A Track Foundation Settlement Assessment Method Based on Electronic Digital Data Processing

Abstract

The present invention relates to the field of track assessment technology, and in particular to a track foundation settlement assessment method based on electrical digital data processing, which is as follows: collecting and processing raw data related to track foundation settlement, storing the processed data, and manually annotating the stored data; performing data enhancement, data synthesis, and sample diversity optimization operations based on the collected data, and then synthesizing electrical digital data samples using a sample interpolation method based on boundary conditions; establishing a track foundation settlement assessment prediction model, processing the collected data and the synthesized data to train the model, obtaining an assessment prediction result, and then evaluating and improving the trained model; providing users with suggestions for settlement control and repair based on the assessment prediction result of the track foundation settlement assessment prediction model, and judging whether to issue an early warning based on the assessment prediction result. This method can improve the accuracy of model evaluation and enhance the robustness and generalization ability of the model.

Description

Technical Field

[0001] The invention relates to the technical field of track assessment, and in particular to a track foundation settlement assessment method based on electrical digital data processing. Background Art

[0002] With the continuous expansion of urban rail systems and the increasing operating load, track foundations will inevitably experience settlement. Traditional monitoring methods mostly rely on analog signals or simple digital acquisition methods, which makes the data lack accuracy, real-time performance, and anti-interference ability, and cannot fully and accurately reflect the actual settlement situation. Existing technologies often use single or localized sensors in the data acquisition process, which are easily affected by environmental noise and interference during data preprocessing. In addition, the data format is not unified and lacks effective time series processing, which makes the settlement information obtained have certain errors and discontinuities.

[0003] On the other hand, traditional settlement assessment methods often use classic neural networks, support vector machines or decision trees in the process of model construction and prediction. These methods are prone to fall into local optimal states when facing high-dimensional, nonlinear and changeable electrical digital data. The parameter initialization and gradient update mechanisms lack sufficient flexibility, which leads to unsatisfactory prediction accuracy and stability of the model. At the same time, in response to the problems of insufficient and uneven data samples, traditional sample synthesis methods mostly use simple linear interpolation, which fails to fully consider the boundary conditions and dynamic change trends in the data, resulting in the inability of the synthesized training data to truly reflect the complex characteristics of track foundation settlement, thus affecting the generalization ability of the model and the actual application effect.

[0004] The prior art has the following deficiencies: (1) Traditional neural networks, support vector machines, and decision trees are prone to falling into local optimal solutions when faced with complex, high-dimensional, and nonlinear digital data, and are unable to fully capture global features, resulting in low classification accuracy and stability. (2) Traditional machine learning models lack adaptive adjustment methods for global and local features during parameter initialization and gradient update, which can easily lead to gradient vanishing or explosion, affecting the convergence speed and training effect of the model. (3) Traditional data synthesis and expansion mainly rely on simple linear interpolation, which makes it difficult to capture the boundary conditions and nonlinear changes in the actual data distribution. As a result, the synthesized new samples often deviate from the characteristics of the actual orbital settlement and cannot effectively solve the data imbalance problem.

[0005] Therefore, the present invention proposes a track foundation settlement assessment method based on electronic digital data processing to solve the above problems. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention develops a track foundation settlement assessment method based on electrical digital data processing. The present invention can improve the accuracy of model assessment and enhance the robustness and generalization ability of the model through a sample interpolation method based on boundary conditions and a fractional-order neural network based on path integral.

[0007] The technical solution to the technical problem of the present invention is a track foundation settlement assessment method based on electronic digital data processing, comprising the following steps: S1. Data collection: collect the original data related to track foundation settlement, process it, store it, and manually annotate it; S2, sample synthesis: based on the collected data, data enhancement, data synthesis, sample diversity optimization operations are performed, and then the sample interpolation method based on boundary conditions is used to synthesize the electrical digital data samples; S3. Learning modeling: Establish a track foundation settlement assessment prediction model, pre-process the collected data and synthesized data, extract key features from the pre-processed data through feature engineering, and then train the track foundation settlement assessment prediction model based on the extracted key features to obtain the assessment prediction results, and then evaluate and improve the trained model; Among them, the track foundation settlement assessment prediction model adopts a fractional-order neural network based on path integral: S4. Recommendations and warnings: Based on the assessment and prediction results of the track foundation settlement assessment and prediction model, suggestions for settlement control and repair are provided to users, and whether to issue a warning is determined based on the assessment and prediction results.

[0008] S1 is as follows: Automatic data collection is performed through sensor equipment. The sensors include settlement detectors, accelerometers and seismic wave detectors. The sensor equipment is arranged at key positions of the track foundation to collect track settlement status information in real time. The collected track foundation settlement related data include ground settlement data, vibration data and surrounding environment change data. The collected data is digitally processed when passing through various sensor equipment and transmitted in the form of discrete and accurate electrical digital data. The electrical digital data is in a digital format suitable for computer processing; Then, the data collected by various sensor devices are processed in time series and formatted, and noise and interference in the data are removed; The processed data is transmitted to the storage system via wired or wireless communication, and each piece of electrical digital data input into the storage system is manually labeled, and a unique identifier is assigned to each piece of electrical digital data. The identifier is the level of settlement status of the track foundation and the data abnormality label. The identifier is the level of settlement status of the track foundation, which is normal settlement, slight settlement and severe settlement.

