A tunnel temperature and humidity monitoring and analysis method and system

Abstract

The present invention relates to the field of data processing, and specifically provides a tunnel temperature and humidity monitoring and analysis method and system, which obtains the x-th tunnel temperature and humidity monitoring data stream of the target monitoring tunnel; determines the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream; loads the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model corresponding to the target data stream type, and obtains the inference temperature and humidity of the target monitoring tunnel at the next observation point of the x observation points. The present invention can improve the accuracy of the temperature and humidity inference analysis of the monitoring tunnel.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for monitoring and analyzing temperature and humidity in a tunnel. Background Art

[0002] In the field of tunnel engineering, the temperature and humidity conditions in the tunnel have a vital impact on the structural safety, equipment operation and personnel comfort of the tunnel. For example, excessive humidity may cause electrical equipment in the tunnel to become damp and damaged, affecting its normal operation and even causing safety accidents; while extreme temperature changes may cause the tunnel structural materials to expand and contract, reducing the stability and durability of the structure. At present, traditional tunnel temperature and humidity monitoring and analysis methods often simply collect and record temperature and humidity data, lack in-depth analysis and effective use of data, and cannot accurately predict the changing trend of tunnel temperature and humidity, making it difficult to meet the actual operation management and safety requirements of the tunnel. Summary of the invention

[0003] In view of this, the present invention provides a tunnel temperature and humidity monitoring and analysis method and system. The technical solution of the embodiment of the present invention is implemented as follows:

[0004] On the one hand, an embodiment of the present invention provides a tunnel temperature and humidity monitoring and analysis method, including: obtaining an x-th tunnel temperature and humidity monitoring data stream of a target monitoring tunnel, wherein the target monitoring tunnel includes multiple tunnel sections, and the x-th tunnel temperature and humidity monitoring data stream includes the actual temperature and humidity of the target monitoring tunnel at the x-th group of observation points, wherein x≥1; clustering the set of example tunnel temperature and humidity monitoring data streams to obtain a data stream type library; or, when the set of example tunnel temperature and humidity monitoring data streams includes K example tunnel temperature and humidity monitoring data streams, extracting sub-data streams of equal length from each of the K example tunnel temperature and humidity monitoring data streams to obtain K sub-data streams, wherein each of the K sub-data streams includes the actual temperature and humidity of one monitoring tunnel among multiple monitoring tunnels at s observation points; clustering the K sub-data streams Cluster, obtain a data stream type library; determine the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, wherein the target data stream type is a data stream type in the data stream type library, and multiple monitoring tunnels include the target monitoring tunnel, or the target monitoring tunnel is different from the multiple monitoring tunnels; load the x-th tunnel temperature and humidity monitoring data stream into a target temperature and humidity inference model corresponding to the target data stream type, and obtain the inferred temperature and humidity of the target monitoring tunnel at the next observation point of the x-th group of observation points, wherein the target temperature and humidity inference model is a model obtained by calibrating the temperature and humidity inference model to be calibrated by a subset of example tunnel temperature and humidity monitoring data streams, and the example tunnel temperature and humidity monitoring data stream subset includes the example tunnel temperature and humidity monitoring data stream corresponding to the target data stream type in the example tunnel temperature and humidity monitoring data stream set.

[0005] On the other hand, an embodiment of the present invention provides a computer system, including a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor implements the steps in the above method when executing the program.

[0006] The beneficial effects of the present invention include at least: in an embodiment of the present invention, the tunnel temperature and humidity monitoring data stream of the monitoring tunnel is loaded into a temperature and humidity inference model corresponding to the data stream type corresponding to the tunnel temperature and humidity monitoring data stream, and the temperature and humidity inference models executed by tunnel temperature and humidity monitoring data streams of different data stream types may be different. In other words, the temperature and humidity inference model corresponding to the data stream type corresponding to the tunnel temperature and humidity monitoring data stream is used to perform inference analysis on the temperature and humidity of the monitoring tunnel, thereby preventing the use of one temperature and humidity inference model to perform inference analysis on the temperature and humidity of different monitoring tunnels, thereby increasing the accuracy of the temperature and humidity inference analysis of the monitoring tunnel.

[0007] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The accompanying drawings herein are incorporated into and constitute a part of the specification. These drawings illustrate embodiments consistent with the present invention and, together with the specification, are used to explain the technical solutions of the present invention.

[0009] Figure 1 A schematic diagram of the implementation flow of a tunnel temperature and humidity monitoring and analysis method provided in an embodiment of the present invention.

[0010] Figure 2 A hardware entity schematic diagram of a computer system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0011] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention are further elaborated in detail below in conjunction with the drawings and embodiments. The described embodiments should not be regarded as limiting the present invention. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present invention.

[0012] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. The terms "first / second / third" are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It is understood that "first / second / third" may be interchanged in a specific order or sequential order where permitted, so that the embodiments of the present invention described herein can be implemented in an order other than that illustrated or described herein.

[0013] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein are only for the purpose of describing the present invention and are not intended to limit the present invention.

[0014] The embodiment of the present invention provides a tunnel temperature and humidity monitoring and analysis method, which can be executed by a processor of a computer system, wherein the computer system can refer to a device with data processing capability, such as a server, a laptop, a tablet computer, or a desktop computer.

[0015] Figure 1 A schematic diagram of the implementation flow of a tunnel temperature and humidity monitoring and analysis method provided by an embodiment of the present invention, such as Figure 1 As shown, the method includes:

[0016] Step S100: Obtain the x-th tunnel temperature and humidity monitoring data stream of the target monitoring tunnel. The target monitoring tunnel includes multiple tunnel segments, and the x-th tunnel temperature and humidity monitoring data stream includes the true temperature and humidity of the target monitoring tunnel at the x-th group of observation points, where x ≥ 1.

[0017] The target monitoring tunnel is a specific tunnel for the computer system to monitor and analyze temperature and humidity. In practical applications, it may be an urban subway tunnel, a highway tunnel, or a railway tunnel, etc. Since a tunnel usually has a certain length, for the convenience of monitoring and management, it is divided into multiple tunnel segments. For example, a 5-kilometer-long highway tunnel can be divided into one tunnel segment every 500 meters, thus obtaining 10 tunnel segments. The temperature and humidity conditions of each tunnel segment may be affected by various factors, such as the ventilation of the tunnel, the temperature and humidity of the surrounding environment, and the heat generated by vehicle passage.

[0018] The tunnel temperature and humidity monitoring data stream is the basic data source for the computer system to analyze temperature and humidity. It records the true temperature and humidity information of the target monitoring tunnel at different observation points. This information can be obtained by real-time collection through temperature and humidity sensors installed in the tunnel. Each tunnel temperature and humidity monitoring data stream corresponds to a specific group of observation points. The observation point is the specific location in the tunnel where the temperature and humidity sensor is placed. For example, in the above 5-kilometer-long highway tunnel, an observation point is set every 50 meters, and each observation point is equipped with a temperature and humidity sensor to monitor the temperature and humidity at that location in real time.

[0019] The x-th tunnel temperature and humidity monitoring data stream represents the true temperature and humidity data of the target monitoring tunnel at the x-th group of observation points obtained by the computer system in a certain order. Here, x is an integer greater than or equal to 1, used to identify different data streams. For example, when x = 1, it means that the computer system obtains the tunnel temperature and humidity monitoring data stream at the first group of observation points; when x = 2, it means obtaining the data stream at the second group of observation points, and so on.

[0020] In practical applications, the computer system can obtain the x-th tunnel temperature and humidity monitoring data stream of the target monitoring tunnel through the following technical means. First, the temperature and humidity sensors are arranged in the target monitoring tunnel. The arrangement of the sensors should be scientifically planned according to the length, structure and monitoring requirements of the tunnel. For example, the density of sensors should be increased at key locations such as the entrance and exit, bends, and vents of the tunnel to ensure that the changes in temperature and humidity in the tunnel can be accurately monitored. The data collected by the temperature and humidity sensors are transmitted to the computer system through the data transmission network. The data transmission network can be a wired network or a wireless network. The wired network has the advantages of stable transmission and high reliability, and is suitable for tunnels with short distances and relatively stable environments; the wireless network has the advantages of easy installation and strong flexibility, and is suitable for tunnels with long distances and complex terrains. Viable wireless networks include Wi-Fi, ZigBee, LoRa, etc.

[0021] After the computer system receives the data transmitted by the temperature and humidity sensor, it preprocesses the data to ensure the accuracy and reliability of the data. Preprocessing includes steps such as data cleaning, data calibration and data normalization. Data cleaning refers to removing noise, outliers and missing values ​​in the data to improve the quality of the data. For example, a sliding average filter algorithm can be used to smooth the data and remove high-frequency noise in the data. Data calibration refers to calibrating the data collected by the sensor to ensure the accuracy of the data. The collected data can be calibrated by comparing it with a standard temperature and humidity sensor. Data normalization refers to unifying the data collected by different sensors to the same scale for subsequent analysis and processing. For example, the minimum-maximum normalization method can be used to normalize the data to the interval [0, 1].

[0022] Assume that the target monitoring tunnel is a subway tunnel with a length of 3 kilometers. It is divided into 10 tunnel sections every 300 meters. An observation point is set every 30 meters in each tunnel section, and a total of 100 observation points are set. Each observation point is equipped with a temperature and humidity sensor for real-time monitoring of the temperature and humidity at that location. The computer system obtains the temperature and humidity data at these observation points in a certain order to form a tunnel temperature and humidity monitoring data stream. When x=1, the computer system obtains the tunnel temperature and humidity monitoring data stream at the first group of observation points, which contains the real temperature and humidity information of the subway tunnel at this group of observation points.

[0023] When acquiring the x-th tunnel temperature and humidity monitoring data stream, the computer system ensures the integrity and accuracy of the data. The integrity of the data can be ensured by setting the data collection cycle and data storage mechanism. For example, the data collection cycle is set to once per minute to ensure that the changes in temperature and humidity in the tunnel can be collected in a timely manner. At the same time, the collected data is stored in the database for subsequent query and analysis.

[0024] In order to verify the accuracy of the acquired tunnel temperature and humidity monitoring data stream, a cross-validation method can be used. For example, multiple backup sensors are set up in the tunnel, and the data collected by the backup sensors are compared with the data collected by the main sensor. If the difference between the two is within the allowable range, the acquired data is considered accurate. In practical applications, data fusion technology can also be used to fuse the data collected by multiple sensors to improve the accuracy and reliability of the data. For example, the weighted average method can be used to fuse the temperature and humidity data collected by multiple sensors to obtain more accurate temperature and humidity values.

[0025] Assume that at a certain moment, at the xth group of observation points, the temperature values ​​collected by the sensor are T1, T2, …, Tn, and the humidity values ​​are H1, H2, …, Hn, where n is the number of observation points in the group. The computer system can use the following formula to calculate the average temperature and average humidity of the group of observation points:

[0026] Average temperature T avg =(T1+T2+…+Tn) / n;

[0027] Average humidity avg =(H1+H2+…+Hn) / n;

[0028] By calculating the average temperature and average humidity, we can more intuitively understand the temperature and humidity conditions of the target monitoring tunnel at this group of observation points.

