A power transmission line tower foundation red soil analysis method based on big data

Abstract

The present application relates to the technical field of soil data analysis, and particularly relates to a method for analyzing red soil of a tower base of a power transmission line based on big data, comprising: collecting red soil data of a tower base area of a power transmission line, and sequentially performing cleaning, denoising, filtering and feature extraction operations on the red soil data to output red soil features, training an initial model according to the red soil features to output a deep learning model; using the deep learning model to analyze the red soil data to output an analysis result, determining the fertilizer application of the red soil of the tower base area of the power transmission line according to the analysis result; determining whether the analysis accuracy of the red soil data meets the requirements based on the loss rate of the red soil data; if the analysis accuracy does not meet the requirements, adjusting the collection frequency of the red soil data; and if the analysis reliability does not meet the requirements, adjusting the learning rate of the deep learning model. The present application improves the analysis accuracy of the red soil data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of soil data analysis, and particularly relates to a method for analyzing red soil at a tower base of a power transmission line based on big data. BACKGROUND

[0002] With the continuous advancement of power infrastructure construction, the stability and safety of the power transmission line, as a key carrier for power transmission, are crucial to ensuring power supply. The tower base of the power transmission line, as the basic structure supporting the line, is long-term in a complex geological environment, and the tower base in the red soil area faces special challenges. Red soil is rich in metal compounds and has strong acidity, high clay content, and complex fertility conditions, which can easily corrode the tower base material and affect the stability of the tower base.

[0003] The traditional method for analyzing red soil at the tower base of the power transmission line relies on manual sampling and experience-based judgment, which is not only inefficient but also difficult to ensure accuracy, and cannot meet the high standard requirements of modern power systems for tower base safety. With the rapid development of big data, sensor technology and deep learning algorithms, although some research attempts to apply them to the field of soil analysis, there are still many problems in the analysis of red soil at the tower base of the power transmission line.

[0004] Chinese Patent Publication No. CN117408430A discloses a soil improvement evaluation system for agricultural planting based on big data, which includes a soil initial collection module, a soil initial detection and analysis module, an improvement method determination module, a regional division and improvement module, a soil secondary collection module, a comparison and analysis module, an evaluation result generation module, a sampling effectiveness evaluation module, a big data platform and a data import module. The soil initial collection module is used to collect samples before the land is improved. The soil initial detection and analysis module is used to detect and analyze the initially collected soil samples to determine the soil index. The improvement method determination module is used to select all schemes for successfully improving the soil according to the determined soil index from the big data platform, and each scheme is marked as Ai, i = 1···n, n is a positive integer. The regional division and improvement module is used to divide the land to be improved into n regions, and each region implements one improvement scheme. The soil secondary collection module is used to determine the number of sampling points in each region, and then collect the samples of the improved land in each region according to the number of sampling points and mix them. The soil secondary collection module is used to detect and analyze the secondary collected soil samples to determine the soil index. It can be seen that the soil improvement evaluation system for agricultural planting based on big data has the problem that the accuracy of the evaluation result is not guaranteed because only the number of sampling points is determined, the standardization of the sampling process and the effectiveness of the samples are not strictly evaluated, and the collected samples cannot truly reflect the soil improvement situation, resulting in inaccurate evaluation results. SUMMARY

[0005] To address this, the present invention provides a big data-based analysis method for red soil of transmission line tower foundations, which overcomes the problem in existing technologies where only the number of sampling points is determined without rigorously evaluating the standardization of the sampling process and the validity of the samples, thus failing to guarantee that the collected samples can truly reflect the soil improvement situation, leading to inaccurate evaluation results.

[0006] To achieve the above objectives, this invention provides a method for analyzing red soil of transmission line tower foundations based on big data, comprising:

[0007] Red soil data of the transmission line tower base area is collected, and the red soil data is cleaned, denoised, filtered and feature extracted in sequence to output red soil features. The initial model is trained based on the red soil features to output a deep learning model.

[0008] The deep learning model is used to analyze the red soil data to output analysis results, and the fertility application of the red soil in the transmission line tower base area is determined based on the analysis results.

[0009] Obtain the data loss rate of red soil data;

[0010] The accuracy of the red soil data analysis is determined based on the loss rate of the red soil data.

