Digital-twin-based method and system for virtual sensor placement of GIS

Abstract

The application provides a GIS virtual sensor point arrangement method and system based on digital twinning, which comprises the following steps: setting different fault types at different positions in a GIS digital twinning gas chamber; arranging multiple groups of virtual sensors in the GIS gas chamber, and numbering the virtual sensor groups; respectively simulating the set fault types based on the GIS digital twinning; drawing a data change curve with fault data as the independent variable and the measured value at the virtual sensor point as the dependent variable; performing linear fitting on the data change curve, and obtaining a slope which reflects the sensitivity of the virtual sensor to the fault data; and using a genetic algorithm to take the obtained slope as an adaptability index to optimize the combination of the virtual sensor points, and obtaining the optimized virtual sensor combination.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of virtual simulation, and particularly relates to a GIS virtual sensor layout method and system based on digital twinning. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute the prior art.

[0003] The position and number of physical sensors installed in actual gas insulated switchgear (GIS) equipment are often limited. And because GIS is a complete and sealed structure, the data obtained by physical sensors is very limited.

[0004] Existing research on sensor layout is usually directed at physical sensors. By establishing a GIS digital twin model, GIS equipment can be digitized to reflect the internal operating state of GIS.

[0005] Due to the complex internal structure of GIS and the large differences in the propagation rules and characteristics of different characteristic parameters in GIS, if virtual sensors are laid out at different positions in the GIS digital twin, the obtained parameter values will differ greatly. And because the GIS digital twin has a large amount of data, placing too many virtual sensors in the GIS will cause data redundancy and affect subsequent state evaluation and other applications. SUMMARY

[0006] To overcome the shortcomings of the prior art, the present application provides a GIS virtual sensor layout method based on digital twinning. The virtual sensor groups laid out have high sensitivity to simulated faults, effectively reducing data redundancy.

[0007] To achieve the above-mentioned purposes, one or more embodiments of the present application provide the following technical solutions:

[0008] In a first aspect, a GIS virtual sensor layout method based on digital twinning is disclosed, comprising:

[0009] Different fault types are set at different positions in the gas chamber of the GIS digital twin;

[0010] Multiple groups of virtual sensors are laid out in the GIS gas chamber, and the virtual sensor groups are numbered;

[0011] The set fault types are simulated based on the GIS digital twin;

[0012] Data change curves are drawn with fault data as the independent variable and the measured values at the virtual sensor points as the dependent variable;

[0013] Linear fitting is performed on the data variation curve, and the slope obtained reflects the sensitivity of the virtual sensor to the fault data;

[0014] The genetic algorithm is used to obtain the slope as the fitness index, and the combination of the virtual sensor distribution points is optimized to obtain the optimized virtual sensor combination.

[0015] As a further technical solution, the fault position in the GIS digital twin gas chamber is randomly selected, and is distributed in the main area of the GIS digital twin gas chamber.

[0016] As a further technical solution, when the GIS digital twin gas chamber is a cylindrical structure, the initial sensor is arranged in a circular ring, and a virtual sensor point is arranged every set angle.

[0017] As a further technical solution, the genetic algorithm is used to obtain the slope as the fitness index, and the combination of the virtual sensor distribution points is optimized, specifically:

[0018] Each group of virtual sensors is divided into a, b, c, and d four areas;

[0019] In each group, one virtual sensor is randomly selected in each of the a, b, c, and d four areas;

[0020] An initial population is generated, the population size, the maximum number of iterations, the crossover ratio, and the selection process are set, and the tournament mechanism is used, and the mutation ratio is set;

[0021] In the setting of the fitness function, the obtained slope is used as the fitness index, and the positions of the virtual sensors that are extremely insensitive are also considered;

[0022] Through genetic algorithm optimization, the optimized virtual sensor combination can be obtained.

[0023] As a further technical solution, it also includes: using support vector machine classification to verify the virtual sensor distribution point method after optimization.

[0024] In the second aspect, a GIS virtual sensor distribution system based on digital twinning is disclosed, which includes:

[0025] The fault type setting module is configured to set different fault types at different positions in the GIS digital twin gas chamber;

[0026] Multiple groups of virtual sensors are arranged in the GIS gas chamber, and the virtual sensor groups are numbered;

[0027] The simulation module is configured to simulate the set fault types based on the GIS digital twin.