[0009] S2 is as follows: Based on the collected data, additional training data samples are synthesized. Sample synthesis includes the following operations: Data enhancement: Transform the collected data, including rotation, scaling, and translation; Synthetic data: Based on the characteristics of the collected data, the data augmentation method is used to synthesize electrical digital data samples similar to the collected data; Sample diversity optimization operation: by analyzing the collected data, the distribution of synthesized electrical digital data samples is adjusted; A sample interpolation method based on boundary conditions is used to synthesize electrical digital data samples: a minority class of electrical digital data samples is used as seed samples, and sample selection is optimized by minimizing the energy function. At the same time, boundary condition analysis is performed on the seed samples, and the distance and direction between the seed samples and the nearest neighbor samples are calculated to dynamically identify boundary conditions.

[0010] In a specific implementation manner, the operation of synthesizing the electrical digital data sample by the sample interpolation method based on the boundary condition is as follows: (1) Select a few quasi-electric digital data samples from the collected data as seed samples, and minimize the energy function to obtain an energy function for evaluating the quality of quasi-electric digital data selection; (2) By calculating the distance and direction between the seed sample and the nearest neighbor sample, the dynamic boundary conditions are identified, and the boundary condition analysis is performed on each seed sample to determine the local structural characteristics of the distribution of the electrical digital data sample; (3) Dynamically adjust the number and direction of the interpolation path and the synthesized electrical digital data samples through interpolation methods and dynamic adjustment vectors based on boundary conditions; (4) Modifying the synthesis strategy according to the influence of the newly synthesized point digital data samples on the overall distribution of the electrical digital data samples, and calculating the dynamic adjustment factor during the synthesis process of the electrical digital data samples; The step of synthesizing the electrical digital data samples by the sample interpolation method based on the boundary condition is repeated multiple times until the number of synthesized new sample data reaches a preset number threshold.

[0011] S3 is as follows: Construct a track foundation settlement assessment prediction model, input the collected data and synthesized data into the track foundation settlement assessment prediction model for training. The specific operations are as follows: Data preprocessing: Read the electrical digital data in the storage system and the electrical digital data samples synthesized by the samples, and perform preprocessing operations on the data. The preprocessing operations include supplementing missing values, removing outliers, and standardizing data; Extract key features: Extract key features of the track foundation settlement process through feature engineering, including foundation compaction, soil moisture, and temperature; Training of the track foundation settlement assessment prediction model: The model is trained through a fractional-order neural network based on path integrals. Specifically, multiple possible decision paths are simulated through path integrals, and a global search of the fractional-order neural network parameters is performed. Random perturbations are performed during the training process using the principle of Brownian motion, and finally the assessment and prediction results of the track foundation settlement are obtained. Evaluation and improvement of the track foundation settlement assessment prediction model: Compare the training results with the actual results, where the actual results are the manually labeled identifiers, evaluate the accuracy of the track foundation settlement assessment prediction model, and adjust and improve the track foundation settlement assessment prediction model based on the evaluation results.

[0012] In a specific implementation manner, the training of the track foundation settlement assessment prediction model is as follows: (1) The weights and biases of the fractional neural network are initialized by using non-integer derivatives in fractional calculus. The calculation formula is as follows: , , , In the formula, represents the first The key features of the training data are represents the first The key features of the training data are Represents the total number of training data input to the fractional-order neural network in the current batch, represents the regularization parameter of the fractional-order neural network, represents the fractional-order neural network Layer weights, represents the fractional-order neural network Layer bias, fractional order neural network layer, , The symbol for a particular distribution is represented by represents the standard normal distribution, represents the fractional-order control parameter, The standard deviation for initializing the fractional neural network; (2) The integral path global search mechanism is used to optimize the fractional-order neural network in the global parameter space. Through dynamic path weighting, the contribution of different paths is adjusted according to the similarity between different input training data and the decision boundary. The calculation formula of the integral path global search mechanism is as follows: , , In the formula, represents the first The key features of the training data are represents the calculation result of the global search mechanism of the integral path, represents the mean of the training data input to the fractional neural network, represents the standard deviation of the training data input to the fractional neural network, represents the first The path weight coefficient corresponding to the training data; (3) The weights and biases of each layer are adjusted through the fractional gradient update strategy, and then the learning rate is adjusted to adapt to the global and global learning needs. The calculation formula for weight update is as follows: , , In the formula, Represents the fractional order neural network The weight update amount of the layer, Represents the fractional order neural network The gradient of the layer's weights, represents the fractional-order control parameter, represents the adjustment coefficient of gradient balance, represents the learning rate of the fractional-order neural network; (4) The global search capability and training stability are improved by using a random perturbation method based on Brownian motion. The calculation formula is as follows: , , , in, Indicates a parameter update operation. represents the updated fractional-order neural network The weight update amount of the layer, represents the Brownian motion disturbance intensity, Represents the fractional order neural network The random perturbation of the layer, represents the random perturbation coefficient, The intensity of particle motion that represents the Brownian motion principle; (5) Through sampling and updating of the integral path global search mechanism, the decision boundary is adjusted using the historical path. The calculation formula is as follows: , , In the formula, Indicates The path integral result of the iteration update is: Indicates The path update intensity control parameter for the iteration, Indicates The path integral result of the iteration is: represents the fractional-order neural network The standard deviation of the paths, The entropy value representing the complexity of the path, represents the regularization parameter, represents the weighted adjustment parameter of the first path, represents the weighted adjustment parameter of the second path; (6) The multi-path decision brought by the global search mechanism of the integrated path and the weight update brought by the fractional gradient update strategy are used to calculate the classification category. The calculation formula is as follows: , In the formula, Represents the classification category result of the fractional-order neural network, that is, the evaluation prediction result, Indicates the category index corresponding to the maximum value. , the classification categories include normal settlement, slight settlement and severe settlement, Indicates normal settlement. Indicates slight subsidence. Indicates severe subsidence. Indicates The weight of each category, Indicates The bias of the categories, is the fractional order neural network The features of the training samples, represents the path weight coefficient; (7) Repeat the above steps (1) to (6) and set a stop iteration condition for the track foundation settlement assessment prediction model until the model training is completed when the stop iteration condition is met.