[0011] If the accuracy of the analysis does not meet the requirements, the collection frequency of the red soil data will be adjusted, or the reliability of the deep learning model will be determined based on the accuracy of the red soil data analysis.

[0012] If the reliability of the analysis does not meet the requirements, the learning rate of the deep learning model is adjusted, or the cutoff frequency of the filter is adjusted based on the response delay of the soil pH sensor.

[0013] The red soil data includes the red soil's pH, nitrogen content, phosphorus content, and water content. The pH of the red soil is detected by the soil pH sensor.

[0014] Furthermore, determining the accuracy of the analysis of the red soil data includes:

[0015] Compare the data loss rate of red soil with the preset first data loss rate;

[0016] If the loss rate of the red soil data is greater than the preset first loss rate, then the accuracy of the red soil data analysis is determined to be unsatisfactory.

[0017] Further, determining the analytical reliability of the deep learning model includes:

[0018] comparing the loss rate of the red soil data with the preset first loss rate and the preset second loss rate respectively;

[0019] If the loss rate of the red soil data is greater than the preset first loss rate and less than or equal to the preset second loss rate, it is preliminarily determined that the analysis reliability of the deep learning model does not meet the requirements, and whether the analysis reliability of the deep learning model meets the requirements is determined according to the accuracy of the red soil data analysis.

[0020] Further, the collection frequency of the red soil data is adjusted, including:

[0021] comparing the loss rate of the red soil data with the preset second loss rate;

[0022] If the loss rate of the red soil data is greater than the preset second loss rate, the collection frequency of the red soil data is increased.

[0023] Further, the increase range of the collection frequency of the red soil data is determined by the difference between the loss rate of the red soil data and the preset second loss rate.

[0024] Further, the learning rate of the deep learning model is adjusted, including:

[0025] comparing the accuracy of the red soil data analysis with the preset first accuracy and the preset second accuracy respectively;

[0026] If the accuracy of the red soil data analysis is less than the preset second accuracy, it is determined that the analysis reliability of the deep learning model does not meet the requirements;

[0027] If the accuracy of the red soil data analysis is greater than the preset first accuracy and less than the preset second accuracy, the learning rate of the deep learning model is reduced;

[0028] If the accuracy of the red soil data analysis is less than or equal to the preset first accuracy, it is preliminarily determined that the environmental stability of the red soil data collection does not meet the requirements, and whether the environmental stability of the red soil data collection meets the requirements is determined according to the response delay time length of the soil pH sensor.

[0029] Further, the reduction range of the learning rate of the deep learning model is determined by the difference between the accuracy of the red soil data analysis and the preset first accuracy.

[0030] Further, the accuracy of the red soil data analysis is the ratio of the number of accurate red soil data analysis to the total number of red soil data analysis.

[0031] Further, the cutoff frequency of the filter is adjusted, including:

[0032] Compare the response delay time length of the soil pH sensor with the preset delay time length;

[0033] If the response delay time length of the soil pH sensor is greater than the preset delay time length, it is determined that the environmental stability of the red soil data collection does not meet the requirements, and the cutoff frequency of the filter is reduced.

[0034] Further, the reduction amplitude of the cutoff frequency of the filter is determined by the difference between the response delay time length of the soil pH sensor and the preset delay time length.

[0035] Compared with the prior art, the method has the beneficial effects that the collection frequency of the red soil data is adjusted according to the loss rate of the red soil data, since part of the data collected by the sensor is lost due to signal interference, the result of analyzing the red soil data is inaccurate, by increasing the collection frequency of the red soil data, more data points can be obtained in the same time, which helps to fill the data vacancy caused by interference, makes the data sequence more complete, and makes the analysis result more reflect the true situation of the red soil, reduces the analysis deviation caused by data loss, the learning rate of the deep learning model is adjusted according to the accuracy of the red soil data analysis, since too many parameters are used to train the model, the model is trained in detail for part of the noise and non-representative features, which causes model overfitting, thereby causing poor generalization ability for new data in application, by reducing the learning rate of the deep learning model, the model can adjust the parameters more carefully, reduce the overfitting of noise and non-representative features, thereby improving the generalization ability of the model on new data, the cutoff frequency of the filter is adjusted according to the response delay time length of the soil pH sensor, since there are many metal compounds in the red soil, which may cause corrosion and aging of the sensor after long-term use, thereby causing inaccurate measurement data of the sensor, by reducing the cutoff frequency of the filter, the filter can better suppress high-frequency noise and retain relatively low-frequency effective measurement signals, thereby improving the accuracy of the measurement data, and improving the analysis accuracy of the red soil data.

