A complex equipment technical support fuzzy comprehensive evaluation method

Abstract

The application discloses a complex equipment technical support fuzzy comprehensive evaluation method, which decomposes safety into 25 aspects, adopts the way of scoring by experts at the present level to score in percentage system, then establishes a fuzzy system and fuzzy rules to obtain the fuzzy score of equipment safety, and on the basis, according to the percentage score data of experts at the present level and the comprehensive evaluation score of safety of equipment after completing a task by experts at the upper level, a kind of radial basis neural network based on positive and negative sine mixed convergence oscillation is established, the weight of the neural network is adaptively trained by using the existing historical data and the data of the fuzzy system, and finally the trained network and the fuzzy system are used to evaluate the safety of the equipment technical support. The method combines the fuzzy system and the neural network well, and solves the problem of low accuracy of the single fuzzy system.

Description

Technical Field

[0001] This invention relates to a fuzzy comprehensive evaluation method for technical support of complex equipment, belonging to the field of equipment reliability assessment and safety assessment prediction. Background Technology

[0002] Fuzzy comprehensive evaluation is a comprehensive evaluation method based on fuzzy mathematics. This method transforms qualitative evaluation into quantitative evaluation based on the membership theory of fuzzy mathematics, that is, it uses fuzzy mathematics to make an overall evaluation of things or objects constrained by multiple factors. It has the characteristics of clear results and strong systematicity, and can better solve fuzzy and difficult-to-quantify problems, making it suitable for solving various uncertain problems. Due to the complexity of evaluation factors, the hierarchical nature of evaluation objects, the fuzziness in evaluation standards, the fuzziness or uncertainty of evaluation influencing factors, and the difficulty in quantifying qualitative indicators, it is difficult to accurately describe objective reality with absolute "either / or" statements. A fuzzy phenomenon of "both / and" often exists, and its description is often expressed using natural language. The biggest characteristic of natural language is its fuzziness, which is difficult to uniformly measure using classical mathematical models. Therefore, the fuzzy comprehensive evaluation method, based on fuzzy sets, comprehensively evaluates the hierarchical status of the evaluated object from multiple indicators. It divides the range of change of the evaluated object, which on the one hand takes into account the hierarchical nature of the object, allowing the fuzziness of evaluation standards and influencing factors to be reflected; on the other hand, it can fully utilize human experience in the evaluation process, making the evaluation results more objective and consistent with reality. Fuzzy comprehensive evaluation can combine qualitative and quantitative factors, expand the amount of information, improve the accuracy of the evaluation, and make the evaluation conclusions credible.

[0003] Safety assessment of complex equipment technical support is a complex process involving numerous stages and high technical content. It is subject to stringent safety requirements, influenced by factors such as high-pressure gases, electronic components, ambient temperature and humidity, temperature and humidity of support equipment, explosive and flammable materials, combustible material safety, theft prevention, explosion-proof equipment in ordinary testing rooms, protective devices in explosive testing rooms, protective devices in flammable testing rooms, grounding, short-circuit sockets, explosion-proof sockets, explosion-proof air conditioning, ventilation in refueling areas, isolation walls, isolation nets, dismantling safety, transportation safety, sub-unit testing safety, installation safety, overall testing safety, technical preparation safety, and departure safety. Therefore, its safety assessment is extremely complex, with intricate hierarchical structures and a high degree of ambiguity in expert scoring. For the reasons mentioned above, this invention fuzzifies the expert scores for various aspects, then establishes a fuzzy system. Based on the security scores provided by the fuzzy system, a radial basis function neural network based on a hybrid sine and cosine convergent oscillatory type is established. The weights of the neural network are adaptively trained using existing historical data superimposed with data from the fuzzy system. Finally, the trained network and the fuzzy system are used to evaluate the technical security of the equipment under evaluation. This method effectively combines fuzzy systems and neural networks, possessing not only high theoretical innovation but also significant engineering practical value. Summary of the Invention

[0004] The technical problem to be solved by the present invention is: a fuzzy comprehensive evaluation method for the technical protection of complex equipment, so as to solve the problem that the accuracy of the fuzzy comprehensive evaluation of security in the prior art is not high.

