similarity aggregation method SAM is used to achieve consistency fusion of expert evaluations, generating a failure mode evaluation matrix, which is expressed in the form of interval intuitionistic fuzzy rough number IVIFRN, including: The similarity function is expressed as: In the formula, Let be the similarity between the evaluation results of the u-th expert and the v-th expert; These are the interval intuitionistic fuzzy rough number evaluations of a certain failure mode by experts u and v, respectively. These are the comprehensive eigenvalues of IVIFRN; According to similarity calculation experts Weighted consistency The expression is: In the formula, For experts The weighted consistency; Let v be the initial weight for the expert. The number of experts participating in the evaluation; According to weighted consistency calculation experts relative consistency The expression is: In the formula, For experts The degree of relative consistency; Total number of experts; The expert consensus coefficient is generated by weighting and summing the weighted consistency and relative consistency. The expression is: In the formula, This represents the expert consensus coefficient. For the weighting factor; The interval intuitionistic fuzzy rough weighted average operator is expressed as follows: In the formula, An interval-valued intuitive fuzzy coarse weighted average operator; The number of evaluation objects participating in the aggregation represents the number of experts; Let be the weight vector of the j-th expert; Based on the above formula, the summary result of expert opinions can be expressed as follows: In the formula, Let j be the consensus coefficient of the expert. The j-th expert is the IVIFRN evaluation of the failure mode.
[0010] Further, step 2 includes: Based on the failure mode evaluation matrix, the IVIFRWG operator is used to fuse multiple expert evaluations to construct a risk factor decision matrix. The interval intuitionistic fuzzy rough weighted geometric operator IVIFRWG is used to subjectively assign weights to risk factors, resulting in a subjective weight vector; The contrast strength and conflict of each risk factor are calculated using the CRITIC method, and the objective weight vector of each risk factor is obtained based on the contrast strength and conflict of each risk factor. Based on the principle of minimum deviation in game theory, the subjective weight vector and the objective weight vector are combined to obtain the comprehensive weight of the risk factors.
[0011] Furthermore, the interval intuitionistic fuzzy rough weighted geometric operator IVIFRWG is used to subjectively weight the risk factors, resulting in a subjective weight vector, including: IVIFRN To perform deblurring, the expression is: In the formula, The value is the defuzzified value; IR(S) is the interval intuitionistic fuzzy rough number. This is the composite value of the membership interval; This is the composite value of the non-membership interval; After defuzzing, the subjective weight vector of the risk operator is calculated, and the expression is: In the formula, Let be the subjective weight of the j-th risk factor; Let be the comprehensive evaluation value of the j-th risk factor; The IVIFRN corresponding to the k-th risk factor; is the score of the k-th risk factor after defuzzification; n is the total number of risk factors.
[0012] Furthermore, the CRITIC method is used to calculate the contrast strength and conflict of each risk factor, and the objective weight vector of each risk factor is obtained based on the contrast strength and conflict of each risk factor, including: The risk factor decision matrix is normalized column-wise to obtain a standardized matrix, expressed as follows: In the formula, These are the elements in the standardized matrix; Let be an element in the risk factor decision matrix, representing the score of the i-th risk factor on the j-th failure mode; i is the risk factor index; j is the failure mode index; The minimum value of the j-th failure mode among all risk factors; Let be the maximum value of the j-th failure mode among all risk factors; m be the total number of risk factors; and n be the total number of failure modes. Based on the standardized matrix, the contrast strength of the risk factors is calculated using the following expression: In the formula, Let be the contrast intensity of the j-th risk factor; The standardized arithmetic mean of the j-th failure mode; Calculate the correlation coefficient matrix among the failure factors. ,in The Pearson correlation coefficient between risk factor j and risk factor k is expressed as: In the formula, Let be the Pearson correlation coefficient between the j-th risk factor and the k-th risk factor; The conflict of risk factors is calculated based on the Pearson correlation coefficient, expressed as follows: In the formula, The conflict of the j-th risk factor; The information content of risk factors is generated by comparing their contrast strength and conflict, expressed as follows: In the formula, Let j be the information content of the j-th risk factor; Based on the information content of the risk factors, the objective weights of the risk factors are calculated, as expressed by: In the formula, The objective weights of risk factors.
