Computational efficiency evaluation method based on radar-based short-term icing prediction

Abstract

The present invention relates to the field of power transmission and distribution technology, and discloses a method for evaluating the computational efficiency of radar short-term icing predictions, so as to accurately screen the optimal calculation scheme. The method comprises the following steps: 1. Building a radar short-term icing prediction computational efficiency evaluation environment, 2. Decoupling the radar short-term icing prediction computational process, 3. Grouping the computational process parameter schemes, 4. Preliminary screening and evaluation of computational schemes, and 5. Optimizing the evaluation of computational schemes. 1. The present invention divides the radar short-term icing prediction into different steps through decoupling calculations, and evaluates the efficiency separately. 2. The present invention has good versatility and can be used for evaluating the computational efficiency of radars in different regions. 3. By adopting the technology of the present invention, it is possible to quickly evaluate the efficiency of existing radar short-term icing warnings, screen the optimal computational scheme, and improve the timeliness of radar short-term icing warnings.

Description

Technical Field

[0001] The present invention relates to the technical field of power transmission and distribution, and in particular to a method for evaluating computational efficiency based on radar-based short-term and impending icing prediction. Background Art

[0002] In recent years, winter transmission lines have frequently suffered from ice cover, threatening the safety of towers, lines, and power supply. To ensure timely anti-icing preparations and understand ice trends, it is imperative to develop efficient and accurate short-term warnings for transmission line icing.

[0003] Currently, radar remote sensing is being conducted on transmission line tower sections prone to ice accumulation. Using the fusion of detection factors, short-term and impending warnings are being developed to estimate the ice thickness of the transmission lines for the next six hours. Due to the large amount of data required for these short-term and impending ice calculations, data processing time can impact the effectiveness of these warnings. Therefore, evaluating the efficiency of radar short-term and impending ice predictions is a prerequisite and key to providing timely line icing forecasts.

[0004] Therefore, it is necessary to study a method for evaluating the efficiency of radar-based short-term icing calculations. By evaluating the efficiency of radar-based short-term icing calculations, the optimal calculation scheme can be screened to achieve rapid output of icing warning results and improve the anti-icing warning capability of the power grid. Summary of the Invention

[0005] The purpose of the present invention is to disclose a method for evaluating the computational efficiency of radar-based short-term icing prediction to accurately screen the optimal computational solution.

[0006] To achieve the above objectives, the present invention proposes a solution: for radar-based extrapolated power grid icing short-term calculation methods, the calculation method is modularly decoupled, the operating parameters of each module are set, the calculation efficiency of each module is preliminarily evaluated, some efficient calculation methods are selected, the calculation scale is further expanded, and the evaluation model is further optimized. The specific steps are as follows:

[0007] 1. Build a computational efficiency evaluation environment for radar icing short-term prediction

[0008] Deploy and install the Linux kernel platform, compile and install parallel computing software, and add the parallel software operating environment to the system environment variables; install and compile the mathematical library, and deploy the radar extrapolated power grid icing short-term calculation program.

[0009] 2. Decoupling of the radar icing short-term prediction calculation process

[0010] The radar icing short-term prediction calculation process is divided into three steps: (1) obtaining radar measurement data from the radar monitoring point, which is recorded as D; (2) performing short-term extrapolation calculation based on the radar measurement data, which is recorded as R; and (3) performing ice thickness inversion based on the extrapolated calculation results, which is recorded as F.

[0011] 3. Grouping of calculation process parameter schemes

[0012] According to the three decoupled processes in step (2), serial computing and parallel computing are used, which are recorded as S and P respectively. Then, eight combinations are obtained, which are recorded as SSS, SSP, SPS, SPP, PSS, PSP, PPS, and PPP respectively.

[0013] 4. Preliminary screening and evaluation of calculation schemes

[0014] The detection range is set to be a plane area with O as the center and M as the radius. For one detection moment, the eight calculation methods set in step 3 are used to evaluate the calculation time respectively, and recorded as t i , i = 1…8. The first three solutions with the shortest time consumption are selected as the preliminary screening results for the next round of optimization screening.

[0015] 5. Optimal evaluation and calculation scheme

[0016] Based on the three computational solutions initially selected in step 4, expand the computational scale by n1, n2, and n3 times, and then evaluate the computational time of each solution. For each scale, assign a weight, descending from shortest to longest, to each solution. Score each of the three computational scales, and calculate the cumulative score F1, F2, and F3 for each solution. The three solutions with the highest scores are then analyzed and determined to be the most efficient solution.

[0017] Preferably, weights are assigned in descending order according to the order of time: the shortest time is scored as 1, the next is scored as 0.8, and the longest is scored as 0.5.

[0018] The present invention has the following beneficial effects:

[0019] 1. The present invention divides radar icing short-term prediction into different steps through decoupling calculation and evaluates the efficiency of each step.

[0020] 2. The present invention has good versatility and can be used to evaluate the calculation efficiency of radars in different regions.

[0021] 3. The technology of the present invention can quickly evaluate the efficiency of existing radar short-term icing warning, select the optimal calculation scheme, and improve the timeliness of radar short-term icing warning.

[0022] The present invention will be described in further detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are intended to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0024] Figure 1 The present invention discloses a method for evaluating the computational efficiency of short-term radar icing prediction based on the flowchart of an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The embodiments of the present invention are described in detail below with reference to the accompanying drawings. However, the present invention can be implemented in many different ways as defined and covered by the claims.