[0028] Take the fault data as the independent variable, and the measured value at the virtual sensor point as the dependent variable to draw a data change curve;

[0029] Linear fitting is performed on the data change curve, and the slope obtained reflects the sensitivity of the virtual sensor at the place to the fault data;

[0030] The optimization module is configured to use a genetic algorithm to obtain the slope as an adaptive index to optimize the combination of virtual sensor point distribution, and obtain the virtual sensor combination after optimization.

[0031] The above one or more technical solutions have the following beneficial effects:

[0032] The GIS internal virtual sensor point distribution method of the present application is different from arranging physical sensors and is not restricted by position and quantity factors.

[0033] The GIS virtual sensor point distribution method based on digital twinning proposed in the present application can arrange virtual sensors according to the characteristics of different fault parameters in GIS. The virtual sensor group arranged has high sensitivity to simulated faults, effectively reduces data redundancy, and provides help for subsequent GIS digital twinning body state evaluation and other applications.

[0034] The advantages of the additional aspects of the present application will be partially given in the following description, partially will become obvious from the following description, or will be understood through the practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0035] The drawings accompanying the specification of the present application form part of the present application and serve to provide further understanding of the present application. The illustrative embodiments of the present application and their description serve to explain the present application. They do not constitute an improper limitation of the present application.

[0036] Figure 1 An initial virtual sensor point distribution schematic diagram for the embodiment of the present application;

[0037] Figure 2 A single-group virtual sensor region division schematic diagram for the embodiment of the present application;

[0038] Figure 3 A virtual sensor point distribution schematic diagram after optimization for the embodiment of the present application;

[0039] Figure 4 A virtual sensor point distribution number comparison schematic diagram before and after optimization for the embodiment of the present application;

[0040] Figure 5 A virtual sensor point distribution method overall flowchart for the embodiment of the present application. DETAILED DESCRIPTION

[0041] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0042] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments consistent with the present application.

[0043] In the case of no conflict, the embodiments in the application and the features in the embodiments can be combined with each other.

[0044] Embodiment one

[0045] The embodiment discloses a GIS virtual sensor distribution method based on digital twinning, selects appropriate fitness indicators according to the characteristics of different characteristic parameters in GIS, and optimizes the distribution position of the virtual sensor in GIS by using a genetic algorithm and support vector machine classification verification, so that the distribution position of the virtual sensor in GIS is optimized, and data redundancy is reduced.

[0046] In the embodiment, the specific method comprises:

[0047] First, different fault types are selected for fault setting at different positions in the GIS digital twinning gas chamber. For example, 10 overheating fault positions are set in the GIS gas chamber, the fault positions are randomly selected, and the main areas in the GIS gas chamber are distributed.

[0048] Then, initial virtual sensors are arranged in the GIS gas chamber. Considering the cylindrical structure of the GIS gas chamber, the initial sensors are arranged in a circular ring. A virtual sensor point is arranged every 30°, and multiple groups of virtual sensors are arranged in the GIS gas chamber. As shown in Figure 1 The initial arrangement of 6 groups of virtual sensors is shown, each group has 12 virtual sensors, and a total of 72 virtual sensors.

[0049] In order to facilitate subsequent selection of virtual sensor groups 1 to 6, each virtual sensor is numbered from 1 to 72. Each group of virtual sensors is divided into a, b, c, and d four areas as shown in Figure 2 .

[0050] After completion, the previously set faults are simulated based on the GIS digital twin. For example, if 10 overheat faults are set, they are simulated ten times based on the GIS digital twin, and the values obtained by all virtual sensors in the 10 simulations are recorded. Taking overheat faults as an example, the overheat temperature at the overheat fault position is set to vary from 60°C to 100°C at intervals of 5°C. For one fault position, a virtual sensor will obtain 9 temperature values at that point. For one fault position, a virtual sensor, with overheat temperature as the independent variable and temperature value at the virtual sensor point as the dependent variable, can draw a temperature change curve. The least squares method is used to linearly fit the temperature curve, and the slope obtained reflects the sensitivity of the virtual sensor at that position to the fault temperature.

