Method and system for automatically checking and optimizing cable laying path based on artificial intelligence

Abstract

The invention discloses a cable laying path automatic checking and optimizing method and system based on artificial intelligence, and relates to the field of electric power engineering, a first path analysis model is established, geographical environment, construction resources, technical constraints and multi-modal sensing data are input, and a second path analysis model is obtained through training by using a sample data set; inputting to-be-planned path data into the model, obtaining a first rationality evaluation value and changes of the first rationality evaluation value of each stage, identifying key driving factors, collecting second related data and inputting the second related data into the model, obtaining a second rationality evaluation value, optimizing model parameters according to an evaluation value difference value, and obtaining an optimized path analysis model. A cellular automaton is used for generating a candidate path scheme, a collision check dynamic optimization algorithm and a digital twinning technology are used for checking and verification, an optimal path scheme is selected, and intelligent design and optimization of a cable laying path are achieved through multi-stage evaluation, key factor analysis and model optimization. And the rationality and the resource utilization rate of path planning are improved.

Description

Technical Field

[0001] This invention relates to the field of power engineering technology, and more specifically, to an automatic verification and optimization method and system for cable laying paths based on artificial intelligence. Background Technology

[0002] In the construction of modern power transmission systems, cable laying is a critical task. The rationality of its route planning is crucial to the safe, stable, and efficient operation of the power system. Traditional cable laying route planning methods mainly rely on manual experience and simple geometric analysis. This approach often has limitations when facing complex geographical environments, diverse construction resource inputs, and strict technical constraints. First, manual planning cannot comprehensively and accurately consider the interaction between various complex factors, such as different terrain features, the distribution of underground obstacles, and the precision of resource allocation. This can easily lead to problems such as high construction difficulty, high cost, and inconvenient later maintenance after the planned route is completed. Second, with the acceleration of urbanization and the continuous growth of electricity demand, the scale and complexity of cable laying projects are increasing. Traditional methods cannot meet the needs of large-scale, high-precision route planning, cannot effectively cope with the rich information brought by multimodal sensing data, and cannot fully utilize this data for in-depth analysis and optimization. In addition, during the construction process, environmental changes and the emergence of various dynamic factors can quickly make the pre-planned route unreasonable. Traditional methods lack effective dynamic verification and optimization mechanisms and cannot adjust in time according to the actual situation, thereby affecting the quality and progress of cable laying projects and even potentially causing safety hazards.

[0003] Existing cable laying path planning methods have problems such as being unable to accurately consider complex factors, being unable to meet the needs of large-scale and high-precision operations, and lacking dynamic verification and optimization mechanisms. Summary of the Invention

[0004] To overcome the problems of existing cable laying path planning methods, such as inability to accurately consider complex factors, difficulty in meeting large-scale and high-precision requirements, and lack of dynamic verification and optimization mechanisms, this invention discloses an artificial intelligence-based automatic verification and optimization method and system for cable laying paths, which can effectively solve the above-mentioned technical problems.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] An AI-based method for automatic verification and optimization of cable laying paths includes the following steps:

[0007] A first path analysis model is established; the inputs of the first path analysis model include geographical environment data of the cable laying area, construction resource input data, technical constraint data, and multimodal perception data.

[0008] Obtain a sample dataset, and use the sample dataset to train the first path analysis model to obtain a second path analysis model; the path planning rationality and resource investment ratio of the second path analysis model are not less than a preset target ratio threshold.

[0009] The relevant data of the cable laying path to be planned is input into the second path analysis model to obtain the first rationality evaluation value of the cable laying path at each stage, and the change of the first rationality evaluation value.

[0010] Based on the changes in the first rationality assessment value, the key driving factors that lead to a decrease in the rationality of cable laying paths are obtained from the relevant data;

[0011] Collect second relevant data on the cable laying path after a preset time, and input the second relevant data and the key driving factors into the second path analysis model to obtain the second rationality evaluation value of the cable laying path;

[0012] Based on the difference between the first rationality assessment value and the second rationality assessment value, the parameters of the second path analysis model are optimized to obtain the optimized path analysis model;

[0013] A number of candidate cable laying path schemes are generated using cellular automata, and a collision check dynamic optimization algorithm is used to check the candidate cable laying path schemes and eliminate path schemes with collision risks.

[0014] A virtual scenario for cable laying is constructed using digital twin technology. The remaining candidate cable laying path schemes are input into the virtual scenario for interactive verification, and the optimal path scheme is selected.

[0015] Preferably, establishing the first path analysis model includes:

[0016] The Analytic Hierarchy Process (AHP) was used to quantify the input geographical environment data, construction resource input data, technical constraint data, and multimodal sensing data of the cable laying area, and to determine the weight coefficients of each data.

[0017] The correlation degree between each data is calculated by a correlation analysis algorithm, and the first evaluation value of the path evaluation stage is obtained based on the correlation degree.

[0018] Cluster the first evaluation value to obtain the evaluation level;

[0019] A first functional expression for the evaluation level is established using a linear regression method, and the parameters of the first functional expression are weighted to obtain the second evaluation value for the path rationality decomposition stage.

[0020] The actual deviation value is calculated based on the second evaluation value and the actual reasonableness value of the cable laying path;

[0021] Based on the actual deviation value, a standard normal distribution curve is constructed, and the standard deviation is calculated based on the standard normal distribution curve to obtain the third evaluation value;

[0022] The third evaluation value is fitted with a quadratic term using the least squares method to obtain the fitting curve;

[0023] The difference sequence nodes are obtained from the fitted curve. The environmental adjustment value is determined based on the coordinates of the difference sequence nodes at the inflection point. The environmental adjustment value is then used to correct the third evaluation value, and the output of the first path analysis model is established.

[0024] Preferably, the second path analysis model includes:

[0025] Obtain initial sample datasets from different cable laying projects;

[0026] The initial sample dataset is augmented using a sample augmentation algorithm to generate several augmented data sets with a similarity greater than a preset similarity threshold to the initial sample dataset.