[0029] The computer system obtains the x-th tunnel temperature and humidity monitoring data stream of the target monitoring tunnel, which is the basic step for tunnel temperature and humidity monitoring and analysis. By reasonably arranging temperature and humidity sensors, selecting appropriate data transmission networks, preprocessing data, and adopting technical means such as data fusion and cross-validation, the acquired data can be ensured to be accurate and complete, providing a reliable basis for subsequent temperature and humidity reasoning analysis. At the same time, by calculating statistics such as average temperature and average humidity, the temperature and humidity conditions in the tunnel can be more intuitively understood, providing strong support for the operation management and safety of the tunnel. In practical applications, the data can also be further analyzed and processed according to specific needs, such as drawing temperature and humidity change curves, and conducting temperature and humidity predictions, so as to better meet the needs of tunnel management.

[0030] Step S200: Determine the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, wherein the target data stream type is a data stream type in a data stream type library, the data stream type library includes multiple data stream types obtained by clustering at least an example tunnel temperature and humidity monitoring data stream set, the example tunnel temperature and humidity monitoring data stream set includes multiple example tunnel temperature and humidity monitoring data streams of multiple monitoring tunnels, each example tunnel temperature and humidity monitoring data stream is the actual temperature and humidity of one monitoring tunnel in multiple monitoring tunnels at a corresponding set of observation points, the multiple monitoring tunnels include a target monitoring tunnel, or the target monitoring tunnel is different from the multiple monitoring tunnels.

[0031] In step S200, the computer system determines the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, wherein the target data stream type is a data stream type in a data stream type library, and the data stream type library is a plurality of data stream types obtained by clustering at least a set of example tunnel temperature and humidity monitoring data streams, and the set of example tunnel temperature and humidity monitoring data streams includes a plurality of example tunnel temperature and humidity monitoring data streams of a plurality of monitoring tunnels, and each example tunnel temperature and humidity monitoring data stream is the actual temperature and humidity of a monitoring tunnel in a plurality of monitoring tunnels at a corresponding set of observation points, and the plurality of monitoring tunnels may include a target monitoring tunnel, or may be different from the target monitoring tunnel.

[0032] The purpose of the computer system determining the target data stream type is to provide a more accurate model matching basis for subsequent temperature and humidity reasoning analysis. Different data stream types reflect different temperature and humidity change patterns, and using the temperature and humidity reasoning model corresponding to the data stream type for analysis can improve the accuracy of temperature and humidity reasoning.

[0033] The example tunnel temperature and humidity monitoring data stream set is the basic data for the computer system to perform clustering operations. It contains the real temperature and humidity data of multiple monitoring tunnels at different observation points. These data can come from tunnels with different geographical locations, different structures, and different usage conditions, such as urban subway tunnels, mountain highway tunnels, etc. By analyzing and clustering these data, the commonalities and differences in temperature and humidity changes in different tunnels can be found, and thus they can be divided into different data stream types.

[0034] The data stream type library is a collection of different data stream types stored and managed by the computer system. It is obtained by clustering the example tunnel temperature and humidity monitoring data stream collection. Clustering is a method of grouping similar data objects into one category. By clustering, the data in the example tunnel temperature and humidity monitoring data stream collection can be divided into multiple different categories, each corresponding to a data stream type. For example, the computer system can divide the example tunnel temperature and humidity monitoring data stream collection into different data stream types such as stable, fluctuating, and mutant types according to the characteristics of the temperature and humidity change trend and fluctuation range.

[0035] When determining the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, the computer system can use different methods. One method is to calculate the commonality measurement results of the x-th tunnel temperature and humidity monitoring data stream and the control monitoring data stream corresponding to each data stream type in the data stream type library, and select the data stream type with the largest commonality measurement result and not less than the reference value as the target data stream type. The commonality measurement result is an indicator to measure the similarity between two data streams. Common commonality measurement methods include Euclidean distance, cosine similarity, etc.

[0036] Another method is to cluster the x-th tunnel temperature and humidity monitoring data stream with the sample tunnel temperature and humidity monitoring data stream set, and then determine the category to which the x-th tunnel temperature and humidity monitoring data stream belongs as the target data stream type. This method can directly classify according to the characteristics of the data, avoiding the selection of the control monitoring data stream and the calculation of the commonality measurement results.

[0037] In practical applications, computer systems can use clustering algorithms to implement clustering operations. Commonly used clustering algorithms include K-means algorithm, DBSCAN algorithm, etc. Taking K-means algorithm as an example, its basic idea is to divide data points into K clusters in an iterative manner, so that the similarity between data points in each cluster is as high as possible, while the similarity between data points in different clusters is as low as possible.

[0038] Assume that the computer system uses the K-means algorithm to cluster the example tunnel temperature and humidity monitoring data stream set, and obtains K data stream types. For the x-th tunnel temperature and humidity monitoring data stream, the computer system can add it to the example tunnel temperature and humidity monitoring data stream set, and then re-run the K-means algorithm to assign the x-th tunnel temperature and humidity monitoring data stream to a suitable cluster, and the data stream type corresponding to the cluster is the target data stream type.

[0039] When determining the target data stream type, the computer system also needs to consider the setting of the reference value. The reference value is a threshold used to determine whether the similarity between the x-th tunnel temperature and humidity monitoring data stream and the control monitoring data stream is high enough. If the commonality measurement result is not less than the reference value, it is considered that the x-th tunnel temperature and humidity monitoring data stream is similar to the data stream type corresponding to the control monitoring data stream, and it can be used as the target data stream type; otherwise, it is necessary to reselect other data stream types or use other methods to determine.

[0040] For example, when the computer system uses the Euclidean distance as a commonality measurement method, the reference value is set to 10. If the Euclidean distance between the x-th tunnel temperature and humidity monitoring data stream and the d-th control monitoring data stream is 8, which is less than the reference value of 10, then it is considered that the similarity between them is high, and the data stream type corresponding to the d-th control monitoring data stream can be used as the target data stream type; if the Euclidean distance is 12, which is greater than the reference value of 10, then it is considered that the similarity between them is low, and other data stream types need to be reselected.

[0041] The computer system determines the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, which is an important step in tunnel temperature and humidity monitoring analysis. By reasonably selecting clustering methods and commonality measurement methods, and setting appropriate reference values, the computer system can accurately determine the target data stream type, providing a more accurate model matching basis for subsequent temperature and humidity reasoning analysis, thereby improving the accuracy and reliability of tunnel temperature and humidity monitoring analysis.

[0042] Step S300: Load the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model corresponding to the target data stream type, and obtain the inferred temperature and humidity of the target monitoring tunnel at the next observation point after the x-th group of observation points, wherein the target temperature and humidity inference model is a model obtained by calibrating the temperature and humidity inference model to be calibrated through a subset of the example tunnel temperature and humidity monitoring data streams, and the example tunnel temperature and humidity monitoring data stream subset includes the example tunnel temperature and humidity monitoring data stream corresponding to the target data stream type in the example tunnel temperature and humidity monitoring data stream set.

[0043] In step S300, the computer system loads the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model corresponding to the target data stream type to obtain the inference temperature and humidity of the target monitoring tunnel at the next observation point of the x-th group of observation points. The target temperature and humidity inference model is a model obtained by calibrating the temperature and humidity inference model to be calibrated by the example tunnel temperature and humidity monitoring data stream subset, and the example tunnel temperature and humidity monitoring data stream subset includes the example tunnel temperature and humidity monitoring data stream corresponding to the target data stream type in the example tunnel temperature and humidity monitoring data stream set.

[0044] The temperature and humidity inference model to be adjusted is an initial model structure, which can be a model based on machine learning or deep learning, such as recurrent neural network (RNN), long short-term memory network (LSTM), etc. These models have strong nonlinear fitting capabilities, can process time series data, and are suitable for inference analysis of tunnel temperature and humidity.

[0045] The example tunnel temperature and humidity monitoring data stream subset is selected from the example tunnel temperature and humidity monitoring data stream set according to the target data stream type. These data have similar temperature and humidity change patterns and can provide targeted training data for the calibration of the target temperature and humidity inference model. The computer system loads the example tunnel temperature and humidity monitoring data stream subset into the temperature and humidity inference model to be calibrated, and continuously adjusts the parameters of the model so that the model can better fit the data, thereby improving the inference accuracy of the model.

[0046] When the x-th tunnel temperature and humidity monitoring data stream is loaded into the target temperature and humidity inference model, the computer system uses the model's reasoning ability to predict the temperature and humidity of the target monitoring tunnel at the next observation point after the x-th group of observation points based on the temperature and humidity data of the x-th group of observation points. This process is similar to predicting future trends based on historical data.

[0047] To illustrate this process, assume that the target monitoring tunnel is an urban subway tunnel, and the xth group of observation points records the temperature and humidity data of multiple locations in the tunnel at a certain moment. The target data stream type is a stable fluctuation type, and the corresponding target temperature and humidity inference model is a calibrated LSTM model. The computer system inputs the xth tunnel temperature and humidity monitoring data stream into the LSTM model, and the model calculates based on the input data and outputs the inferred temperature and humidity of the tunnel at the next observation point after the xth group of observation points.

[0048] In practical applications, the computer system can use the following technical means to implement step S300. First, the temperature and humidity inference model to be calibrated needs to be trained and adjusted. Optimization algorithms such as stochastic gradient descent (SGD) and adaptive moment estimation (Adam) can be used to update the parameters of the model so that the loss function of the model is minimized. The loss function can select the mean square error (MSE). During the training process, the computer system divides the subset of the example tunnel temperature and humidity monitoring data stream into a training set and a validation set. The training set is used to update the parameters of the model, and the validation set is used to evaluate the performance of the model. By continuously adjusting the parameters of the model, the loss function of the model on the validation set is minimized, thereby obtaining the target temperature and humidity inference model.

[0049] When the x-th tunnel temperature and humidity monitoring data stream is loaded into the target temperature and humidity inference model, the computer system preprocesses the input data. For example, the data is normalized so that the data value range is between [0, 1], which can improve the training efficiency and inference accuracy of the model. The normalization process can use the minimum-maximum normalization method.

[0050] The computer system also needs to consider the generalization ability of the model. In order to avoid overfitting of the model, regularization methods such as L1 and L2 regularization can be used during training. The loss function of L2 regularization can be expressed as: ; Where L is the original loss function, is the regularization coefficient, are the parameters of the model and m is the number of parameters.

[0051] After obtaining the inferred temperature and humidity, the computer system can evaluate and verify the inference results according to actual needs. For example, the inferred temperature and humidity can be compared with the actually measured temperature and humidity to calculate the error between the two. If the error is within an acceptable range, the inference result is considered reliable; otherwise, the target temperature and humidity inference model needs to be further adjusted and optimized. The computer system can also apply the inferred temperature and humidity to the operation management and safety assurance of the tunnel. For example, if the inferred temperature and humidity exceed the safety threshold, the computer system can issue an alarm in time to remind relevant personnel to take measures, such as strengthening ventilation, adjusting the air-conditioning system, etc., to ensure that the temperature and humidity environment in the tunnel meets safety requirements.