[0013] Further, step 3 includes: Based on the comprehensive weights of the failure mode evaluation matrix and risk factors, the dominance of each failure mode relative to each risk factor is calculated. Calculate the overall dominance of each failure mode based on its dominance. The overall dominance of each failure mode is summarized, and the positive and negative currents of each failure mode are calculated. Calculate the net flow for each failure mode based on positive and negative flow; Based on the net current value of each failure mode, the failure modes are sorted to generate a risk priority sequence for each risk mode.
[0014] Furthermore, based on the failure mode evaluation matrix and the comprehensive weights of risk factors, the dominance of each failure mode relative to each risk factor is calculated, expressed as follows: In the formula, Failure mode relative to failure mode Dominance under risk factor j; Let be the overall weight of the j-th risk factor; These are the attenuation control parameters; The distance between two IVIFRN evaluations; This is the loss attenuation coefficient.
[0015] Furthermore, the overall dominance of each failure mode is calculated based on its dominance, expressed as follows: In the formula, Failure mode relative to failure mode The overall dominance; n is the total number of failure modes; Let represent the dominance of the j-th failure mode.
[0016] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: This invention discloses a failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy rough numbers. This method effectively characterizes the fuzziness, hesitation, and uncertainty in expert evaluation, enhancing the consistency of expression in group decision-making. It can comprehensively determine the weights of risk factors by integrating subjective experience and objective data, improving the coordination and robustness of the weighting results. It can improve the risk differentiation capability of failure modes through a non-compensatory ranking mechanism, and enhance the interpretability of the ranking results by utilizing positive flow, negative flow, and net flow. It is applicable to the identification and assessment of leakage risks and related failure modes in liquefied natural gas storage tank systems, providing support for risk warning, maintenance decisions, and safety management. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a schematic diagram of the overall process of the method of the present invention; Figure 2 This is a schematic diagram of failure mode identification in the leakage scenario of the storage tank system in this embodiment; Figure 3 This is a schematic diagram of the aggregated expert evaluation results for each failure mode under each risk factor in this embodiment; Figure 4 This is a diagram showing the comparison of the subjective weight, objective weight, and comprehensive weight of risk factors in this embodiment; Figure 5 This is a schematic diagram of the positive flow, negative flow, and net flow results of the failure modes in this embodiment; Figure 6 This is a schematic diagram showing the overall net flow and risk priority ranking results of each failure mode in this embodiment. Detailed Implementation
[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0020] The purpose of this invention is to provide a failure mode evaluation method for tank systems based on interval intuitionistic fuzzy roughness numbers, aiming to solve or improve at least one of the above-mentioned technical problems.
[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0022] like Figure 1 As shown, this invention provides a failure mode evaluation method for tank systems based on interval intuitionistic fuzzy roughness numbers, including: The relevant failure modes of liquefied natural gas (LNG) storage tank systems were obtained, and a risk assessment model for failure modes of LNG storage tank systems was established based on the failure modes as evaluation objects, focusing on the occurrence (O), severity (S), detectability (D), maintainability (M), and economic impact (P) of risk factors.
[0023] like Figure 2 As shown, based on the structural composition, operating conditions and accident mechanisms of the liquefied natural gas storage tank system, the failure modes are identified as severe mechanical failure FM1, auxiliary pipeline rupture FM2, design defect risk level FM3, tank overflow FM4, overpressure rupture FM5, natural disaster risk level FM6, and low-pressure rupture FM7.