[0013] The effects provided in the content of the invention are only the effects of the embodiments, rather than all the effects of the invention. The above technical solution has the following advantages or beneficial effects: (1) The present invention adopts a fractional-order neural network modeling based on path integrals, utilizes the principle of fractional-order calculus, and initializes and updates the gradient of the neural network parameters through non-integer derivatives. This solves the problem of insufficient diversity in parameter initialization of traditional networks when processing data with large changes and noise interference; In addition, the present invention adopts a path integral global search mechanism to simulate multiple possible decision paths, effectively broadening the global search capability of the model for high-dimensional, nonlinear electrical digital data, avoiding falling into local optimality, thereby greatly improving the accuracy and stability of classification decisions; Moreover, the present invention combines the principle of Brownian motion and adds random perturbations in the gradient update process, which helps the model to jump out of the local minimum, accelerates convergence, and ensures that the overall training process is more robust.

[0014] (2) The present invention adopts a sample synthesis technology based on boundary conditions. To solve the problem of minority samples and uneven data distribution in electrical digital data, the present invention selects representative seed samples by minimizing the energy function and dynamically adjusts the interpolation direction and amplitude based on boundary condition analysis (calculating the nearest neighbor distance and direction), so that the synthesized new samples are closer to the physical laws of actual track settlement. In addition, the present invention uses a dynamic adjustment factor to control the sample synthesis strategy to avoid excessive density or sparse data in some areas, thereby improving the balance of the overall samples and further enhancing the robustness and generalization ability of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0016] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0017] Figure 2 Comparison of the accuracy of different methods in track settlement assessment.

[0018] Figure 3 The performance indicators of different methods are compared.

[0019] Figure 4 The error distribution comparison diagram of different models.

[0020] Figure 5 Comparison chart of interpolation directions using different methods.

[0021] Figure 6 It is a comparison diagram of feature space distribution density. DETAILED DESCRIPTION

[0022] In order to clearly illustrate the technical features of the present invention, the present invention is described in detail below through specific implementation methods and in conjunction with the accompanying drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the present invention. In order to simplify the disclosure of the present invention, the components and settings of specific examples are described below.

[0023] Example 1 A method for evaluating track foundation settlement based on electronic digital data processing comprises the following steps: S1. Data collection: collect the original data related to track foundation settlement, process it, store it, and manually annotate it; S2, sample synthesis: based on the collected data, data enhancement, data synthesis, sample diversity optimization operations are performed, and then the sample interpolation method based on boundary conditions is used to synthesize the electrical digital data samples; S3. Learning modeling: Establish a track foundation settlement assessment prediction model, pre-process the collected data and synthesized data, extract key features from the pre-processed data through feature engineering, and then train the track foundation settlement assessment prediction model based on the extracted key features to obtain the assessment prediction results, and then evaluate and improve the trained model; Among them, the track foundation settlement assessment prediction model adopts a fractional-order neural network based on path integral: S4. Recommendations and warnings: Based on the assessment and prediction results of the track foundation settlement assessment and prediction model, suggestions for settlement control and repair are provided to users, and whether to issue a warning is determined based on the assessment and prediction results.

[0024] S1 is as follows: Automatic data collection is performed through sensor equipment. The sensors include settlement detectors, accelerometers and seismic wave detectors. The sensor equipment is arranged at key positions of the track foundation to collect track settlement status information in real time. The collected track foundation settlement related data include ground settlement data, vibration data and surrounding environment change data. The collected data is digitally processed when passing through various sensor equipment and transmitted in the form of discrete and precise electrical digital data. The electrical digital data is in a digital format suitable for computer processing, such as binary form, decimal form, etc.; Then, the data collected by various sensor devices are processed in time series and formatted, and noise and interference in the data are removed; The processed data is transmitted to the storage system via wired or wireless communication, and each piece of electrical digital data input into the storage system is manually labeled, and a unique identifier is assigned to each piece of electrical digital data. The identifier is the level of the settlement state of the track foundation and the data abnormality label. The identifier is the level of the settlement state of the track foundation, which is normal settlement, slight settlement and severe settlement, respectively, to ensure the uniqueness and traceability of each piece of electrical digital data. The format of the electrical digital data input into the storage system is CSV, JSON, XML, etc.; The format of data storage includes but is not limited to: timestamp, sensor ID, sampling value and other key information; For example, the collected data on track foundation settlement are as follows: Ra is the timestamp, which indicates the exact time of data collection (usually a time value accurate to seconds or milliseconds); Da is the sensor ID, which uniquely identifies the sensor from which the data comes; Pa is the sensor installation location, which describes the sensor's geographical location or coordinates (such as coordinate system location, track segment location, etc.); Va is the sampling value, which indicates the value of the measurement result. For example, the displacement sensor may collect the position change value, and the acceleration sensor collects the acceleration value. Ta is the ambient temperature, which describes the temperature of the data collection environment and affects the accuracy of the sensor; Ha is the ambient humidity, which describes the humidity of the data collection environment; Sa is the voltage signal strength, the voltage amplitude at the sensor input, which affects the measurement accuracy of the sensor; Fa is the measurement error, which indicates the error caused by sensor accuracy or environmental interference; Ca is the calibration status, indicating whether the sensor is in the correct calibration state (e.g., the calibration error is less than a certain threshold); Ua is the noise factor, which describes the impact of possible noise in the signal on the measurement result.