[0052] The process of the computer system executing step S300 involves multiple links such as model training, data preprocessing, and evaluation of reasoning results. By reasonably selecting the model structure, optimization algorithm and regularization method, and effectively preprocessing the data, the computer system can accurately obtain the inferred temperature and humidity of the target monitoring tunnel at the next observation point after the xth group of observation points, providing strong support for the operation management and safety of the tunnel. At the same time, continuously optimizing and improving the target temperature and humidity inference model can further improve the accuracy and reliability of reasoning and meet the needs of practical applications. In future research and applications, more data analysis technologies and sensor data can be combined to further improve the level of tunnel temperature and humidity monitoring and analysis.

[0053] As an implementation manner, in step S200, determining the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream includes:

[0054] Step S210: when the data stream type library includes A data stream types obtained by clustering the example tunnel temperature and humidity monitoring data stream set, and the A data stream types are matched one-to-one with the A control monitoring data streams, determine the commonality measurement results of the x-th tunnel temperature and humidity monitoring data stream and each control monitoring data stream in the A control monitoring data streams, respectively, and obtain A commonality measurement results, wherein the d-th control monitoring data stream in the A control monitoring data streams is a data stream determined based on the example tunnel temperature and humidity monitoring data stream corresponding to the d-th data stream type in the example tunnel temperature and humidity monitoring data stream set, and the A data stream types include the d-th data stream type, wherein A≥2, 1≤d≤A;

[0055] Step S220: When the kth commonality measurement result among A commonality measurement results is the largest and the kth commonality measurement result is not less than the reference value, the target data stream type is determined as the data stream type corresponding to the kth commonality measurement result among the A data stream types, where 1≤k≤A.

[0056] In step S210, under the above conditions, the computer system determines the commonality measurement results of the x-th tunnel temperature and humidity monitoring data stream and each of the A control monitoring data streams, thereby obtaining A commonality measurement results. The control monitoring data stream here is a data stream determined based on the example tunnel temperature and humidity monitoring data stream corresponding to the d-th data stream type in the example tunnel temperature and humidity monitoring data stream set, and the value range of d is a natural number between 1 and A, and A is greater than or equal to 2. The commonality measurement result is used to measure the similarity between two data streams, and is a key indicator for determining which control monitoring data stream the x-th tunnel temperature and humidity monitoring data stream is most similar to.

[0057] In order to calculate the commonality measurement results, the computer system can use a variety of methods, one of which is the Euclidean distance. Euclidean distance is the straight-line distance between two points in multidimensional space. For two data streams of the same length, the Euclidean distance can measure the degree of difference between them. The smaller the Euclidean distance, the more similar the two data streams are.

[0058] Another commonly used method is cosine similarity, which measures the similarity between two vectors by calculating the cosine value of the angle between them. The value of cosine similarity ranges from -1 to 1. The closer the value is to 1, the more similar the two data streams are.

[0059] In step S220, after obtaining A commonality measurement results, the computer system finds the largest k-th commonality measurement result, and when the k-th commonality measurement result is not less than the reference value, the target data stream type is determined as the data stream type corresponding to the k-th commonality measurement result in the A data stream types. The reference value is a preset threshold value used to determine whether the similarity between the x-th tunnel temperature and humidity monitoring data stream and the control monitoring data stream is high enough.

[0060] After determining the target data stream type, the computer system can use the target temperature and humidity inference model corresponding to the type to perform temperature and humidity inference analysis. Since the target data stream type reflects the characteristics and change pattern of the x-th tunnel temperature and humidity monitoring data stream, the use of the corresponding target temperature and humidity inference model can more accurately predict the temperature and humidity of subsequent observation points.

[0061] Computer systems can use statistical analysis methods to detect and process outliers. For example, by calculating the mean and standard deviation of the data stream, data points that deviate from the mean by more than a certain multiple of the standard deviation are considered outliers and deleted or corrected. In addition, methods such as sliding window filtering can be used to smooth the data and reduce the impact of noise.

[0062] When choosing a commonality measurement method, the computer system makes a choice based on the characteristics of the data and actual needs. Euclidean distance is suitable for situations where the data distribution is relatively uniform and the data dimension is low; cosine similarity focuses more on measuring the directional similarity of the data and is suitable for situations where the data dimension is high and the size difference of the data is large.

[0063] The computer system can also use multiple commonality measurement methods for comprehensive analysis to improve the accuracy of determining the target data stream type. For example, the Euclidean distance and cosine similarity can be calculated simultaneously, and then a weighted average is performed based on the results of the two to obtain a comprehensive commonality measurement result.

[0064] Steps S210-S220 are an effective method for the computer system to determine the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream. By calculating the commonality measurement results and comparing them with the reference values, the computer system can accurately classify the x-th tunnel temperature and humidity monitoring data stream into the appropriate data stream type, providing strong support for subsequent temperature and humidity reasoning analysis.

[0065] As another implementation, step S200, determining the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, includes:

[0066] Step S201: clustering the example tunnel temperature and humidity monitoring data stream set and the x-th tunnel temperature and humidity monitoring data stream to obtain B data stream types, where B>1;

[0067] Step S202: Determine the target data stream type as the data stream type corresponding to the xth tunnel temperature and humidity monitoring data stream among the B data stream types.

[0068] Clustering is a process of grouping data objects into multiple classes or clusters, so that data objects in the same cluster have higher similarity, while data objects in different clusters have lower similarity. In the analysis of tunnel temperature and humidity monitoring, clustering can help the computer system discover the characteristics and patterns of different tunnel temperature and humidity monitoring data streams, thereby classifying them into different data stream types.

[0069] Computer systems can use a variety of clustering algorithms to implement clustering operations, one of which is the K-means algorithm. The K-means algorithm is an iterative clustering algorithm. Its basic idea is to minimize the sum of the distances from each data point to its cluster center by continuously adjusting the position of the cluster center. The specific steps are as follows: First, the computer system randomly selects K data points as the initial cluster centers, where K is the number of data stream types expected, that is, B; then, the computer system calculates the distance from each data point to the K cluster centers, and assigns each data point to the cluster with the nearest cluster center; then, the computer system recalculates the cluster center of each cluster, that is, the mean of all data points in the cluster; finally, the computer system repeats the above steps of assigning and updating the cluster center until the position of the cluster center no longer changes or the preset maximum number of iterations is reached.

[0070] Assume that the computer system has a set of 100 example tunnel temperature and humidity monitoring data streams, and the x-th tunnel temperature and humidity monitoring data stream, for a total of 101 data streams. The computer system hopes to obtain B=5 data stream types by clustering. The computer system first randomly selects 5 data streams as the initial cluster centers, and then calculates the Euclidean distance of each data stream to these 5 cluster centers. The computer system assigns each data stream to the cluster with the closest cluster center. For example, if the x-th tunnel temperature and humidity monitoring data stream is closest to the third cluster center, then it is assigned to the third cluster. After that, the computer system recalculates the cluster center of each cluster. For example, for the third cluster, the value of each dimension of the new cluster center is the average value of the corresponding dimension of all data streams in the cluster. The computer system repeats this process until the cluster center no longer changes or the maximum number of iterations is reached.

[0071] Another commonly used clustering algorithm is the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm. The DBSCAN algorithm is a density-based clustering algorithm that divides data points with sufficient density into a cluster and treats data points in areas with lower density as noise points. The core concepts of the DBSCAN algorithm include core points, boundary points, and noise points. A core point refers to a point that contains at least MinPts data points in its neighborhood; a boundary point refers to a data point that is not in the neighborhood of a core point but in the neighborhood of a core point; a noise point refers to a data point that is neither a core point nor a boundary point. The specific steps of the DBSCAN algorithm are as follows: the computer system first starts from an unvisited data point and calculates the number of data points in its neighborhood; if the data point is a core point, it takes this point as the starting point and recursively expands all core points in its neighborhood to form a cluster; if the data point is not a core point, it is marked as a noise point or a boundary point; the computer system repeats this process until all data points have been visited.

[0072] Before clustering, the computer system also needs to preprocess the data to ensure the accuracy of clustering. The preprocessing steps include data cleaning, data normalization, etc. Data cleaning refers to removing noise, outliers, and missing values ​​from the data. For example, the computer system can detect and remove outliers by setting thresholds. If the temperature and humidity values ​​in a data stream exceed the normal range, it will be regarded as an outlier and processed. Data normalization refers to converting data to a unified scale to avoid the impact of data scale differences on clustering results.

[0073] After completing the clustering operation, the computer system obtains B data stream types. Step S202 requires the computer system to determine the data stream type to which the x-th tunnel temperature and humidity monitoring data stream belongs, which is the target data stream type. The computer system can determine the corresponding target data stream type by checking the cluster to which the x-th tunnel temperature and humidity monitoring data stream is assigned. For example, in the above-mentioned K-means algorithm example, if the x-th tunnel temperature and humidity monitoring data stream is ultimately assigned to the third cluster, then the data stream type represented by the third cluster is the target data stream type.

[0074] Different data stream types reflect different tunnel temperature and humidity change patterns. Using the target temperature and humidity inference model corresponding to the target data stream type can more accurately predict the temperature and humidity of subsequent observation points. For example, if the target data stream type represents a relatively stable temperature and humidity change pattern, then the corresponding target temperature and humidity inference model can be constructed based on the stationary time series analysis method; if the target data stream type represents a relatively drastic temperature and humidity change pattern, then the corresponding target temperature and humidity inference model may need to adopt a more complex nonlinear model.

[0075] Different clustering algorithms are suitable for different data distributions and characteristics. The computer system selects the appropriate algorithm based on the specific conditions of the example tunnel temperature and humidity monitoring data stream set and the xth tunnel temperature and humidity monitoring data stream. At the same time, the parameter settings in the clustering algorithm will also have an important impact on the clustering results. The computer system can determine the optimal parameter values ​​through multiple experiments and evaluations. For example, in the K-means algorithm, the choice of K value will directly affect the clustering effect. The computer system can use the Elbow Method to determine the appropriate K value. The Elbow Method calculates the intra-cluster sum of squared errors (SSE) under different K values, and observes the curve of SSE changing with K value, and selects the K value corresponding to the point with the largest slope change in the curve as the optimal value. The calculation formula for the sum of squared errors within the cluster is: , where K is the number of clusters, is the i-th cluster, x is the data point in the cluster, is the cluster center of the ith cluster.

[0076] As an implementation manner, in step S200, before determining the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, the method may further include:

[0077] Step S101: clustering the example tunnel temperature and humidity monitoring data stream set to obtain a data stream type library; or when the example tunnel temperature and humidity monitoring data stream set includes K example tunnel temperature and humidity monitoring data streams, extracting sub-data streams of equal length from each of the K example tunnel temperature and humidity monitoring data streams to obtain K sub-data streams, wherein each of the K sub-data streams includes the actual temperature and humidity of one monitoring tunnel among multiple monitoring tunnels at s observation points; clustering the K sub-data streams to obtain a data stream type library.

