A safety analysis method for equipment under multi-task mode

Abstract

The present invention discloses a safety analysis method under a multi-task mode of equipment. The equipment is decomposed into multiple task modes according to the tasks. Each task mode is scored by experts according to six characteristics: duration, occupied space, frequency of use, economic value generated, overall operation complexity, and accident consequences. The weight coefficient of each mode on the above six characteristics is given, as well as the safety score and weight score of each mode in each stage in eleven aspects, including equipment structure, temperature system, grounding facilities, power distribution system, fire protection facilities, test equipment, equipment operation and inspection personnel, command decision-making personnel, support and maintenance personnel, working environment, and safety education. Then, a triangular anti-saturation basis function neural network is established, and an anti-saturation weight adjustment law is designed to train the network according to historical data. Finally, the final safety analysis data of the equipment technology under the multi-task mode is obtained by solving the problem.

Description

Technical Field

[0001] The present invention relates to the field of safety assessment of equipment under multi-task mode, and in particular to a safety assessment method of equipment under multi-task mode based on anti-saturation basis function neural network estimation and parameter training. Background Art

[0002] A complex multi-task modal equipment operation safety analysis is significantly different from the conventional multi-process equipment safety analysis. The characteristics of multi-task modality are that each task mode lasts for a different time, has a different background, and uses a different space. For multi-process operations, each process is inevitable, but the safety analysis is more complicated due to the large number of links. However, some modes of multi-task modal equipment operations may exist for several years, but some modes may only operate for a few days. Therefore, each task mode itself may have a very complex process. In general, its safety analysis requires a new method. Based on the above background reasons, the present invention proposes a method of using expert historical data of different task modes to construct a class of anti-saturation basis function neural networks, train the network so that the network error meets the requirements, and then input the current task modal data to solve the total data of the safety analysis. The experiment shows that the proposed method has a good safety assessment effect, so the present invention has a high practical value.

[0003] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to ordinary technicians in the field. Summary of the invention

[0004] The purpose of the present invention is to provide a security analysis method under a multi-task mode of a device, thereby overcoming the problems of low security assessment accuracy and slow network assessment speed caused by the limitations and defects of traditional assessment technologies.

[0005] According to one aspect of the present invention, a method for security analysis in a multi-task mode of a device is provided, comprising the following steps:

[0006] Step S10, decomposing the equipment into multiple task modes according to the tasks, and each task mode is scored by experts according to six characteristics: duration, occupied space, frequency of use, economic value generated, overall operation complexity, and accident consequences, and the weight coefficient of each mode on the above six characteristics is given. The absolute weight coefficient of each mode is solved comprehensively, and then the relative weight coefficient of each mode is solved nonlinearly;

[0007] Step S20, for each task mode, according to the six stages of on-site preparation, single-machine sub-test, comprehensive operation, overall assembly, overall test, and operation return, each stage is scored by experts according to the five dimensions of duration, occupied space, frequency of use, economic value generated, and overall operation complexity, and the dimensional weight coefficients of each stage in different dimensions are given at the same time, and then the total dimensional weight coefficients of each stage are obtained by comprehensive analysis, and then the total dimensional weight coefficients are processed to obtain the relative weight coefficients of each stage; at the same time, the experts score and give the safety scores and weight scores of each stage in eleven aspects, including equipment structure, temperature system, grounding facilities, power distribution system, fire protection facilities, test equipment, equipment operation and inspection personnel, command and decision-making personnel, maintenance personnel, working environment, and safety education, so as to solve the safety analysis data of the entire task mode;

[0008] Step S30, based on the relative weight coefficient past data of the existing task mode, the relative weight coefficient past data of each stage under the task mode, the safety score and weight score past data of eleven aspects under the task mode, the safety analysis past data of the task mode, and the total safety analysis past data under the multi-task mode of the equipment, a triangular anti-saturation basis function-based neural network is established, and a triangular anti-saturation basis function-type weight adjustment law is designed to train the network. After the given error requirement is met, the network training is stopped;

[0009] Step S40, according to the above-mentioned trained anti-saturation basis function-based neural network, the relative weight coefficient of the existing task mode, the relative weight coefficient data of each stage under the task mode, the safety score and weight score data of eleven aspects under the task mode, and the safety analysis data of the task mode are input into the trained network to obtain the final safety analysis data of the equipment technology under the multi-task mode.