[0005] The technical solution adopted in this invention is: a fuzzy comprehensive evaluation method for technical support of complex equipment, which includes the following steps:

[0006] Step S10 involves assessing the safety of the technical support process for each piece of equipment across multiple systems, categorized into 25 aspects: high-pressure gas, electronic components, ambient temperature and humidity, temperature and humidity of support equipment, temperature and humidity of explosives, temperature and humidity of flammable materials, safety against combustibles, anti-theft safety, explosion-proof equipment in ordinary testing rooms, protective devices in explosive testing rooms, protective devices in flammable testing rooms, grounding, short-circuit sockets, explosion-proof sockets, explosion-proof air conditioning, ventilation in refueling areas, isolation walls, isolation nets, disassembly safety, transportation safety, sub-unit testing safety, installation safety, overall testing safety, technical preparation safety, and departure safety. Experts at this level then summarize the scores (out of 100) to obtain the total safety score for the current equipment across these 25 aspects, denoted as 'a'. ij , represents the safety score of the j-th aspect of the technical support process for the i-th piece of equipment.

[0007] Step S20: Define the concepts of 25 aspects of equipment security scores and the concept of equipment security fuzzy scores, define fuzzy rules, establish a fuzzy system, and run the fuzzy system to obtain the equipment security fuzzy scores.

[0008] Step S30: Based on the safety scores of the experts at this level for 25 aspects of the technical support process for n pieces of equipment, an adaptive radial basis function neural network is established. First, for the j-th aspect of the technical support process for the i-th piece of equipment, 70 neural network node center values ​​are selected respectively. Then, the deviation data of the network center point is obtained by comparing with these values. The node sensitivity intervals of the neural network are set, and the absolute value of the deviation is transformed and time decayed. The deviation is then multiplied with the sine transformation of the deviation data of the network center point to obtain the radial basis coefficients of the convergent oscillating network. Then, the coefficients are multiplied with the absolute value exponential integer fractional order mixed decay function of the deviation data of the network center point to obtain the radial basis function of the neural network. Then, the corresponding neural network weights are multiplied and the 70 nodes are accumulated to obtain the comprehensive safety output of the neural network for the j-th aspect of the i-th piece of equipment. Finally, the 25 aspects are accumulated and superimposed with the fuzzy safety score of the i-th piece of equipment to obtain the total evaluation data of the safety score of the neural network for the technical support of the i-th piece of equipment.

[0009] Step S40: The network training error data is obtained by comparing the total safety score of the i-th equipment with the comprehensive safety assessment score of the i-th equipment after the completion of the mission by the superior expert; then, an adaptive adjustment law based on sine and cosine hybrid convergent oscillation weight is designed based on the network training error data, and the network weights are adaptively updated by integral operation; when the network training error converges to the region near 0, training and weight update are stopped.

[0010] Step S50: Based on the safety score data of 25 aspects of the technical support process for the equipment under evaluation by the experts at this level, the data is substituted into the fuzzy system and then into the trained neural network to obtain the total evaluation data of the safety score of the technical support for the equipment under evaluation by the neural network.

[0011] In one exemplary embodiment of the present invention, the concepts of 25 aspects of equipment security scores and equipment security fuzzy scores are defined, fuzzy rules are defined, a fuzzy system is established, and the fuzzy system is run to obtain the equipment security fuzzy scores, including:

[0012] First, define the input variable as the security score a of the j-th aspect for the i-th piece of equipment. ij The fuzzy concepts can be mainly divided into the following five categories:

[0013] a ij ={HBMSO};

[0014] Where H represents the input variable aij Very large, B represents the input variable a ij Larger, M represents the input variable a ij For medium, S represents the input variable a. ij Smaller, O indicates input variable a ij Approximately 0; specifically as follows:

[0015] When 80≤|a ij When |<100, j=1,2,3,…,25, we consider a ij Very large;

[0016] When 60≤|a ij When |<80, j=1,2,3,…,25, we consider a ij Larger;

[0017] When 40≤|a ij When |<60, j=1,2,3,…,25, we consider a ij medium;

[0018] When 20≤|a ij When |<40, j=1,2,3,…,25, we consider a ij Smaller;

[0019] When 0≤|a ij When |<20, j=1,2,3,…,25, we consider a ij Approaching 0;

[0020] Secondly, define the fuzzy score d for the safety of the output equipment. i (a ij The fuzzy concept of ) is also divided into the following five fuzzy concepts, namely

[0021] d i (a ij )={H' B' M' S' O'}

[0022] Where H' represents the fuzzy score of equipment safety d i (a ij B' represents the fuzzy score d for equipment safety. i (a ij M' represents the larger value, and M' represents the fuzzy score d for equipment safety. i (a ij S' represents the fuzzy score for equipment safety, where S' represents the medium level. i (a ij ) represents a smaller value, and O' represents the fuzzy score d for equipment safety. i (a ij The value is close to 0; the specific value will be selected during the design process.