[0024] Step 1: Using the interval intuitionistic fuzzy set IVIFNS, the linguistic evaluation information of experts regarding each failure mode and risk factor of the liquefied natural gas storage tank system is converted into corresponding interval intuitionistic fuzzy representations. An improved consensus mechanism is used to determine expert weights, and the evaluation results of each expert are aggregated to obtain a failure mode evaluation matrix represented in the form of interval intuitionistic fuzzy rough numbers IVIFRN, including: Step 11, construct a language scale based on interval intuitionistic fuzzy sets IVIFNS, including: There are nine linguistic variables: Very High (EH), Very High (VH), High (H), Relatively High (RH), Moderate (M), Relatively Low (RL), Low (L), Very Low (VL), and Very Low (EL). The correspondence between the linguistic variables of expert evaluation and IVIFNS is shown in Table 1.
[0025] Table 1. Correspondence between linguistic variables and IVIFNS
[0026] To unify the perceptions and expressions of risk levels of various failure modes in liquefied natural gas storage tank systems among different experts, a risk language evaluation set was pre-constructed, and language terms such as "extremely low", "relatively low", "medium", "relatively high" and "extremely high" were mapped to corresponding interval intuitive fuzzy numbers, so as to convert qualitative language evaluations into calculable quantitative information.
[0027] Step 12: Obtain expert language evaluations. Based on the language scale, convert the language evaluation results into corresponding interval intuitionistic fuzzy numbers to generate an initial evaluation matrix, including: Suppose we have a set Z of language evaluations from N experts, i.e. The order of the elements is ,in Represents a language evaluation in Z; Z is represented in the form of an IVIFN set as follows: In the formula, For language evaluation Interval intuitive fuzzy numbers; This is the lower bound of the membership degree; This is the upper bound of the membership degree; This is the lower bound of the non-membership degree; This is the upper bound of the non-membership degree; but The lower approximation Approximate with the above It can be determined as follows: In the formula, The following approximation includes all items that explicitly belong to the evaluation level. A collection of objects; It is the set of all objects in the set of all language evaluations; , is the lower union of the upper approximation, and represents the set All ratings less than or equal to Element; The lower union of the upper approximation, representing the set All ratings are greater than or equal to Element; Improved IVIFRN lower limit and upper limit The expression is: In the formula, To improve the IVIFRN lower bound, by aggregating less than or equal to All relevant information was calculated; To be downward and concentrated, satisfying the condition The total number of language assessments; For the evaluation of language involved in the calculation; To improve the IVIFRN upper bound, by aggregating greater than All relevant information was calculated; To be upward and concentrated, satisfying the condition The total number of language assessments; For language evaluation IVIFRN value; Then the interval value intuition fuzzy class Improved IVIFRN for: In the formula, To generate the final improved interval intuitionistic fuzzy rough number, the lower bound is used. and upper limit Common definition; , , , These are the arithmetic mean of the components corresponding to the lower and upper limits, respectively; In this embodiment, five experts, E1, E2, E3, E4 and E5, were invited to conduct linguistic evaluations of failure modes FM1-FM7 under various risk factors, including occurrence (O), severity (S), detectability (D), maintainability (M) and economic impact (P). The linguistic evaluation results were converted into corresponding interval intuition fuzzy numbers based on the linguistic scale, thereby forming the initial evaluation matrix for each expert.
[0028] Step 13: Calculate expert weights based on the similarity and consistency levels among the language evaluation results of each expert, and generate an expert weight vector, including: According to k experts The scores are shown in Table 2. The overall expert score is calculated. Table 2 Expert Weighting Criteria and Scores
[0029] The individual weight of the expert is calculated by normalizing the overall score, as shown in the expression: In the formula, Expert weighting; Total score for experts; Number the experts; The number of experts.
[0030] In this embodiment, considering the differences among experts in terms of engineering experience, professional background and consistency of judgment, the expert weights are calculated based on the similarity and consistency level between the evaluation results of each expert, resulting in expert weight vectors w1, w2, w3, w4, and w5, to reflect the degree of contribution of different experts in group decision-making.