[0010] In an exemplary embodiment of the present invention, the equipment is decomposed into multiple task modes according to the task. Each task mode is scored by experts according to the six characteristics of duration, occupied space, frequency of use, economic value generated, overall operation complexity, and accident consequences to give the weight coefficient of each mode on the above six characteristics. The absolute weight coefficient of each mode is solved comprehensively, and then the relative weight coefficient of each mode is solved nonlinearly, including:

[0011]

[0012]

[0013] Among them, H j is the relative weight coefficient of the jth task mode, n is the total number of modes, G j is the absolute weight coefficient of the jth task mode; A jis the weight coefficient of the jth mode on the duration feature; B j is the weight coefficient of the jth mode in the occupied space feature; C j is the weight coefficient of the jth mode in the use frequency feature; D j is the weight coefficient of the jth mode in the generated economic value characteristics; E j is the weight coefficient of the jth mode in the overall operation complexity characteristics; F j is the weight coefficient of the jth mode in the accident consequence characteristics.

[0014] In an exemplary embodiment of the present invention, for each task mode, according to the six stages of on-site preparation, single-machine sub-test, comprehensive operation, overall assembly, overall test, and operation return, each stage is scored by experts according to the five dimensions of duration, occupied space, frequency of use, economic value generated, and overall operation complexity. At the same time, the dimensional weight coefficients of each stage in different dimensions are given, and then the total dimensional weight coefficients of each stage are obtained by synthesis, and then the total dimensional weight coefficients are processed to obtain the relative weight coefficients of each stage; at the same time, the safety scores and weight scores of eleven aspects of equipment structure, temperature system, grounding facilities, power distribution system, fire protection facilities, test equipment, equipment operation and inspection personnel, command decision-making personnel, maintenance personnel, working environment, and safety education are given by experts, so as to solve the safety analysis data of the entire task mode, including:

[0015]

[0016]

[0017]

[0018] Among them, i = 1, 2, 3, 4, 5, 6 respectively represent the six stages of approach preparation, single machine sub-test, comprehensive operation, overall assembly, overall test, and operation return of each task mode; a i represents the dimension weight coefficient of each stage according to the duration analysis; b i represents the dimension weight coefficient of each stage according to the occupied space analysis; c i Represents the dimension weight coefficient of each stage according to the frequency of use analysis; d i Represents the dimension weight coefficient of each stage according to the economic value analysis; e i Indicates the dimension weight coefficient of each stage analyzed according to the overall task complexity; g i Represents the total dimension weight coefficient of each stage; h i Represents the relative weight coefficient of each stage; o 1iIndicates the safety score of each stage in terms of equipment structure, p 1i Indicates the weight score of each stage in terms of equipment structure; 2i Indicates the safety score of each stage in terms of temperature system, p 2i Indicates the weight score of each stage in the temperature system; 3i Indicates the safety score of grounding facilities at each stage, p 3i Indicates the weight score of each stage in terms of grounding facilities; 4i represents the safety score of the power distribution system at each stage, p 4i Indicates the weight score of each stage in the power distribution system; 5i Indicates the safety score of fire protection facilities at each stage, p 5i Indicates the weight score of fire protection facilities in each stage; 6i Indicates the safety score of the test equipment at each stage, p 6i Indicates the weight score of each stage in terms of test equipment; 7i Indicates the safety score of equipment operation and inspection personnel at each stage, p 7i Indicates the weight score of equipment operation and inspection personnel at each stage; 8i represents the safety score of the command decision-makers at each stage, p 8i Indicates the weight score of each stage in commanding decision-makers; 9i represents the safety score of each stage in terms of protecting maintenance personnel, p 9i Indicates the weight score of each stage in terms of ensuring maintenance personnel; 10i Indicates the safety score of the working environment at each stage, p 10i Indicates the weight of each stage in terms of the working environment; 11i Indicates the safety score of each stage in safety education, p 11i Indicates the weight of safety education in each stage; T j Represents the safety analysis data of the j-th task mode.