[0023] When 0.8 < |d i (a ij When | < 1, the fuzzy score d for equipment safety is considered to be... i (a ij It's very big;

[0024] When 0.6 < |d i (a ij When | < 0.8, the fuzzy score d for equipment safety is considered to be... i (a ij (larger)

[0025] When 0.4 < |d i (a ij When | < 0.6, the fuzzy score d for equipment safety is considered to be... i (a ij )medium;

[0026] When 0.2 < |d i (a ij When | < 0.4, the fuzzy score d for equipment safety is considered to be... i (a ij Smaller;

[0027] When 0 < |d i (a ij When | < 0.2, the fuzzy score d for equipment safety is considered to be... i (a ij The value is almost zero;

[0028] Finally, for all 25 aspects j = 1, 2, 3, ..., 25, 5 fuzzy rules are defined for each aspect as follows:

[0029] When input variable a ij When the value is large, the fuzzy score d for equipment safety is considered to be... i (a ij It's very big;

[0030] When input variable a ij Generally, when the value is large, the fuzzy score d for equipment safety is considered to be... i (a ij (larger)

[0031] When input variable a ij The score is considered moderate, with the equipment safety fuzzy score d being considered the optimal value. i (a ij )medium;

[0032] When input variable a ij When the value is small, the fuzzy score d for equipment safety is considered to be... i (a ij Smaller;

[0033] When input variable a ij When the value is almost zero, the fuzzy score d for equipment safety is considered to be... i (a ij The value is almost zero;

[0034] Then, the fuzzy system is run to obtain the fuzzy security score d for the i-th piece of equipment. i (a ij ).

[0035] In one exemplary embodiment of the present invention, based on the safety scores of experts at this level for 25 aspects of the technical support process for n pieces of equipment, an adaptive radial basis function neural network is established, and the safety fuzzy score of the i-th piece of equipment is superimposed to obtain the total evaluation data of the neural network for the safety score of the technical support of the i-th piece of equipment, including:

[0036] ε ijk =a ij -f jk ;

[0037]

[0038]

[0039]

[0040]

[0041] Where f jk Let be the center value of the 70 neural network nodes in the j-th aspect, which is a constant; where k = 1, 2, ..., 70, ε ijk For network center point deviation data; σ j Let γ be the node sensitivity region of the neural network in the j-th aspect, and γ be a constant parameter. 0ijk The radial basis coefficients of the convergent oscillatory network are given, where t is the time signal; c1 and c2 are constant parameters; and γ is the spectral density. ijk u is the radial basis function of the neural network; ij For the neural network's comprehensive output on the security of the i-th equipment in the j-th aspect, k jk p represents the weights of the neural network. i This represents the total security score assessment data for the technical support of the i-th piece of equipment by the neural network.

[0042] In one exemplary embodiment of the present invention, network training error data is obtained by comparing the total safety score evaluation data of the i-th equipment by the neural network with the comprehensive safety evaluation score of the i-th equipment after the completion of the mission by the superior expert; then, based on the network training error data, an adaptive adjustment law based on a sine and cosine hybrid convergent oscillation weight is designed, including:

[0043] ei =p i -b i ;

[0044]

[0045] k jk (n+1)=k jk (n)+d jk ;

[0046] Where e i b is the network training error data. i The score for the comprehensive safety assessment of the i-th piece of equipment after it completes its mission, as determined by higher-level experts; d jk This is based on an adaptive adjustment law for weights of a hybrid sine and cosine convergent oscillation type; jk This is a constant parameter used to adjust the convergence speed of neural network weights.

