Sluice engineering safety risk assessment early warning system

Abstract

The invention relates to the technical field of water gate maintenance, and discloses a water gate engineering safety risk assessment and early warning system which integrates five modules, namely a data analysis module, a real-time monitoring module, a state comparison module, a risk assessment module and an early warning notification module. The future safety state of the sluice project is predicted through a machine learning model, key safety parameters are monitored in real time, real-time data and predicted data are compared and analyzed, and a curve graph is drawn and visually displayed. And the risk assessment module calculates the difference degree of the two groups of curves and identifies risk points. And the early warning notification module comprehensively evaluates the safety state of the sluice project according to the risk point and the safety evaluation rule, and timely sends an evaluation result and early warning information to a manager so as to quickly take countermeasures. According to the system, the intelligent level of safety management of the water gate engineering is improved, early discovery and early warning of safety risks are realized, and safe operation of the water gate engineering is effectively guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of sluice maintenance, and in particular to a sluice engineering safety risk assessment and early warning system. Background Art

[0002] As an important part of water conservancy facilities, sluice projects undertake multiple tasks such as regulating water flow, flood control and drainage, irrigation and water supply, etc., and play a vital role in ensuring national water resource security and promoting sustainable economic and social development. However, since sluice projects are exposed to the natural environment for a long time, they are affected by various factors such as water erosion, geological changes, and human damage, and their safety status is always facing severe challenges. Traditional sluice project safety management mainly relies on regular inspections and maintenance, which is not only inefficient, but also difficult to detect and predict potential safety risks in a timely manner.

[0003] Most existing sluice project safety monitoring systems lack the ability to conduct in-depth data analysis, predict future safety status, and comprehensively assess risks. This is specifically reflected in the following aspects:

[0004] Low data utilization: Although sluice projects generate a large amount of historical operation data, environmental monitoring data and structural design parameters during operation, these data are often stored in isolation and are not effectively integrated and utilized, and cannot provide comprehensive and accurate information support for safety assessment.

[0005] Lack of predictive capabilities: The existing monitoring system mainly focuses on the collection and display of real-time data, and lacks the ability to predict future trends in security status changes, resulting in managers being unable to take preventive measures in advance to deal with possible security risks.

[0006] Incomplete risk assessment: Traditional risk assessment methods often rely on experience and judgment, lack a scientific and systematic assessment system, and are difficult to accurately identify potential safety hazards. The reliability and accuracy of the assessment results need to be improved.

[0007] Untimely early warning response: Due to the lack of an effective early warning mechanism, when safety anomalies occur in the sluice project, the management personnel are often unable to be notified in time, resulting in missing the best time to deal with the problem and causing unnecessary losses. Summary of the invention

[0008] The purpose of the present invention is to provide a sluice engineering safety risk assessment and early warning system to solve the problems raised in the above background technology.

[0009] To achieve the above object, the present invention provides the following technical solution: a sluice engineering safety risk assessment and early warning system, the system comprising:

[0010] The data analysis module is used to use machine learning models to combine the historical operation data, environmental monitoring data and structural design parameters of the sluice project to predict the trend of the safety status of the sluice project in the future;

[0011] Real-time monitoring module, used to monitor the real-time safety parameters of the sluice project, including data on water level, flow, gate opening, structural stress, seepage conditions and surrounding geological changes, and pre-process the original monitoring data to eliminate erroneous data and noise interference;

[0012] The status comparison module obtains the real-time safety parameter data output by the real-time monitoring module and the predicted safety status data output by the data analysis module, draws the real-time and predicted curves of each set of safety parameter data over time, and performs time synchronization processing on the two sets of curves and displays them in the same coordinate system;

[0013] The risk assessment module is used to calculate the difference between the two curves generated by each set of safety parameter data, and determine whether the difference exceeds the acceptable range based on the preset risk threshold. If it exceeds, it is marked as a risk point;

[0014] The early warning notification module is used to conduct a comprehensive assessment of the safety status of the sluice project based on the identified risk points and preset safety assessment rules, determine whether there are potential safety risks, and send the assessment results and early warning information to management personnel.

[0015] Preferably, the data analysis module includes:

[0016] Data integration unit, which is used to collect historical operation records, environmental monitoring reports, structural design documents and maintenance records of the sluice project to build a comprehensive database;

[0017] The data preprocessing unit is responsible for cleaning, format conversion and normalization of the data in the comprehensive database;

[0018] The feature mining unit extracts key data that affects the safety of the sluice gate from the preprocessed data and constructs a feature set;

[0019] The model building unit, based on the feature set and using machine learning algorithms, builds and trains a sluice project safety status prediction model to predict the changing trend of the future safety status.