[0019] In an exemplary embodiment of the present invention, based on the relative weight coefficient past data of the existing task mode, the relative weight coefficient past data of each stage under the task mode, the safety score and weight score past data of eleven aspects under the task mode, the safety analysis past data of the task mode, and the total safety analysis past data under the multi-task mode of the device, a triangular anti-saturation basis function neural network is established, and a triangular anti-saturation weight adjustment law is designed to train the network. After reaching a given error requirement, stopping the training of the network includes:

[0020]

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027]

[0028] Among them, H aj is the relative weight coefficient of the jth task mode in the past data; h ai Represents the relative weight coefficient of each stage in the past data; a1i Indicates the past data of safety scores in equipment structure at each stage, p a1i Indicates the historical data of weight scores in equipment structure at each stage; a2i Indicates the safety score of each stage in the temperature system, historical data, p a2i Indicates the historical data of weight scores of each stage in the temperature system; a3i Indicates the historical data of the safety score of grounding facilities at each stage, p a3i Indicates the historical data of weight scores of grounding facilities in each stage; a4i Indicates the historical data of the safety score of the power distribution system at each stage, p a4i Represents the historical data of weight scores in the power distribution system at each stage; a5i Indicates the past data of safety scores of fire protection facilities in each stage, p a5i Indicates the historical data of weight scores of fire protection facilities in each stage; a6i Indicates the past data of the safety scores of the test equipment at each stage, p a6i Indicates the historical data of weight scores of test equipment in each stage; a7i Indicates the past data of safety scores of equipment operators and inspectors at each stage, p a7i Indicates the historical data of weight scores of equipment operation and inspection personnel at each stage; a8i represents the past data of the safety scores of the command decision-makers at each stage, p a8i Represents the historical data of weight scores of command decision-makers at each stage; a9iIndicates the past data of the safety score of maintenance personnel in each stage, p a9i Represents the historical data of the weight score for maintaining maintenance personnel at each stage; a10i Indicates the historical data of safety scores in the working environment at each stage, p a10i Indicates the historical data of the weight score of each stage in the working environment; a11i Indicates the past data of safety scores in safety education at each stage, p a11i Represents the historical data of the weight score of safety education in each stage; T aj represents the past data of the safety analysis data of the jth task mode; e represents the network training error; y a Indicates the total past data of security analysis in multi-tasking mode of the device; The triangular anti-saturation basis function weights for the main output of the network; The triangular anti-saturation basis function weights are auxiliary outputs of the network weights; is the triangular anti-saturation basis function weight output by the network's task mode; represents the main output of the network; Represents the weighted auxiliary output of the network; represents the task-mode output of the network; represents the total output of the network; is the triangle anti-saturation weight adjustment law of the main output of the network; k a1i is the network main output gain adaptive speed adjustment coefficient, which is a constant value; is the triangle anti-saturation weight adjustment law of the network weight auxiliary output; k eji is the gain adaptation speed adjustment coefficient of the network weight auxiliary output, which is a constant value; is the triangular anti-saturation weight adjustment law of the network's task mode output; k cji is the gain adaptive speed adjustment coefficient of the network's task mode output, which is a constant value; when e is less than the given error range, stop training the network.

[0029] In an exemplary embodiment of the present invention, according to the above-mentioned trained anti-saturation basis function-based neural network, the relative weight coefficient of the existing task mode, the relative weight coefficient data of each stage under the task mode, the safety score and weight score data of eleven aspects under the task mode, and the safety analysis data of the task mode are input into the trained network, and the final safety analysis data of the equipment technology under the multi-task mode is obtained, including:

[0030]

[0031]

[0032]

[0033] y=y1+y2+y3;

[0034] Among them, y1 represents the main output of the trained network after the current task modal data is input; y2 represents the auxiliary output of the network weight trained after the current task modal data is input; y3 represents the task modal output of the trained network after the current task modal data is input; y represents the total output of the trained network after the current task modal data is input, which is the final safety analysis data under the multi-task mode of the equipment technology.

[0035] Beneficial Effects

[0036] The present invention discloses a method for analyzing the safety of a device under a multi-task mode, and its main innovations are as follows: First, since the task mode segmentation can avoid the time and space overlap of the process segmentation, the safety is segmented according to the task mode, which is essentially different from the traditional segmentation according to the process, and can better match the actual safety mechanism of the device, making the safety analysis more in line with the actual situation. Second, a triangular anti-saturation basis neural network is designed, in which the trigonometric function is combined with anti-saturation, which can simulate the complex nonlinearity in the safety assessment and has dual anti-saturation characteristics, which greatly limits the divergence of the network, thereby making the network training faster and more convenient. Moreover, the complex frequency characteristics of the trigonometric function can make the network training have random ergodicity, which greatly shortens the network training time. As a result, the method of the present invention has good safety evaluation effect and rapidity.