[0047] In one exemplary embodiment of the present invention, based on the security score data of 25 aspects of the technical support process for the equipment to be evaluated by the experts at this level, the data is substituted into a fuzzy system and then into a trained neural network to obtain the total security score evaluation data of the neural network for the technical support of the equipment to be evaluated, including:

[0048] ε Djk =a Dj -f jk ;

[0049]

[0050]

[0051]

[0052]

[0053] Where a Dj The data includes safety scores for 25 aspects of the technical support process for the equipment to be evaluated; ε Djk For the network center point deviation data of the equipment to be evaluated; γ 0Djk γ represents the radial basis coefficient of the convergent oscillatory network of the equipment to be evaluated. Djk u is the radial basis function of the neural network of the equipment to be evaluated. Dj For the comprehensive safety output of the neural network for the j-th aspect of the equipment to be evaluated, d D (a Dj p represents the fuzzy score for the safety of the equipment to be evaluated. D This is the total evaluation data for the safety score of the technical support for the equipment being evaluated, provided by the neural network.

[0054] Beneficial effects of the present invention

[0055] Compared with existing technologies, the method adopted in this invention has the following three major innovations. First, based on the background characteristics and actual situation of rapid technical support for complex equipment, it decomposes the equipment into 25 aspects: high-pressure gas, electronic devices, ambient temperature and humidity, equipment temperature and humidity, explosive material temperature and humidity, flammable material temperature and humidity, combustible material safety, anti-theft safety, explosion-proof equipment in ordinary testing rooms, protective devices in explosive testing rooms, protective devices in flammable testing rooms, grounding, short-circuit sockets, explosion-proof sockets, explosion-proof air conditioning, ventilation in refueling areas, isolation walls, isolation nets, disassembly safety, transportation safety, sub-unit testing safety, installation safety, overall testing safety, technical preparation safety, and departure safety. A fuzzy system is established to obtain the fuzzy relationship between expert scores and equipment safety. Second, a radial basis function neural network based on a sine / cosine hybrid convergent oscillatory type is superimposed on the fuzzy system. This network can adaptively train the weights of the neural network using existing historical data superimposed with data from the fuzzy system, improving the accuracy of the fuzzy system evaluation. Thirdly, a method is proposed to achieve rapid convergence of neural network weights by using a weight adaptive adjustment law based on a sine and cosine hybrid convergent oscillation. This can accelerate the network training process, speed up the weight convergence, and reduce the probability of network weight divergence during training. At the same time, by introducing data from a fuzzy system as a basis, the initial blindness of neural network training is avoided, and the convergence time of the neural network is accelerated. Attached Figure Description

[0056] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention. It is obvious that the drawings described below are merely some embodiments of the invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0057] Figure 1 This is a flowchart of a fuzzy comprehensive evaluation method for technical support of complex equipment;

[0058] Figure 2 It is the input fuzzy membership function image of the fuzzy system provided by the embodiment of the present invention;

[0059] Figure 3 It is the output fuzzy membership function image of the fuzzy system of the method provided in the embodiments of the present invention;

[0060] Figure 4 This is a graph showing the convergence of network training error according to the method provided in this embodiment of the invention;

[0061] Figure 5 The neural network weight k is the method provided in the embodiments of the present invention. 11The convergence diagram. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of the present invention clearer, specific embodiments are described below, with reference to the appendix. Figure 1 The present invention will be further described in detail below.

[0063] Example 1: Using the rapid technical support process of 350 pieces of equipment stored in a warehouse, a fuzzy system is established. The neural network is trained using the safety scores of local experts on 25 aspects of the rapid technical support process for the 350 pieces of equipment, and the comprehensive safety assessment scores of higher-level experts on the equipment after completing its mission. This serves as a case study to illustrate a fuzzy comprehensive assessment method for the technical support of complex equipment. The method includes the following steps:

[0064] Step S10 can be broken down into the following two parts. The first step involves assessing the safety of the technical support process for each of the 350 pieces of equipment in 25 aspects: high-pressure gas, electronic components, ambient temperature and humidity, temperature and humidity of support equipment, temperature and humidity of explosives, temperature and humidity of flammable materials, safety against combustibles, anti-theft safety, explosion-proof equipment in ordinary testing rooms, protective devices in explosive testing rooms, protective devices in flammable testing rooms, grounding, short-circuit sockets, explosion-proof sockets, explosion-proof air conditioning, ventilation in the filling area, isolation walls, isolation nets, disassembly safety, transportation safety, sub-unit testing safety, installation safety, overall testing safety, technical preparation safety, and departure safety.