[0078] Step S101 is an operation performed before step S200 determines the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, and its purpose is to build a data stream type library to provide a basis for subsequent data stream classification. There are two implementation methods for this step. One is to directly cluster the example tunnel temperature and humidity monitoring data stream set to obtain a data stream type library; the other is when the example tunnel temperature and humidity monitoring data stream set contains K example tunnel temperature and humidity monitoring data streams, first extract sub-data streams of equal length from each example tunnel temperature and humidity monitoring data stream to obtain K sub-data streams, and then cluster the K sub-data streams to obtain a data stream type library.

[0079] In the first embodiment, the computer system directly clusters the set of example tunnel temperature and humidity monitoring data streams. The set of example tunnel temperature and humidity monitoring data streams is composed of multiple example tunnel temperature and humidity monitoring data streams of multiple monitoring tunnels, which record the actual temperature and humidity of different monitoring tunnels at corresponding observation points. Clustering is the process of classifying similar data objects into one category. Through clustering, the computer system can discover the potential patterns and structures in the data and classify data streams with similar temperature and humidity change characteristics into the same type. The computer system can use a variety of clustering algorithms to achieve clustering, such as K-means algorithm, hierarchical clustering algorithm, etc. Taking the K-means algorithm as an example, it is an iterative clustering method. Its basic idea is to minimize the sum of the distances from each data point to its cluster center by continuously adjusting the position of the cluster center. Specifically, the computer system first randomly selects K data points as the initial cluster centers, where K is the number of data flow types expected to be obtained; then calculates the distance from each example tunnel temperature and humidity monitoring data flow to the K cluster centers, and assigns it to the cluster with the closest cluster center; then recalculates the cluster center of each cluster, that is, the mean of all data flows in the cluster; repeats the above steps of assigning and updating cluster centers until the cluster center no longer changes or reaches the preset maximum number of iterations. For example, assuming that the computer system has a set of 100 example tunnel temperature and humidity monitoring data flows, and hopes to obtain 5 types of data flows, the K-means algorithm can be used for clustering.

[0080] In the second embodiment, the computer system first extracts sub-data streams of equal length from each example tunnel temperature and humidity monitoring data stream. The purpose of this is to ensure that the data of subsequent clustering operations are comparable, because different example tunnel temperature and humidity monitoring data streams may have different lengths, and direct clustering may affect the accuracy of the results. Each sub-data stream contains the actual temperature and humidity of the corresponding monitoring tunnel at s observation points. For example, assuming that each example tunnel temperature and humidity monitoring data stream records the temperature and humidity data of 100 observation points, the computer system can choose to extract the data of 20 consecutive observation points as a sub-data stream, that is, s=20.

[0081] After extracting the sub-data streams, the computer system clusters the K sub-data streams to obtain a data stream type library. In actual operation, the computer system can use a variety of methods to measure the similarity between sub-data streams. For example, the Euclidean distance is used to measure the distance between two data objects. The smaller the Euclidean distance, the more similar the two sub-data streams are. In addition to the Euclidean distance, other methods such as cosine similarity can also be used to measure the similarity between sub-data streams.

[0082] When performing clustering operations, the computer system can classify similar sub-data streams into the same type based on the similarity measurement results between the sub-data streams. For example, if the computer system uses the K-means algorithm to cluster K sub-data streams, it will continuously adjust the position of the cluster center so that the similarity between sub-data streams in the same cluster is as high as possible, and the similarity between sub-data streams in different clusters is as low as possible. Ultimately, the different clusters obtained by the computer system correspond to different data stream types, and these data stream types constitute a data stream type library.

[0083] In order to evaluate the quality of clustering results, computer systems can use some evaluation indicators, such as the silhouette coefficient. The silhouette coefficient comprehensively considers the compactness within the cluster and the separation between clusters. Its value range is between -1 and 1. The closer the value is to 1, the better the clustering effect. The calculation formula of the silhouette coefficient is , where a(i) is the average distance from sample i to other samples in its cluster, and b(i) is the average distance from sample i to the samples in the nearest neighbor cluster.

[0084] Step S101 constructs a data stream type library for the computer system by clustering the example tunnel temperature and humidity monitoring data stream set or its extracted sub-data streams. This data stream type library is the basis for subsequently determining the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, and can help the computer system to more accurately classify the tunnel temperature and humidity monitoring data stream, thereby providing more effective support for the reasoning analysis of tunnel temperature and humidity. In practical applications, the computer system selects appropriate clustering methods and parameters according to the specific circumstances to ensure the quality and effectiveness of the data stream type library. At the same time, by continuously optimizing and improving the clustering process, the accuracy and reliability of tunnel temperature and humidity monitoring analysis can be improved, providing a more powerful decision-making basis for tunnel operation management and safety assurance.

[0085] As an implementation mode, in step S101, clustering the K sub-data streams to obtain a data stream type library includes:

[0086] Step S1011: respectively determining the commonality measurement results between each sub-data stream in the K sub-data streams and each sub-data stream in the K sub-data streams, and obtaining a commonality measurement result array with a dimension of K*K, wherein the commonality measurement result between the e-th sub-data stream in the K sub-data streams and the f-th sub-data stream in the K sub-data streams is the commonality measurement result between the s real temperatures and humidity in the e-th sub-data stream and the s real temperatures and humidity in the f-th sub-data stream, and e and f are natural numbers not less than 1 and not greater than K at the same time;

[0087] Step S1012: Determine a classification indication array with a dimension of 1*K based on the K*K commonality measurement result array, wherein the 1*K classification indication array represents the data stream type corresponding to each sub-data stream in the K sub-data streams, wherein the data stream type library includes different data stream types indicated by the classification indication array.

[0088] Step S1011 requires the computer system to determine the commonality measurement results between each sub-data stream in the K sub-data streams and each sub-data stream in the K sub-data streams, respectively, so as to obtain a commonality measurement result array with a dimension of K * K. The commonality measurement results here are used to measure the similarity between the two sub-data streams, which is the basis for subsequent clustering operations. Each sub-data stream contains the real temperature and humidity of the corresponding monitoring tunnel at s observation points. The commonality measurement result between the e-th sub-data stream and the f-th sub-data stream is the commonality measurement result between the s real temperatures and humidities in the two sub-data streams, and e and f are natural numbers not less than 1 and not greater than K. The computer system can use a variety of methods to calculate the commonality measurement results, such as Euclidean distance and cosine similarity.

[0089] The computer system calculates the commonality measurement results between each sub-data stream and other sub-data streams, and finally obtains a commonality measurement result array with a dimension of K * K. Each row and each column of this array corresponds to a sub-data stream, and each element in the array represents the commonality measurement result between the corresponding two sub-data streams. For example, the element in the e-th row and the f-th column of the array represents the commonality measurement result between the e-th sub-data stream and the f-th sub-data stream.

[0090] In step S1012, the computer system determines a classification indication array of dimension 1 * K based on the K * K commonality measurement result array, wherein the array represents the data stream type corresponding to each sub-data stream in the K sub-data streams, and the data stream type library includes different data stream types indicated by the classification indication array.

[0091] The computer system can use a hierarchical clustering algorithm to determine the classification indicator array based on the commonality measurement result array. The hierarchical clustering algorithm is a method for clustering based on the similarity between data. It can be divided into two types: agglomerative and divisive. The agglomerative hierarchical clustering algorithm starts with each data point as a separate cluster, and then gradually merges the clusters with the highest similarity until the preset number of clusters is reached or all data points are merged into one cluster; the divisive hierarchical clustering algorithm starts with all data points in one cluster and gradually splits them into smaller clusters.

[0092] Taking the agglomerative hierarchical clustering algorithm as an example, the computer system first treats each sub-data stream as a separate cluster, then finds the two clusters with the highest similarity based on the commonality measurement result array and merges them into a new cluster. Next, the commonality measurement result array is updated to calculate the similarity between the new cluster and other clusters. This process is repeated until the preset number of clusters is reached. In this process, the computer system can record the number of the cluster to which each sub-data stream belongs, and finally form a classification indicator array.

[0093] Another method is to use the K-means algorithm. The K-means algorithm is an iterative clustering algorithm. Its basic idea is to minimize the sum of the distances from each data point to its cluster center by continuously adjusting the position of the cluster center. The computer system first randomly selects K cluster centers, where K is the number of data stream types expected to be obtained. Then, based on the array of commonality measurement results, each sub-data stream is assigned to the cluster with the nearest cluster center. Next, the cluster center of each cluster is recalculated, that is, the mean of all sub-data streams in the cluster. Repeat this process until the cluster center no longer changes or the preset maximum number of iterations is reached. Finally, the computer system can generate a classification indicator array based on the number of the cluster to which each sub-data stream belongs.

[0094] The computer system can improve the quality of clustering by calculating the silhouette coefficient of each sub-data stream and adjusting the clustering results. For example, if the silhouette coefficient of a sub-data stream is low, it means that it may be incorrectly assigned to a cluster. The computer system can try to reallocate the sub-data stream or adjust the clustering parameters, such as the number of clusters.

[0095] Steps S1011-S1012 are important steps for the computer system to cluster K sub-data streams to obtain a data stream type library. By calculating the commonality measurement results between the sub-data streams, a K * K commonality measurement result array is obtained, and then a 1 * K classification indication array is determined based on the array. The computer system can classify the K sub-data streams into different data stream types, thereby constructing a data stream type library. In practical applications, the computer system reasonably selects the commonality measurement method and clustering algorithm, and evaluates and adjusts the clustering results through evaluation indicators to ensure the quality and effectiveness of the data stream type library. At the same time, continuously optimizing and improving these methods and techniques can further improve the level of tunnel temperature and humidity monitoring and analysis, and provide more accurate decision-making basis for tunnel operation management and safety assurance.

[0096] As an implementation mode, step S300, loading the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model corresponding to the target data stream type, and obtaining the inference temperature and humidity of the target monitoring tunnel at the next observation point of the x-th group of observation points, includes:

[0097] Step S310: Load the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model to obtain the inference temperature and humidity expectation and inference temperature and humidity dispersion of the target monitoring tunnel at the next observation point of the x-th group of observation points;

[0098] Step S320: Determine the inferred temperature and humidity of the target monitoring tunnel at the next observation point of the xth group of observation points based on the inferred temperature and humidity expectation and the inferred temperature and humidity dispersion.

[0099] The target temperature and humidity inference model is a model obtained by calibrating the temperature and humidity inference model to be calibrated by using a subset of the sample tunnel temperature and humidity monitoring data streams. The sample tunnel temperature and humidity monitoring data stream subset includes the sample tunnel temperature and humidity monitoring data streams corresponding to the target data stream type in the sample tunnel temperature and humidity monitoring data stream set. This model can predict the temperature and humidity conditions of subsequent observation points based on the input tunnel temperature and humidity monitoring data streams.

[0100] In step S310, the computer system inputs the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model. The inference temperature and humidity expectation is the average estimated value of the target monitoring tunnel temperature and humidity at the next observation point after the x-th group of observation points predicted by the model, which represents the temperature and humidity conditions that the model considers to be most likely to occur. The inference temperature and humidity dispersion measures the uncertainty of the predicted temperature and humidity, reflecting the extent to which the actual temperature and humidity may deviate from the inference temperature and humidity expectation.

