An iterative local search method for solving the area coverage problem of distribution network sensors

By establishing a mathematical optimization model and adjusting the circular area radius method, the problem of uneven coverage of the distribution network sensor area is solved, the coverage rate and monitoring effect are improved, and the occurrence of heating accidents in line equipment is reduced.

CN114626171BActive Publication Date: 2024-11-19NANJING UNIV OF POSTS & TELECOMM
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210237170.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-10
Publication Date
2024-11-19
Estimated Expiration
2042-03-10

AI Technical Summary

Technical Problem

The existing technology is difficult to effectively solve the problem of sensor area coverage in the distribution network, resulting in uneven network coverage and low coverage rate, which affects the safety monitoring of power line equipment.

Method used

By establishing a mathematical optimization model and introducing a method of adjusting the circular area radius, an iterative local search method is used to optimize the sensor layout and area radius to improve the coverage effect.

Benefits of technology

The uniformity and coverage rate of the distribution network sensor area coverage effect are achieved, the monitoring effect of power line equipment is improved, and the occurrence of heating accidents in line equipment is reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114626171B_ABST
    Figure CN114626171B_ABST
Patent Text Reader

Abstract

An iterative local search method for solving the problem of distribution network sensor area coverage includes the following steps: establishing a mathematical optimization model for the distribution network sensor circular area coverage problem; randomly generating an initial solution; executing an iterative local search method; if the iterative local search method obtains a legal solution, calling an area adjustment method to adjust the radius of the circular area; when the algorithm reaches the maximum time limit, outputting the obtained optimal solution. This method can achieve extremely high coverage of the distribution network sensor area by establishing a continuous optimization model and designing an adjustment method for the circular area radius. This method has fast calculation speed and high solution accuracy. In the actual construction process of the distribution network sensor, a high-quality sensor layout plan can be obtained within a relatively short time frame, thereby improving the monitoring capability of the sensor in the distribution network and reducing the construction cost.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention is applicable to the technical field of distribution network sensor networks and provides an iterative local search method for solving the problem of distribution network sensor area coverage. Background Art

[0002] Wireless Sensor Networks (WSN) are mobile self-organizing networks composed of a large number of miniature, low-cost and energy-resource-constrained sensor nodes to achieve effective monitoring of the target area. Appropriate network deployment is crucial for wireless sensor networks. A good network layout can significantly improve network node coverage, reduce network energy loss, and increase network robustness.

[0003] With the development of science and technology and the popularization and application of sensor networks, wireless sensor networks play an increasingly important role in the field of smart distribution networks. In recent years, substation decompression accidents caused by heating of line equipment have brought huge economic losses to power companies, and have also had a certain impact on the reputation and safety of power companies. At present, the heating of line equipment has become an urgent problem to ensure the safe operation of transmission and distribution lines. Therefore, power companies should increase their attention to preventing the heating of transmission and distribution line equipment, and use wireless temperature sensors to effectively monitor the temperature of power grid lines, so as to promptly discover the heating problem of line equipment, and then take effective strategies to solve the problems found. Summary of the invention

[0004] The present invention can make the coverage effect of the distribution network sensor area coverage problem more uniform and achieve a better coverage rate by establishing a mathematical optimization model and introducing a method for adjusting the circular area radius.

[0005] An iterative local search method for solving the distribution network sensor area coverage problem includes the following steps:

[0006] Step 1: Establish a coordinate system based on the distribution network sensor area, construct an evaluation function for evaluating the size of the uncovered area, and establish a mathematical optimization model for the distribution network sensor area coverage problem;

[0007] Step 2: Input the number of small circles N representing sensors, the initial radius R0 of the circular area to be covered and the radius r of the small circle; set parameters, including the maximum time limit MaxTime, the maximum search depth MaxNoImp1, the accuracy ρ of the local search algorithm, and the criterion E for determining a legal solution legal , the perturbation intensity δ of the perturbation operator, the maximum search depth Q and the penalty factor α;