[0025] It should be noted that this embodiment is only used to illustrate one data format and type of the present invention. In actual applications, the attributes of data are usually more than 10 attributes, and the number of attributes of data may reach dozens or even hundreds.

[0026] S2 is as follows: Based on the collected data, additional training data samples are synthesized. Sample synthesis includes the following operations: Data enhancement: Transform the collected data, including rotation, scaling, and translation, to synthesize more variant data and improve the adaptability of the track foundation settlement assessment prediction model to different scenarios; Synthetic data: Based on the characteristics of the collected data, the data augmentation method is used to synthesize electrical digital data samples similar to the collected data, thereby enriching the data samples used for training; Sample diversity optimization operation: By analyzing the collected data and adjusting the distribution of the synthesized electrical digital data samples, the track foundation settlement assessment prediction model can effectively handle different categories and edge cases in the data, further improving the accuracy and reliability of the prediction; A sample interpolation method based on boundary conditions is used to synthesize electrical digital data samples: a minority class of electrical digital data samples is used as seed samples, and the sample selection is optimized by minimizing the energy function to make the selected samples more representative. At the same time, the boundary conditions of the seed samples are analyzed, and the distance and direction between them and the nearest neighbor samples are calculated. The boundary conditions are dynamically identified, and the interpolation direction and amplitude of subsequent electrical digital data sample synthesis are accurately controlled, making the synthesized electrical digital data samples more representative and diverse, thereby improving the quality of electrical digital data and the generalization ability of the model.

[0027] In a specific implementation manner, the operation of synthesizing the electrical digital data sample by the sample interpolation method based on the boundary condition is as follows: (1) According to the type of manual annotation, a few electrical-like digital data samples are selected from the collected data as seed samples, and the energy function used to evaluate the quality of electrical-like digital data selection is obtained by minimizing the energy function. The calculation formula is as follows: , In the formula, represents the energy function used to evaluate the quality of the selection of electrical-like digital data, represents the selected minority class electronic digital data sample set, Represents the total number of digital data samples, Indicates The feature vector of the digital data sample, represents the mean vector of the characteristics of the digital data samples, Indicates the distribution density adjustment weight, set , represents the L2 norm, represents the boundary complexity adjustment weight, setting , Indicates The average distance from the feature vector of a digital data sample to its nearest neighbor sample set, represents the nearest neighbor sample set; Among them, the energy function comprehensively considers the distribution density and boundary complexity of the electrical digital data samples in the feature space, so that the selected minority electrical digital data samples can represent the key areas, thereby improving the diversity of the electrical digital data samples synthesized by subsequent expansion; for example, the signal from the ground displacement sensor may be less in the small deformation area of ​​the track foundation, while the strain sensor and accelerometer have more sampling data in the severe settlement area. The traditional method easily leads to the neglect of these low-density areas during data expansion. The present method minimizes the energy function to ensure that the selected seed samples cover different states of the track foundation (such as micro-settlement, stability, and severe settlement), thereby improving the effectiveness of data expansion and ensuring that a more uniform sample synthesis effect can be obtained even when the electrical digital data is unevenly distributed, thereby improving the robustness of the classification model; (2) By calculating the distance and direction between the seed sample and the nearest neighbor sample, the dynamic boundary conditions are identified, and the boundary condition analysis is performed on each seed sample to determine the local structural characteristics of the distribution of the electrical digital data sample. The calculation formula is as follows: , , In the formula, represents the dynamic adjustment vector based on boundary conditions, Indicates The feature vector of the digital data sample, Indicates The nearest neighbor sample set of a digital data sample, Indicates The digital data samples and The weight between the digital data samples is used to measure the The importance between a digital data sample and its neighboring digital data samples, Represents the attenuation parameter, set , Indicates The feature vector of the digital data sample, , , Indicates the total number of digital data samples Three different indexes; By analyzing the boundary conditions of each seed sample, it is convenient to accurately control the interpolation direction and interpolation amplitude in the subsequent synthesis of electrical digital data samples; for example, the displacement signal of the ground displacement sensor may change smoothly, while the signals of the strain sensor and the accelerometer may show mutations or periodic oscillations. Traditional data expansion methods often use simple linear interpolation, which easily causes the distribution of synthetic data to deviate from the actual track state. The present invention can capture the changing trend of track settlement data in different areas by calculating the dynamic boundary adjustment vector of the seed sample. In the stable area of ​​the track foundation, the displacement signal is relatively gentle and the interpolation amplitude should be small, while in the violent settlement area, the interpolation amplitude should be larger to ensure that the synthesized data can reflect the real characteristics of the track state, avoid the problem that the samples synthesized by the traditional interpolation method do not conform to the actual track settlement mode, thereby improving the rationality of data expansion; (3) The number and direction of the interpolation path and the synthesized electrical digital data samples are dynamically adjusted through the interpolation method and the dynamic adjustment vector based on the boundary conditions. The calculation formula is as follows: , , In the formula, represents the synthesized new electrical digital data sample, represents the interpolation coefficient, Represents a random number between 0 and 1. represents the dynamic adjustment factor; Dynamic adjustment can improve the limitations of traditional linear interpolation methods and adaptively expand the distribution of local electrical digital data. For example, in track settlement monitoring, the data distribution of accelerometers is usually highly discrete, and simple interpolation may cause the synthesized samples to deviate from the actual track state. This method dynamically adjusts the interpolation direction and amplitude to ensure that the expanded data can more realistically reflect the changes in track settlement. If a data point collected by a ground displacement sensor shows that the track foundation has undergone a large deformation in a certain area, and the change trend of the nearest neighbor sample is relatively slow, then the newly synthesized interpolation sample will be interpolated along the main direction of the track deformation according to the distribution characteristics of its neighboring samples, and will not deviate from the actual track state, ensuring the rationality and diversity of data expansion. (4) According to the influence of the newly synthesized point digital data samples on the overall distribution of the electrical digital data samples, the synthesis strategy is modified, and the dynamic adjustment factor is calculated during the synthesis process of the electrical digital data samples. The calculation is shown as follows: , In the formula, represents the dynamic adjustment factor, Indicates the adjustment rate parameter, Represents the number of dimensions of the electrical digital data sample, Indicates the current synthesized digital data sample in The distribution probability in the dimension, Indicates the target digital data sample at The distribution probability in the dimension, Indicates the operation of taking the minimum value; By calculating the dynamic adjustment factor, it is possible to avoid excessive or insufficient expansion of certain areas and maintain a balanced distribution of electrical digital data. A key problem with traditional data expansion methods is that they may cause the data density in some areas to be too high and the data in other areas to be too little, thus affecting the learning effect of the model. For example, there are fewer track foundation settlement anomalies in a certain area of ​​the electrical digital data, and the traditional data expansion method simply copies the existing samples. Then, the model may be affected by the uneven data distribution during the training process, resulting in an increase in the misclassification rate. This method dynamically adjusts the interpolation strategy by calculating the deviation between the newly synthesized samples and the target distribution, so that the new samples can be more evenly distributed in the entire feature space. If the settlement data of some track areas is over-concentrated during the sample synthesis process, the data distribution is made more balanced by reducing the interpolation amplitude of the area or increasing the interpolation ratio of other areas, thereby improving the generalization ability of the model. Finally, the step of synthesizing the electrical digital data samples using the sample interpolation method based on the boundary conditions is repeated multiple times until the number of synthesized new sample data reaches a preset number threshold.