[0101] The computer system can use various types of models as the target temperature and humidity inference model, such as a model based on probability statistics or a machine learning model. Taking the Gaussian process regression model as an example, it is a commonly used probability model that can naturally output the predicted mean (i.e., the expected value of the inferred temperature and humidity) and variance (related to the inferred temperature and humidity dispersion). For the Gaussian process regression model, assuming that the x-th tunnel temperature and humidity monitoring data stream is , the target temperature and humidity inference model can be expressed as a Gaussian process:

[0102] ;

[0103] Where m(X) is the mean function and k(X, X') is the covariance function. Given the training data (a subset of the sample tunnel temperature and humidity monitoring data stream) and the input X x , the model can calculate the expected value of temperature and humidity at the next observation point of the xth group of observation points and inference temperature and humidity dispersion (variance ). Infer the expected temperature and humidity The most likely temperature and humidity values ​​are obtained based on the model's fitting and prediction of the data, and the inferred temperature and humidity dispersion represents the range of uncertainty around this expectation.

[0104] Taking the long short-term memory network (LSTM), a variant of the recurrent neural network (RNN), as an example, during the training process, the LSTM model learns the time dependency of the tunnel temperature and humidity data. When the computer system inputs the x-th tunnel temperature and humidity monitoring data stream into the trained LSTM model, the model outputs a prediction result. In order to obtain the expected and discrete temperature and humidity, the computer system can use the Monte Carlo Dropout method. Specifically, when predicting, Dropout (a technique to prevent overfitting) is turned on multiple times for forward propagation to obtain multiple predicted values. The mean of these predicted values ​​can be used as the expected temperature and humidity, and their standard deviation can be used as the discrete temperature and humidity.

[0105] In step S320, the computer system determines the inference temperature and humidity based on the inference temperature and humidity expectation and the inference temperature and humidity dispersion. The inference temperature and humidity expectation provides a most likely temperature and humidity value, but due to the uncertainty in the actual situation, the inference temperature and humidity dispersion reflects the magnitude of this uncertainty. By comprehensively considering these two factors, the computer system can obtain a more reasonable inference temperature and humidity.

[0106] Different methods can be used to combine the inference temperature and humidity expectation and the inference temperature and humidity dispersion to determine the inference temperature and humidity. One feasible method is to make certain adjustments based on the inference temperature and humidity dispersion on the basis of the inference temperature and humidity expectation. For example, the dispersion can be weighted considering the needs of actual applications and risk preferences. If you want to get a relatively conservative prediction result, you can make a certain shift in the direction of larger dispersion based on the inference temperature and humidity expectation; if you pay more attention to the average situation, you can directly use the inference temperature and humidity expectation as the inference temperature and humidity.

[0107] Assume that the expected temperature and humidity is , the inferred temperature and humidity dispersion is , the computer system can determine the inference temperature and humidity according to the preset rules For example, you can set an adjustment factor , the calculation formula of temperature and humidity is inferred as: ;in The value of can be adjusted according to the actual situation. =0, the inferred temperature and humidity are equal to the expected inferred temperature and humidity; if >0, the inference temperature and humidity will shift toward a larger value based on the expected inference temperature and humidity; if <0, it will shift towards smaller values.

[0108] In actual applications, the computer system determines the appropriate adjustment method and parameters according to different scenarios and requirements. For example, in tunnel environments that are sensitive to temperature and humidity, a more conservative prediction may be required, that is, a larger value; in the case of higher tolerance to temperature and humidity changes, a smaller value can be set value.

[0109] When executing steps S310-S320, the computer system also considers the performance evaluation and verification of the model. Historical data can be used for cross-validation, and the data can be divided into a training set and a test set. The target temperature and humidity inference model is trained on the training set, and the prediction performance of the model is evaluated on the test set. For example, evaluation indicators include mean square error (MSE), mean absolute error (MAE), etc. In addition, the computer system can also perform targeted optimization of the target temperature and humidity inference model according to different target data stream types. Different data stream types may have different temperature and humidity change patterns, such as stable, fluctuating, etc. For stable data stream types, the model can pay more attention to the mean and trend of historical data; for fluctuating data stream types, the model may need to capture data changes more flexibly.

[0110] As an implementation manner, step S320, determining the inferred temperature and humidity of the target monitoring tunnel at the next observation point of the xth group of observation points according to the inferred temperature and humidity expectation and the inferred temperature and humidity dispersion, includes:

[0111] Step S321: taking the inferred temperature and humidity as equal to the sum of the expected inferred temperature and humidity and the target eccentricity adjustment result, wherein the target eccentricity adjustment result is the multiplication result between the preset discrete eccentricity coefficient and the discreteness of the inferred temperature and humidity.

[0112] The expected value of inference temperature and humidity is the average estimated value of the target monitoring tunnel temperature and humidity at the next observation point of the xth group of observation points predicted by the target temperature and humidity inference model based on the input xth tunnel temperature and humidity monitoring data stream, which represents the temperature and humidity conditions that the model believes are most likely to occur. The inference temperature and humidity dispersion measures the uncertainty of the predicted temperature and humidity, reflecting the extent to which the actual temperature and humidity may deviate from the expected value of inference temperature and humidity. The discrete eccentricity coefficient is a pre-set parameter used to control the extent to which the expected value of inference temperature and humidity is adjusted according to the inference temperature and humidity dispersion.

[0113] The computer system calculates the final inference temperature and humidity through the formula "inference temperature and humidity = inference temperature and humidity expectation + target eccentricity adjustment result", that is, "inference temperature and humidity = inference temperature and humidity expectation + discrete eccentricity coefficient × inference temperature and humidity dispersion". The value of the discrete eccentricity coefficient has an important influence on the calculation result of the inference temperature and humidity. If the discrete eccentricity coefficient is positive, the inference temperature and humidity will shift towards a larger value based on the inference temperature and humidity expectation; if the discrete eccentricity coefficient is negative, the inference temperature and humidity will shift towards a smaller value; if the discrete eccentricity coefficient is zero, the inference temperature and humidity is equal to the inference temperature and humidity expectation.

[0114] For example, it is assumed that the computer system obtains the expected inference temperature and humidity of the target monitoring tunnel at the next observation point of the xth group of observation points through the target temperature and humidity inference model as 25°C, the inference temperature and humidity dispersion is 3°C, and the preset discrete eccentricity coefficient is 0.5. According to the formula of step S321, the target eccentricity adjustment result is 0.5×3=1.5°C, then the inference temperature and humidity is 25+1.5=26.5°C. If the discrete eccentricity coefficient becomes -0.5, the target eccentricity adjustment result is -0.5×3=-1.5°C, and the inference temperature and humidity is 25-1.5=23.5°C.

[0115] In actual applications, the computer system reasonably sets the discrete eccentricity coefficient according to different scenarios and needs. In some scenarios with strict requirements on temperature and humidity and requiring more conservative predictions, such as the operating environment of some tunnel equipment that is sensitive to temperature and humidity, the computer system can set a larger positive discrete eccentricity coefficient to ensure that the predicted temperature and humidity can cover the possible higher temperature and humidity conditions, and prepare countermeasures in advance. On the contrary, in scenarios with a higher tolerance for temperature and humidity changes, the discrete eccentricity coefficient can be set smaller or even zero, and the predicted value closer to the average situation can be used as the inferred temperature and humidity.

[0116] The computer system can use the following technical means to determine the discrete eccentricity coefficient. One method is to perform statistical analysis based on historical data. The computer system can collect a large amount of historical tunnel temperature and humidity data, as well as the corresponding model prediction results, and analyze the deviation between the actual temperature and humidity and the expected inference temperature and humidity. According to the statistical characteristics of these deviations, such as mean, standard deviation, etc., a suitable discrete eccentricity coefficient is determined so that the inference temperature and humidity calculated by step S321 is closer to the actual value. Another method is to combine expert experience. Domain experts have an in-depth understanding of the changing laws of tunnel temperature and humidity and actual application needs. The computer system can consult experts and set the discrete eccentricity coefficient according to the experts' suggestions.

[0117] The computer system can also adopt a method of dynamically adjusting the discrete eccentricity coefficient. In different time periods or under different tunnel operation conditions, the changing characteristics and uncertainty of temperature and humidity may change. The computer system can monitor the relevant parameters of the tunnel in real time, such as ventilation volume, vehicle flow, etc., and dynamically adjust the discrete eccentricity coefficient according to the changes in these parameters. For example, when the ventilation volume increases, the uncertainty of temperature and humidity may decrease, and the computer system can reduce the discrete eccentricity coefficient accordingly; when the vehicle flow increases, the heat generated may increase the uncertainty of temperature and humidity, and the computer system increases the discrete eccentricity coefficient.

[0118] Step S321 provides a computer system with an effective method for combining the expected value of inference temperature and humidity with the discreteness of inference temperature and humidity to determine the final inference temperature and humidity. By reasonably setting the discrete eccentricity coefficient, the computer system can flexibly adjust the calculation results of the inference temperature and humidity according to different scenarios and needs, so as to more accurately predict the temperature and humidity conditions of the target monitoring tunnel, and provide a more reliable decision-making basis for the operation management and safety of the tunnel. At the same time, the use of technical means such as statistical analysis based on historical data, combined with expert experience and dynamic adjustment to determine the discrete eccentricity coefficient can further improve the accuracy and adaptability of the inference temperature and humidity.

[0119] As an implementation mode, in step S300, before loading the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model corresponding to the target data stream type and obtaining the inference temperature and humidity of the target monitoring tunnel at the next observation point of the x-th group of observation points, the method further includes:

[0120] Step S301: Debug the temperature and humidity inference model to be calibrated through the example tunnel temperature and humidity monitoring data stream subset until the target debugging error corresponding to the temperature and humidity inference model to be calibrated meets the preset error requirement, stop debugging, and use the temperature and humidity inference model to be calibrated when debugging is stopped as the target temperature and humidity inference model, wherein the target debugging error is the debugging error between the example inference temperature and humidity and the example true temperature and humidity, the example inference temperature and humidity is the inference temperature and humidity of one of the multiple monitoring tunnels at an observation point determined by the temperature and humidity inference model to be calibrated based on the example tunnel temperature and humidity monitoring data stream subset loaded thereto, and the example true temperature and humidity is the true temperature and humidity of one of the multiple monitoring tunnels at an observation point determined in advance.

[0121] In step S301, the computer system debugs the temperature and humidity inference model to be calibrated through the example tunnel temperature and humidity monitoring data stream subset, and stops debugging until the target debugging error corresponding to the temperature and humidity inference model to be calibrated meets the preset error requirement, and uses the temperature and humidity inference model to be calibrated when debugging is stopped as the target temperature and humidity inference model, wherein the target debugging error is the debugging error between the example inference temperature and humidity and the example true temperature and humidity, the example inference temperature and humidity is the inference temperature and humidity of one of the multiple monitoring tunnels at an observation point determined by the temperature and humidity inference model to be calibrated based on the example tunnel temperature and humidity monitoring data stream subset loaded thereto, and the example true temperature and humidity is the true temperature and humidity of one of the multiple monitoring tunnels at the observation point determined in advance.