[0037] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present invention, and together with the specification are used to explain the principles of the present invention. Obviously, the accompanying drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0039] Figure 1 It is a flow chart of a safety analysis method under a multi-tasking mode of a device provided by the present invention.

[0040] Figure 2 is a network training error curve of the method provided by the embodiment of the present invention;

[0041] Figure 3is a triangular anti-saturation basis function type weight curve of the network main output of the method provided in the embodiment of the present invention;

[0042] Figure 4 It is a triangular anti-saturation basis function type weight curve of the network weight auxiliary output of the method provided in the embodiment of the present invention;

[0043] Figure 5 It is a triangular anti-saturation basis function type weight curve of the task mode output of the network of the method provided by the embodiment of the present invention. DETAILED DESCRIPTION

[0044] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present invention will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present invention. However, those skilled in the art will appreciate that the technical solutions of the present invention may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid blurring out various aspects of the present invention.

[0045] The present invention provides a method for training a network by designing an anti-saturation weight adjustment law through a triangular anti-saturation basis neural network, decomposing equipment into multiple task modes according to tasks, and each task mode is divided into six characteristics according to duration, occupied space, frequency of use, economic value generated, overall operation complexity, and accident consequences, as well as six stages of preparation operation, single-machine sub-test, comprehensive operation, overall assembly, overall test, and operation return. Experts score and give safety analysis data of each stage of each task mode, and substitute the data into the trained network to solve and obtain the final safety analysis data of the equipment multi-task mode.

[0046] Next, a security analysis method for a device in a multi-task mode according to the present invention will be further explained and illustrated in conjunction with the accompanying drawings. Figure 1 As shown, the security analysis method of a device in a multi-task mode may include the following steps:

[0047] Step S10, decompose the equipment into multiple task modes according to the task. Each task mode is scored by experts according to six characteristics: duration, occupied space, frequency of use, economic value generated, overall operation complexity, and accident consequences. The weight coefficient of each mode on the above six characteristics is given, and the absolute weight coefficient of each mode is solved comprehensively, and then the relative weight coefficient of each mode is solved nonlinearly.

[0048] Specifically, the tasks are first decomposed into n task modes. Secondly, experts score each task mode according to six characteristics: duration, space occupied, frequency of use, economic value generated, overall operation complexity, and accident consequences, and give each mode a weight coefficient on the above six characteristics, which are recorded as A and B. j , B j , C j , D j 、E j 、F j .

[0049] Then, the absolute weight coefficient of each mode is calculated as follows:

[0050]

[0051] Where n is the total number of modes, G j is the absolute weight coefficient of the jth task mode; A j is the weight coefficient of the jth mode on the duration feature; B j is the weight coefficient of the jth mode in the occupied space feature; C j is the weight coefficient of the jth mode in the use frequency feature; D j is the weight coefficient of the jth mode in the generated economic value characteristics; E j is the weight coefficient of the jth mode in the overall operation complexity characteristics; F j is the weight coefficient of the jth mode in the accident consequence characteristics.

[0052] Finally, the relative weight coefficient of each mode is solved nonlinearly according to the following formula:

[0053]

[0054] Among them, H j is the relative weight coefficient of the j-th task mode.

[0055] Step S20, for each task mode, according to the six stages of on-site preparation work, single-machine sub-testing, comprehensive operation, overall assembly, overall testing, and operation return, each stage is scored according to five dimensions: duration, occupied space, frequency of use, economic value generated, and overall operation complexity. At the same time, the dimensional weight coefficients of each stage in different dimensions are given, and then the total dimensional weight coefficients of each stage are obtained by comprehensive analysis, and then the total dimensional weight coefficients are processed to obtain the relative weight coefficients of each stage; at the same time, the experts score and give the safety scores and weight scores of each stage in eleven aspects, including equipment structure, temperature system, grounding facilities, power distribution system, fire-fighting facilities, test equipment, equipment operation and inspection personnel, command and decision-making personnel, support and maintenance personnel, working environment, and safety education, so as to solve the safety analysis data of the entire task mode.

[0056] Specifically, each task mode is divided into six stages: on-site preparation work, single-machine sub-testing, comprehensive operation, overall assembly, overall testing, and operation return.