[0028] S3 is as follows: Construct a track foundation settlement assessment prediction model, input the collected data and synthesized data into the track foundation settlement assessment prediction model for training. The specific operations are as follows: Data preprocessing: Read the electrical digital data in the storage system and the electrical digital data samples synthesized by the samples, and perform preprocessing operations on the data. The preprocessing operations include supplementing missing values, removing outliers, and standardizing data; Extract key features: Extract key features of the track foundation settlement process through feature engineering, including foundation compaction, soil moisture, and temperature; Training of the track foundation settlement assessment prediction model: The model is trained through a fractional-order neural network based on path integrals. Specifically, multiple possible decision paths are simulated through path integrals, and a global search of the fractional-order neural network parameters is performed. Random perturbations are performed during the training process using the principle of Brownian motion, and finally the assessment and prediction results of the track foundation settlement are obtained. Evaluation and improvement of the track foundation settlement assessment prediction model: Compare the training results with the actual results, where the actual results are the manually labeled identifiers, evaluate the accuracy of the track foundation settlement assessment prediction model, and adjust and improve the track foundation settlement assessment prediction model based on the evaluation results.

[0029] In a specific implementation manner, the training of the track foundation settlement assessment prediction model is as follows: (1) Initializing the weights and biases of the fractional-order neural network by using non-integer derivatives in fractional calculus can solve the problem of insufficient diversity in parameter initialization of traditional fractional-order neural networks, especially when there are large changes and noise in the data (such as sudden changes in track settlement). It can better capture the complexity of track settlement data. The calculation formula is as follows: , , , In the formula, represents the first The key features of the training data are represents the first The key features of the training data are Represents the total number of training data input to the fractional-order neural network in the current batch, represents the regularization parameter of the fractional-order neural network, set , represents the fractional-order neural network Layer weights, represents the fractional-order neural network Layer bias, fractional order neural network layer, , The symbol for a particular distribution is represented by represents the standard normal distribution, represents the fractional-order control parameter, The standard deviation of the initialization of the fractional-order neural network is set to ; Among them, the key features include several data, such as sensor installation position Pa, sampling value Va, ambient temperature Ta, ambient humidity Ha, voltage signal strength Sa, measurement error Fa, calibration status Ca, noise coefficient Ua, etc.; (2) The integral path global search mechanism is used to optimize the fractional-order neural network in the global parameter space. Through dynamic path weighting, the contribution of different paths is adjusted according to the similarity between different input training data and the decision boundary. The calculation formula of the integral path global search mechanism is as follows: , , In the formula, represents the first The key features of the training data are represents the calculation result of the global search mechanism of the integral path, represents the mean of the training data input to the fractional neural network, represents the standard deviation of the training data input to the fractional neural network, represents the first The path weight coefficient corresponding to the training data; The integral path global search mechanism is adopted to enhance the global search capability of nonlinear and high-dimensional electrical digital data through the simulation of multiple paths, so as to solve the problem of being easily trapped in the local optimum and enable the fractional-order neural network to find the best solution in the global parameter space. The electrical digital data usually exhibits complex nonlinear relationships. For example, the track settlement process is affected by multiple factors, and the signal provided by each sensor is also high-dimensional. The integral path global search mechanism can prevent the model from falling into the local optimum and help the neural network find a more suitable solution in the global parameter space. In this case, through dynamic path weighting, the model can adjust the contribution of different paths according to the similarity between different data samples and the decision boundary, so as to more accurately reflect the real state of track settlement. (3) The weights and biases of each layer are adjusted through the fractional gradient update strategy, and then the learning rate is adjusted to adapt to the global and global learning needs. The calculation formula for weight update is as follows: , , In the formula, Represents the fractional order neural network The weight update amount of the layer, Represents the fractional order neural network The gradient of the layer's weights, represents the fractional-order control parameter, Indicates the adjustment coefficient of gradient balance, set , Represents the learning rate of the fractional-order neural network, set =0.3; The fractional order gradient update strategy is adopted to adjust the weight and bias of each layer through fractional order differentials, so as to avoid the problem that the learning rate of the fractional order neural network is difficult to take into account the global and local precision when processing complex nonlinear electrical digital data. The convergence speed can be flexibly adjusted under diverse distributions. In electrical digital data, different data characteristics of sensors (such as acceleration changes or irregular displacement fluctuations) may require