[0122] The temperature and humidity inference model to be adjusted is an initial model structure. It can be a model based on machine learning or deep learning, such as a neural network, support vector machine, etc. These models have certain generalization capabilities, but their prediction results may not be accurate before debugging. The sample tunnel temperature and humidity monitoring data stream subset is selected from the sample tunnel temperature and humidity monitoring data stream set. These data correspond to the target data stream type and have similar temperature and humidity change patterns, which can provide targeted training data for model debugging.

[0123] The target debugging error is a key indicator for measuring model performance. It reflects the degree of difference between the temperature and humidity predicted by the model and the actual temperature and humidity. The feasible methods for calculating the target debugging error include mean square error (MSE), mean absolute error (MAE), etc. The preset error requirement is a threshold set by the computer system to determine whether the model has achieved an acceptable performance level. When the target debugging error is less than or equal to this threshold, the computer system believes that the model has been debugged and can be used as a target temperature and humidity inference model.

[0124] In the actual debugging process, the computer system uses an iterative method to optimize the temperature and humidity inference model to be adjusted. In each iteration, the computer system loads a subset of the sample tunnel temperature and humidity monitoring data stream into the model. The model generates sample inference temperature and humidity based on the input data, and then the computer system calculates the target debugging error between the sample inference temperature and humidity and the sample actual temperature and humidity. If the target debugging error does not meet the preset error requirements, the computer system will adjust the parameters of the model to reduce the error. This process will be repeated until the target debugging error meets the preset error requirements.

[0125] For example, suppose that the computer system has a temperature and humidity inference model to be calibrated, and the sample tunnel temperature and humidity monitoring data stream subset contains 100 samples, each of which records the temperature and humidity data of a monitoring tunnel at an observation point. The computer system first inputs these samples into the model, and the model outputs the sample inference temperature and humidity. Assuming that the mean square error is used as the calculation method for the target debugging error, the computer system calculates that the current mean square error is 10. The preset error requirement is that the mean square error is less than or equal to 2. Since the current mean square error is greater than the preset error requirement, the computer system will adjust the parameters of the model, such as the weights and biases in the neural network. Then, the computer system inputs the sample tunnel temperature and humidity monitoring data stream subset into the adjusted model again and recalculates the mean square error. After multiple iterations, the computer system finally reduces the mean square error to 1.5, which is less than the preset error requirement. At this time, the computer system stops debugging and uses the model as the target temperature and humidity inference model.

[0126] In order to debug the model, the computer system can use a variety of optimization algorithms, such as stochastic gradient descent (SGD), adaptive moment estimation (Adam), etc.

[0127] During the debugging process, we need to pay attention to the problems of overfitting and underfitting. Overfitting refers to the phenomenon that the model performs well on the training data but performs poorly on unseen data. To prevent overfitting, computer systems can use regularization methods such as L1 and L2 regularization. L2 regularization limits the size of model parameters by adding a regularization term to the target debugging error. Its formula is:

[0128] ;

[0129] Where L is the original target debugging error, is the regularization coefficient, is the parameter of the model, and m is the number of parameters. Underfitting means that the model cannot fit the training data well, which is manifested as the target debugging error cannot be reduced to an acceptable level. When underfitting occurs, the computer system can consider increasing the complexity of the model, such as increasing the number of layers or neurons of the neural network.

[0130] The computer system can also use a cross-validation method to evaluate the performance of the model. Cross-validation divides the sample tunnel temperature and humidity monitoring data stream subset into multiple subsets, using one of the subsets as a validation set each time and the remaining subsets as training sets. Through multiple iterations, the computer system can obtain the performance indicators of the model on different validation sets, thereby more comprehensively evaluating the generalization ability of the model.

[0131] Step S301 is a key step for the computer system to convert the temperature and humidity reasoning model to be adjusted into the target temperature and humidity reasoning model. By continuously adjusting the parameters of the model so that the target debugging error meets the preset error requirements, the computer system can improve the prediction accuracy and generalization ability of the model. During the debugging process, the reasonable selection of optimization algorithms, handling of overfitting and underfitting problems, and the use of cross-validation and other technical means can ensure that the model has good performance in practical applications and provide strong support for the accurate reasoning of tunnel temperature and humidity.

[0132] As an implementation mode, step S301, debugging the temperature and humidity inference model to be calibrated by using a subset of the example tunnel temperature and humidity monitoring data stream, includes:

[0133] The following steps are used to debug the temperature and humidity inference model to be calibrated for the gth time, where g ≥ 2:

[0134] Step S3011: Load the g-th example tunnel temperature and humidity monitoring data stream used in the g-th debugging into the temperature and humidity inference model determined in the g-1-th debugging to obtain the example inference temperature and humidity of the g-th debugging, wherein the example inference temperature and humidity of the g-th debugging is the inference temperature and humidity of one of the multiple monitoring tunnels at the g+1-th observation point;

[0135] Step S3012: determining a debugging error between the example inference temperature and humidity of the g-th debugging and the example real temperature and humidity of the g-th debugging, and obtaining a target debugging error of the g-th debugging, wherein the example real temperature and humidity of the g-th debugging is the real temperature and humidity of one of the multiple monitoring tunnels at the g+1-th observation point determined in advance;

[0136] Step S3013: when the target debugging error of the g-th debugging does not meet the preset error requirement, optimize the model parameter variables in the temperature and humidity reasoning model determined by the g-1-th debugging to obtain the temperature and humidity reasoning model determined by the g-th debugging;

[0137] Step S3014: When the target debugging error of the g-th debugging meets the preset error requirement, the debugging is stopped, and the temperature and humidity inference model obtained by the g-1-th debugging is determined as the target temperature and humidity inference model.

[0138] In step S3011, the computer system loads the g-th example tunnel temperature and humidity monitoring data stream used in the g-th debugging into the temperature and humidity inference model determined in the g-1-th debugging, and obtains the example inference temperature and humidity of the g-th debugging, where the example inference temperature and humidity of the g-th debugging is the inference temperature and humidity of one monitoring tunnel among multiple monitoring tunnels at the g+1-th observation point. The example tunnel temperature and humidity monitoring data stream is selected from the subset of example tunnel temperature and humidity monitoring data streams, which record the actual temperature and humidity conditions of the monitoring tunnel at different observation points. The temperature and humidity inference model determined in the g-1-th debugging is the model obtained after the previous g-1 debugging, and its parameters have been optimized to a certain extent during the previous g-1 debugging process.

[0139] In actual operation, the computer system can use various types of temperature and humidity inference models, such as neural network models, time series models, etc. Taking the neural network model as an example, when the computer system inputs the g-th example tunnel temperature and humidity monitoring data stream into the neural network model determined by the g-1th debugging, the model will perform a series of calculations based on the input data, including weighted summation between neurons, application of activation functions, etc., and finally output the example inference temperature and humidity at the g+1th observation point.

[0140] Step S3012 requires the computer system to determine the debugging error between the example reasoning temperature and humidity of the g-th debugging and the example real temperature and humidity of the g-th debugging, and obtain the target debugging error of the g-th debugging, where the example real temperature and humidity of the g-th debugging is the real temperature and humidity of one of the multiple monitoring tunnels at the g+1-th observation point determined in advance. The target debugging error is an important indicator for measuring the accuracy of model prediction, and feasible calculation methods include mean square error (MSE), mean absolute error (MAE), etc.

[0141] In step S3013, when the target debugging error of the g-th debugging does not meet the preset error requirement, the computer system optimizes the model parameters in the temperature and humidity reasoning model determined by the g-1-th debugging to obtain the temperature and humidity reasoning model determined by the g-th debugging. The preset error requirement is a threshold set in advance by the computer system to determine whether the prediction accuracy of the model reaches an acceptable level. When the target debugging error is greater than this threshold, it means that the performance of the model is not good enough and needs further optimization.

[0142] Computer systems can use a variety of optimization algorithms to adjust model parameters, such as stochastic gradient descent (SGD), adaptive moment estimation (Adam), etc. Taking the stochastic gradient descent algorithm as an example, its basic idea is to update the model parameters along the negative gradient direction of the target debugging error to gradually reduce the error. Its update formula is:

[0143] ;

[0144] in, is the model parameter at the tth iteration, is the learning rate, which controls the step size of each parameter update. is the gradient of the target error with respect to the model parameters. For example, in a linear regression model In, θ 0 and θ 1 It is the parameter of the model. The computer system will calculate θ based on the target debugging error. 0 and θ 1 The gradient of , and then update the values ​​of these two parameters according to the above formula.

[0145] In step S3014, when the target debugging error of the g-th debugging meets the preset error requirement, the computer system stops debugging and determines the temperature and humidity reasoning model obtained by the g-1-th debugging as the target temperature and humidity reasoning model. This means that the prediction accuracy of the model has reached an acceptable level and can be used for actual temperature and humidity reasoning.

[0146] As an implementation mode, step S3011, loading the g-th example tunnel temperature and humidity monitoring data stream used in the g-th debugging into the temperature and humidity reasoning model determined in the g-1-th debugging, and obtaining the example reasoning temperature and humidity of the g-th debugging, includes:

[0147] Step S30111: Load the g-th example tunnel temperature and humidity monitoring data stream used in the g-th debugging into the temperature and humidity inference model determined in the g-1-th debugging, and obtain the inference temperature and humidity expectation and inference temperature and humidity discreteness of one monitoring tunnel in multiple monitoring tunnels at the g+1-th observation point;

[0148] Step S30112: Determine the example inference temperature and humidity for the g-th debugging according to the inference temperature and humidity expectation and the inference temperature and humidity dispersion of one monitoring tunnel among the multiple monitoring tunnels at the g+1-th observation point.

[0149] In step S30111, the computer system loads the g-th example tunnel temperature and humidity monitoring data stream used in the g-th debugging into the temperature and humidity inference model determined in the g-1-th debugging, and obtains the inference temperature and humidity expectation and inference temperature and humidity dispersion of one monitoring tunnel among multiple monitoring tunnels at the g+1-th observation point. The inference temperature and humidity expectation represents the average predicted value of the model for the temperature and humidity at the g+1-th observation point, which is the most likely temperature and humidity condition obtained based on the model's learning and analysis of historical data; the inference temperature and humidity dispersion reflects the uncertainty of the prediction result, and reflects the extent to which the actual temperature and humidity may deviate from the inference temperature and humidity expectation.

[0150] The computer system can use various types of models to perform calculations in this step, such as the Gaussian process regression model. In addition to the Gaussian process regression model, the computer system can also use the long short-term memory network (LSTM), a variant of the recurrent neural network (RNN), for calculations. LSTM can process time series data and capture long-term dependencies in the data. During the training process, the LSTM model learns the time-dependent pattern of the tunnel temperature and humidity data. When the computer system inputs the g-th example tunnel temperature and humidity monitoring data stream into the trained LSTM model, the model outputs a prediction result. In order to obtain the inference temperature and humidity expectation and the inference temperature and humidity dispersion, the computer system can use the Monte Carlo Dropout method. Specifically, when predicting, Dropout (a technique to prevent overfitting) is turned on multiple times for forward propagation to obtain multiple prediction values. The mean of these prediction values ​​can be used as the inference temperature and humidity expectation, and their standard deviation can be used as the inference temperature and humidity dispersion.