[0031] Step 14: The similarity aggregation method SAM is used to perform consistency fusion of the expert evaluations, generating a failure mode evaluation matrix, which is expressed in the form of interval intuitionistic fuzzy rough number IVIFRN, including: The similarity function is expressed as: In the formula, The similarity between the evaluation results of the u-th expert and the v-th expert is denoted by a value between 0 and 1. The closer the value is to 1, the more consistent the opinions of the two experts are. These are the interval intuitionistic fuzzy roughness number (IVIFRN) evaluations of a certain failure mode by experts u and v, respectively. These are the comprehensive eigenvalues of IVIFRN; The above uses a variant of Euclidean distance to calculate the difference between two IVIFRNs. The smaller the difference, the higher the similarity.
[0032] According to similarity calculation experts Weighted consistency The expression is: In the formula, For experts The weighted consistency score measures the average similarity between expert t and all other experts, taking into account the weights of the other experts. Let v be the initial weight for the expert. The number of experts participating in the evaluation; According to weighted consistency calculation experts relative consistency The expression is: In the formula, For experts The degree of relative consistency; Total number of experts; The expert consensus coefficient is generated by weighting and summing the weighted consistency and relative consistency. The expression is: In the formula, This represents the expert consensus coefficient. The β value is a weighting factor that indicates the degree of importance that decision-makers attach to the weight of experts. If β=1, the decision-makers only consider the initial weight (status) of the experts and ignore the consensus of opinions. If β=0, the decision-makers only consider the consensus of the experts and ignore the initial status. The Interval Intuitive Fuzzy Rough Weighted Average (IVIFRWA) operator is expressed as follows: In the formula, An interval-valued intuitive fuzzy coarse weighted average operator; The number of evaluation objects participating in the aggregation represents the number of experts; Let be the weight vector of the j-th expert; Based on the above formula, the summary result of expert opinions can be expressed as follows: In the formula, Let j be the consensus coefficient of the expert. The j-th expert is the IVIFRN evaluation of the failure mode.
[0033] like Figure 3 As shown, after aggregation, an interval intuitionistic fuzzy rough number evaluation matrix is obtained for failure modes FM1-FM7 under five risk factors: occurrence (O), severity (S), detectability (D), maintainability (M), and economic impact (P). Each element in the evaluation matrix is represented in the form of interval intuitionistic fuzzy rough numbers, which can simultaneously characterize the fuzziness, hesitation, and interval uncertainty in expert evaluation, providing basic data for subsequent risk factor weight calculation and failure mode ranking.
[0034] Step 2: Based on the failure mode evaluation matrix, the subjective weights of risk factors are calculated using the interval intuitionistic fuzzy rough weighted geometric operator, and the objective weights of risk factors are calculated using the CRITIC method. A combined weighting model is then constructed using the game theory minimum deviation principle to obtain the comprehensive weights of each risk evaluation indicator, including: Step 21: Based on the failure mode evaluation matrix, use the IVIFRWG operator to fuse multiple expert evaluations and construct a risk factor decision matrix. The expression for the IVIFRWG operator is: In the formula, For interval intuitionistic fuzzy rough weighted geometric operators; This is the evaluation by the j-th expert; Let j be the weight of the expert. Let be the interval intuitionistic fuzzy rough number for the j-th expert evaluation; Step 22: Subjectively weight the risk factors using the Interval Intuitive Fuzzy Rough Weighted Geometric Operator (IVIFRWG) to obtain the subjective weight vector, including: IVIFRN To perform deblurring, the expression is: In the formula, The value is the defuzzified value. The larger the value, the more important the risk factor is in the eyes of the experts. IR(S) is the interval intuitionistic fuzzy rough number, which represents the mathematical expression of the expert's evaluation of a certain risk factor. This is the composite value of the membership degree interval, obtained by averaging the upper and lower limits of the original IVIFRN membership degree. This is the composite value of the non-membership intervals, obtained by averaging the upper and lower limits of the non-membership of the original IVIFRN. After defuzzing, the subjective weight vector of the risk operator is calculated, and the expression is: In the formula, Let be the subjective weight of the j-th risk factor, representing the relative importance of the risk factor in the expert's subjective judgment; Let be the comprehensive evaluation value of the j-th risk factor; The IVIFRN corresponding to the k-th risk factor; is the score of the k-th risk factor after defuzzification; n is the total number of risk factors.