different learning rates to adapt to global and local learning needs. The fractional order gradient update can flexibly adjust the learning rate, avoid the gradient vanishing or exploding problems that may occur in traditional methods, and accelerate the convergence speed of the model. (4) The global search capability and training stability are improved by using a random perturbation method based on Brownian motion. The calculation formula is as follows: , , , in, Indicates a parameter update operation. represents the updated fractional-order neural network The weight update amount of the layer, Represents the Brownian motion perturbation intensity, set , Represents the fractional order neural network The random perturbation of the layer, represents the random perturbation coefficient, set , The intensity of particle motion that represents the Brownian motion principle, ; A random perturbation method based on Brownian motion is adopted. The Brownian motion refers to the never-ending irregular motion of particles suspended in liquid or gas. According to this principle, the weights are randomly adjusted after each gradient update to solve the problems of falling into local minima and gradient instability, and to improve the global search capability and training stability. In the task of track sedimentation assessment, electrical digital data may be affected by noise, especially in the real-time acquisition process, the accuracy of sensor data may be disturbed by the external environment. The Brownian motion method can help the model escape the local minimum through random perturbations, improve its global search capability, and enable the model to be trained more stably when facing complex and irregular sedimentation patterns. (5) Through sampling and updating of the integral path global search mechanism, the decision boundary is adjusted using the historical path. The calculation formula is as follows: , , In the formula, Indicates The path integral result of the iteration update is: Indicates The path update intensity control parameter for the iteration, Indicates The path integral result of the iteration is: represents the fractional-order neural network The standard deviation of the paths, The entropy value representing the complexity of the path, represents the regularization parameter, Represents the weighted adjustment parameter of the first path, set , Represents the weighted adjustment parameter of the second path, set ; By sampling and updating the path integral, the decision boundary is adjusted using the historical path to solve the balance problem between global optimization and local adaptability in classification decision-making, continuously correct the classification boundary, and improve the overall robustness. For different types of track settlement, the model can perform multi-path adaptive adjustment based on the input data of the sensor, not only considering the global information, but also dynamically adjusting the local classification boundary, making the evaluation result more accurate. (6) The multi-path decision brought by the global search mechanism of the integrated path and the weight update brought by the fractional gradient update strategy are used to calculate the classification category. The calculation formula is as follows: , In the formula, Represents the classification category result of the fractional-order neural network, that is, the evaluation prediction result, Indicates the category index corresponding to the maximum value. , the classification categories include normal settlement, slight settlement and severe settlement, Indicates normal settlement. Indicates slight subsidence. Indicates severe subsidence. Indicates The weight of each category, Indicates The bias of the categories, is the fractional order neural network The features of the training samples, Represents the path weight coefficient, set ; The multi-path decision-making brought by the integrated path integral and the flexible weight update brought by the fractional calculus are used to calculate the classification categories. In the track foundation settlement assessment task, the electrical digital data from different locations and different sensors may present different distributions. Traditional classifiers may find it difficult to handle these complex feature distributions. The multi-path decision-making mechanism of the path integral enables the model to dynamically adjust the classification decision boundary, thus improving the overall classification accuracy and robustness when facing complex or extreme settlement situations. (7) Repeat the above steps (1) to (6) and set a stop iteration condition for the track foundation settlement assessment prediction model until the model training is completed when the stop iteration condition is met; The preset condition for stopping iteration is reaching a preset maximum number of iterations, and the preset maximum number of iterations in the present invention is set to 1000 times.

[0030] Example 2 like Figure 2 As shown in the figure, in order to verify the effect of the technical solution of the present invention, the accuracy performance of different methods in the track settlement assessment task is evaluated. Accuracy is the core indicator for measuring the performance of the classification model, which directly reflects the model's ability to recognize the track settlement state. During the experiment, the accuracy of different methods in multiple test rounds was sampled and a line graph was drawn. The experimental results showed that the accuracy of the fractional-order neural network based on path integral was always stable between 92% and 95%, while the accuracy of the traditional neural network, support vector machine and decision tree was between 85% and 88%, 82% and 86% and 80% and 85%, respectively, and showed large fluctuations, indicating that the present technology is not only significantly superior to the traditional method in accuracy, but also has stronger stability and can more reliably classify the track settlement state. The reason is that the path integral method can simulate a variety of possible decision paths, thereby avoiding the local optimal problem caused by the nonlinear characteristics of the data, making the classification decision more comprehensive and accurate.