[0151] Step S30112 requires the computer system to determine the example reasoning temperature and humidity for the g-th debugging based on the reasoning temperature and humidity expectation and the reasoning temperature and humidity dispersion of one of the multiple monitoring tunnels at the g+1-th observation point. The purpose of this step is to comprehensively consider the average predicted value and uncertainty of the reasoning temperature and humidity to obtain a more reasonable example reasoning temperature and humidity.

[0152] The computer system can use a variety of methods to combine the inference temperature and humidity expectations and the inference temperature and humidity dispersion. One feasible method is to make certain adjustments to the inference temperature and humidity expectations according to the actual situation, and the adjustment range is related to the inference temperature and humidity dispersion. For example, the computer system can set an adjustment coefficient to increase or decrease the inference temperature and humidity expectations based on the size of the inference temperature and humidity dispersion.

[0153] In actual applications, the computer system determines the appropriate adjustment method and parameters according to different scenarios and needs. For example, in a tunnel environment with strict temperature and humidity requirements, such as a tunnel for storing precision instruments, the computer system may set a larger positive adjustment coefficient to ensure that the predicted temperature and humidity can cover the possible higher temperature and humidity conditions, and prepare countermeasures in advance; in scenarios with a higher tolerance for temperature and humidity changes, the adjustment coefficient can be set to a smaller value or even zero, and the predicted value closer to the average situation is used as an example to infer temperature and humidity.

[0154] The computer system may use a statistical analysis method based on historical data to determine the adjustment coefficient. By analyzing a large amount of historical data, the computer system may understand the relationship between the actual range of temperature and humidity changes and the expected and discrete values ​​of the inference temperature and humidity predicted by the model, thereby determining a suitable adjustment coefficient so that the example inference temperature and humidity calculated in step S30112 is closer to the actual value.

[0155] In addition, the computer system can also determine the adjustment method and parameters based on expert experience. Domain experts have an in-depth understanding of the changing laws of tunnel temperature and humidity and actual application needs. The computer system can consult experts and set the adjustment coefficient and select the appropriate adjustment method based on their suggestions.

[0156] Steps S30111-S30112 are important steps for the computer system to obtain the example reasoning temperature and humidity of the g-th debugging when the temperature and humidity reasoning model to be adjusted is debugged for the g-th time. By inputting the g-th example tunnel temperature and humidity monitoring data stream into the temperature and humidity reasoning model determined by the g-1-th debugging, the reasoning temperature and humidity expectation and the reasoning temperature and humidity dispersion are obtained, and the example reasoning temperature and humidity are determined based on these two indicators. The computer system can more accurately evaluate the performance of the model and provide a more reliable basis for further debugging of the model. In practical applications, the computer system selects a suitable model, determines reasonable adjustment methods and parameters, and combines historical data and expert experience to improve the accuracy and reliability of the example reasoning temperature and humidity, so as to better optimize the temperature and humidity reasoning model and provide strong support for the accurate prediction and monitoring of tunnel temperature and humidity. At the same time, continuously improving and perfecting the implementation methods of these two steps can further improve the overall level of tunnel temperature and humidity monitoring and analysis and meet the actual needs under different tunnel environments.

[0157] As an implementation manner, step S30112, based on the expected inference temperature and humidity and the discreteness of the inference temperature and humidity of one monitoring tunnel among the multiple monitoring tunnels at the g+1th observation point, determines the example inference temperature and humidity of the gth debugging, including:

[0158] Step S301121: The example inferred temperature and humidity of the g-th debugging is taken as equal to the sum of the expected inferred temperature and humidity of one monitoring tunnel among multiple monitoring tunnels at the g+1-th observation point and the example eccentricity adjustment result, wherein the example eccentricity adjustment result is the multiplication result of the preset discrete eccentricity coefficient and the discreteness of the inferred temperature and humidity of one monitoring tunnel among multiple monitoring tunnels at the g+1-th observation point.

[0159] Step S301121 requires the computer system to treat the example reasoning temperature and humidity of the g-th debugging as the sum of the expected reasoning temperature and humidity of one monitoring tunnel among multiple monitoring tunnels at the g+1-th observation point and the example eccentricity adjustment result, where the example eccentricity adjustment result is the multiplication result between the preset discrete eccentricity coefficient and the discreteness of the reasoning temperature and humidity of one monitoring tunnel among multiple monitoring tunnels at the g+1-th observation point. This step is for the computer system to determine the specific calculation method of the example reasoning temperature and humidity of the g-th debugging after obtaining the expected reasoning temperature and humidity and the discreteness of the reasoning temperature and humidity. The core of this step is to comprehensively consider the average temperature and humidity predicted by the model and the uncertainty of the prediction, and adjust the prediction result by introducing the discrete eccentricity coefficient to obtain the example reasoning temperature and humidity that is more in line with the actual situation.

[0160] The inference temperature and humidity expectation is the most likely temperature and humidity value at the g+1th observation point predicted by the temperature and humidity inference model determined by the g-1th debugging based on the gth example tunnel temperature and humidity monitoring data stream. It is a central estimate given by the model based on historical data and learned patterns. The inference temperature and humidity dispersion reflects the uncertainty of the prediction results, indicating that the actual temperature and humidity may fluctuate around the inference temperature and humidity expectation. The discrete eccentricity coefficient is a pre-set parameter used to control the adjustment of the inference temperature and humidity expectation based on the inference temperature and humidity dispersion. It reflects the computer system's handling of prediction uncertainty and its expectation of actual temperature and humidity deviation from expectations.

[0161] Expressed in terms of formula, the temperature and humidity inference of the g-th debugging example It can be expressed as: ,in is the expected value of temperature and humidity at the g+1th observation point, is the preset discrete eccentricity coefficient, is the inferred temperature and humidity dispersion at the g+1th observation point.

[0162] The value of the discrete eccentricity coefficient has an important impact on the calculation results of the example reasoning temperature and humidity. When the discrete eccentricity coefficient is a positive number, the example reasoning temperature and humidity will add an adjustment value related to the discreteness on the basis of the expected reasoning temperature and humidity, which means that the computer system is more inclined to consider the possibility of high temperature and humidity; when the discrete eccentricity coefficient is a negative number, the example reasoning temperature and humidity will be lower than the expected reasoning temperature and humidity, reflecting the concern about the possibility of low temperature and humidity; when the discrete eccentricity coefficient is zero, the example reasoning temperature and humidity is equal to the expected reasoning temperature and humidity, that is, the deviation caused by the uncertainty of the prediction is not considered.

[0163] The computer system can use a variety of technical means to determine the discrete eccentricity coefficient. One method is based on statistical analysis of historical data. The computer system can collect a large amount of historical tunnel temperature and humidity data and the corresponding model prediction results, and analyze the deviation between the actual temperature and humidity and the expected inference temperature and humidity. By calculating the mean, standard deviation and other statistics of these deviations, a suitable discrete eccentricity coefficient is determined so that the adjusted example inference temperature and humidity are closer to the actual value. For example, if historical data shows that the actual temperature and humidity are often higher than the expected inference temperature and humidity, then a positive discrete eccentricity coefficient can be set.

[0164] Another method is to combine expert experience. Domain experts have a deep understanding of the temperature and humidity variation patterns in tunnels and actual application scenarios, and they can provide reasonable discrete eccentricity coefficient recommendations based on actual conditions. For example, for a tunnel with poor ventilation conditions, experts may recommend setting a larger positive discrete eccentricity coefficient to cope with possible high temperature and high humidity conditions.

[0165] The computer system can also dynamically adjust the discrete eccentricity coefficient. The changing characteristics and uncertainty of temperature and humidity may change in different time periods, different weather conditions or different tunnel operation states. The computer system can monitor relevant parameters in real time, such as the ventilation volume in the tunnel, the traffic volume, external weather conditions, etc., and dynamically adjust the discrete eccentricity coefficient according to the changes in these parameters. For example, when the ventilation volume increases, the uncertainty of temperature and humidity may decrease, and the computer system can reduce the discrete eccentricity coefficient accordingly; when the traffic volume increases, the heat generated may increase the uncertainty of temperature and humidity, and the computer system will increase the discrete eccentricity coefficient.

[0166] Step S301121 provides a scientific and reasonable method for the computer system to determine the example reasoning temperature and humidity for the g-th debugging. By introducing the discrete eccentricity coefficient, the expected and discrete degree of reasoning temperature and humidity are comprehensively considered, making the prediction result more valuable for practical reference. By reasonably setting the discrete eccentricity coefficient and combining technical means such as historical data statistical analysis, expert experience and dynamic adjustment, the computer system can continuously optimize the calculation of example reasoning temperature and humidity, improve the accuracy and reliability of the temperature and humidity reasoning model, and provide more powerful support for the operation management and safety of the tunnel.

[0167] An embodiment of the present invention provides a computer system, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor implements part or all of the steps in the above method when executing the program.

[0168] The embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, some or all of the steps in the above method are implemented. The computer-readable storage medium may be transient or non-transient.

[0169] An embodiment of the present invention provides a computer program, including computer-readable codes. When the computer-readable codes are run in a computer device, a processor in the computer device executes some or all of the steps for implementing the above method.

[0170] An embodiment of the present invention provides a computer program product, the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and when the computer program is read and executed by a computer, some or all of the steps in the above method are implemented. The computer program product can be implemented specifically by hardware, software or a combination thereof. In some embodiments, the computer program product is specifically embodied as a computer storage medium, and in other embodiments, the computer program product is specifically embodied as a software product, such as a software development kit (SDK), etc.

[0171] It should be noted here that the description of the various embodiments above tends to emphasize the differences between the various embodiments, and the same or similar aspects can be referenced to each other. The description of the above device, storage medium, computer program and computer program product embodiments is similar to the description of the above method embodiment, and has similar beneficial effects as the method embodiment. For technical details not disclosed in the embodiments of the device, storage medium, computer program and computer program product of the present invention, please refer to the description of the method embodiment of the present invention for understanding.

[0172] Figure 2A hardware entity diagram of a computer system provided by an embodiment of the present invention is as follows Figure 2 As shown, the hardware entity of the computer system 1000 includes: a processor 1001 and a memory 1002, wherein the memory 1002 stores a computer program that can be run on the processor 1001, and the processor 1001 implements the steps in the method of any of the above embodiments when executing the program.

[0173] The memory 1002 stores computer programs that can be run on the processor. The memory 1002 is configured to store instructions and applications executable by the processor 1001. It can also cache data to be processed or processed by the processor 1001 and various modules in the computer system 1000 (for example, image data, audio data, voice communication data, and video communication data). This can be achieved through flash memory (FLASH) or random access memory (Random Access Memory, RAM).

[0174] When the processor 1001 executes the program, the steps of any of the above tunnel temperature and humidity monitoring and analysis methods are implemented. The processor 1001 generally controls the overall operation of the computer system 1000.

[0175] An embodiment of the present invention provides a computer storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the steps of the tunnel temperature and humidity monitoring and analysis method in any of the above embodiments.