[0035] like Figure 4 As shown, the subjective weights are: O=0.192, S=0.233, D=0.197, M=0.144, P=0.235. This set of weights mainly reflects the expert's perception of the importance of each risk factor, with P and S having relatively high subjective weights.
[0036] Step 23: Calculate the contrast strength and conflict of each risk factor using the CRITIC method. Based on the contrast strength and conflict of each risk factor, obtain the objective weight vector of the risk factors, including: The risk factor decision matrix is normalized column-wise to obtain a standardized matrix. For benefit-type indicators, linear normalization is used. In the formula, These are the elements in the standardized matrix; Let be an element in the risk factor decision matrix, representing the score of the i-th risk factor on the j-th failure mode; i is the risk factor index; j is the failure mode index; The minimum value of the j-th failure mode among all risk factors; Let be the maximum value of the j-th failure mode among all risk factors; m be the total number of risk factors; and n be the total number of failure modes. Based on the standardized matrix, the contrast strength of the risk factors is calculated using the following expression: In the formula, Let be the contrast strength of the j-th risk factor, i.e., the standard deviation; The standardized arithmetic mean of the j-th failure mode; Calculate the correlation coefficient matrix among the failure factors. ,in The Pearson correlation coefficient between risk factor j and risk factor k is expressed as: In the formula, Let be the Pearson correlation coefficient between the j-th risk factor and the k-th risk factor. The closer it is to 1, the stronger the positive correlation and the smaller the conflict. The conflict of risk factors is calculated based on the Pearson correlation coefficient, expressed as follows: In the formula, The conflict of the j-th risk factor; The information content of risk factors is generated by comparing their contrast strength and conflict, expressed as follows: In the formula, Let j be the information content of the j-th risk factor; Based on the information content of the risk factors, the objective weights of the risk factors are calculated, as expressed by: In the formula, The objective weights of risk factors.
[0037] like Figure 4 The objective weights shown are: O=0.261, S=0.181, D=0.285, M=0.118, P=0.155. These results indicate that the objective weights corresponding to D and O are relatively high, suggesting that these two risk factors have a stronger ability to distinguish information within the evaluation matrix.
[0038] Step 24: Based on the principle of minimum deviation in game theory, combine the subjective weight vector and the objective weight vector to obtain a compromise and stable comprehensive weight of risk factors, expressed as: In the formula, For comprehensive weighting; like Figure 4 The comprehensive weight results are as follows: O=0.226, S=0.207, D=0.241, M=0.131, P=0.195. The comprehensive weight results show that D has the largest comprehensive weight (0.241), indicating its high influence in the failure mode risk assessment of liquefied natural gas storage tank systems; M has the smallest comprehensive weight (0.131), indicating its relatively weak influence. These comprehensive weights will be used as inputs for subsequent failure mode risk prioritization.
[0039] Step 3: Based on the combined weights of the failure mode evaluation matrix and risk factors, the ExpTODIM-PROMETHEEII method is used to rank the risk priorities of each failure mode, outputting a risk priority sequence of failure modes, including: Step 31: Based on the failure mode evaluation matrix and the comprehensive weights of risk factors, calculate the dominance of each failure mode relative to each risk factor, expressed as: In the formula, Failure mode relative to failure mode Dominance under risk factor j; Let be the overall weight of the j-th risk factor; This is a decay control parameter used to adjust the decay rate of the exponential function, reflecting the decision-maker's sensitivity to advantage differences. The distance between two IVIFRN evaluations; This is the loss attenuation coefficient, used to amplify the negative effects of loss; Step 32: Calculate the overall dominance of each failure mode based on its dominance, expressed as: In the formula, Failure mode relative to failure mode The overall advantage, representing the failure mode's overall strength after considering all risk factors, is... Failure Mode Comparison A higher degree of risk; Step 33: Summarize the overall dominance of each failure mode and calculate the positive and negative currents for each failure mode. The expression is: In the formula, Failure mode The mainstream; Failure mode Negative current; It is a set of failure modes; This represents the total number of failure modes.