[0008] Step 3: Set the current search time Time = 0;

[0009] Step 4: Randomly generate an initial solution S0 consisting of a layout X0 of N small circles and an initial radius R0 of the circular area, that is, S0 = (X0, R0), and then use the quasi-Newton algorithm to locally optimize it and record its potential energy function value The specific steps are as follows:

[0010] Step 4.1: Randomly place N points in the circular area with an initial radius R0 as the centers of N small circles;

[0011] Step 4.2: Use the quasi-Newton algorithm to locally optimize the layout of N circles and record its potential energy function value

[0012] Step 5: If the search time Time does not reach the maximum time limit MaxTime, go to step 6, otherwise go to step 9;

[0013] Step 6: Execute the iterative local search method ILS, the steps are as follows:

[0014] Step 6.1: Record the initial solution S0 as the current solution S1, and set the current search depth Iterations1 = 0;

[0015] Step 6.2: For the N circle center coordinates (x i ,y i )i=1,2,...,N performs a uniform random perturbation with a perturbation intensity of δ, that is, x i =x i +δ·θ,y i =y i +δ·θ, i=1,2,...,N, where θ is a random number in the interval [-1,1];

[0016] Step 6.3: Use the quasi-Newton algorithm to locally optimize the perturbed layout. The resulting solution is recorded as solution S2, and its potential energy function value is recorded.

[0017] Step 6.4: If the potential energy function value of solution S2 is updated Less than the potential energy function value of the current solution S1 Then go to step 6.5, otherwise go to step 6.6;

[0018] Step 6.5: Optimal solution S * Updated to S2, that is, S * =S2, And set the current search depth Iterations1 = 0;

[0019] Step 6.6: The existing solution S1 is better, and it is used to update the optimal solution S * , that is, S * =S1, And increase the current search depth Iterations1 by 1;

[0020] Step 6.7: If the current search depth Iterations1 is less than the maximum search depth MaxNoImp1, go to step 6.2, otherwise go to step 6.8;

[0021] Step 6.8: Output the optimal solution S * ;

[0022] Step 7: If the optimal solution S * is a legal solution, that is Then go to step 8, otherwise go to step 4;

[0023] Step 8: Based on the penalty function method, the radius and layout of the circular area (i.e. S * =(X * ,R * )) Make adjustments and save the current solution as the optimal solution S * , the specific steps are as follows:

[0024] Step 8.1: Enter a legal optimal solution S * ;

[0025] Step 8.2: Set the current search depth Iterations2 = 0;

[0026] Step 8.3: If the current search depth Iterations2 is less than the maximum search depth Q, go to step 8.4, otherwise go to step 8.6;

[0027] Step 8.4: Adjust the optimal solution S * The radius of the full circular area is α, and the penalty factor α is enlarged, that is, α = 10·α;

[0028] Step 8.5: Increase the current search depth Iterations2 by 1, and then go to step 8.3;

[0029] Step 8.6: Output the optimal solution S * , go to step 4 (reinitialize to prevent falling into a local trap);

[0030] Step 9: Output the optimal solution S * .

[0031] Furthermore, the mathematical optimization model E of the distribution network sensor area coverage problem described in step 1 is R (X) and layout X, which are specifically defined as follows:

[0032] The distribution network sensor area coverage problem is described as follows: in a circular area with a radius of R, N sensors with a sensing radius of r are placed, that is, the sensors are represented by N small circles with a radius of r, and the small circles are required to cover the entire circular area. The goal is to maximize the radius R of the circular area.