[0036] Further, the method adjusts the collection frequency of the red soil data by setting a preset first loss rate and a preset second loss rate, since part of the data collected by the sensor is lost due to signal interference, the result of analyzing the red soil data is inaccurate, by increasing the collection frequency of the red soil data, more data points can be obtained in the same time, which helps to fill the data vacancy caused by interference, makes the data sequence more complete, and makes the analysis result more reflect the true situation of the red soil, reduces the analysis deviation caused by data loss, and further improves the analysis accuracy of the red soil data.

[0037] Further, the method disclosed in the present application adjusts the learning rate of the deep learning model by setting the preset first accuracy and the preset second accuracy. Due to too many parameters for training the model, the model is trained in detail for part of the noise and the features without representativeness, which leads to overfitting of the model, thereby leading to poor generalization ability of the model for new data. By reducing the learning rate of the deep learning model, the model can be adjusted more carefully, and overfitting for noise and features without representativeness is reduced, thereby improving the generalization ability of the model for new data, and further improving the analysis accuracy of the red soil data.

[0038] Further, the method disclosed in the present application adjusts the cutoff frequency of the filter by setting the preset delay duration. Due to more metal compounds in the red soil, the sensor may be corroded and aged after long-term use, thereby leading to inaccurate measurement data of the sensor. By reducing the cutoff frequency of the filter, the filter can better suppress high-frequency noise and retain relatively low-frequency effective measurement signals, thereby improving the accuracy of the measurement data, and further improving the analysis accuracy of the red soil data. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 The overall flowchart of the analysis method of the red soil of the power transmission line tower foundation based on big data according to the embodiment of the present application is shown in the figure.

[0040] Figure 2 The logic flowchart of the adjustment process of the red soil data collection frequency of the analysis method of the red soil of the power transmission line tower foundation based on big data according to the embodiment of the present application is shown in the figure.

[0041] Figure 3 The logic flowchart of the adjustment process of the learning rate of the deep learning model of the analysis method of the red soil of the power transmission line tower foundation based on big data according to the embodiment of the present application is shown in the figure.

[0042] Figure 4 The logic flowchart of the adjustment process of the cutoff frequency of the filter of the analysis method of the red soil of the power transmission line tower foundation based on big data according to the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0043] In order to make the purpose and advantages of the present application more clear and explicit, the present application is further described below in combination with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present application, and do not limit the present application.

[0044] The preferred embodiments of the present application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application, and are not used to limit the protection scope of the present application.

[0045] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0046] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0047] Please see Figure 1 , Figure 2 , Figure 3 as well as Figure 4 The diagrams shown are, respectively, an overall flowchart of the big data-based analysis method for red soil in transmission line tower foundations according to an embodiment of the present invention, a logical flowchart of the process for adjusting the acquisition frequency of red soil data, a logical flowchart of the process for adjusting the learning rate of the deep learning model, and a logical flowchart of the process for adjusting the cutoff frequency of the filter. The present invention provides a big data-based analysis method for red soil in transmission line tower foundations, comprising:

[0048] Step S1: Collect red soil data of the transmission line tower base area, and perform cleaning, noise reduction, filtering and feature extraction operations on the red soil data in sequence to output red soil features. Train the initial model based on the red soil features to output a deep learning model.

[0049] Step S2: Use the deep learning model to analyze the red soil data and output the analysis results, and determine the fertility application of the red soil in the transmission line tower base area based on the analysis results;

[0050] Step S3: Obtain the loss rate of red soil data;

[0051] Step S4: Determine whether the accuracy of the red soil data analysis meets the requirements based on the loss rate of the red soil data;

[0052] Step S5: If the accuracy of the analysis does not meet the requirements, the collection frequency of the red soil data is adjusted, or the reliability of the deep learning model is determined based on the accuracy of the red soil data analysis.