[0027] The sample dataset is obtained based on the initial sample dataset and the expanded data;

[0028] The first path analysis model is trained using the sample dataset to obtain the first weight parameters of each input indicator of the first path analysis model;

[0029] The first weight parameter is used to assign weights to the input indicators of the first path analysis model to obtain the second path analysis model.

[0030] The generation of several augmented data sets with a similarity greater than a preset similarity threshold to the initial sample dataset includes:

[0031] The initial sample dataset is normalized by construction phase to obtain phase-normalized data, and the Kendall rank correlation coefficient is calculated based on the phase-normalized data.

[0032] The sampling weights are determined using the Kendall rank correlation coefficient, and the stage normalized data are calculated based on the sampling weights to obtain the first stage data.

[0033] The moving average method is used to obtain the numerical central trend of the first stage data, and the probability distribution is fitted to the numerical central trend to obtain the second stage data.

[0034] Based on the data from the second stage, a construction stage time series curve is constructed. The least squares method is used to establish a prediction function for the cable laying progress, construction quality and resource utilization rate in the construction stage time series curve. The third stage data is obtained based on the prediction function.

[0035] A time-series evaluation curve is constructed based on the data from the third stage. The mean and variance of the time-series evaluation curve are calculated. The time-series evaluation curve is then segmented to obtain the slope of the segmented curve.

[0036] The initial sample dataset is expanded based on the slope of the segmented curve and the preset similarity threshold to generate expanded data.

[0037] Preferably, the first weight parameters for obtaining each input indicator of the first path analysis model include:

[0038] Based on the scale of the cable laying project corresponding to the sample dataset, the sample dataset is subjected to stratified sampling to obtain the training sample dataset and the test sample dataset.

[0039] The training sample input data of the first path analysis model is obtained from the training sample dataset, and the correlation between the training sample input data is calculated to obtain the covariance matrix;

[0040] The initial weight parameters are obtained based on the eigenvalues ​​of the covariance matrix.

[0041] A linear programming function is constructed based on the initial weight parameters to obtain the fitted curve;

[0042] The initial weight parameters are optimized based on the fitted curve and the actual values ​​of the training sample data to obtain optimized weight parameters;

[0043] The test sample dataset is input into the first path analysis model that is weighted using the optimized weight parameters to obtain the verification output values.

[0044] The verification output values ​​are fitted with a normal distribution to the test sample dataset to obtain the error parameters;

[0045] Based on the optimized weight parameters and the error parameters, the first weight parameters of each input index of the first path analysis model are obtained.

[0046] Preferably, obtaining the first reasonableness assessment value of the cable laying path at each stage, and the changes in the first reasonableness assessment value, includes:

[0047] Collect relevant data on the cable laying route to be planned at different construction stages; the relevant data includes geographical environment data, construction resource input data, and technical constraint data;

[0048] In the second path analysis model, the correlation matrix of the relevant data is calculated, and cluster analysis is performed on the correlation matrix to obtain the matching parameters of the relevant data;

[0049] The construction effect data in the relevant data is subjected to exponential smoothing to obtain a construction effect data sequence;

[0050] Calculate the difference between each point in the construction effect data sequence to obtain the stage node sequence;

[0051] Based on the ratio parameters and construction effect data of each sequence point in the stage node sequence, the first rationality evaluation value of each stage is obtained;

[0052] The first rationality assessment value is fitted with least squares and segmented according to the fitting result to obtain the slope sequence;

[0053] Based on the fluctuation range of the slope sequence, a key interval is obtained, and the first change point of the ratio parameter and the second change point of the construction effect data are extracted from the key interval.

[0054] Based on the first change point and the second change point, the changes in the first rationality assessment value of the cable laying path at each stage are obtained.

[0055] Preferably, the key driving factors derived from the relevant data that lead to a decrease in the rationality of cable laying paths include:

[0056] Based on the changes in the first reasonableness assessment value, relevant data for the decreasing range are obtained from the relevant data; the absolute value of the difference between the reasonableness assessment value of the first stage corresponding to the relevant data for the decreasing range and the reasonableness assessment values ​​of the remaining second stages is greater than a preset reasonableness assessment value difference threshold.

[0057] Calculate the correlation coefficients of the geographic environment data, construction resource input data, and technical constraint data of the data related to the descent interval, and group the data related to the descent interval according to the correlation coefficients to obtain several data groups of the descent interval;

[0058] Principal component analysis was performed on the construction effect data in the relevant data of the descent interval to obtain the feature vector;

[0059] The month-on-month change rate is calculated based on the feature vector, and significant change indicators are obtained from the construction effect data of the relevant data in the decreasing interval based on the month-on-month change rate and the preset month-on-month change rate threshold.

[0060] A correlation analysis is performed on the data set in the decreasing interval and the significant change index to obtain the correlation strength value;

[0061] Based on the correlation strength value, key driving factors are obtained; the key driving factors include key geographical environment data, key construction resource input data, and key technical constraint data that lead to a decrease in the rationality of cable laying path.

[0062] Preferably, obtaining the correlation strength value includes:

[0063] The difference between the data set in the decreasing interval and the significant change index and the preset benchmark data are calculated to obtain the data fluctuation amplitude sequence and the effect fluctuation amplitude sequence.

[0064] Calculate the correlation between the data fluctuation amplitude sequence and the effect fluctuation amplitude sequence to obtain the correlation curve;

[0065] The correlation curve is divided into several intervals according to a preset correlation threshold.

[0066] Based on the aforementioned intervals, interval features are obtained; each interval corresponds to one interval feature.

[0067] Based on the interval characteristics, the data group of the decreasing interval and the significant change index are divided into a strongly correlated group and a weakly correlated group;

[0068] Establish a decision tree for the strongly correlated group, split the nodes of the decision tree according to a preset information gain criterion, and extract the numerical distribution pattern from the nodes;

[0069] Based on the numerical distribution pattern, the correlation coefficient is calculated to obtain the correlation strength value.

[0070] Preferably, optimizing the parameters of the second path analysis model to obtain the optimized path analysis model includes:

[0071] Calculate the difference between the first rationality assessment value and the second rationality assessment value to obtain the rationality change fitting curve within the preset time period;

[0072] Based on the fluctuations of the fitted curve, abnormal stages are identified from the relevant data across all stages.