[0040] Step 34: Calculate the net current for each failure mode based on positive and negative currents. The expression is: In the formula, For net flow; like Figure 5 As shown in the diagram. Positive flow represents the advantage of a particular failure mode relative to other failure modes, negative flow represents the disadvantage of a particular failure mode relative to other failure modes, and the overall net flow reflects the overall risk priority of that failure mode. Figure 5 This indicates that there are significant differences in the distribution of each failure mode in the positive and negative flow, demonstrating that the method of the present invention can effectively distinguish the relative risk levels of different failure modes.
[0041] Step 35: Sort each failure mode according to its net current value and generate a risk priority sequence for each risk mode.
[0042] The higher the net flow value, the more attention needs to be paid to this failure mode.
[0043] like Figure 6 As shown. The combined net flow rates for each failure mode are: FM1=0.02136, FM2=0.62428, FM3=-0.54239, FM4=0.28391, FM5=0.00805, FM6=-0.30425, FM7=-0.07485.
[0044] The resulting risk priority ranking is: FM2 > FM4 > FM1 > FM5 > FM7 > FM6 > FM3. The corresponding rankings are: FM2 ranked 1st, FM4 ranked 2nd, FM1 ranked 3rd, FM5 ranked 4th, FM7 ranked 5th, FM6 ranked 6th, and FM3 ranked 7th.
[0045] The ranking results above show that the overall net flow rate of the auxiliary pipeline rupture FM2 is the largest, indicating that it has the highest risk priority and should be the focus of risk prevention and control in the liquefied natural gas storage tank system. The risk priority of the tank overflow FM4 and the serious mechanical failure FM1 is the second highest and should also be the focus of monitoring and control during system operation, maintenance and safety management. The overall net flow rate of the design defect risk level FM3 is the smallest, and its risk priority is relatively low.
[0046] This paper, based on the traditional FMEA framework, integrates IVIFRN, SAM, IVIFRWG, CRITIC, game theory, and ExpTODIM–PROMETHEE II methods to construct an improved FMEA risk ranking model suitable for high-risk, highly coupled LNG storage tank systems. The model systematically improves traditional methods from three levels: information modeling, weight determination, and ranking decision. Case studies show that the proposed model can achieve uncompensated total order ranking of failure modes, with the result being FM2>FM4>FM1>FM5>FM7>FM6>FM3. Among these, the rupture of the auxiliary pipeline (FM2) and overloading / overpressure operation of the storage tank (FM4) are identified as critical high-risk modes, a conclusion consistent with engineering practice. Compared with traditional FMEA methods, the proposed model demonstrates significant advantages in uncertainty representation, the rationality of subjective and objective weight integration, and the stability and discriminative power of risk ranking. It can more reliably identify key risk sources in complex and uncertain environments, providing an effective tool for risk management and decision support in LNG storage tank systems.
[0047] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0048] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. A failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy rough numbers, characterized in that, include: Step 1: Using the interval intuitionistic fuzzy set IVIFNS, the linguistic evaluation information of experts on each failure mode and risk factor of the liquefied natural gas storage tank system is converted into the corresponding interval intuitionistic fuzzy representation. The expert weights are determined through an improved consensus mechanism, and the evaluation results of each expert are aggregated to obtain the failure mode evaluation matrix represented in the form of interval intuitionistic fuzzy rough number IVIFRN. Step 2: Based on the failure mode evaluation matrix, the subjective weights of risk factors are calculated using the interval intuitionistic fuzzy rough weighted geometric operator, the objective weights of risk factors are calculated using the CRITIC method, and a combined weighting model is constructed by combining the game theory minimum deviation principle to obtain the comprehensive weights of each risk evaluation index. Step 3: Based on the combined weights of the failure mode evaluation matrix and risk factors, the ExpTODIM-PROMETHEE II method is used to rank the risk priorities of each failure mode and output the risk priority sequence of the failure modes.