[0053] If the analysis reliability does not meet the requirements, the learning rate of the deep learning model is adjusted, or the cutoff frequency of the filter is adjusted based on the response delay length of the soil pH sensor.

[0054] The red soil data includes the pH of the red soil, the nitrogen content of the red soil, the phosphorus content of the red soil, and the water content of the red soil, and the pH of the red soil is detected by the soil pH sensor.

[0055] Specifically, the red soil characteristics include the bulk density of the red soil, the water content of the red soil, and the pH of the red soil.

[0056] Specifically, the deep learning model includes a convolutional neural network, a recurrent neural network, and a multilayer perceptron.

[0057] Specifically, the analysis results include the fertility of the red soil, the sedimentation risk of the red soil, and the vegetation growth suitability of the red soil.

[0058] Specifically, the filter includes a low-pass filter, a high-pass filter, and a band-stop filter, and the preferred embodiment is a low-pass filter.

[0059] In implementation, the method adjusts the acquisition frequency of the red soil data according to the loss rate of the red soil data. Since the sensor collecting the red soil data is interfered by signals, some data is lost, resulting in inaccurate analysis results of the red soil data. By increasing the acquisition frequency of the red soil data, more data points can be obtained in the same time, which helps to fill the data gap caused by interference, makes the data sequence more complete, and makes the analysis results more accurately reflect the true situation of the red soil. The learning rate of the deep learning model is adjusted according to the accuracy of the red soil data analysis. Since too many parameters are used to train the model, the model is trained in detail for some noise and non-representative features, resulting in overfitting of the model, which reduces the generalization ability of the model for new data. By reducing the learning rate of the deep learning model, the model can adjust the parameters more carefully, reduce the overfitting of noise and non-representative features, and improve the generalization ability of the model for new data. The cutoff frequency of the filter is adjusted according to the response delay length of the soil pH sensor. Since there are many metal compounds in the red soil, the sensor may be corroded and aged after long-term use, resulting in inaccurate measurement data of the sensor. By reducing the cutoff frequency of the filter, the filter can better suppress high-frequency noise and retain relatively low-frequency effective measurement signals, thereby improving the accuracy of the measurement data and improving the analysis accuracy of the red soil data.

[0060] Specifically, the analysis accuracy of the red soil data is determined, including:

[0061] Compare the loss rate of the red soil data with a preset first loss rate;

[0062] If the loss rate of the red soil data is greater than the preset first loss rate, it is determined that the analysis accuracy of the red soil data does not meet the requirements.

[0063] Specifically, the analysis reliability of the deep learning model is determined, including:

[0064] The loss rate of the red soil data is compared with the preset first loss rate and a preset second loss rate, respectively.

[0065] If the loss rate of the red soil data is greater than the preset first loss rate and less than or equal to the preset second loss rate, it is preliminarily determined that the analysis reliability of the deep learning model does not meet the requirements, and whether the analysis reliability of the deep learning model meets the requirements is determined according to the accuracy of the red soil data analysis.

[0066] It can be understood that the three intervals divided by the preset first loss rate and the preset second loss rate correspond to three situations, respectively.

[0067] The first interval is that the loss rate of the red soil data is less than or equal to the preset first loss rate, and the corresponding situation is that it is determined that the analysis accuracy of the red soil data meets the requirements.

[0068] The second interval is that the loss rate of the red soil data is greater than the preset first loss rate and less than or equal to the preset second loss rate, and the corresponding situation is that due to too many parameters for training the model, the model has been trained in detail for part of the noise and non-representative features, resulting in overfitting of the model, thereby resulting in poor generalization ability for new data in application.

[0069] The third interval is that the loss rate of the red soil data is greater than the preset second loss rate, and the corresponding situation is that due to signal interference of the sensor for collecting the red soil data, part of the data is lost, thereby resulting in inaccurate results when analyzing the red soil data.

[0070] In implementation, the preset first loss rate is generally selected in the range of [0.2%, 0.4%], and the preset second loss rate is generally selected in the range of [0.5%, 0.7%].