[0033] The distribution network sensor area coverage problem is mathematically modeled, and a Cartesian coordinate system is established with the center of the area as the coordinate origin (0,0). It is assumed that the coordinates of the center of the small circle of the i-th (i=1,2,...,N) sensor in the area are (x i ,y i ), the radius of each small circle is r. For a given circular area radius R, in order to measure the degree of coverage of the area, the following evaluation function is constructed to evaluate the size of the uncovered area:

[0034]

[0035] where X∈R 2N is the coordinate vector of the center of the circle, that is, X=(x1,y1,x2,y2,...,x N ,y N ), S tot represents the area to be covered, S ij represents the overlapping area between the i-th circle and the j-th circle, S ijk represents the common overlapping area of ​​the three circles i, j, and k, S i0 Indicates the area of ​​the i-th circle beyond the area to be covered, S 0ij represents the part where the overlapping area of ​​the i-th circle and the j-th circle exceeds the area to be covered, and S ij , S ijk , S i0 and S 0ij Depends on the circular coordinate vector X; potential energy E R (X) indicates the uncovered area, E R (X) = 0 means that the area to be covered has been completely covered and the current solution is a legal solution. Therefore, in order to find a legal area coverage, the potential energy function E needs to be minimized. R (X), find an E R The solution X where (X)=0.

[0036] Furthermore, the method for adjusting the radius of the circular area described in step 8.4 is specifically described as follows:

[0037] Given a legal solution S = (X, R), the purpose of the circular area radius adjustment method is to adjust the area radius R and the layout X so that the radius R is locally maximized while maintaining the legality of the layout X. To this end, starting from the given initial legal solution S = (X, R), the local optimization function Eα (X, R); the unconstrained minimization technique of the sequence is used, and the constrained minimization problem is transformed into a series of unconstrained minimization problems, namely

[0038] Minmize E α (X, R) = α·E(X, R)-R 2

[0039] In the potential energy function, R is considered as a variable, α is the penalty factor, and each value of α defines an unconstrained minimization problem; by minimizing a series of E α (X,R), as α reaches a large number, the algorithm converges to a legal layout.

[0040] The beneficial effect achieved by the present invention is that the method of the present invention can make the coverage effect of the distribution network sensor area coverage problem more uniform and achieve a better coverage rate by establishing a mathematical optimization model and designing a method for adjusting the circular area radius. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of a mathematical optimization model in a specific embodiment of the present invention.

[0042] Figure 2 It is a diagram of the system architecture type in a specific embodiment of the present invention.

[0043] Figure 3 It is a technical route flow chart of a specific embodiment of the present invention.

[0044] Figure 4 It is a diagram showing the coverage effect of a random initial sensor layout in a specific embodiment of the present invention.

[0045] Figure 5 It is a flow chart of an iterative local search method in a specific embodiment of the present invention.

[0046] Figure 6 It is a flow chart of a method for adjusting the radius of a circular area in a specific embodiment of the present invention.

[0047] Figure 7 It is a diagram showing the coverage effect of the optimized sensor layout in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0048] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings.

[0049] Taking into account the convergence, uniformity and stability of the algorithm, in order to improve the coverage of the distribution network sensor network, the present invention provides an iterative local search method for solving the distribution network sensor area coverage problem. Figure 3As shown, a flowchart of an iterative local search method for solving the area coverage problem of distribution network sensors provided by an embodiment of the present invention is shown.

[0050] In this embodiment, in order to have a clearer understanding of the present invention, a brief description is given here. The present invention provides an iterative local search method for solving the distribution network sensor area coverage problem, which is specifically described as follows:

[0051] Step 1: Establish a mathematical optimization model E for the distribution network sensor area coverage problem R (X).

[0052] A Cartesian coordinate system is established with the center of the region as the origin (0,0), such as Figure 1 Assume that the coordinates of the center of the i-th (i=1,2,...,N) small circle in the region are (x i ,y i ), the radius of each small circle is r. For a given circular area radius R, in order to measure the degree of coverage of the area, the following evaluation function is constructed to evaluate the size of the uncovered area:

[0053]