[0031] Example 3 like Figure 3 As shown, in order to further verify the overall performance of the model, the relationship between accuracy and error is used to measure the advantages and disadvantages of different methods. The experiment uses a bar graph to display the accuracy and error distribution of different methods, which is convenient for intuitive comparison. The experimental results show that the method of the present invention has the highest accuracy (0.95) and the lowest error (0.05); in comparison, the accuracy of the traditional neural network is 0.88, and the error is 0.12, the accuracy of the support vector machine is 0.85, and the error is 0.15. The decision tree has the lowest accuracy of 0.83 and the largest error of 0.17. It further shows that the method of the present invention can not only more accurately evaluate the track settlement state, but also effectively control the classification error and more accurately adapt to different modes of track settlement.

[0032] Example 4 like Figure 4 As shown, in order to further analyze the distribution of model errors and verify the stability of different methods when facing real data, traditional error analysis methods may only focus on the average error, but in the track settlement assessment task, the error fluctuation range is equally important. The KDE kernel density estimation diagram is used to visualize the error distribution of different methods and observe the concentration of errors. The experimental results show that the error distribution of the fractional-order neural network based on the integral path used in the present invention is the most concentrated, and most of the error values ​​are concentrated around 0.05, while the error distribution of other methods is obviously more dispersed. The error mean of the traditional neural network is about 0.12, the error mean of the support vector machine is 0.15, and the error distribution of the decision tree is the most discrete, with an error mean of 0.17. This proves that the method of the present invention can provide a more stable prediction effect when facing complex, high-dimensional electrical digital data, and avoid the problem of excessive classification uncertainty.

[0033] Example 5 like Figure 5 As shown, in order to verify that the effect of the dynamic adjustment interpolation method of the present invention is better than the traditional linear interpolation method, experiments are carried out. From the experimental results, it can be seen that the interpolation direction of the traditional method diverges randomly, while the present invention makes the interpolation direction (blue arrow) consistent with the main direction of the actual track deformation through boundary condition analysis, ensuring that the synthetic sample conforms to the actual physical laws. Therefore, the dynamic adjustment interpolation method is better.

[0034] Example 6 like Figure 6 As shown, the present invention also conducts a comparative experiment on the distribution density of different methods in the feature space. The experimental results show that the synthetic samples of the traditional method are concentrated in the middle of the feature space, while the present invention successfully expands the sample coverage of the low-density area (on both sides) while maintaining the main distribution, verifying the effectiveness of the boundary condition analysis.

[0035] Although the above describes the specific implementation mode of the invention in conjunction with the drawings, it is not intended to limit the scope of protection of the invention. Based on the technical solution of the present invention, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present invention.

Claims

1. A method for evaluating track foundation settlement based on electronic digital data processing, characterized in that: S1. Data collection: collect the original data related to track foundation settlement, process it, store it, and manually annotate it; S2, sample synthesis: based on the collected data, data enhancement, data synthesis, sample diversity optimization operations are performed, and then the sample interpolation method based on boundary conditions is used to synthesize the electrical digital data samples; S3. Learning modeling: Establish a track foundation settlement assessment prediction model, pre-process the collected data and synthesized data, extract key features from the pre-processed data through feature engineering, and then train the track foundation settlement assessment prediction model based on the extracted key features to obtain the assessment prediction results, and then evaluate and improve the trained model; Among them, the track foundation settlement assessment prediction model adopts a fractional-order neural network based on path integral; S4. Recommendations and warnings: Based on the assessment and prediction results of the track foundation settlement assessment and prediction model, suggestions for settlement control and repair are provided to users, and whether to issue a warning is determined based on the assessment and prediction results.

2. A track foundation settlement assessment method based on electronic digital data processing according to claim 1, characterized in that: S1 is as follows: Automatic data collection is performed through sensor equipment. The sensors include settlement detectors, accelerometers and seismic wave detectors. The sensor equipment is arranged at key positions of the track foundation to collect track settlement status information in real time. The collected track foundation settlement related data include ground settlement data, vibration data and surrounding environment change data. The collected data is digitally processed when passing through various sensor equipment and transmitted in the form of discrete and accurate electrical digital data. The electrical digital data is in a digital format suitable for computer processing; Then, the data collected by various sensor devices are processed in time series and formatted, and noise and interference in the data are removed; The processed data is transmitted to the storage system via wired or wireless communication, and each piece of electrical digital data input into the storage system is manually labeled, and a unique identifier is assigned to each piece of electrical digital data. The identifier is the level of settlement status of the track foundation and the data abnormality label. The identifier is the level of settlement status of the track foundation, which is normal settlement, slight settlement and severe settlement.

3. A track foundation settlement assessment method based on electronic digital data processing according to claim 2, characterized in that: S2 is as follows: Based on the collected data, additional training data samples are synthesized. Sample synthesis includes the following operations: Data enhancement: Transform the collected data, including rotation, scaling, and translation; Synthetic data: Based on the characteristics of the collected data, the data augmentation method is used to synthesize electrical digital data samples similar to the collected data; Sample diversity optimization operation: by analyzing the collected data, the distribution of synthesized electrical digital data samples is adjusted; A sample interpolation method based on boundary conditions is used to synthesize electrical digital data samples: a minority class of electrical digital data samples is used as seed samples, and sample selection is optimized by minimizing the energy function. At the same time, boundary condition analysis is performed on the seed samples, and the distance and direction between the seed samples and the nearest neighbor samples are calculated to dynamically identify boundary conditions.