[0071] Preferably, the preferred embodiment of the preset first loss rate is 0.3%, and the preferred embodiment of the preset second loss rate is 0.6%.

[0072] Specifically, the loss rate of the red soil data is the ratio of the amount of lost data of the red soil data to the total amount of the red soil data.

[0073] In the implementation, the method provided by the application determines the analysis accuracy of the red soil data by setting the preset first loss rate and the preset second loss rate, reduces the influence of inaccurate determination of the analysis accuracy of the red soil data on the analysis stability of the red soil data, and further improves the analysis accuracy of the red soil data.

[0074] Specifically, the collection frequency of the red soil data is adjusted, including:

[0075] The loss rate of the red soil data is compared with the preset second loss rate;

[0076] If the loss rate of the red soil data is greater than the preset second loss rate, the collection frequency of the red soil data is increased.

[0077] Specifically, the increase range of the collection frequency of the red soil data is determined by the difference between the loss rate of the red soil data and the preset second loss rate.

[0078] Specifically, when the difference between the loss rate of the red soil data and the preset second loss rate is within 0.2%, the collection frequency of the red soil data is increased to 1.2 times of the original; when the difference between the loss rate of the red soil data and the preset second loss rate exceeds 0.2%, the collection frequency of the red soil data is increased by 1 time / minute for each 0.1% exceeding the 1.2 times of the original, for example, the difference between the loss rate of the red soil data and the preset second loss rate is 0.4%, the current collection frequency of the red soil data is 5 times / minute, and the increased collection frequency of the red soil data is 5*1.2+1*2=8 times / minute.

[0079] In the implementation, the method provided by the application adjusts the collection frequency of the red soil data by setting the preset first loss rate and the preset second loss rate. Due to the signal interference on the sensor collecting the red soil data, part of the data is lost, which leads to inaccurate results when the red soil data is analyzed. By increasing the collection frequency of the red soil data, more data points can be obtained in the same time, which helps to fill the data vacancy caused by the interference, makes the data sequence more complete, makes the analysis result more reflect the real situation of the red soil, reduces the analysis deviation caused by the data loss, and further improves the analysis accuracy of the red soil data.

[0080] Specifically, the learning rate of the deep learning model is adjusted, including:

[0081] The accuracy of the red soil data analysis is compared with the preset first accuracy and the preset second accuracy, respectively;

[0082] If the accuracy of the red soil data analysis is less than the preset second accuracy, it is determined that the analysis reliability of the deep learning model does not meet the requirements.

[0083] If the accuracy of the red soil data analysis is greater than the preset first accuracy and less than the preset second accuracy, the learning rate of the deep learning model is reduced.

[0084] If the accuracy of the red soil data analysis is less than or equal to the preset first accuracy, it is preliminarily determined that the environmental stability of the red soil data collection does not meet the requirements, and whether the environmental stability of the red soil data collection meets the requirements is determined according to the response delay length of the soil acidity-alkalinity sensor.

[0085] It can be understood that the three intervals divided by the preset first accuracy and the preset second accuracy correspond to three situations respectively.

[0086] The first interval is that the accuracy of the red soil data analysis is less than or equal to the preset first accuracy, and the corresponding situation is that because there are many metal compounds in the red soil, the sensor may be corroded and aged after long-term use, thereby causing the measurement data of the sensor to be inaccurate.

[0087] The second interval is that the accuracy of the red soil data analysis is greater than the preset first accuracy and less than the preset second accuracy, and the corresponding situation is that because too many parameters are used to train the model, the model is trained in detail for part of the noise and features that are not representative, resulting in overfitting of the model, thereby causing poor generalization ability of the model to new data in application.

[0088] The third interval is that the accuracy of the red soil data analysis is greater than or equal to the preset second accuracy, and the corresponding situation is that it is determined that the analysis reliability of the deep learning model meets the requirements.

[0089] In implementation, the preset first accuracy is generally selected in the range of [93%, 95%], and the preset second accuracy is generally selected in the range of [96%, 98%].

[0090] Preferably, the preferred embodiment of the preset first accuracy is 94%, and the preferred embodiment of the preset second accuracy is 97%.