[0054] where X∈R 2N is the coordinate vector of the center of the circle, that is, X=(x1,y1,x2,y2,...,x N ,y N ), S tot represents the area to be covered (i.e. a known quantity), S ij represents the overlapping area between the i-th circle and the j-th circle, S ijk represents the common overlapping area of ​​the three circles i, j, and k, S i0 Indicates the area of ​​the i-th circle beyond the area to be covered, S 0ij represents the part where the overlapping area of ​​the i-th circle and the j-th circle exceeds the area to be covered, and S ij , S ijk , S i0 and S 0ij depends on the circular coordinate vector X. It should be noted that the potential energy E R (X) represents the uncovered area, so E R (X) = 0 means that the area to be covered has been completely covered and the current solution is a legal solution. Therefore, in order to find a legal area coverage, it is only necessary to minimize the potential energy function E R (X), find an E R The solution X where (X)=0.

[0055] Step 2: Input the number of sensors N = 23, the initial radius of the distribution network to be covered R0 = 50 meters and the sensor's perception radius r = 10 meters, set the algorithm parameters, including the maximum search time limit MaxTime = 2 hours, the maximum search depth MaxNoImp1 = 100 and MaxNoImp2 = 100, and the local search method accuracy ρ = 10 -12 , the criterion for determining a legal solution is E legal =10 -20 , the perturbation strength of the perturbation operator δ = 0.8, the maximum search depth Q = 10 and the penalty factor α = 10 6 .

[0056] Step 3: Set the current search time Time=0.

[0057] Step 4: Based on the current radius R, an initial solution S0 is randomly generated, where the initial solution S0 is composed of the layout X0 of N small circles and the initial radius R0 of the circular area, that is, S0 = (x1, y1, x2, y2, ..., x 23 ,y 23 ,R0), and then use the quasi-Newton algorithm to locally optimize it and record its potential energy function value The specific steps are as follows:

[0058] Step 4.1: Randomly place 23 points X0 = (x1, y1, x2, y2, ..., x 23 ,y 23 ), obtain the initial position information of each sensor, the coordinates of any sensor are (x i ,y i )i,j=1,2,...,23. The specific initial layout effect diagram is as follows Figure 4 shown.

[0059] Step 4.2: Use the quasi-Newton algorithm to locally optimize the layout of 23 circles and record its potential energy function value

[0060] Step 5: If the search time Time does not reach the maximum time limit of 2 hours, go to step 6, otherwise go to step 9.

[0061] Step 6: Execute the iterative local search method (ILS). The specific iterative local search method flowchart is as follows: Figure 5 As shown, the steps are as follows:

[0062] Step 6.1: Record the initial solution S0 as the current solution S1, and set the current search depth Iterations1 = 0.

[0063] Step 6.2: For the 23 circle center coordinates (x i ,y i )i=1,2,...,23, a total of 46 coordinate variables are randomly perturbed with a perturbation intensity of 0.8, that is, x i =x i +0.8·θ,y i =y i +0.8·θ, i=1,2,...,23, where θ is a randomly generated real number in the interval [-1,1].

[0064] Step 6.3: Use the quasi-Newton algorithm to locally optimize the coordinates of the 23 sensors after the disturbance. The solution is recorded as solution S2, and its potential energy function value is recorded.

[0065] Step 6.4: If the potential energy function value of solution S2 is updated Less than the potential energy function value of the current solution S1 Then go to step 6.5, otherwise go to step 6.6.

[0066] Step 6.5: Optimal solution S * Updated to S2, that is, S * =S2, And set the current search depth Iterations1 = 0.

[0067] Step 6.6: The existing solution S1 is better, and it is used to update the optimal solution S * , that is, S * =S1, And increase the current search depth Iterations1 by 1.

[0068] Step 6.7: If the current search depth Iterations1 is less than the maximum search depth 100, go to step 6.2, otherwise go to step 6.8.

[0069] Step 6.8: Output the optimal solution S * .

[0070] Step 7: If the optimal solution S * is a legal solution, that is Then go to step 8, otherwise go to step 4.