4. A track foundation settlement assessment method based on electronic digital data processing according to claim 3, characterized in that: The operation of synthesizing electrical digital data samples by the sample interpolation method based on boundary conditions is as follows: (1) Select a few quasi-electric digital data samples from the collected data as seed samples, and minimize the energy function to obtain an energy function for evaluating the quality of quasi-electric digital data selection; (2) By calculating the distance and direction between the seed sample and the nearest neighbor sample, the dynamic boundary conditions are identified, and the boundary condition analysis is performed on each seed sample to determine the local structural characteristics of the distribution of the electrical digital data sample; (3) Dynamically adjust the number and direction of the interpolation path and the synthesized electrical digital data samples through interpolation methods and dynamic adjustment vectors based on boundary conditions; (4) Modifying the synthesis strategy according to the influence of the newly synthesized point digital data samples on the overall distribution of the electrical digital data samples, and calculating the dynamic adjustment factor during the synthesis process of the electrical digital data samples; The step of synthesizing the electrical digital data samples by the sample interpolation method based on the boundary condition is repeated multiple times until the number of synthesized new sample data reaches a preset number threshold.

5. A track foundation settlement assessment method based on electronic digital data processing according to claim 6, characterized in that: S3 is as follows: Construct a track foundation settlement assessment prediction model, input the collected data and synthesized data into the track foundation settlement assessment prediction model for training. The specific operations are as follows: Data preprocessing: Read the electrical digital data in the storage system and the electrical digital data samples synthesized by the samples, and perform preprocessing operations on the data. The preprocessing operations include supplementing missing values, removing outliers, and standardizing data; Extract key features: Extract key features of the track foundation settlement process through feature engineering, including foundation compaction, soil moisture, and temperature; Training of the track foundation settlement assessment prediction model: The model is trained through a fractional-order neural network based on path integrals. Specifically, multiple possible decision paths are simulated through path integrals, and a global search of the fractional-order neural network parameters is performed. Random perturbations are performed during the training process using the principle of Brownian motion, and finally the assessment and prediction results of the track foundation settlement are obtained. Evaluation and improvement of the track foundation settlement assessment prediction model: Compare the training results with the actual results, where the actual results are the manually labeled identifiers, evaluate the accuracy of the track foundation settlement assessment prediction model, and adjust and improve the track foundation settlement assessment prediction model based on the evaluation results.

6. A method for evaluating track foundation settlement based on electronic digital data processing according to claim 5, characterized in that: The training of the track foundation settlement assessment prediction model is as follows: (1) The weights and biases of the fractional neural network are initialized by using non-integer derivatives in fractional calculus. The calculation formula is as follows: ， ， ， In the formula, represents the first The key features of the training data are represents the first The key features of the training data are Represents the total number of training data input to the fractional-order neural network in the current batch, represents the regularization parameter of the fractional-order neural network, represents the fractional-order neural network Layer weights, represents the fractional-order neural network Layer bias, fractional order neural network layer, , The symbol for a particular distribution is represented by represents the standard normal distribution, represents the fractional-order control parameter, The standard deviation for initializing the fractional neural network; (2) The integral path global search mechanism is used to optimize the fractional-order neural network in the global parameter space. Through dynamic path weighting, the contribution of different paths is adjusted according to the similarity between different input training data and the decision boundary. The calculation formula of the integral path global search mechanism is as follows: ， ， In the formula, represents the first The key features of the training data are represents the calculation result of the global search mechanism of the integral path, represents the mean of the training data input to the fractional neural network, represents the standard deviation of the training data input to the fractional neural network, represents the first The path weight coefficient corresponding to the training data; (3) The weights and biases of each layer are adjusted through the fractional gradient update strategy, and then the learning rate is adjusted to adapt to the global and global learning needs. The calculation formula for weight update is as follows: ， ， In the formula, Represents the fractional order neural network The weight update amount of the layer, Represents the fractional order neural network The gradient of the layer's weights, represents the fractional-order control parameter, represents the adjustment coefficient of gradient balance, Represents the learning rate of the fractional-order neural network; (4) The global search capability and training stability are improved by using a random perturbation method based on Brownian motion. The calculation formula is as follows: ， ， ， in, Indicates a parameter update operation. represents the updated fractional-order neural network The weight update amount of the layer, represents the Brownian motion disturbance intensity, Represents the fractional order neural network The random perturbation of the layer, represents the random perturbation coefficient, The intensity of particle motion that represents the Brownian motion principle; (5) Through sampling and updating of the integral path global search mechanism, the decision boundary is adjusted using the historical path. The calculation formula is as follows: ， ， In the formula, Indicates The path integral result of the iteration update is: Indicates The path update intensity control parameter for the iteration, Indicates The path integral result of the iteration is: represents the fractional-order neural network The standard deviation of the paths, The entropy value representing the complexity of the path, represents the regularization parameter, represents the weighted adjustment parameter of the first path, represents the weighted adjustment parameter of the second path; (6) The multi-path decision brought by the global search mechanism of the integrated path and the weight update brought by the fractional gradient update strategy are used to calculate the classification category. The calculation formula is as follows: ， In the formula, Represents the classification category result of the fractional-order neural network, that is, the evaluation prediction result, Indicates the category index corresponding to the maximum value. , the classification categories include normal settlement, slight settlement and severe settlement, Indicates normal settlement. Indicates slight subsidence. Indicates severe subsidence. Indicates The weight of each category, Indicates The bias of the categories, is the fractional order neural network The features of the training samples, represents the path weight coefficient; (7) Repeat the above steps (1) to (6) and set a stop iteration condition for the track foundation settlement assessment prediction model until the model training is completed when the stop iteration condition is met.

Images