[0065] The second step involves the local experts assigning scores to the 25 aspects of the 350 pieces of equipment on a 100-point scale. The total safety score for these 25 aspects is then calculated and denoted as 'a'. ij , represents the safety score of the j-th aspect of the technical support process for the i-th piece of equipment.

[0066] Step S20: Define the concepts of 25 aspects of equipment security scores and the concept of equipment security fuzzy scores, define fuzzy rules, establish a fuzzy system, and run the fuzzy system to obtain the equipment security fuzzy scores.

[0067] Specifically, it can be broken down into the following four steps.

[0068] The first step is to define the input variable as the security score a of the j-th aspect for the i-th piece of equipment. ij The fuzzy concepts can be mainly divided into the following five categories:

[0069] a ij ={HBMSO};

[0070] Where H represents the input variable a ijVery large, B represents the input variable a ij Larger, M represents the input variable a ij For medium, S represents the input variable a. ij Smaller, O indicates input variable a ij Approximately 0; specifically as follows:

[0071] When 80≤|a ij When |<100, j=1,2,3,…,25, we consider a ij Very large;

[0072] When 60≤|a ij When |<80, j=1,2,3,…,25, we consider a ij Larger;

[0073] When 40≤|a ij When |<60, j=1,2,3,…,25, we consider a ij medium;

[0074] When 20≤|a ij When |<40, j=1,2,3,…,25, we consider a ij Smaller;

[0075] When 0≤|a ij When |<20, j=1,2,3,…,25, we consider a ij Close to 0.

[0076] Specifically, the input fuzzy membership function of this fuzzy system is as follows: Figure 2 As shown.

[0077] The second step is to define the fuzzy score d for the safety of the output equipment. i (a ij The fuzzy concept of ) is also divided into the following five fuzzy concepts, namely

[0078] d i (a ij )={H' B' M' S' O'};

[0079] Where H' represents the fuzzy score of equipment safety d i (a ij B' represents the fuzzy score d for equipment safety. i (a ij M' represents the larger value, and M' represents the fuzzy score d for equipment safety. i (a ij S' represents the fuzzy score for equipment safety, where S' represents the medium level. i (a ij ) represents a smaller value, and O' represents the fuzzy score d for equipment safety. i (aij The value is close to 0; the specific value will be selected during the design process.

[0080] When 0.8 < |d i (a ij When ) < |1, the fuzzy score d for equipment safety is considered to be... i (a ij It's very big;

[0081] When 0.6 < |d i (a ij When | < 0.8, the fuzzy score d for equipment safety is considered to be... i (a ij (larger)

[0082] When 0.4 < |d i (a ij When ) < |0.6, the fuzzy score d for equipment safety is considered to be... i (a ij )medium;

[0083] When 0.2 < |d i (a ij When ) < |0.4, the fuzzy score d of equipment safety is considered to be i (a ij Smaller;

[0084] When 0 < |d i (a ij When | < 0.2, the fuzzy score d for equipment safety is considered to be... i (a ij The value is almost zero.

[0085] Specifically, the output fuzzy membership function of this fuzzy system is as follows: Figure 3 As shown.

[0086] The third step is to define 5 fuzzy rules for each of the 25 aspects where j = 1, 2, 3, ..., 25, as follows:

[0087] When input variable a ij When the value is large, the fuzzy score d for equipment safety is considered to be... i (a ij It's very big;

[0088] When input variable a ij Generally, when the value is large, the fuzzy score d for equipment safety is considered to be... i (a ij (larger)

[0089] When input variable a ij The score is considered moderate, with the equipment safety fuzzy score d being considered the optimal value. i (a ij )medium;

[0090] When input variable a ij When the value is small, the fuzzy score d for equipment safety is considered to be... i (a ij Smaller;

[0091] When input variable a ij When the value is almost zero, the fuzzy score d for equipment safety is considered to be... i (a ij The value is almost zero;

[0092] The fourth step is to run the fuzzy system to obtain the security fuzzy score of the i-th piece of equipment, denoted as d. i (a ij ).

[0093] Step S30: Based on the safety scores of the experts at this level for 25 aspects of the technical support process for n pieces of equipment, establish an adaptive radial basis function neural network. Specifically, this can be broken down into the following five steps.