[0071] Step 8: Based on the penalty function method, the radius and layout of the circular area (i.e. S * =(X * ,R * )) Make adjustments and save the current solution as the optimal solution S * The specific flow chart of the method for adjusting the radius of the circular area is as follows: Figure 6 , the steps are as follows:

[0072] Step 8.1: Enter a legal optimal solution S * .

[0073] Step 8.2: Set the current search depth Iterations2 = 0.

[0074] Step 8.3: If the current search depth Iterations2 is less than the maximum search depth 10, go to step 8.4, otherwise go to step 8.6.

[0075] Step 8.4: For the optimal solution S * Adjust the radius of the circular area and enlarge the penalty factor α, that is, let α = 10·α. The method of adjusting the radius of the circular area is specifically implemented as follows: use the quasi-Newton algorithm to calculate the potential energy function E α (X,R)=10 6 E(X,R)-R 2 Perform a local descent search, which is described as follows:

[0076] Given a legal solution S = (X, R), the purpose of the circular area radius adjustment method is to adjust the area radius R and the layout X so that the radius R is locally maximized while maintaining the legality of the layout X. To this end, starting from the given initial legal solution S = (X, R), the local optimization function E α (X, R). However, it is difficult to solve the constrained optimization problem directly using the local optimization method. The present invention adopts the sequence unconstrained minimization technology, and the constrained minimization problem is transformed into a series of unconstrained minimization problems, namely

[0077] Minmize E α (X, R) = α·E(X, R)-R 2

[0078] In the potential energy function, R is a variable, α is a penalty factor, and each value of α defines an unconstrained minimization problem. α (X,R), as α reaches a large number, the algorithm converges to a legal layout.

[0079] Step 8.5: Increase the current search depth Iterations2 by 1, and then go to step 8.3.

[0080] Step 8.6: Output the optimal solution S * , go to step 4 (reinitialize to prevent falling into a local trap).

[0081] Step 9: Output the optimal solution S * The final coverage effect is shown below: Figure 7 shown.

Claims

1. An iterative local search method for solving the area coverage problem of distribution network sensors, characterized by: The method comprises the following steps: Step 1: Establish a coordinate system based on the distribution network sensor area, construct an evaluation function for evaluating the size of the uncovered area, and establish a mathematical optimization model for the distribution network sensor area coverage problem; Step 2: Input the number of small circles N representing sensors, the initial radius R0 of the circular area to be covered and the radius r of the small circle; set parameters, including the maximum time limit MaxTime, the maximum search depth MaxNo Im p1, the accuracy ρ of the local search algorithm, and the criterion E for determining a legal solution legal , the perturbation intensity δ of the perturbation operator, the maximum search depth Q and the penalty factor α; Step 3: Set the current search time Time = 0; Step 4: Randomly generate an initial solution S0 consisting of a layout X0 of N small circles and an initial radius R0 of the circular area, that is, S0 = (X0, R0), and then use the quasi-Newton algorithm to locally optimize it and record its potential energy function value The specific steps are as follows: Step 4.1: Randomly place N points in the circular area with an initial radius R0 as the centers of N small circles; Step 4.2: Use the quasi-Newton algorithm to locally optimize the layout of N circles and record its potential energy function value Step 5: If the search time Time does not reach the maximum time limit MaxTime, go to step 6, otherwise go to step 9; Step 6: Execute the iterative local search method ILS, the steps are as follows: Step 6.1: Record the initial solution S0 as the current solution S1, and set the current search depth Iterations1 = 0; Step 6.2: For the N circle center coordinates (x i ,y i ), i = 1, 2, ..., N performs a uniform random perturbation with a perturbation intensity of δ, that is, x i =x i +δ·θ,y i =y i +δ·θ, i=1,2,...,N, where θ is a random number in the interval [-1,1]; Step 6.3: Use the quasi-Newton algorithm to locally optimize the perturbed layout. The resulting solution is recorded as solution S2, and its potential energy function value is recorded. Step 6.4: If the potential energy function value of solution S2 is updated Less than the potential energy function value of the current solution S1 Then go to step 6.5, otherwise go to step 6.6; Step 6.5: Optimal solution S * Updated to S2, that is, S * =S2, And set the current search depth Iterations1 = 0; Step 6.6: The existing solution S1 is better, and it is used to update the optimal solution S * , that is, S * =S1, And increase the current search depth Iterations1 by 1; Step 6.7: If the current search depth Iterations1 is less than the maximum search depth MaxNo Im p1, go to step 6.2, otherwise go to step 6.8; Step 6.8: Output the optimal solution S * ; Step 7: If the optimal solution S * is a legal solution, that is Then go to step 8, otherwise go to step 4; Step 8: Adjust the radius and layout of the circular area based on the penalty function method, that is, S * =(X * ,R * ) and save the current solution as the optimal solution S * , the specific steps are as follows: Step 8.1: Enter a legal optimal solution S * ; Step 8.2: Set the current search depth Iterations2 = 0; Step 8.3: If the current search depth Iterations2 is less than the maximum search depth Q, go to step 8.4, otherwise go to step 8.6; Step 8.4: Adjust the optimal solution S * The radius of the full circular area is increased, and the penalty factor α is enlarged, that is, α = 10·α; Step 8.5: Increase the current search depth Iterations2 by 1, and then go to step 8.3; Step 8.6: Output the optimal solution S * , go to step 4; Step 9: Output the optimal solution S * .