[0073] The similarity of the parameters of the second path analysis model corresponding to the abnormal stage is calculated to obtain the parameter similarity.

[0074] Based on the parameter similarity and the preset parameter similarity threshold, the parameters are adjusted within the iteration interval to obtain the adjusted parameters;

[0075] The parameters of the second path analysis model are optimized using the aforementioned adjustment parameters to obtain the optimized path analysis model.

[0076] An AI-based automatic cable laying path verification and optimization system includes:

[0077] The model building module is used to build a first path analysis model; the inputs of the first path analysis model include geographical environment data of the cable laying area, construction resource input data, technical constraint data, and multimodal perception data.

[0078] The model training module is used to acquire a sample dataset and train the first path analysis model using the sample dataset to obtain a second path analysis model; the path planning rationality and resource investment ratio of the second path analysis model are not less than a preset target ratio threshold.

[0079] The first evaluation value calculation module is used to input relevant data of the cable laying path to be planned into the second path analysis model to obtain the first rationality evaluation value of the cable laying path at each stage, and the change of the first rationality evaluation value.

[0080] The key driving factor acquisition module is used to obtain the key driving factors that cause the rationality of cable laying path to decrease from the relevant data based on the changes in the first rationality assessment value.

[0081] The second evaluation value calculation module is used to collect second relevant data of the cable laying path after a preset time, input the second relevant data and the key driving factors into the second path analysis model, and obtain the second rationality evaluation value of the cable laying path.

[0082] The path planning management module is used to optimize the parameters of the second path analysis model based on the difference between the first rationality assessment value and the second rationality assessment value, so as to obtain the optimized path analysis model.

[0083] The path generation module is used to generate several candidate cable laying path schemes using cellular automata.

[0084] The collision verification module is used to verify the candidate cable laying path schemes using a collision verification dynamic optimization algorithm, and to eliminate path schemes with collision risks.

[0085] The digital twin verification module is used to construct a virtual scenario of cable laying using digital twin technology. The remaining candidate cable laying path schemes are input into the virtual scenario for interactive verification, and the optimal path scheme is selected.

[0086] Compared with existing technologies, the advantages of this invention are as follows: This invention integrates multi-dimensional data such as the geographical environment of the cable laying area, construction resource input, technical constraints, and multimodal perception through the established path analysis model. It also considers various complex factors in cable laying path planning through steps such as quantifying and determining weight coefficients using the analytic hierarchy process (AHP) and calculating correlation values ​​using correlation analysis. Compared with traditional methods that rely solely on experience or partial data, this invention can more accurately assess path rationality. Furthermore, it trains a second path analysis model on a large amount of sample data using artificial intelligence technology and expands the sample dataset using a sample augmentation algorithm, enabling the model to handle large-scale data. To ensure high-precision evaluation results, cellular automata are used when generating candidate path schemes, combined with collision check dynamic optimization algorithms and digital twin technology for verification, improving the accuracy and reliability of path planning and meeting the high-precision path planning requirements of large-scale cable laying projects. By analyzing the rationality evaluation values ​​of each stage of the planned path, key driving factors are identified, and the model parameters are optimized using subsequently collected data, realizing dynamic verification and optimization of cable laying path planning. Unlike traditional static planning methods, this approach can adjust and optimize path schemes in a timely manner according to changes in actual conditions, improving the rationality and adaptability of cable laying paths. Attached Figure Description

[0087] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other embodiments can be derived from the provided drawings without creative effort.

[0088] Figure 1 This is a diagram illustrating the steps of the method of the present invention;

[0089] Figure 2 This is a system structure diagram of the present invention. Detailed Implementation

[0090] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.

[0091] To better illustrate this embodiment, some parts in the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions;

[0092] It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings.

[0093] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0094] Example 1

[0095] An AI-based method for automatic verification and optimization of cable laying paths includes the following steps:

[0096] A first path analysis model is established; the inputs of the first path analysis model include geographical environment data of the cable laying area, construction resource input data, technical constraint data, and multimodal perception data.

[0097] Obtain a sample dataset, and use the sample dataset to train the first path analysis model to obtain a second path analysis model; the path planning rationality and resource investment ratio of the second path analysis model are not less than a preset target ratio threshold.

[0098] The relevant data of the cable laying path to be planned is input into the second path analysis model to obtain the first rationality evaluation value of the cable laying path at each stage, and the change of the first rationality evaluation value.

[0099] Based on the changes in the first rationality assessment value, the key driving factors that lead to a decrease in the rationality of cable laying paths are obtained from the relevant data;

[0100] Collect second relevant data on the cable laying path after a preset time, and input the second relevant data and the key driving factors into the second path analysis model to obtain the second rationality evaluation value of the cable laying path;

[0101] Based on the difference between the first rationality assessment value and the second rationality assessment value, the parameters of the second path analysis model are optimized to obtain the optimized path analysis model;

[0102] A number of candidate cable laying path schemes are generated using cellular automata, and a collision check dynamic optimization algorithm is used to check the candidate cable laying path schemes and eliminate path schemes with collision risks.

[0103] A virtual scenario for cable laying is constructed using digital twin technology. The remaining candidate cable laying path schemes are input into the virtual scenario for interactive verification, and the optimal path scheme is selected.

[0104] The establishment of the first path analysis model includes:

[0105] The Analytic Hierarchy Process (AHP) was used to quantify the input geographical environment data, construction resource input data, technical constraint data, and multimodal sensing data of the cable laying area, and to determine the weight coefficients of each data.

[0106] The correlation degree between each data is calculated by a correlation analysis algorithm, and the first evaluation value of the path evaluation stage is obtained based on the correlation degree.

[0107] Cluster the first evaluation value to obtain the evaluation level;

[0108] A first functional expression for the evaluation level is established using a linear regression method, and the parameters of the first functional expression are weighted to obtain the second evaluation value for the path rationality decomposition stage.