[0051] Then the genetic algorithm is used, and the obtained slope is used as the fitness index. The specific process is shown in the flowchart of Figure 5 The specific implementation method is as follows: three groups are randomly selected from the six groups of virtual sensors, and one virtual sensor is randomly selected from the a, b, c, and d regions in each group as shown in Figure 2 The initial parent population is generated, the initial parent population size is set to 300, the maximum number of iterations is set to 100, the crossover ratio is set to 80%, the tournament mechanism is used in the selection process, and the mutation ratio is set to 2%. Then the number of child populations is 240, and the initial parent population and the child population are merged to form a new population. In the setting of the fitness function, the slopes of the fitting curves of all virtual sensor positions in the population are added as the fitness index, and if there are extremely insensitive virtual sensor positions, the fitness index is penalized. Then the new population is sorted from high to low according to the fitness index, and the top 300 groups are selected as the new parent population to be recycled until the maximum number of iterations is reached.

[0052] The optimized virtual sensor combination can be obtained through the genetic algorithm optimization. The optimized virtual sensor distribution is shown in Figure 3 The optimized virtual sensor group numbers are 2, 3, and 5. The optimized single virtual sensor numbers are 13, 17, 19, 21, 25, 27, 31, 33, 49, 53, 54, and 57. As shown in Figure 4 The number of virtual sensor distribution points is reduced from 72 to 12 after optimization, and the number of data groups is also reduced from 72 to 12. After optimization, insensitive virtual sensors are removed, and data redundancy is reduced.

[0053] Then support vector machine classification is used to verify the optimized virtual sensor distribution method. Taking overheat faults as an example, first set A and B fault positions in the GIS chamber, and the fault temperatures are set from 50°C to 100°C at intervals of 5°C. As shown in Table 1, the fault degree is divided into 1 overheat warning and 2 severe overheat according to the temperature range.

[0054] Table 1 fault degree classification

[0055]

[0056] As shown in Table 2, according to the fault location and the fault degree of the fault point, the fault type is divided into four types, 1 represents overheating warning at two locations, 2 represents overheating warning at fault location A and serious overheating at fault location B, 3 represents serious overheating at fault location A and overheating warning at fault location B, and 4 represents serious overheating at both locations A and B.

[0057] Table 2 fault degree and location classification

[0058]

[0059] Then, support vector machine classification is used to train the model by using the fault simulation data obtained by the optimized virtual sensor group, and test whether the optimized virtual sensor group can judge the severity and location of the fault prone location. Under the above set overheating fault condition, the test set accuracy rate reaches 96%, which is more than 90%, and the optimized virtual sensor distribution meets the requirements. If the test set accuracy rate is too low, less than 80%, the initial arrangement of the virtual sensor is adjusted, the number of virtual sensor groups is increased, the number of virtual sensors in a single circle is increased, or the position of the virtual sensor group is changed. The overall flow chart of the GIS virtual sensor distribution method based on digital twinning is shown in Figure 5 .

[0060] Example two

[0061] The purpose of this embodiment is to provide a computer device, including a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to realize the steps of the above method.

[0062] Example three

[0063] The purpose of this embodiment is to provide a computer readable storage medium.

[0064] A computer readable storage medium has a computer program stored thereon, which is executed by a processor to perform the steps of the above method.

[0065] Example four

[0066] The purpose of this embodiment is to provide a GIS virtual sensor distribution system based on digital twinning, which includes:

[0067] The fault type setting module is configured to set different fault types at different locations in the GIS digital twin gas chamber;

[0068] A plurality of virtual sensor groups are arranged in the GIS gas chamber, and the virtual sensor groups are numbered;

[0069] The simulation module is configured to simulate the set fault types respectively based on the GIS digital twin;

[0070] Taking the fault data as the independent variable and the measured value at the virtual sensor point as the dependent variable, a data change curve is drawn;

[0071] The data change curve is linearly fitted, and the slope obtained reflects the sensitivity of the virtual sensor at the point to the fault data;

[0072] The optimization module is configured to use a genetic algorithm to optimize the combination of virtual sensor point arrangement based on the obtained slope as the fitness index, and obtain the optimized virtual sensor combination.

[0073] The steps and methods involved in the devices of embodiments two, three and four correspond to embodiment one, and the specific implementation can refer to the relevant description part of embodiment one. The term "computer readable storage medium" should be understood to include a single medium or multiple media of one or more instruction sets; it should also be understood to include any medium that can store, encode or carry instruction sets for execution by a processor and make the processor execute any method in the present application.