2. The iterative local search method for solving the distribution network sensor area coverage problem according to claim 1, characterized in that: Mathematical optimization model E of the distribution network sensor area coverage problem described in step 1 R (X) and layout X, which are specifically defined as follows: The distribution network sensor area coverage problem is described as follows: in a circular area with a radius of R, N sensors with a sensing radius of r are placed, that is, the sensors are represented by N small circles with a radius of r, and the small circles are required to cover the entire circular area. The goal is to maximize the radius R of the circular area. The distribution network sensor area coverage problem is mathematically modeled, and a Cartesian coordinate system is established with the center of the area as the origin (0,0). Assume that the coordinates of the center of the small circle of the i-th sensor in the area are (x i ,y i ), i = 1, 2, ..., N, the radius of each small circle is r, for a given circular area radius R, in order to measure the degree of coverage of the area, the following evaluation function is constructed to evaluate the size of the uncovered area: where X∈R 2N is the coordinate vector of the center of the circle, that is, X=(x1,y1,x2,y2,...,x N ,y N ), S tot represents the area to be covered, S ij represents the overlapping area between the i-th circle and the j-th circle, S ijk represents the common overlapping area of ​​the three circles i, j, and k, S i0 Indicates the area of ​​the i-th circle beyond the area to be covered, S 0ij represents the part where the overlapping area of ​​the i-th circle and the j-th circle exceeds the area to be covered, and S ij , S ijk , S i0 and S 0ij Depends on the circular coordinate vector X; potential energy E R (X) indicates the uncovered area, E R (X) = 0 means that the area to be covered has been completely covered and the current solution is a legal solution. Therefore, in order to find a legal area coverage, the potential energy function E needs to be minimized. R (X), find an E R The solution X where (X)=0.

3. The iterative local search method for solving the distribution network sensor area coverage problem according to claim 1, characterized in that: Step 8.4 adjusts the circular area radius method, and its specific description is as follows: Given a legal solution S = (X, R), the purpose of the circular area radius adjustment method is to adjust the area radius R and the layout X so that the radius R is locally maximized while maintaining the legality of the layout X. To this end, starting from the given initial legal solution S = (X, R), the local optimization function E α (X, R); the unconstrained minimization technique of the sequence is used, and the constrained minimization problem is transformed into a series of unconstrained minimization problems, namely Minmize E α (X,R)=α·E(X,R)-R 2 In the potential energy function, R is considered as a variable, α is the penalty factor, and each value of α defines an unconstrained minimization problem; by minimizing a series of E α (X,R), as α reaches a large number, the algorithm converges to a legal layout.