[0094] The first step is to select the center values ​​of 70 neural network nodes based on the j-th aspect of the technical support process for the i-th piece of equipment; then, compare these values ​​to obtain the following network center point deviation data:

[0095] ε ijk =a ij -f jk ;

[0096] Where f jk Let be the center value of the 70 neural network nodes in the j-th aspect, which is a constant; where k = 1, 2, ..., 70, ε ijk This is the deviation data for the network center point.

[0097] The second step is to set the node sensitivity intervals of the neural network, perform absolute value transformation on the deviations and time decay, and then multiply them by the sine transform of the deviation data at the network center point to obtain the radial basis coefficients of the convergent oscillating network as follows:

[0098]

[0099] Where σ j Let γ be the node sensitivity region of the neural network in the j-th aspect, and γ be a constant parameter. 0ijk Here, t represents the radial basis coefficients of the convergent oscillatory network, t is the time signal, and c1 is a constant parameter, chosen to be c1 = 2.5.

[0100] The third step involves multiplying the radial basis function of the neural network by the absolute value exponential integer fractional order mixed decay function of the network center point deviation data, resulting in the following:

[0101]

[0102] Where c2 is a constant parameter, chosen as c2 = 0.8, γ ijk is the radial basis function of the neural network.

[0103] The fourth step involves multiplying the radial basis function of the neural network by the corresponding neural network weights and accumulating the results across 70 nodes. The resulting comprehensive output of the neural network regarding the security of the i-th equipment in the j-th aspect is as follows:

[0104]

[0105] Where u ij For the neural network's comprehensive output on the security of the i-th equipment in the j-th aspect, k jk These are the weights of the neural network.

[0106] The fifth step involves summing and superimposing the fuzzy security scores of the i-th equipment across the 25 aspects to obtain the total security score assessment data of the neural network for the technical support of the i-th equipment, as follows:

[0107]

[0108] Where p i This represents the total security score assessment data for the technical support of the i-th piece of equipment by the neural network.

[0109] Step S40 can be broken down into the following four sub-steps. Step 1: Based on the comparison between the total safety score assessment data of the i-th equipment obtained by the neural network and the comprehensive safety assessment score of the i-th equipment after the completion of the mission by the superior expert, the network training error data is obtained as follows:

[0110] e i =p i -b i ;

[0111] Where e i b is the network training error data. i This is the comprehensive safety assessment score given by higher-level experts after the i-th piece of equipment completes its mission.

[0112] The second step involves designing an adaptive adjustment law for weights based on a hybrid sinusoidal and cosine convergent oscillation type, according to the network training error data:

[0113]

[0114] d jk This is based on an adaptive adjustment law for weights of a hybrid sine and cosine convergent oscillation type; jk For constant parameters, choose l. jk =0.05, used to adjust the convergence speed of neural network weights.

[0115] The third step involves adaptively updating the network weights through integral operations as follows:

[0116] k jk (n+1)=k jk (n)+d jk ;

[0117] Where k jk These are the weights of the neural network.

[0118] The fourth step is to stop training and weight updates when the network training error converges to the region near 0.

[0119] The convergence curve of the network training error is as follows: Figure 4 As shown. The neural network weights k 11 The convergence curve is as follows Figure 5 As shown.

[0120] Step S50 can be broken down into the following six sub-steps. The first step is to use the safety score data (a) from the 25 aspects of the technical support process for the equipment being evaluated, as determined by the experts at this level. Dj Substituting the values ​​of [92,78,83,68,87,82,73,69,87,94,96,87,74,83,89,88,76,82,85,84,82,80,85,86,97] into the fuzzy system, we obtain the fuzzy score d for the safety of the equipment to be evaluated. D (a Dj = 0.8465.

[0121] The second step involves substituting the security score data from the aforementioned 25 aspects into the trained neural network to calculate the network center point deviation data for the equipment to be evaluated, as follows:

[0122] ε Djk =a Dj -f jk ;

[0123] The third step is to calculate the radial basis coefficients of the convergent oscillatory network of the equipment to be evaluated as follows:

[0124]

[0125] The fourth step is to calculate the radial basis functions of the neural network of the equipment to be evaluated.

[0126]

[0127] The fifth step is to calculate the comprehensive safety output of the neural network for the j-th aspect of the equipment being evaluated.