[0057] Then, each stage is scored by experts according to the five dimensions of duration, occupied space, frequency of use, economic value generated, and overall operation complexity. The dimensional weight coefficients of each stage in different dimensions are given at the same time, and the total dimensional weight coefficient of each stage is solved according to the following formula:

[0058]

[0059] Among them, i=1, 2, 3, 4, 5, 6 represent the six stages of approach preparation, single machine sub-test, comprehensive operation, overall assembly, overall test, and operation return of each task mode respectively; a i represents the dimension weight coefficient of each stage according to the duration analysis; b i represents the dimension weight coefficient of each stage according to the occupied space analysis; c i Represents the dimension weight coefficient of each stage according to the frequency of use analysis; d i Represents the dimension weight coefficient of each stage according to the economic value analysis; e i Indicates the dimension weight coefficient of each stage analyzed according to the overall task complexity; g i Represents the total dimension weight coefficient of each stage.

[0060] Secondly, the total dimension weight coefficient is processed to obtain the relative weight coefficient of each stage as follows:

[0061]

[0062] where h i Represents the relative weight coefficient of each stage.

[0063] Finally, the experts scored and gave the safety scores and weights of eleven aspects at each stage, including equipment structure, temperature system, grounding facilities, power distribution system, fire protection facilities, test equipment, equipment operation and inspection personnel, command and decision-making personnel, support and maintenance personnel, working environment, and safety education, so as to solve the safety analysis data of the entire task mode as follows:

[0064]

[0065] Among them 1i Indicates the safety score of each stage in terms of equipment structure, p 1i Indicates the weight score of each stage in terms of equipment structure; 2i Indicates the safety score of each stage in terms of temperature system, p 2i Indicates the weight score of each stage in the temperature system; 3i Indicates the safety score of grounding facilities at each stage, p 3i Indicates the weight score of each stage in terms of grounding facilities; 4i represents the safety score of the power distribution system at each stage, p 4i Indicates the weight score of each stage in the power distribution system; 5i Indicates the safety score of fire protection facilities at each stage, p 5i Indicates the weight score of fire protection facilities in each stage; 6i Indicates the safety score of the test equipment at each stage, p 6i Indicates the weight score of each stage in terms of test equipment; 7i Indicates the safety score of equipment operation and inspection personnel at each stage, p 7i Indicates the weight score of equipment operation and inspection personnel at each stage; 8i represents the safety score of the command decision-makers at each stage, p 8i Indicates the weight score of each stage in commanding decision-makers; 9i represents the safety score of each stage in terms of protecting maintenance personnel, p 9i Indicates the weight score of each stage in terms of ensuring maintenance personnel; 10i Indicates the safety score of the working environment at each stage, p 10i Indicates the weight of each stage in terms of the working environment; 11i Indicates the safety score of each stage in safety education, p 11i Indicates the weight of safety education in each stage; T j Represents the safety analysis data of the j-th task mode.

[0066] Step S30, based on the past data of relative weight coefficients of existing task modes, the past data of relative weight coefficients of each stage under the task mode, the past data of safety scores and weight scores of eleven aspects under the task mode, the past data of safety analysis of the task mode, and the past total data of safety analysis under the multi-task mode of the equipment, a triangular anti-saturation basis function-based neural network is established, and a triangular anti-saturation basis function-type weight adjustment law is designed to train the network. After the given error requirement is met, the network training is stopped.

[0067] Specifically, a class of triangular anti-saturation basis function neural networks is first established as follows:

[0068]

[0069]

[0070]

[0071] Among them, H aj is the relative weight coefficient of the jth task mode in the past data; h ai Represents the relative weight coefficient of each stage in the past data; a1i Indicates the past data of safety scores in equipment structure at each stage, p a1i Indicates the historical data of weight scores in equipment structure at each stage; a2i Indicates the safety score of each stage in the temperature system, historical data, p a2i Indicates the historical data of weight scores of each stage in the temperature system; a3i Indicates the historical data of the safety score of grounding facilities at each stage, p a3i Indicates the historical data of weight scores of grounding facilities in each stage; a4i Indicates the historical data of the safety score of the power distribution system at each stage, p a4i Represents the historical data of weight scores in the power distribution system at each stage; a5i Indicates the past data of safety scores of fire protection facilities in each stage, p a5i Indicates the historical data of weight scores of fire protection facilities in each stage; a6i Indicates the past data of the safety scores of the test equipment at each stage, p a6i Indicates the historical data of weight scores of test equipment in each stage; a7i Indicates the past data of safety scores of equipment operators and inspectors at each stage, p a7i Indicates the historical data of weight scores of equipment operation and inspection personnel at each stage; a8irepresents the past data of the safety scores of the command decision-makers at each stage, p a8i Represents the historical data of weight scores of command decision-makers at each stage; a9i Indicates the past data of the safety score of maintenance personnel in each stage, p a9i Represents the historical data of the weight score for maintaining maintenance personnel at each stage; a10i Indicates the historical data of safety scores in the working environment at each stage, p a10i Indicates the historical data of the weight score of each stage in the working environment; a11i Indicates the past data of safety scores in safety education at each stage, p a11i Represents the historical data of the weight score of safety education in each stage; T aj Represents the past data of safety analysis data of the j-th task mode; The triangular anti-saturation basis function weights for the main output of the network; The triangular anti-saturation basis function weights are auxiliary outputs of the network weights; is the triangular anti-saturation basis function weight output by the network's task mode; represents the main output of the network; Represents the weighted auxiliary output of the network; Represents the task-modal output of the network.