[0026] Example 1

[0027] This embodiment takes a 220kV line radar icing warning module in Hunan as an example to disclose a method for evaluating the efficiency of short-term and impending prediction of radar icing. Figure 1 As shown, the following steps are included:

[0028] 1. Build a computational efficiency evaluation environment for radar icing short-term prediction

[0029] Deploy and install the Linux kernel platform, using Centos7.9; configure MPICH installation parameters, compile and install the parallel computing software MPICH, and add the MPICH runtime environment to the system environment variables; install and compile the mathematical library GotoBLAS, generate two mathematical calculation process dependency files, libgoto2.a and libgoto.so, and deploy the radar extrapolated power grid icing short-term calculation program.

[0030] 2. Decoupling of the radar icing short-term prediction calculation process

[0031] The radar icing short-term prediction calculation process is divided into three steps: (1) obtaining radar measurement data from the radar monitoring point, which is recorded as D; (2) performing short-term extrapolation calculation based on the radar measurement data, which is recorded as R; and (3) performing ice thickness inversion based on the extrapolated calculation results, which is recorded as F.

[0032] 3. Grouping of calculation process parameter schemes

[0033] According to the three decoupled processes in step (2), serial computing and parallel computing are used, which are recorded as S and P respectively. Then, eight combinations are obtained, which are recorded as SSS, SSP, SPS, SPP, PSS, PSP, PPS, and PPP respectively.

[0034] 4. Preliminary screening and evaluation of calculation schemes

[0035] The detection range is set to a plane area with a radius of 10 kilometers and a 50# tower of a 220kV line as the center. For one detection moment, the eight calculation methods set in step 3 are used to evaluate the calculation time.

[0036] Table 1: Time consumption of different solutions

[0037] SSS SSP SPS SPP PSS PSP PPS PPP Time (seconds) 18.2 15.2 10.8 11.3 14.4 15 13.1 10.9

[0038] Then the preliminary screening results are SPS, SPP and PPP, and the next round of optimization screening is carried out.

[0039] 5. Optimal evaluation and calculation scheme

[0040] Based on the three computational solutions initially screened in step 4, scale the computational scale by 10, 50, and 100 times, and then evaluate the computational time of the three solutions. For each scale, the shortest time is scored as 1, the second fastest as 0.8, and the longest as 0.5.

[0041] Table 2: Calculation time of the amplified scheme

[0042] SPS SPP PPP 10 times 90.1 105.4 95.5 50 times 483.2 535.2 502.8 100 times 933.3 1104.4 996.6 Score 3 1.5 2.4

[0043] Comparing the scores of the three schemes, the SPS scheme scored the highest, that is, data collection adopted a serial method, extrapolation calculation adopted a parallel method, and inversion calculation adopted a serial method, which is the most efficient.

[0044] In summary: 1. This embodiment decouples computations to divide radar short-term icing prediction into distinct steps, evaluating their efficiency separately. 2. This embodiment is highly versatile and can be used to evaluate the computational efficiency of radars in different regions. 3. Using the techniques in this embodiment, the efficiency of existing radar short-term icing warnings can be rapidly evaluated, the optimal computational solution can be selected, and the timeliness of radar short-term icing warnings can be improved.

[0045] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for evaluating computational efficiency of radar-based short-term icing prediction, characterized in that: include: Step 1: Build a computational efficiency evaluation environment for radar icing short-term prediction Deploy and install the Linux kernel platform, compile and install parallel computing software, and add the parallel software runtime environment to the system environment variables; install and compile the mathematical library, and deploy the radar extrapolation power grid icing short-term calculation program; Step 2: Decoupling the radar icing short-term prediction calculation process The calculation process of radar icing short-term prediction is divided into three steps: (1) obtaining radar measurement data of radar monitoring points, and the process is recorded as D; (2) Perform short-term extrapolation calculation based on radar measurement data, and the process is recorded as R; (3) Invert the ice thickness based on the extrapolated calculation results, and the process is recorded as F; Step 3: Grouping of calculation process parameter schemes According to the three decoupled processes in step 2, two schemes, serial computing and parallel computing, are adopted, which are recorded as S and P respectively; Eight combinations are obtained, which are recorded as SSS, SSP, SPS, SPP, PSS, PSP, PPS, and PPP; Step 4: Preliminary screening and evaluation of calculation schemes The detection range is set to be a plane area with O as the center and M as the radius. For one detection moment, the eight calculation methods set in step 3 are used to evaluate the calculation time respectively, and recorded as t i , i = 1…8; select the first three solutions with the shortest time consumption as the preliminary screening results and proceed to the next round of optimization screening; Step 5: Optimize the evaluation calculation scheme Based on the three calculation schemes initially screened in step 4, the calculation scale is expanded by n1, n2, and n3 times, and then the calculation time of the three schemes is evaluated. For each type of scale, weights are assigned from large to small according to the order of time from short to long; then the scores are calculated for the three types of calculation scales respectively, and the cumulative score values ​​F1, F2, and F3 of each scheme are calculated. Then, the three highest-scoring schemes are analyzed and regarded as the most efficient scheme.

2. The method according to claim 1, characterized in that The weights are assigned in descending order according to the time from short to long: the shortest time is scored as 1, the second is scored as 0.8, and the longest is scored as 0.5.

Images