[0109] The actual deviation value is calculated based on the second evaluation value and the actual reasonableness value of the cable laying path;

[0110] Based on the actual deviation value, a standard normal distribution curve is constructed, and the standard deviation is calculated based on the standard normal distribution curve to obtain the third evaluation value;

[0111] The third evaluation value is fitted with a quadratic term using the least squares method to obtain the fitting curve;

[0112] The difference sequence nodes are obtained from the fitted curve. The environmental adjustment value is determined based on the coordinates of the difference sequence nodes at the inflection point. The environmental adjustment value is then used to correct the third evaluation value, and the output of the first path analysis model is established.

[0113] The obtained second path analysis model includes:

[0114] Obtain initial sample datasets from different cable laying projects;

[0115] The initial sample dataset is augmented using a sample augmentation algorithm to generate several augmented data sets with a similarity greater than a preset similarity threshold to the initial sample dataset.

[0116] The sample dataset is obtained based on the initial sample dataset and the expanded data;

[0117] The first path analysis model is trained using the sample dataset to obtain the first weight parameters of each input indicator of the first path analysis model;

[0118] The first weight parameter is used to assign weights to the input indicators of the first path analysis model to obtain the second path analysis model.

[0119] The generation of several augmented data sets with a similarity greater than a preset similarity threshold to the initial sample dataset includes:

[0120] The initial sample dataset is normalized by construction phase to obtain phase-normalized data, and the Kendall rank correlation coefficient is calculated based on the phase-normalized data.

[0121] The sampling weights are determined using the Kendall rank correlation coefficient, and the stage normalized data are calculated based on the sampling weights to obtain the first stage data.

[0122] The moving average method is used to obtain the numerical central trend of the first stage data, and the probability distribution is fitted to the numerical central trend to obtain the second stage data.

[0123] Based on the data from the second stage, a construction stage time series curve is constructed. The least squares method is used to establish a prediction function for the cable laying progress, construction quality and resource utilization rate in the construction stage time series curve. The third stage data is obtained based on the prediction function.

[0124] A time-series evaluation curve is constructed based on the data from the third stage. The mean and variance of the time-series evaluation curve are calculated. The time-series evaluation curve is then segmented to obtain the slope of the segmented curve.

[0125] The initial sample dataset is expanded based on the slope of the segmented curve and the preset similarity threshold to generate expanded data.

[0126] The first weight parameters for each input indicator of the first path analysis model include:

[0127] Based on the scale of the cable laying project corresponding to the sample dataset, the sample dataset is subjected to stratified sampling to obtain the training sample dataset and the test sample dataset.

[0128] The training sample input data of the first path analysis model is obtained from the training sample dataset, and the correlation between the training sample input data is calculated to obtain the covariance matrix;

[0129] The initial weight parameters are obtained based on the eigenvalues ​​of the covariance matrix.

[0130] A linear programming function is constructed based on the initial weight parameters to obtain the fitted curve;

[0131] The initial weight parameters are optimized based on the fitted curve and the actual values ​​of the training sample data to obtain optimized weight parameters;

[0132] The test sample dataset is input into the first path analysis model that is weighted using the optimized weight parameters to obtain the verification output values.

[0133] The verification output values ​​are fitted with a normal distribution to the test sample dataset to obtain the error parameters;

[0134] Based on the optimized weight parameters and the error parameters, the first weight parameters of each input index of the first path analysis model are obtained.

[0135] The first reasonableness assessment value of the cable laying path at each stage, and the changes in the first reasonableness assessment value, include:

[0136] Collect relevant data on the cable laying route to be planned at different construction stages; the relevant data includes geographical environment data, construction resource input data, and technical constraint data;

[0137] In the second path analysis model, the correlation matrix of the relevant data is calculated, and cluster analysis is performed on the correlation matrix to obtain the matching parameters of the relevant data;

[0138] The construction effect data in the relevant data is subjected to exponential smoothing to obtain a construction effect data sequence;

[0139] Calculate the difference between each point in the construction effect data sequence to obtain the stage node sequence;

[0140] Based on the ratio parameters and construction effect data of each sequence point in the stage node sequence, the first rationality evaluation value of each stage is obtained;

[0141] The first rationality assessment value is fitted with least squares and segmented according to the fitting result to obtain the slope sequence;

[0142] Based on the fluctuation range of the slope sequence, a key interval is obtained, and the first change point of the ratio parameter and the second change point of the construction effect data are extracted from the key interval.

[0143] Based on the first change point and the second change point, the changes in the first rationality assessment value of the cable laying path at each stage are obtained.

[0144] The key driving factors that lead to a decrease in the rationality of cable laying paths, as identified from the relevant data, include:

[0145] Based on the changes in the first reasonableness assessment value, relevant data for the decreasing range are obtained from the relevant data; the absolute value of the difference between the reasonableness assessment value of the first stage corresponding to the relevant data for the decreasing range and the reasonableness assessment values ​​of the remaining second stages is greater than a preset reasonableness assessment value difference threshold.

[0146] Calculate the correlation coefficients of the geographic environment data, construction resource input data, and technical constraint data of the data related to the descent interval, and group the data related to the descent interval according to the correlation coefficients to obtain several data groups of the descent interval;

[0147] Principal component analysis was performed on the construction effect data in the relevant data of the descent interval to obtain the feature vector;

[0148] The month-on-month change rate is calculated based on the feature vector, and significant change indicators are obtained from the construction effect data of the relevant data in the decreasing interval based on the month-on-month change rate and the preset month-on-month change rate threshold.

[0149] A correlation analysis is performed on the data set in the decreasing interval and the significant change index to obtain the correlation strength value;

[0150] Based on the correlation strength value, key driving factors are obtained; the key driving factors include key geographical environment data, key construction resource input data, and key technical constraint data that lead to a decrease in the rationality of cable laying path.

[0151] The obtained correlation strength value includes:

[0152] The difference between the data set in the decreasing interval and the significant change index and the preset benchmark data are calculated to obtain the data fluctuation amplitude sequence and the effect fluctuation amplitude sequence.

[0153] Calculate the correlation between the data fluctuation amplitude sequence and the effect fluctuation amplitude sequence to obtain the correlation curve;

[0154] The correlation curve is divided into several intervals according to a preset correlation threshold.

[0155] Based on the aforementioned intervals, interval features are obtained; each interval corresponds to one interval feature.