[0072] Secondly, using the past data of relative weight coefficients of existing task modes, the past data of relative weight coefficients of each stage under the task mode, the past data of safety scores and weight scores of eleven aspects under the task mode, the past data of safety analysis of task modes, and the past total data of safety analysis under the multi-task mode of equipment, the network error training is solved as follows:

[0073]

[0074]

[0075] Among them, e represents the network training error; y a Indicates the total past data of security analysis in multi-tasking mode of the device; Represents the total output of the network.

[0076] Finally, the network training error is used to design an anti-saturation weight adjustment rule to train the network as follows:

[0077]

[0078]

[0079]

[0080] in is the triangular anti-saturation basis function weight adjustment law of the main output of the network; k a1i is the network main output gain adaptive speed adjustment coefficient, which is a constant value. The detailed design is shown in the case implementation below; is the triangular anti-saturation basis function weight adjustment law of the network weight auxiliary output; k eji is the gain adaptation speed adjustment coefficient of the network weight auxiliary output, which is a constant value. The detailed design is shown in the case implementation below; is the triangular anti-saturation basis function weight adjustment law of the network's task mode output; k cji is the gain adaptive speed adjustment coefficient of the network's task mode output, which is a constant value. The detailed design is shown in the case implementation below. When the network training error e is less than the given value, the training is stopped.

[0081] Step S40, according to the above-mentioned trained anti-saturation basis function-based neural network, the relative weight coefficient of the existing task mode, the relative weight coefficient data of each stage under the task mode, the safety score and weight score data of eleven aspects under the task mode, and the safety analysis data of the task mode are input into the trained network to obtain the final safety analysis data of the equipment technology under the multi-task mode.

[0082] Specifically, according to the triangular anti-saturation basis function weights of the main output of the network, the triangular anti-saturation basis function weights of the auxiliary output of the network weight, and the triangular anti-saturation basis function weights of the task mode output of the network, the relative weight coefficient H of the existing task mode is substituted. i , the relative weight coefficient data h of each stage under the task mode i , safety scores of eleven aspects under task mode ji With weight score data p ji , safety analysis data of task mode T j , the final equipment technical preparation process safety conclusion data y is as follows:

[0083]

[0084]

[0085]

[0086] y=y1+y2+y3;

[0087] in The triangular anti-saturation basis function weights for the main output of the network; The triangular anti-saturation basis function weights are auxiliary outputs of the network weights; is the triangular anti-saturation basis function weight of the network's task modal output; y1 represents the main output of the network trained after the current task modal data is input; y2 represents the auxiliary output of the network weight trained after the current task modal data is input; y3 represents the task modal output of the network trained after the current task modal data is input; y represents the total output of the network trained after the current task modal data is input, which is the final safety analysis data of the equipment technology under multi-task mode.

[0088] Case implementation and computer simulation results analysis

[0089] In step S10, n=3 is selected, and the device is decomposed into three task modes: short-term task mode, medium-term task mode, and long-term task mode according to the task.

[0090] In step S20, each task is decomposed and scored by experts according to the six stages, five dimensions and eleven aspects described in the content of the invention.