[0128]

[0129] Step 6: Summarize the output of the fuzzy system to obtain the total evaluation data of the safety score of the technical support for the equipment under evaluation by the neural network, as follows:

[0130]

[0131] Where a Dj The data includes safety scores for 25 aspects of the technical support process for the equipment to be evaluated; ε Djk For the network center point deviation data of the equipment to be evaluated; γ 0Djk γ represents the radial basis coefficient of the convergent oscillatory network of the equipment to be evaluated. Djk u is the radial basis function of the neural network of the equipment to be evaluated. Dj For the comprehensive safety output of the neural network for the j-th aspect of the equipment to be evaluated, d D (a Dj p represents the fuzzy score for the safety of the equipment to be evaluated. D This represents the total safety score evaluation data for the technical support of the equipment under evaluation, obtained through a neural network. The final result p... D The value was 87.65, therefore the conclusion was that the safety was good.

[0132] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A fuzzy comprehensive evaluation method for technical support of complex equipment, characterized by the following steps: Step S10 involves assessing the safety of the technical support process for each piece of equipment across multiple systems, categorized into 25 aspects: high-pressure gas, electronic components, ambient temperature and humidity, temperature and humidity of support equipment, temperature and humidity of explosives, temperature and humidity of flammable materials, safety against combustibles, anti-theft safety, explosion-proof equipment in ordinary testing rooms, protective devices in explosive testing rooms, protective devices in flammable testing rooms, grounding, short-circuit sockets, explosion-proof sockets, explosion-proof air conditioning, ventilation in refueling areas, isolation walls, isolation nets, disassembly safety, transportation safety, sub-unit testing safety, installation safety, overall testing safety, technical preparation safety, and departure safety. Experts at this level then summarize the scores (out of 100) to obtain the total safety score for the current equipment across these 25 aspects, denoted as 'a'. ij , represents the safety score of the j-th aspect of the technical support process for the i-th piece of equipment; Step S20: Define the concepts of 25 aspects of equipment security scores and the concept of equipment security fuzzy scores, define fuzzy rules, establish a fuzzy system, and run the fuzzy system to obtain the equipment security fuzzy scores as follows: First, define the input variable as the security score a of the j-th aspect for the i-th piece of equipment. ij The fuzzy concepts are divided into the following five fuzzy concepts, namely: a ij ={H B M S O}; Where H represents the input variable a ij Very large, B represents the input variable a ij Larger, M represents the input variable a ij For medium, S represents the input variable a. ij Smaller, O indicates input variable a ij Approximately 0; specifically as follows: When 80≤|a ij When |<100, j=1,2,3,…,25, we consider a ij Very large; When 60≤|a ij When |<80, j=1,2,3,…,25, we consider a ij Larger; When 40≤|a ij When |<60, j=1,2,3,…,25, we consider a ij medium; When 20≤|a ij When |<40, j=1,2,3,…,25, we consider a ij Smaller; When 0≤|a ij When |<20, j=1,2,3,…,25, we consider a ij Approaching 0; Secondly, define the fuzzy score d for the safety of the output equipment. i (a ij The fuzzy concept of ) is also divided into the following five fuzzy concepts, namely d i (a ij )={H' B' M' S' O'} Where H' represents the fuzzy score of equipment safety d i (a ij B' represents the fuzzy score d for equipment safety. i (a ij M' represents the larger value, and M' represents the fuzzy score d for equipment safety. i (a ij S' represents the fuzzy score for equipment safety, where S' represents the medium level. i (a ij ) represents a smaller value, and O' represents the fuzzy score d for equipment safety. i (a ij The value is close to 0; the specific value will be selected during the design process. When 0.8 < |d i (a ij When ) < |1, the fuzzy score d for equipment safety is considered to be... i (a ij It's very big; When 0.6 < |d i (a ij When | < 0.8, the fuzzy score d for equipment safety is considered to be... i (a ij (larger) When 0.4 < |d i (a ij When | < 0.6, the fuzzy score d for equipment safety is considered to be... i (a ij )medium; When 0.2 < |d i (a ij When ) < |0.4, the fuzzy score d of equipment safety is considered to be i (a ij Smaller; When 0 < |d i (a ij When ) < |0.2, the fuzzy score d of equipment safety is considered to be i (a ij The value is almost zero; Finally, for all 25 aspects j = 1, 2, 3, ..., 25, 5 fuzzy rules are defined for each aspect as follows: When input variable a ij When the value is large, the fuzzy score d for equipment safety is considered to be... i (a ij It's very big; When input variable a ij Generally, when the value is large, the fuzzy score d