[0091] In implementation, the method provided by the application determines the analysis reliability of the deep learning model by setting the preset first accuracy and the preset second accuracy, reduces the influence of inaccurate determination of the analysis reliability of the deep learning model on the analysis accuracy of the red soil data, and further improves the analysis accuracy of the red soil data.

[0092] Specifically, the reduction range of the learning rate of the deep learning model is determined by the difference between the accuracy of the red soil data analysis and the preset first accuracy.

[0093] Specifically, when the difference between the accuracy of red soil data analysis and the preset first accuracy is within 3%, the learning rate of the deep learning model is reduced to 0.9 times of the original; when the difference between the accuracy of red soil data analysis and the preset first accuracy exceeds 3%, the learning rate of the deep learning model is reduced by 0.002 for each 1% exceeded, based on the reduction to 0.9 times of the original, for example, when the difference between the accuracy of red soil data analysis and the preset first accuracy is 5%, and the current learning rate of the deep learning model is 0.01, the learning rate of the deep learning model after reduction is 0.01*0.9-0.002*2=0.005.

[0094] In implementation, the method of the present application adjusts the learning rate of the deep learning model by setting the preset first accuracy and the preset second accuracy. Due to too many parameters for training the model, the model is trained in detail for part of the noise and non-representative features, leading to overfitting of the model, thereby leading to poor generalization ability of the model for new data in application. By reducing the learning rate of the deep learning model, the model can adjust the parameters more carefully, reduce the overfitting for noise and non-representative features, thereby improving the generalization ability of the model for new data, and further improving the analysis accuracy of red soil data.

[0095] Specifically, the accuracy of red soil data analysis is the ratio of the number of accurate red soil data analysis to the total number of red soil data analysis.

[0096] Specifically, adjusting the cutoff frequency of the filter comprises:

[0097] Comparing the response delay duration of the soil pH sensor with the preset delay duration;

[0098] If the response delay duration of the soil pH sensor is greater than the preset delay duration, it is determined that the environmental stability of red soil data collection does not meet the requirements, and the cutoff frequency of the filter is reduced.

[0099] It can be understood that the two intervals divided by the preset delay duration correspond to two situations respectively:

[0100] The first interval is that the response delay duration of the soil pH sensor is less than or equal to the preset delay duration, and the corresponding situation is that it is determined that the environmental stability of red soil data collection meets the requirements;

[0101] The second interval is that the response delay duration of the soil pH sensor is greater than the preset delay duration, and the corresponding situation is that the sensor may be corroded and aged after long-term use due to the large amount of metal compounds in red soil, thereby leading to inaccurate measurement data of the sensor.

[0102] In implementation, the preset delay duration is generally selected in the range of [2min, 3min].

[0103] Preferably, the preferred embodiment of the preset delay duration is 2.5min.

[0104] In implementation, the method of the present application determines the environmental stability of red soil data collection by setting a preset delay duration, which reduces the impact of inaccurate determination of the environmental stability of red soil data collection on the analysis accuracy of red soil data, and further improves the analysis accuracy of red soil data.

[0105] Specifically, the reduction amplitude of the cutoff frequency of the filter is determined by the difference between the response delay duration of the soil pH sensor and the preset delay duration.

[0106] Specifically, when the difference between the response delay duration of the soil pH sensor and the preset delay duration is within 2min, the cutoff frequency of the filter is reduced to 0.92 times the original; when the difference between the response delay duration of the soil pH sensor and the preset delay duration exceeds 2min, the cutoff frequency of the filter is reduced by 3Hz for every 1min in addition to the reduction to 0.92 times the original, for example, when the difference between the response delay duration of the soil pH sensor and the preset delay duration is 4min, the current cutoff frequency of the filter is 50Hz, and the reduced cutoff frequency of the filter is 40Hz.

[0107] In implementation, the method of the present application adjusts the cutoff frequency of the filter by setting a preset delay duration. Since there are many metal compounds in red soil, the sensor may corrode and age after long-term use, resulting in inaccurate measurement data. By reducing the cutoff frequency of the filter, the filter can better suppress high-frequency noise and retain relatively low-frequency effective measurement signals, thereby improving the accuracy of measurement data and further improving the analysis accuracy of red soil data.