2. The failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy rough numbers according to claim 1, characterized in that, Step 1 includes: Construct a language scale based on interval intuitionistic fuzzy sets IVIFNS; Obtain expert language evaluations, convert the language evaluation results into corresponding interval intuition fuzzy numbers based on the language scale, and generate an initial evaluation matrix; The expert weights are calculated based on the similarity and consistency levels among the language evaluation results of each expert, and an expert weight vector is generated. The similarity aggregation method SAM is used to achieve consistency fusion of expert evaluations, generate a failure mode evaluation matrix, and represent it in the form of interval intuitionistic fuzzy rough number IVIFRN.
3. The failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy rough numbers according to claim 2, characterized in that, The process of obtaining expert language evaluations involves converting the language evaluation results into corresponding interval intuitionistic fuzzy numbers based on a language scale, and generating an initial evaluation matrix, including: The set of expert language evaluations is represented in the form of an IVIFN set as follows: In the formula, For language evaluation Interval intuitive fuzzy numbers; This is the lower bound of the membership degree; This is the upper bound of the membership degree; This is the lower bound of the non-membership degree; This is the upper bound of the non-membership degree; but The lower approximation Approximate with the above It can be determined as follows: In the formula, The following approximation is used; It is the set of all objects in the language evaluation set; , is the lower union of the upper approximation; It is the lower union of the upper approximation; Improved IVIFRN lower limit and upper limit The expression is: In the formula, For the improved IVIFRN lower limit; To be downward and concentrated, satisfying the condition The total number of language assessments; For the evaluation of language involved in the calculation; For the improved IVIFRN upper limit; To be upward and concentrated, satisfying the condition The total number of language assessments; For language evaluation IVIFRN value; Then the interval value intuition fuzzy class Improved IVIFRN for: In the formula, The resulting improved interval is intuitively fuzzy and coarse. , , , These are the arithmetic mean of the components corresponding to the lower and upper limits, respectively.
4. The failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy rough numbers according to claim 2, characterized in that, The similarity aggregation method (SAM) is used to fuse the expert evaluations for consistency, generating a failure mode evaluation matrix, which is expressed in the form of interval intuitionistic fuzzy rough number (IVIFRN), including: The similarity function is expressed as: In the formula, Let be the similarity between the evaluation results of the u-th expert and the v-th expert; These are the interval intuitionistic fuzzy rough number evaluations of a certain failure mode by experts u and v, respectively. These are the comprehensive eigenvalues of IVIFRN; According to similarity calculation experts Weighted consistency The expression is: In the formula, For experts The weighted consistency; Let v be the initial weight for the expert. The number of experts participating in the evaluation; According to weighted consistency calculation experts relative consistency The expression is: In the formula, For experts The degree of relative consistency; Total number of experts; The expert consensus coefficient is generated by weighting and summing the weighted consistency and relative consistency. The expression is: In the formula, This represents the expert consensus coefficient. For the weighting factor; The interval intuitionistic fuzzy rough weighted average operator is expressed as follows: In the formula, An interval-valued intuitive fuzzy coarse weighted average operator; The number of evaluation objects participating in the aggregation represents the number of experts; Let be the weight vector of the j-th expert; Based on the above formula, the summary result of expert opinions can be expressed as follows: In the formula, Let j be the consensus coefficient of the expert. The j-th expert is the IVIFRN evaluation of the failure mode.
5. The failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy roughness number according to claim 1, characterized in that, Step 2 includes: Based on the failure mode evaluation matrix, the IVIFRWG operator is used to fuse multiple expert evaluations to construct a risk factor decision matrix. The interval intuitionistic fuzzy rough weighted geometric operator IVIFRWG is used to subjectively assign weights to risk factors, resulting in a subjective weight vector; The contrast strength and conflict of each risk factor are calculated using the CRITIC method, and the objective weight vector of each risk factor is obtained based on the contrast strength and conflict of each risk factor. Based on the principle of minimum deviation in game theory, the subjective weight vector and the objective weight vector are combined to obtain the comprehensive weight of the risk factors.