for equipment safety is considered to be... i (a ij (larger) When input variable a ij The score is considered moderate, with the equipment safety fuzzy score d being considered the optimal value. i (a ij )medium; When input variable a ij When the value is small, the fuzzy score d for equipment safety is considered to be... i (a ij Smaller; When input variable a ij When the value is almost zero, the fuzzy score d for equipment safety is considered to be... i (a ij The value is almost zero; Then, the fuzzy system is run to obtain the fuzzy security score d for the i-th piece of equipment. i (a ij ); Step S30: Based on the safety scores of the experts at this level for 25 aspects of the technical support process for n pieces of equipment, an adaptive radial basis function neural network is established. First, for the j-th aspect of the technical support process for the i-th piece of equipment, 70 neural network node center values ​​are selected respectively. Then, the deviation data of the network center point is obtained by comparing with these values. The node sensitivity intervals of the neural network are set, and the absolute value of the deviation is transformed and time decayed. Then, the deviation is multiplied by the sine transformation of the deviation data of the network center point to obtain the radial basis coefficients of the convergent oscillating network. Then, it is multiplied by the absolute value exponential integer fractional order mixed decay function of the deviation data of the network center point to obtain the radial basis function of the neural network. Then, it is multiplied by the corresponding neural network weights and accumulated by the 70 nodes to obtain the comprehensive safety output of the neural network for the j-th aspect of the i-th piece of equipment. Finally, the 25 aspects are accumulated and the fuzzy safety score of the i-th piece of equipment is superimposed to obtain the total evaluation data of the safety score of the neural network for the technical support of the i-th piece of equipment as follows: e ijk =a ij -f jk ; Where f jk Let be the center value of the 70 neural network nodes in the j-th aspect, which is a constant; where k = 1, 2, ..., 70, ε ijk For network center point deviation data; σ j Let γ be the node sensitivity region of the neural network in the j-th aspect, and γ be a constant parameter. 0ijk The radial basis coefficients of the convergent oscillatory network are given, where t is the time signal; c1 and c2 are constant parameters; and γ is the spectral density. ijk u is the radial basis function of the neural network; ij For the neural network's comprehensive output on the security of the i-th equipment in the j-th aspect, k jk p represents the weights of the neural network. i This represents the total safety score assessment data for the technical support of the i-th piece of equipment provided by the neural network. Step S40: The network training error data is obtained by comparing the total safety score of the i-th equipment with the comprehensive safety score of the i-th equipment after the superior expert has completed the task. Then, based on the network training error data, an adaptive adjustment law for weights based on a hybrid sinusoidal and cosine convergent oscillation is designed, and the network weights are adaptively updated through integral operations. When the network training error converges to the region near 0, training and weight updates are stopped. e i =p i -b i ; k jk (n+1)=k jk (n)+d jk ; Where e i b is the network training error data. i The score for the comprehensive safety assessment of the i-th piece of equipment after it completes its mission, as determined by higher-level experts; d jk This is based on a weight adaptive adjustment law of sine and cosine hybrid convergent oscillation. l jk These are constant parameters used to adjust the convergence speed of neural network weights; Step S50: Based on the safety score data of 25 aspects of the technical support process for the equipment under evaluation by the experts at this level, the data is substituted into the fuzzy system and then into the trained neural network to obtain the total safety score evaluation data of the neural network for the technical support of the equipment under evaluation as follows: e Djk =a Dj -f jk ; Where a Dj The data includes safety scores for 25 aspects of the technical support process for the equipment to be evaluated; ε Djk For the network center point deviation data of the equipment to be evaluated; γ 0Djk γ represents the radial basis coefficient of the convergent oscillatory network of the equipment to be evaluated. Djk u is the radial basis function of the neural network of the equipment to be evaluated. Dj For the comprehensive safety output of the neural network for the j-th aspect of the equipment to be evaluated, d D (a Dj p represents the fuzzy score for the safety of the equipment to be evaluated. D This is the total evaluation data for the safety score of the technical support for the equipment being evaluated, provided by the neural network.

Images