[0156] Based on the interval characteristics, the data group of the decreasing interval and the significant change index are divided into a strongly correlated group and a weakly correlated group;

[0157] Establish a decision tree for the strongly correlated group, split the nodes of the decision tree according to a preset information gain criterion, and extract the numerical distribution pattern from the nodes;

[0158] Based on the numerical distribution pattern, the correlation coefficient is calculated to obtain the correlation strength value.

[0159] The optimization of the parameters of the second path analysis model to obtain the optimized path analysis model includes:

[0160] Calculate the difference between the first rationality assessment value and the second rationality assessment value to obtain the rationality change fitting curve within the preset time period;

[0161] Based on the fluctuations of the fitted curve, abnormal stages are identified from the relevant data across all stages.

[0162] The similarity of the parameters of the second path analysis model corresponding to the abnormal stage is calculated to obtain the parameter similarity.

[0163] Based on the parameter similarity and the preset parameter similarity threshold, the parameters are adjusted within the iteration interval to obtain the adjusted parameters;

[0164] The parameters of the second path analysis model are optimized using the aforementioned adjustment parameters to obtain the optimized path analysis model.

[0165] An electronic device includes: a memory and at least one processor, the memory storing instructions, wherein at least one processor invokes the instructions in the memory to cause the device to perform the steps of the method described above.

[0166] A computer-readable storage medium having instructions stored thereon, which, when executed by a processor, implement the steps of the method described above.

[0167] For specific implementation details, please refer to [link / reference]. Figure 1 The study collected geographical environmental data on the urban cable laying area, including the route, depth, and material of underground pipelines (water supply, drainage, gas, communications, etc.), road width, traffic flow, pedestrian flow, and basic information of surrounding buildings and underground space utilization. Construction resource input data covered construction equipment such as excavators and cranes, including their types, quantities, and performance parameters, as well as the number, job types, and skill levels of construction personnel. Technical constraint data included cable voltage levels, insulation types, allowable bending radii, and minimum safe distances from existing pipelines. Multimodal sensing data came from ground radar scan images, underground pipeline detection instrument data, and real-time monitoring videos of the construction site. Using the analytic hierarchy process (AHP), power, municipal, and construction experts scored the importance of each type of data, determining the weight coefficient of each data point in the path analysis. For example, the complexity of underground pipelines was weighted at 0.35, and road traffic flow at 0. The correlation degree between various data is calculated using correlation analysis algorithms, such as a correlation of 0.8 between the complexity of underground pipelines and construction difficulty, and a correlation of 0.75 between road traffic volume and construction window period. Based on the correlation degree, the first evaluation value range of 0-100 points is obtained in the path assessment stage. The first evaluation value is clustered into high risk (70-100 points), medium risk (40-69 points), and low risk (0-39 points) levels. Linear regression is used to establish a function expression between the evaluation level and each data point, and weighted to obtain the second evaluation value. The deviation value is calculated by combining the second evaluation value with the actual rationality value. A standard normal distribution curve is constructed and the standard deviation is calculated to obtain the third evaluation value. The least squares quadratic term is used to fit the curve to obtain the fitting curve. The difference sequence nodes are extracted from the curve to determine the environmental adjustment value. The third evaluation value is corrected to obtain the output of the first path analysis model, i.e., the rationality assessment score.

[0168] Initial sample datasets were obtained from previous urban cable laying projects, including project location, scale, construction conditions, routes, and evaluation results. A sample augmentation algorithm was used to normalize the initial sample data by construction stage, and Kendall's rank correlation coefficient was calculated to determine sampling weights, resulting in the first-stage data. Moving averages were used to obtain the numerical central trend, and a probability distribution was fitted to obtain the second-stage data. A construction stage time-series curve was constructed, and a prediction function was established using the least squares method to obtain the third-stage data. Mean-variance time-series evaluation curves were calculated, and the initial sample dataset was expanded to generate 240 expanded datasets based on the slope of the segmented curves and a similarity threshold of 0.75. The expanded sample datasets were divided into training and testing sets. The covariance matrix of the training set was calculated to obtain initial weight parameters. A linear programming function was constructed to fit the curve, and the weight parameters were optimized by comparing them with actual values. The test set was input into the weighted first path analysis model, and a normal distribution was fitted to obtain error parameters. The first weight parameters for each input indicator were then obtained, resulting in the second path analysis model. Its path planning rationality and resource input ratio reached 0.68, with a preset target ratio threshold of 0.65.

[0169] For the planned urban cable laying routes, relevant data were collected at different construction stages (pipeline relocation, road excavation, cable laying, and road surface restoration), including the density of underground pipelines, accessibility of construction equipment, and technical requirements for cable laying. A correlation matrix was calculated in the second path analysis model, and cluster analysis was used to obtain the allocation parameters. The construction effect data was exponentially smoothed to obtain a data sequence, and the differences were calculated to obtain the stage node sequence. Based on the allocation parameters and construction effect data, the first rationality assessment value for each stage was obtained, such as 75 points for pipeline relocation, 60 points for road excavation, 65 points for cable laying, and 80 points for road surface restoration. Least squares fitting was performed to segment the data to obtain a slope sequence. Based on the slope fluctuation range, key intervals were determined, and the change points were extracted to obtain the changes in the rationality assessment value. It was found that the rationality increased from the road excavation to the cable laying stage.

[0170] Based on the changes in the rationality assessment value, data from the road excavation to the cable laying stage were selected as the relevant data for the declining interval. The difference between the rationality assessment value of this stage and other stages exceeded a preset threshold of 10. The correlation coefficients of geographical environment data in this interval, such as underground obstacle density, construction resource input data, laying equipment power, technical constraint data, and cable turning radius, were calculated to obtain multiple data groups. For construction effect data, such as laying efficiency and failure rate, principal component analysis was used to obtain eigenvectors. The year-on-year change rate was calculated based on the eigenvectors to identify significant change indicators. Correlation analysis was used to obtain the correlation strength value, and the key driving factors were determined to be underground obstacle density (geographical environment data), laying equipment power (construction resource input data), and cable turning radius (technical constraint data).