6. The failure mode evaluation method for tank systems based on interval intuitionistic fuzzy roughness number according to claim 5, characterized in that, The method employs the interval intuitionistic fuzzy rough weighted geometric operator IVIFRWG to subjectively weight risk factors, resulting in a subjective weight vector, which includes: IVIFRN To perform deblurring, the expression is: In the formula, The value is the defuzzified value; IR(S) is the interval intuitionistic fuzzy rough number. This is the composite value of the membership interval; This is the composite value of the non-membership interval; After defuzzing, the subjective weight vector of the risk operator is calculated, and the expression is: In the formula, Let be the subjective weight of the j-th risk factor; Let be the comprehensive evaluation value of the j-th risk factor; The IVIFRN corresponding to the k-th risk factor; is the score of the k-th risk factor after defuzzification; n is the total number of risk factors.
7. The failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy rough numbers according to claim 5, characterized in that, The method of calculating the contrast strength and conflict of each risk factor using the CRITIC method, and obtaining the objective weight vector of each risk factor based on the contrast strength and conflict of each risk factor, includes: The risk factor decision matrix is normalized column-wise to obtain a standardized matrix, expressed as follows: In the formula, These are the elements in the standardized matrix; Let be an element in the risk factor decision matrix, representing the score of the i-th risk factor on the j-th failure mode; i is the risk factor index; j is the failure mode index; The minimum value of the j-th failure mode among all risk factors; Let be the maximum value of the j-th failure mode among all risk factors; m be the total number of risk factors; and n be the total number of failure modes. Based on the standardized matrix, the contrast strength of the risk factors is calculated using the following expression: In the formula, Let be the contrast intensity of the j-th risk factor; The standardized arithmetic mean of the j-th failure mode; Calculate the correlation coefficient matrix among the failure factors. ,in The Pearson correlation coefficient between risk factor j and risk factor k is expressed as: In the formula, Let be the Pearson correlation coefficient between the j-th risk factor and the k-th risk factor; The conflict of risk factors is calculated based on the Pearson correlation coefficient, expressed as follows: In the formula, The conflict of the j-th risk factor; The information content of risk factors is generated by comparing their contrast strength and conflict, expressed as follows: In the formula, Let j be the information content of the j-th risk factor; Based on the information content of the risk factors, the objective weights of the risk factors are calculated, as expressed by: In the formula, The objective weights of risk factors.
8. The failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy rough numbers according to claim 1, characterized in that, Step 3 includes: Based on the comprehensive weights of the failure mode evaluation matrix and risk factors, the dominance of each failure mode relative to each risk factor is calculated. Calculate the overall dominance of each failure mode based on its dominance. The overall dominance of each failure mode is summarized, and the positive and negative currents of each failure mode are calculated. Calculate the net flow for each failure mode based on positive and negative flow; Based on the net current value of each failure mode, the failure modes are sorted to generate a risk priority sequence for each risk mode.
9. The failure mode evaluation method for storage tank systems based on interval intuitionistic fuzzy roughness number according to claim 8, characterized in that, The dominance of each failure mode relative to each risk factor is calculated based on the comprehensive weights of the failure mode evaluation matrix and risk factors, expressed as follows: In the formula, Failure mode relative to failure mode Dominance under risk factor j; Let be the overall weight of the j-th risk factor; These are the attenuation control parameters; The distance between two IVIFRN ratings; This is the loss attenuation coefficient.
10. The failure mode evaluation method for tank systems based on interval intuitionistic fuzzy roughness number according to claim 8, characterized in that, The overall dominance of each failure mode is calculated based on its dominance, expressed as follows: In the formula, Failure mode relative to failure mode The overall dominance; n is the total number of failure modes; Let represent the dominance of the j-th failure mode.