[0074] Those skilled in the art should understand that each module or step of the present application described above can be realized by a general computer device, alternatively, they can be realized by program codes executable by a computing device, so that they can be stored in a storage device for execution by a computing device, or they can be respectively made into individual integrated circuit modules, or a plurality of modules or steps among them can be made into a single integrated circuit module to realize. The present application is not limited to any specific combination of hardware and software.

[0075] Although the specific embodiments of the present application are described above in combination with the accompanying drawings, it is not a limitation on the scope of protection of the present application, and those skilled in the art should understand that various modifications or changes made by those skilled in the art on the basis of the technical solutions of the present application without creative labor are still within the scope of protection of the present application.

Claims

1. A GIS virtual sensor layout method based on digital twinning, characterized in that, The method comprises the following steps: Different fault types are set at different positions in the GIS digital twin gas chamber; A plurality of groups of virtual sensors are arranged in the GIS gas chamber, and the groups of virtual sensors are numbered; The set fault types are simulated based on the GIS digital twin; Data change curves are drawn with fault data as independent variables and measured values at virtual sensor points as dependent variables; Linear fitting is performed on the data change curves, and the obtained slope reflects the sensitivity of the virtual sensor at the position to the fault data; A genetic algorithm is used to obtain the slope as an adaptability index to optimize the combination of virtual sensor points, and an optimized virtual sensor combination is obtained; The genetic algorithm is used to obtain the slope as an adaptability index to optimize the combination of virtual sensor points, specifically as follows: Each group of virtual sensors is divided into four regions a, b, c, and d; In each group, one virtual sensor is randomly selected in each of the four regions a, b, c, and d; An initial population is generated, and the population size, maximum number of iterations, crossover ratio, and mutation ratio are set; In the setting of the adaptability function, the obtained slope is used as an adaptability index, and extremely insensitive virtual sensor positions are also considered; An optimized virtual sensor combination can be obtained through genetic algorithm optimization.

2. The GIS virtual sensor layout method based on digital twinning of claim 1, wherein, The fault positions in the GIS digital twin gas chamber are randomly selected, and are distributed in the main regions of the GIS digital twin gas chamber. 3.The GIS virtual sensor layout method based on digital twinning of claim 1, wherein, When the GIS digital twin gas chamber is in a cylindrical structure, the initial sensors are arranged in a circular ring, and a virtual sensor point is arranged every certain angle.

4. The GIS virtual sensor siting method based on digital twinning of claim 1, wherein, The method further comprises: using a support vector machine classification to verify the optimized virtual sensor point arrangement.

5. The GIS virtual sensor layout system based on digital twinning, characterized in that, The method comprises the following steps: A fault type setting module is configured to set different fault types at different positions in the GIS digital twin gas chamber; A plurality of groups of virtual sensors are arranged in the GIS gas chamber, and the groups of virtual sensors are numbered; A simulation module is configured to simulate the set fault types based on the GIS digital twin; Data change curves are drawn with fault data as independent variables and measured values at virtual sensor points as dependent variables; Linear fitting is performed on the data change curves, and the obtained slope reflects the sensitivity of the virtual sensor at the position to the fault data; An optimization module is configured to use a genetic algorithm to obtain the slope as an adaptability index to optimize the combination of virtual sensor points, and obtain an optimized virtual sensor combination; The genetic algorithm is used to obtain the slope as an adaptability index to optimize the combination of virtual sensor points, specifically as follows: Each group of virtual sensors is divided into four regions a, b, c, and d; In each group, one virtual sensor is randomly selected in each of the four regions a, b, c, and d; An initial population is generated, and the population size, maximum number of iterations, crossover ratio, and mutation ratio are set; In the setting of the adaptability function, the obtained slope is used as an adaptability index, and extremely insensitive virtual sensor positions are also considered; An optimized virtual sensor combination can be obtained through genetic algorithm optimization.

6. The digital-twin-based GIS virtual sensor siting system of claim 5, wherein, The fault type in the simulation module is an overheat fault.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1-4 when executing the program.

8. A computer readable storage medium having stored thereon a computer program, wherein the program, when executed by a processor, performs the steps of the method according to any one of claims 1-4.

Images