[0171] After a preset time, second relevant data on the cable laying path are collected, including new underground pipeline layouts and actual use of construction resources. The second relevant data and key driving factors are input into the second path analysis model to obtain a second rationality assessment value. For example, if the assessment value increases from 60 points to 68 points from road excavation to cable laying, the difference between the first and second rationality assessment values ​​is calculated as 8 points, and a rationality change fitting curve is obtained. Abnormal stages are screened out from all construction stages, and the similarity of the model parameters corresponding to the abnormal stages is calculated. The parameters are adjusted within the iteration interval according to the similarity and the preset threshold of 0.7 to obtain the adjusted parameters. The optimized path analysis model improves the rationality of path planning and the ratio of resource input to 0.75.

[0172] Using cellular automata, based on the urban underground pipeline network layout rules, such as cell state transition rules (if four adjacent cells are in a safe laying state, the current cell is converted to a layable state) and construction constraints (such as cable laying must not damage major traffic arteries), 35 candidate cable laying path schemes are generated.

[0173] A collision-checking dynamic optimization algorithm was used to simulate dynamic collisions during cable laying. A collision distance threshold of 0.4 meters was set to eliminate 10 path schemes with collision risks. A virtual scenario of urban cable laying was constructed using digital twin technology, including virtual underground pipelines, construction equipment, and cables. The remaining 25 candidate path schemes were input into the virtual scenario for interactive verification. The laying process under different construction conditions was simulated in the virtual environment, and indicators such as path adaptability, construction difficulty, and impact on traffic were monitored. Considering factors such as rationality evaluation value, construction cost, and impact on the urban environment, the optimal path scheme was finally selected. This scheme achieved a rationality evaluation value of 92 points, reduced construction costs, and minimal interference with urban traffic.

[0174] Example 2

[0175] An AI-based automatic cable laying path verification and optimization system includes:

[0176] The model building module is used to build a first path analysis model; the inputs of the first path analysis model include geographical environment data of the cable laying area, construction resource input data, technical constraint data, and multimodal perception data.

[0177] The model training module is used to acquire a sample dataset and train the first path analysis model using the sample dataset to obtain a second path analysis model; the path planning rationality and resource investment ratio of the second path analysis model are not less than a preset target ratio threshold.

[0178] The first evaluation value calculation module is used to input relevant data of the cable laying path to be planned into the second path analysis model to obtain the first rationality evaluation value of the cable laying path at each stage, and the change of the first rationality evaluation value.

[0179] The key driving factor acquisition module is used to obtain the key driving factors that cause the rationality of cable laying path to decrease from the relevant data based on the changes in the first rationality assessment value.

[0180] The second evaluation value calculation module is used to collect second relevant data of the cable laying path after a preset time, input the second relevant data and the key driving factors into the second path analysis model, and obtain the second rationality evaluation value of the cable laying path.

[0181] The path planning management module is used to optimize the parameters of the second path analysis model based on the difference between the first rationality assessment value and the second rationality assessment value, so as to obtain the optimized path analysis model.

[0182] The path generation module is used to generate several candidate cable laying path schemes using cellular automata.

[0183] The collision verification module is used to verify the candidate cable laying path schemes using a collision verification dynamic optimization algorithm, and to eliminate path schemes with collision risks.

[0184] The digital twin verification module is used to construct a virtual scenario of cable laying using digital twin technology. The remaining candidate cable laying path schemes are input into the virtual scenario for interactive verification, and the optimal path scheme is selected.

[0185] Please see Figure 2 The model building module establishes the first path analysis model. The input includes geographical environmental data of the cross-river area, including riverbed topography, water flow velocity, riverbed geology, shipping routes, etc.; construction resource input data, including large laying vessels, number of divers, underwater robots, etc.; technical constraint data, including cable waterproof performance, tensile strength, safe distance from shipping, etc.; and multimodal perception data, including underwater sonar detection data, real-time water surface monitoring video, etc.

[0186] The model training module collects the initial sample datasets from 30 previous cross-river cable laying projects, expands them to 90 data points using a sample augmentation algorithm, and forms a complete sample dataset to train the first path analysis model, resulting in the second path analysis model. The rationality of its path planning and the proportion of resource input are not less than the preset target proportion threshold of 0.6.

[0187] The first evaluation value calculation module inputs relevant data on the planned cross-river cable laying route into the second path analysis model. Data is collected at different construction stages, such as riverbed preparation, cable laying, fixing and protection, and water surface restoration, including the complexity of the riverbed geology, the performance of the construction vessels, and the technical parameters of the cable. The model calculates the first rationality evaluation value for each stage, such as 70 points for riverbed preparation, 60 points for cable laying, 75 points for fixing and protection, and 85 points for water surface restoration. The model also analyzes the changes in the rationality evaluation values, finding that the rationality is lower during the cable laying stage.

[0188] The key driver acquisition module filters data from the cable laying stage as the declining interval based on the change in the rationality assessment value. It calculates the correlation coefficients of geographical environment data in this interval, such as the intensity of the riverbed undercurrent, construction resource input data, such as the efficiency of divers, and technical constraint data, such as the cable sealing performance, to obtain multiple data groups. For construction effect data, such as laying progress and failure rate, principal component analysis is used to obtain eigenvectors. The year-on-year change rate is calculated based on the eigenvectors to identify significant change indicators. Correlation analysis is used to obtain the correlation strength value, and the key driver factors are determined to be the intensity of the riverbed undercurrent (geographical environment data), the efficiency of divers (construction resource input data), and the cable sealing performance (technical constraint data).

[0189] After a preset time, the second evaluation value calculation module collects second relevant data on the cable laying path, including new riverbed geological scanning data and actual use of construction resources. The second relevant data and key driving factors are input into the second path analysis model to obtain the second rationality evaluation value. For example, the evaluation value of the cable laying stage increases from 60 points to 66 points.

[0190] The route planning management module calculates the difference of 6 points between the first and second rationality assessment values, obtains the rationality change fitting curve, screens out abnormal stages from all construction stages, calculates the similarity of the model parameters corresponding to the abnormal stages, and adjusts the parameters within the iteration interval according to the similarity and the preset threshold of 0.72 to obtain the adjusted parameters. The optimized route analysis model improves the ratio of route planning rationality to resource input to 0.7.