[0091] In step S30, the triangular anti-saturation basis function weights of the main output of the network are set The initial value of is 3; set the triangular anti-saturation basis function weight of the network weight auxiliary output The initial value is -5; set the triangular anti-saturation basis function weights of the network's task mode output The initial value of is 5;

[0092] Set the network main output gain adaptive speed adjustment coefficient k a1i =0.05; Set the gain adaptive speed adjustment coefficient k of the network weight auxiliary output eji =0.03, set the gain adaptive speed adjustment coefficient k of the network's task mode output cji =0.01; the final network training error curve is as follows Figure 2 As shown; the triangular anti-saturation basis function weight curve of the main output of the network is as follows Figure 3 As shown; the triangular anti-saturation basis function weights of the network weight auxiliary output are as follows Figure 4 As shown; the triangular anti-saturation basis function weights of the network's task mode output are as follows Figure 5 As shown. Figure 3 , Figure 4 and Figure 5 It can be seen that the final weights are stable around -6, 15 and 5. Figure 2 It can be seen that the network training error finally stabilizes at around 0, indicating that after good training, the network has become completely stable and the error has gradually decreased to 0.

[0093] In step S40, the relative weight coefficient H of the existing task mode is substituted i , the relative weight coefficient data h of each stage under the task mode i , safety scores of eleven aspects under task mode ji With weight score data p ji , safety analysis data of task mode T j ,According to the trained network calculation, the final safety analysis data of the equipment technology under the multi-task mode is 0.951, and the safety assessment conclusion is excellent, which is consistent with the final result.