[0108] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the accompanying drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to related technical features without deviating from the principles of the present application, and the technical solutions after these changes or replacements will fall within the protection scope of the present application.

Claims

1. A method for analyzing red soil at a power transmission line tower base based on big data, characterized in that, The method comprises the following steps: Collecting red soil data of a tower foundation area of a power transmission line, and sequentially performing cleaning, denoising, filtering and feature extraction operations on the red soil data to output red soil features, and training an initial model according to the red soil features to output a deep learning model; Using the deep learning model to analyze the red soil data to output an analysis result, and determining the fertilizer application of the red soil of the tower foundation area of the power transmission line according to the analysis result; Obtaining a loss rate of the red soil data; Determining whether the analysis accuracy of the red soil data meets the requirements based on the loss rate of the red soil data; If the analysis accuracy does not meet the requirements, adjusting the collection frequency of the red soil data, or determining whether the analysis reliability of the deep learning model meets the requirements based on the accuracy of the red soil data analysis; If the analysis reliability does not meet the requirements, adjusting the learning rate of the deep learning model, or adjusting the cutoff frequency of the filter based on the response delay time length of the soil pH sensor; The red soil data includes the pH of the red soil, the nitrogen content of the red soil, the phosphorus content of the red soil, and the water content of the red soil, and the pH of the red soil is detected by the soil pH sensor; Determining the analysis accuracy of the red soil data comprises: Comparing the loss rate of the red soil data with a preset first loss rate; If the loss rate of the red soil data is greater than the preset first loss rate, it is determined that the analysis accuracy of the red soil data does not meet the requirements; Determining the analysis reliability of the deep learning model comprises: Comparing the loss rate of the red soil data with the preset first loss rate and a preset second loss rate, respectively; If the loss rate of the red soil data is greater than the preset first loss rate and less than or equal to the preset second loss rate, it is preliminarily determined that the analysis reliability of the deep learning model does not meet the requirements, and whether the analysis reliability of the deep learning model meets the requirements is determined according to the accuracy of the red soil data analysis; Adjusting the collection frequency of the red soil data comprises: Comparing the loss rate of the red soil data with the preset second loss rate; If the loss rate of the red soil data is greater than the preset second loss rate, the collection frequency of the red soil data is increased; Adjusting the learning rate of the deep learning model comprises: Comparing the accuracy of the red soil data analysis with a preset first accuracy and a preset second accuracy, respectively; If the accuracy of the red soil data analysis is less than the preset second accuracy, it is determined that the analysis reliability of the deep learning model does not meet the requirements; If the accuracy of the red soil data analysis is greater than the preset first accuracy and less than the preset second accuracy, the learning rate of the deep learning model is decreased; If the accuracy of the red soil data analysis is less than or equal to the preset first accuracy, it is preliminarily determined that the environmental stability of the red soil data collection does not meet the requirements, and whether the environmental stability of the red soil data collection meets the requirements is determined according to the response delay time length of the soil pH sensor; Adjusting the cutoff frequency of the filter comprises: Comparing the response delay time length of the soil pH sensor with a preset delay time length; If the response delay duration of the soil pH sensor is greater than the preset delay duration, it is determined that the environmental stability of the red soil data collection does not meet the requirements, and the cutoff frequency of the filter is reduced.

2. The big data based analysis method of power transmission line tower foundation red soil according to claim 1, characterized in that, The increase amplitude of the red soil data collection frequency is determined by the difference between the loss rate of the red soil data and the preset second loss rate.

3. The big data based analysis method of power transmission line tower foundation red soil according to claim 2, characterized in that, The decrease amplitude of the learning rate of the deep learning model is determined by the difference between the accuracy of the red soil data analysis and the preset first accuracy.

4. The big data based analysis method of power transmission line tower foundation red soil according to claim 3, characterized in that, The accuracy of the red soil data analysis is the ratio of the number of accurate red soil data analysis to the total number of red soil data analysis.

5. The big data based analysis method of power transmission line tower foundation red soil according to claim 4, characterized in that, The decrease amplitude of the cutoff frequency of the filter is determined by the difference between the response delay duration of the soil pH sensor and the preset delay duration.

Images