[0191] The path generation module uses cellular automata to generate 20 candidate cable laying path schemes based on the riverbed topography and water flow patterns, such as cell state transition rules: if three adjacent cells are in a safe laying state and the water flow velocity is less than 2m / s, then the current cell is converted to a layable state and construction constraints, such as cable laying must not affect the main shipping channel.

[0192] The collision verification module uses a dynamic optimization algorithm to simulate dynamic collisions during cable laying. The collision distance threshold is set to 0.6 meters. The module verifies 20 candidate paths and eliminates 6 path schemes that pose a collision risk.

[0193] The digital twin verification module uses digital twin technology to construct a virtual scenario for laying cross-river cables, including virtual riverbed topography, laying vessels, and cables. The remaining 14 candidate route schemes are input into the virtual scenario for interactive verification. The laying process under different hydrological conditions is simulated in the virtual environment, and indicators such as the adaptability of the route, construction difficulty, and impact on shipping are monitored. Considering factors such as the rationality evaluation value, construction cost, and impact on shipping safety, the optimal route scheme is finally selected. This scheme has a rationality evaluation value of 95 points, reduces construction costs, and minimizes interference with shipping safety.

[0194] The same or similar labels correspond to the same or similar parts;

[0195] The terms used to describe positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.

[0196] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. For those skilled in the art, other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all implementation methods here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the claims of the present invention.

Claims

1. An automatic verification and optimization method for cable laying paths based on artificial intelligence, characterized in that, Includes the following steps: A first path analysis model is established; the inputs of the first path analysis model include geographical environment data of the cable laying area, construction resource input data, technical constraint data, and multimodal perception data. Obtain a sample dataset, and use the sample dataset to train the first path analysis model to obtain a second path analysis model; the path planning rationality and resource investment ratio of the second path analysis model are not less than a preset target ratio threshold. The relevant data of the cable laying path to be planned are input into the second path analysis model to obtain the first rationality evaluation value of the cable laying path at each stage, and the change of the first rationality evaluation value. Based on the changes in the first rationality assessment value, the key driving factors that lead to a decrease in the rationality of cable laying paths are obtained from the relevant data; Collect second relevant data on the cable laying path after a preset time, and input the second relevant data and the key driving factors into the second path analysis model to obtain the second rationality evaluation value of the cable laying path; Based on the difference between the first rationality assessment value and the second rationality assessment value, the parameters of the second path analysis model are optimized to obtain the optimized path analysis model; A number of candidate cable laying path schemes are generated using cellular automata, and a collision check dynamic optimization algorithm is used to check the candidate cable laying path schemes and eliminate path schemes with collision risks. A virtual scenario for cable laying is constructed using digital twin technology. The remaining candidate cable laying path schemes are input into the virtual scenario for interactive verification, and the optimal path scheme is selected.

2. The method according to claim 1, characterized in that, The establishment of the first path analysis model includes: The Analytic Hierarchy Process (AHP) was used to quantify the input geographical environment data, construction resource input data, technical constraint data, and multimodal sensing data of the cable laying area, and to determine the weight coefficients of each data. The correlation degree between each data is calculated by a correlation analysis algorithm, and the first evaluation value of the path evaluation stage is obtained based on the correlation degree. Cluster the first evaluation value to obtain the evaluation level; A first functional expression for the evaluation level is established using a linear regression method, and the parameters of the first functional expression are weighted to obtain the second evaluation value for the path rationality decomposition stage. The actual deviation value is calculated based on the second evaluation value and the actual reasonableness value of the cable laying path; Based on the actual deviation value, a standard normal distribution curve is constructed, and the standard deviation is calculated based on the standard normal distribution curve to obtain the third evaluation value; The third evaluation value is fitted with a quadratic term using the least squares method to obtain the fitting curve; The difference sequence nodes are obtained from the fitted curve. The environmental adjustment value is determined based on the coordinates of the difference sequence nodes at the inflection point. The environmental adjustment value is then used to correct the third evaluation value, and the output of the first path analysis model is established.

3. The method according to claim 1, characterized in that, The obtained second path analysis model includes: Obtain initial sample datasets from different cable laying projects; The initial sample dataset is augmented using a sample augmentation algorithm to generate several augmented data sets with a similarity greater than a preset similarity threshold to the initial sample dataset. The sample dataset is obtained based on the initial sample dataset and the expanded data; The first path analysis model is trained using the sample dataset to obtain the first weight parameters of each input indicator of the first path analysis model; The first weight parameter is used to assign weights to the input indicators of the first path analysis model to obtain the second path analysis model.

4. The method according to claim 3, characterized in that, The generation of several augmented data sets with a similarity greater than a preset similarity threshold to the initial sample dataset includes: The initial sample dataset is normalized by construction phase to obtain phase-normalized data, and the Kendall rank correlation coefficient is calculated based on the phase-normalized data. The sampling weights are determined using the Kendall rank correlation coefficient, and the stage normalized data are calculated based on the sampling weights to obtain the first stage data. The moving average method is used to obtain the numerical central trend of the first stage data, and the probability distribution is fitted to the numerical central trend to obtain the second stage data. Based on the data from the second stage, a construction stage time series curve is constructed. The least squares method is used to establish a prediction function for the cable laying progress, construction quality and resource utilization rate in the construction stage time series curve. The third stage data is obtained based on the prediction function. A time-series evaluation curve is constructed based on the data from the third stage. The mean and variance of the time-series evaluation curve are calculated. The time-series evaluation curve is then segmented to obtain the slope of the segmented curve. The initial sample dataset is expanded based on the slope of the segmented curve and the preset similarity threshold to generate expanded data.