[0176] It should be noted here that the description of the above storage medium and device embodiments is similar to the description of the above method embodiments, and has similar beneficial effects as the method embodiments. For technical details not disclosed in the storage medium and device embodiments of the present invention, please refer to the description of the method embodiments of the present invention for understanding. The above processor can be at least one of a target application integrated circuit (Application Specific Integrated Circuit, ASIC), a digital signal processor (Digital Signal Processor, DSP), a digital signal processing device (Digital Signal Processing Device, DSPD), a programmable logic device (Programmable Logic Device, PLD), a field programmable gate array (Field Programmable Gate Array, FPGA), a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, and a microprocessor. It can be understood that the electronic device that realizes the function of the above processor can also be other, and the embodiment of the present invention is not specifically limited.

[0177] The above-mentioned computer storage medium / memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory (Flash Memory), a magnetic surface memory, an optical disk, or a compact disc read-only memory (CD-ROM) and the like; it can also be various terminals including one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0178] It should be understood that the "one embodiment" or "an embodiment" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present invention. Therefore, "in one embodiment" or "in one embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in various embodiments of the present invention, the size of the sequence number of the above-mentioned steps / processes does not mean the order of execution, and the execution order of each step / process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention. The above-mentioned sequence number of the embodiment of the present invention is only for description and does not represent the advantages and disadvantages of the embodiment. It should be noted that in this article, the term "includes", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or device including the element.

[0179] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0180] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0181] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0182] A person skilled in the art can understand that all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, etc., various media that can store program codes.

[0183] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the relevant technology. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0184] The above description is only an implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A tunnel temperature and humidity monitoring and analysis method, characterized in that: include: Acquire an x-th tunnel temperature and humidity monitoring data stream of a target monitoring tunnel, wherein the target monitoring tunnel includes a plurality of tunnel sections, and the x-th tunnel temperature and humidity monitoring data stream includes the actual temperature and humidity of the target monitoring tunnel at an x-th group of observation points, wherein x≥1; Clustering the example tunnel temperature and humidity monitoring data stream set to obtain a data stream type library; or, when the example tunnel temperature and humidity monitoring data stream set includes K example tunnel temperature and humidity monitoring data streams, extracting a sub-data stream of equal length from each of the K example tunnel temperature and humidity monitoring data streams to obtain K sub-data streams, wherein each of the K sub-data streams includes the actual temperature and humidity of one monitoring tunnel among the multiple monitoring tunnels at s observation points; clustering the K sub-data streams to obtain a data stream type library; Determine a target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream, wherein the target data stream type is a data stream type in the data stream type library, and the plurality of monitoring tunnels include the target monitoring tunnel, or the target monitoring tunnel is different from the plurality of monitoring tunnels; Loading the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model corresponding to the target data stream type, obtaining the inference temperature and humidity of the target monitoring tunnel at the next observation point of the x-th group of observation points, wherein the target temperature and humidity inference model is a model obtained by calibrating the temperature and humidity inference model to be calibrated through a subset of example tunnel temperature and humidity monitoring data streams, and the example tunnel temperature and humidity monitoring data stream subset includes the example tunnel temperature and humidity monitoring data stream corresponding to the target data stream type in the set of example tunnel temperature and humidity monitoring data streams; The target temperature and humidity inference model is a Gaussian process regression model. Assuming that the x-th tunnel temperature and humidity monitoring data stream is X x , the target temperature and humidity reasoning model is expressed as: ； Where m(X) is the mean function and k(X, X') is the covariance function. Given a subset of the tunnel temperature and humidity monitoring data stream and input X x , calculate the expected temperature and humidity μ and the dispersion of the inference temperature and humidity at the next observation point of the xth group of observation points. The expected temperature and humidity μ is the most likely temperature and humidity value obtained based on the model's fitting and prediction of the data, while the dispersion of the inference temperature and humidity represents the uncertainty range around the expectation; The step of clustering the K sub-data streams to obtain a data stream type library includes: Determine the commonality measurement results between each sub-data stream in the K sub-data streams and each sub-data stream in the K sub-data streams respectively, and obtain a commonality measurement result array with a dimension of K*K, wherein the commonality measurement result between the e-th sub-data stream in the K sub-data streams and the f-th sub-data stream in the K sub-data streams is the commonality measurement result between the s real temperatures and humidity in the e-th sub-data stream and the s real temperatures and humidity in the f-th sub-data stream, and e and f are natural numbers not less than 1 and not greater than K; Based on the K*K commonality measurement result array, determine a classification indication array with a dimension of 1*K, wherein the 1*K classification indication array represents the data stream type corresponding to each sub-data stream in the K sub-data streams, and wherein the data stream type library includes different data stream types indicated by the classification indication array.

2. The method according to claim 1, characterized in that The determining of the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream includes: When the data stream type library includes A data stream types obtained by clustering the example tunnel temperature and humidity monitoring data stream set, and the A data stream types are matched one-to-one with A control monitoring data streams, determine the commonality measurement results of the x-th tunnel temperature and humidity monitoring data stream and each of the A control monitoring data streams, respectively, to obtain A commonality measurement results, wherein the d-th control monitoring data stream in the A control monitoring data streams is a data stream determined based on the example tunnel temperature and humidity monitoring data stream corresponding to the d-th data stream type in the example tunnel temperature and humidity monitoring data stream set, and the A data stream types include the d-th data stream type, wherein A≥2, 1≤d≤A; When the kth commonality measurement result among the A commonality measurement results is the largest and the kth commonality measurement result is not less than a reference value, the target data stream type is determined as the data stream type corresponding to the kth commonality measurement result among the A data stream types, wherein 1≤k≤A; Alternatively, the determining the target data stream type corresponding to the x-th tunnel temperature and humidity monitoring data stream includes: Clustering the example tunnel temperature and humidity monitoring data stream set and the x-th tunnel temperature and humidity monitoring data stream to obtain B data stream types, where M>1; The target data stream type is determined as the data stream type corresponding to the xth tunnel temperature and humidity monitoring data stream among the B data stream types.

3. The method according to claim 1, characterized in that The step of loading the x-th tunnel temperature and humidity monitoring data stream into a target temperature and humidity inference model corresponding to the target data stream type, and obtaining the inference temperature and humidity of the target monitoring tunnel at a subsequent observation point of the x-th group of observation points, comprises: Loading the x-th tunnel temperature and humidity monitoring data stream into the target temperature and humidity inference model, obtaining the inference temperature and humidity expectation and inference temperature and humidity dispersion of the target monitoring tunnel at the next observation point of the x-th group of observation points; The inferred temperature and humidity of the target monitoring tunnel at the latter observation point of the x-th group of observation points is determined according to the inferred temperature and humidity expectation and the inferred temperature and humidity dispersion.

4. The method according to claim 3, characterized in that Determining the inferred temperature and humidity of the target monitoring tunnel at the next observation point of the x-th group of observation points according to the inferred temperature and humidity expectation and the inferred temperature and humidity dispersion includes: The inferred temperature and humidity are taken as equal to the sum of the expected inferred temperature and humidity and a target eccentricity adjustment result, wherein the target eccentricity adjustment result is the multiplication result between a preset discrete eccentricity coefficient and the discreteness of the inferred temperature and humidity.

5. The method according to claim 1, characterized in that Before loading the x-th tunnel temperature and humidity monitoring data stream into a target temperature and humidity inference model corresponding to the target data stream type and obtaining the inference temperature and humidity of the target monitoring tunnel at a subsequent observation point of the x-th group of observation points, the method further includes: The temperature and humidity inference model to be calibrated is debugged through the example tunnel temperature and humidity monitoring data stream subset, and debugging is stopped until the target debugging error corresponding to the temperature and humidity inference model to be calibrated meets the preset error requirement, and the temperature and humidity inference model to be calibrated when debugging is stopped is used as the target temperature and humidity inference model, wherein the target debugging error is the debugging error between the example inference temperature and humidity and the example true temperature and humidity, the example inference temperature and humidity is the inference temperature and humidity of one of the multiple monitoring tunnels at an observation point determined by the temperature and humidity inference model to be calibrated based on the example tunnel temperature and humidity monitoring data stream subset loaded thereto, and the example true temperature and humidity is the true temperature and humidity of one of the multiple monitoring tunnels at the one observation point determined in advance.

6. The method according to claim 5, characterized in that The debugging of the temperature and humidity inference model to be calibrated by using the example tunnel temperature and humidity monitoring data stream subset includes: The temperature and humidity inference model to be calibrated is debugged for the gth time using the following steps, where g≥2: Loading the g-th example tunnel temperature and humidity monitoring data stream used in the g-th debugging into the temperature and humidity inference model determined in the g-1-th debugging, obtaining the example inference temperature and humidity of the g-th debugging, wherein the example inference temperature and humidity of the g-th debugging is the inference temperature and humidity of one of the multiple monitoring tunnels at the g+1-th observation point; Determine a debugging error between the example inference temperature and humidity of the g-th debugging and the example real temperature and humidity of the g-th debugging, and obtain the target debugging error of the g-th debugging, wherein the example real temperature and humidity of the g-th debugging is the real temperature and humidity of one of the multiple monitoring tunnels at the g+1-th observation point determined in advance; When the target debugging error of the g-th debugging does not meet the preset error requirement, optimizing the model parameter variables in the temperature and humidity reasoning model determined by the g-1-th debugging to obtain the temperature and humidity reasoning model determined by the g-th debugging; When the target debugging error of the g-th debugging meets the preset error requirement, debugging is stopped, and the temperature and humidity reasoning model obtained by the g-1-th debugging is determined as the target temperature and humidity reasoning model.

7. The method according to claim 6, characterized in that The step of loading the g-th example tunnel temperature and humidity monitoring data stream used in the g-th debugging into the temperature and humidity inference model determined in the g-1-th debugging to obtain the example inference temperature and humidity of the g-th debugging includes: Loading the g-th example tunnel temperature and humidity monitoring data stream used in the g-th debugging into the temperature and humidity inference model determined in the g-1-th debugging, and obtaining the inference temperature and humidity expectation and inference temperature and humidity dispersion of one monitoring tunnel among the multiple monitoring tunnels at the g+1-th observation point; The example inferred temperature and humidity for the g-th debugging is determined according to the inferred temperature and humidity expectation and the inferred temperature and humidity dispersion of one monitoring tunnel among the multiple monitoring tunnels at the g+1-th observation point.

8. The method according to claim 7, characterized in that The determining the example inferred temperature and humidity of the g-th debugging according to the inferred temperature and humidity expectation and the inferred temperature and humidity dispersion of one of the multiple monitoring tunnels at the g+1-th observation point includes: The example inferred temperature and humidity of the g-th debugging is taken as equal to the sum of the expected inferred temperature and humidity of one of the multiple monitoring tunnels at the g+1-th observation point and the example eccentricity adjustment result, wherein the example eccentricity adjustment result is the multiplication result of a preset discrete eccentricity coefficient and the discreteness of the inferred temperature and humidity of one of the multiple monitoring tunnels at the g+1-th observation point.

9. A computer system comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, wherein: When the processor executes the program, the steps in the method according to any one of claims 1 to 8 are implemented.

Images