Claims

1. A security analysis method for a device in a multi-task mode, characterized by the following steps: Step S10, the equipment is decomposed into multiple task modes according to the task. Each task mode is scored by experts according to the six characteristics of duration, occupied space, frequency of use, economic value generated, overall operation complexity, and accident consequences to give the weight coefficient of each mode on the above six characteristics. The absolute weight coefficient of each mode is solved comprehensively, and then the relative weight coefficient of each mode is solved by nonlinear method, including: Among them, H j is the relative weight coefficient of the jth task mode, n is the total number of modes, G j is the absolute weight coefficient of the jth task mode; A j is the weight coefficient of the jth mode on the duration feature; B j is the weight coefficient of the jth mode in the occupied space feature; C j is the weight coefficient of the jth mode in the use frequency feature; D j is the weight coefficient of the jth mode in the generated economic value characteristics; E j is the weight coefficient of the jth mode in the overall operation complexity characteristics; F j is the weight coefficient of the jth mode in the accident consequence characteristics; Step S20, for each task mode, according to the six stages of on-site preparation, single-machine sub-test, comprehensive operation, overall assembly, overall test, and operation return, each stage is scored by experts according to the five dimensions of duration, occupied space, frequency of use, economic value generated, and overall operation complexity, and the dimensional weight coefficients of each stage in different dimensions are given at the same time, and then the total dimensional weight coefficients of each stage are obtained by comprehensive analysis, and then the total dimensional weight coefficients are processed to obtain the relative weight coefficients of each stage; at the same time, the experts score and give the safety scores and weight scores of each stage in eleven aspects, including equipment structure, temperature system, grounding facilities, power distribution system, fire protection facilities, test equipment, equipment operation and inspection personnel, command decision-making personnel, maintenance personnel, working environment, and safety education, so as to solve the safety analysis data of the entire task mode, including: Among them, i = 1, 2, 3, 4, 5, 6 respectively represent the six stages of approach preparation, single machine sub-test, comprehensive operation, overall assembly, overall test, and operation return of each task mode; a i represents the dimension weight coefficient of each stage according to the duration analysis; b i represents the dimension weight coefficient of each stage according to the occupied space analysis; c i Represents the dimension weight coefficient of each stage according to the frequency of use analysis; d i Represents the dimension weight coefficient of each stage according to the economic value analysis; e i Indicates the dimension weight coefficient of each stage analyzed according to the overall task complexity; g i Represents the total dimension weight coefficient of each stage; h i Represents the relative weight coefficient of each stage; o 1i Indicates the safety score of each stage in terms of equipment structure, p 1i Indicates the weight score of each stage in terms of equipment structure; 2i Indicates the safety score of each stage in terms of temperature system, p 2i Indicates the weight score of each stage in the temperature system; 3i Indicates the safety score of grounding facilities at each stage, p 3i Indicates the weight score of each stage in terms of grounding facilities; 4i represents the safety score of the power distribution system at each stage, p 4i Indicates the weight score of each stage in the power distribution system; 5i Indicates the safety score of fire protection facilities at each stage, p 5i Indicates the weight score of fire protection facilities in each stage; 6i Indicates the safety score of the test equipment at each stage, p 6i Indicates the weight score of each stage in terms of test equipment; 7i Indicates the safety score of equipment operation and inspection personnel at each stage, p 7i Indicates the weight score of equipment operation and inspection personnel at each stage; 8i represents the safety score of the command decision-makers at each stage, p 8i Indicates the weight score of each stage in commanding decision-makers; 9i represents the safety score of each stage in terms of protecting maintenance personnel, p 9i Indicates the weight score of each stage in terms of ensuring maintenance personnel; 10i Indicates the safety score of the working environment at each stage, p 10i Indicates the weight of each stage in terms of the working environment; 11i Indicates the safety score of each stage in safety education, p 11i Indicates the weight of safety education in each stage; T j Represents the safety analysis data of the jth task mode; Step S30, based on the relative weight coefficient past data of the existing task mode, the relative weight coefficient past data of each stage under the task mode, the safety score and weight score past data of eleven aspects under the task mode, the safety analysis past data of the task mode, and the total safety analysis past data of the equipment under the multi-task mode, a triangular anti-saturation basis function neural network is established, and an anti-saturation weight adjustment law is designed to train the network. After the given error requirement is met, the network training is stopped, and then the relative weight coefficient of the existing task mode, the relative weight coefficient data of each stage under the task mode, the safety score and weight score data of eleven aspects under the task mode, and the safety analysis data of the task mode are input into the trained network to obtain the final equipment technology safety analysis data under the multi-task mode, including: y=y1+y2+y3; Among them, H aj is the relative weight coefficient of the jth task mode in the past data; h ai Represents the relative weight coefficient of each stage in the past data; a1i Indicates the past data of safety scores in equipment structure at each stage, p a1i Indicates the historical data of weight scores in equipment structure at each stage; a2i Indicates the safety score of each stage in the temperature system, historical data, p a2i Indicates the historical data of weight scores of each stage in the temperature system; a3i Indicates the historical data of the safety score of grounding facilities at each stage, p a3i Indicates the historical data of weight scores of grounding facilities in each stage; a4i Indicates the historical data of the safety score of the power distribution system at each stage, p a4i Represents the historical data of weight scores in the power distribution system at each stage; a5i Indicates the past data of safety scores of fire protection facilities in each stage, p a5i Indicates the historical data of weight scores of fire protection facilities in each stage; a6i Indicates the past data of the safety scores of the test equipment at each stage, p a6i Indicates the historical data of weight scores of test equipment in each stage; a7i Indicates the past data of safety scores of equipment operators and inspectors at each stage, p a7i Indicates the historical data of weight scores of equipment operation and inspection personnel at each stage; a8i represents the past data of the safety scores of the command decision-makers at each stage, p a8i Represents the historical data of weight scores of command decision-makers at each stage; a9i Indicates the past data of the safety score of maintenance personnel in each stage, p a9i Represents the historical data of the weight score for maintaining maintenance personnel at each stage; a10i Indicates the historical data of safety scores in the working environment at each stage, p a10i Indicates the historical data of the weight score of each stage in the working environment; a11i Indicates the past data of safety scores in safety education at each stage, p a11i Represents the historical data of the weight score of safety education in each stage; T aj represents the past data of the safety analysis data of the jth task mode; e represents the network training error; y a Indicates the total past data of security analysis in multi-tasking mode of the device; The triangular anti-saturation basis function weights for the main output of the network; The triangular anti-saturation basis function weights are auxiliary outputs of the network weights; is the triangular anti-saturation basis function weight output by the network's task mode; represents the main output of the network; Represents the weighted auxiliary output of the network; represents the task-mode output of the network; represents the total output of the network; is the triangular anti-saturation basis function weight adjustment law of the main output of the network; k a1i is the network main output gain adaptive speed adjustment coefficient, which is a constant value; is the triangular anti-saturation basis function weight adjustment law of the network weight auxiliary output; k eji is the gain adaptation speed adjustment coefficient of the network weight auxiliary output, which is a constant value; is the triangular anti-saturation basis function weight adjustment law of the network's task mode output; k cji is the gain adaptive speed adjustment coefficient of the network's task modal output, which is a constant; y1 represents the main output of the trained network after the current task modal data is input; y2 represents the auxiliary output of the trained network weight after the current task modal data is input; y3 represents the task modal output of the trained network after the current task modal data is input; y represents the total output of the trained network after the current task modal data is input, which is the final safety analysis data of the equipment technology under multi-task mode.

Images