5. The method according to claim 3, characterized in that, The first weight parameters for each input indicator of the first path analysis model include: Based on the scale of the cable laying project corresponding to the sample dataset, the sample dataset is subjected to stratified sampling to obtain the training sample dataset and the test sample dataset. The training sample input data of the first path analysis model is obtained from the training sample dataset, and the correlation between the training sample input data is calculated to obtain the covariance matrix; The initial weight parameters are obtained based on the eigenvalues ​​of the covariance matrix. A linear programming function is constructed based on the initial weight parameters to obtain the fitted curve; The initial weight parameters are optimized based on the fitted curve and the actual values ​​of the training sample data to obtain optimized weight parameters; The test sample dataset is input into the first path analysis model that is weighted using the optimized weight parameters to obtain the verification output values. The verification output values ​​are fitted with a normal distribution to the test sample dataset to obtain the error parameters; Based on the optimized weight parameters and the error parameters, the first weight parameters of each input index of the first path analysis model are obtained.

6. The method according to claim 1, characterized in that, The first reasonableness assessment value of the cable laying path at each stage, and the changes in the first reasonableness assessment value, include: Collect relevant data on the cable laying route to be planned at different construction stages; the relevant data includes geographical environment data, construction resource input data, and technical constraint data; In the second path analysis model, the correlation matrix of the relevant data is calculated, and cluster analysis is performed on the correlation matrix to obtain the matching parameters of the relevant data; The construction effect data in the relevant data is subjected to exponential smoothing to obtain a construction effect data sequence; Calculate the difference between each point in the construction effect data sequence to obtain the stage node sequence; Based on the ratio parameters and construction effect data of each sequence point in the stage node sequence, the first rationality evaluation value of each stage is obtained; The first rationality assessment value is fitted with least squares and segmented according to the fitting result to obtain the slope sequence; Based on the fluctuation range of the slope sequence, a key interval is obtained, and the first change point of the ratio parameter and the second change point of the construction effect data are extracted from the key interval. Based on the first change point and the second change point, the changes in the first rationality assessment value of the cable laying path at each stage are obtained.

7. The method according to claim 1, characterized in that, The key driving factors that lead to a decrease in the rationality of cable laying paths, as identified from the relevant data, include: Based on the changes in the first reasonableness assessment value, relevant data for the decreasing range are obtained from the relevant data; the absolute value of the difference between the reasonableness assessment value of the first stage corresponding to the relevant data for the decreasing range and the reasonableness assessment values ​​of the remaining second stages is greater than a preset reasonableness assessment value difference threshold. Calculate the correlation coefficients of the geographic environment data, construction resource input data, and technical constraint data of the data related to the descent interval, and group the data related to the descent interval according to the correlation coefficients to obtain several data groups of the descent interval; Principal component analysis was performed on the construction effect data in the relevant data of the descent interval to obtain the feature vector; The month-on-month change rate is calculated based on the feature vector, and significant change indicators are obtained from the construction effect data of the relevant data in the decreasing interval based on the month-on-month change rate and the preset month-on-month change rate threshold. A correlation analysis is performed on the data set in the decreasing interval and the significant change index to obtain the correlation strength value; Based on the correlation strength value, key driving factors are obtained; the key driving factors include key geographical environment data, key construction resource input data, and key technical constraint data that lead to a decrease in the rationality of cable laying path.

8. The method according to claim 7, characterized in that, The obtained correlation strength value includes: The difference between the data set in the decreasing interval and the significant change index and the preset benchmark data are calculated to obtain the data fluctuation amplitude sequence and the effect fluctuation amplitude sequence. Calculate the correlation between the data fluctuation amplitude sequence and the effect fluctuation amplitude sequence to obtain the correlation curve; The correlation curve is divided into several intervals according to a preset correlation threshold. Based on the aforementioned intervals, interval features are obtained; each interval corresponds to one interval feature. Based on the interval characteristics, the data group of the decreasing interval and the significant change index are divided into a strongly correlated group and a weakly correlated group; Establish a decision tree for the strongly correlated group, split the nodes of the decision tree according to a preset information gain criterion, and extract the numerical distribution pattern from the nodes; Based on the numerical distribution pattern, the correlation coefficient is calculated to obtain the correlation strength value.

9. The method according to claim 1, characterized in that, The optimization of the parameters of the second path analysis model to obtain the optimized path analysis model includes: Calculate the difference between the first rationality assessment value and the second rationality assessment value to obtain the rationality change fitting curve within the preset time period; Based on the fluctuations of the fitted curve, abnormal stages are identified from the relevant data across all stages. The similarity of the parameters of the second path analysis model corresponding to the abnormal stage is calculated to obtain the parameter similarity. Based on the parameter similarity and the preset parameter similarity threshold, the parameters are adjusted within the iteration interval to obtain the adjusted parameters; The parameters of the second path analysis model are optimized using the aforementioned adjustment parameters to obtain the optimized path analysis model.

10. An AI-based automatic cable laying path verification and optimization system, characterized in that, include: The model building module is used to build a first path analysis model; the inputs of the first path analysis model include geographical environment data of the cable laying area, construction resource input data, technical constraint data, and multimodal perception data. The model training module is used to acquire a sample dataset and train the first path analysis model using the sample dataset to obtain a second path analysis model; the path planning rationality and resource investment ratio of the second path analysis model are not less than a preset target ratio threshold. The first evaluation value calculation module is used to input relevant data of the cable laying path to be planned into the second path analysis model to obtain the first rationality evaluation value of the cable laying path at each stage, and the change of the first rationality evaluation value. The key driving factor acquisition module is used to obtain the key driving factors that cause the rationality of cable laying path to decrease from the relevant data based on the changes in the first rationality assessment value. The second evaluation value calculation module is used to collect second relevant data of the cable laying path after a preset time, input the second relevant data and the key driving factors into the second path analysis model, and obtain the second rationality evaluation value of the cable laying path. The path planning management module is used to optimize the parameters of the second path analysis model based on the difference between the first rationality assessment value and the second rationality assessment value, so as to obtain the optimized path analysis model. The path generation module is used to generate several candidate cable laying path schemes using cellular automata. The collision verification module is used to verify the candidate cable laying path schemes using a collision verification dynamic optimization algorithm, and to eliminate path schemes with collision risks. The digital twin verification module is used to construct a virtual scenario of cable laying using digital twin technology. The remaining candidate cable laying path schemes are input into the virtual scenario for interactive verification, and the optimal path scheme is selected.

Images