[0020] Preferably, the data preprocessing unit uses a normalization algorithm to normalize the data so that data with different characteristics have similar scales; specifically, for each numerical data, the following formula is used for normalization:

[0021]

[0022] Where x is the original data, min(x) and max(x) are the minimum and maximum values ​​of the parameter data, respectively, and x' is the normalized data.

[0023] Preferably, the feature mining unit adopts the Pearson correlation coefficient method to calculate the linear correlation coefficient between each data item in the comprehensive database and the sluice safety parameter; specifically, for each pair of data (X, Y), where X is the data item and Y is the sluice safety parameter, the correlation analysis subunit calculates the Pearson correlation coefficient r between them, and the calculation formula is:

[0024]

[0025] Where n represents the number of data points, X i and Y i are the X and Y values ​​of the ith data point, and are the average values ​​of X and Y respectively;

[0026] Based on the calculated Pearson correlation coefficient, a correlation threshold is set, and data items with correlation coefficients with sluice safety parameters higher than the threshold are selected as key features to construct a feature set.

[0027] Preferably, the safety status prediction model adopts a long short-term memory network LSTM structure, which is composed of an input layer, multiple LSTM hidden layers and an output layer connected in sequence;

[0028] The input layer is responsible for receiving the multivariate time series data characteristics of the sluice project, including water level changes, flow data, gate operation records, structural stress monitoring values, and environmental factors;

[0029] The LSTM hidden layer consists of one or more layers of LSTM units. Each layer of LSTM units contains a forget gate, an input gate, and an output gate to capture and remember long-term dependencies and complex patterns in time series data.

[0030] The output layer predicts the changing trend of the safety status of the sluice project in the future through the fully connected layer based on the output of the LSTM hidden layer.

[0031] Preferably, the step of the model building unit training the safety status prediction model includes:

[0032] Step 1: Divide the feature set into a training set and a validation set for model training and performance evaluation;

[0033] Step 2: Use the training set data to iteratively train the LSTM model. In each iteration, a batch of input data is sent to the LSTM model, and the values ​​of the input layer, LSTM hidden layer, and output layer are calculated using the following formula, and the memory units and hidden states are updated to finally obtain the predicted output sequence:

[0034] Input layer: input data x i With the weight matrix W x and the bias vector b x Multiply to get the output of the input layer

[0035] LSTM hidden layer: according to the output of the input layer The hidden state h at the previous moment t-1 and the memory cell state c t-1 , calculate the hidden state h at the current moment through the forget gate, input gate and output gate of LSTM t and the memory cell state c t ;

[0036] Output layer: The output h of the LSTM hidden layer t With the weight matrix W y and the bias vector b y Multiply them together to get the predicted output sequence

[0037] During training, regularly use the validation set to evaluate model performance and adjust the learning rate or stop training early as needed to prevent overfitting.

[0038] Step 4: After training is completed, save the weights and biases of the LSTM model and deploy the model.

[0039] Preferably, during the training process of the safety status prediction model, the mean square error (MSE) algorithm is used to evaluate the model performance, and the specific implementation steps include:

[0040] Step 1: After each training iteration, use the validation set data to evaluate the performance of the LSTM model.

[0041] Step 2: Calculate the mean square error (MSE) between the predicted output of the LSTM model and the actual output of the validation set. The calculation formula for MSE is:

[0042]

[0043] Among them, N is the number of samples in the validation set, y i is the actual output value of the i-th sample in the validation set, is the predicted output value of the LSTM model for the i-th sample;

[0044] Step 3: Evaluate the performance of the LSTM model based on the calculated MSE value. The smaller the MSE value, the higher the prediction accuracy of the model and the better the performance.

[0045] Step 4: Based on the changing trend of the MSE value and the preset performance threshold, decide whether to continue training the model, adjust the model parameters, or stop training early.

[0046] Preferably, the risk assessment module calculates the difference between two curves in each set of safety parameter data, wherein the calculation of the difference includes the distance between the curves and the slope change rate.

[0047] Preferably, the risk assessment module presets one or more risk thresholds, and compares the calculated difference with the preset risk threshold to determine whether the difference exceeds an acceptable range; if the difference exceeds the preset risk threshold, the group of data is automatically marked as a risk point, and the detailed information of the risk point is recorded, including the difference value, the corresponding security parameters and the timestamp.

[0048] Preferably, the early warning notification module receives the risk point information marked by the risk assessment module; analyzes the received risk point information according to preset safety assessment rules, evaluates the overall safety status of the sluice project, determines whether there are potential safety risks, and generates an assessment result report.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] Through the data analysis module, this system integrates and utilizes the historical operation data, environmental monitoring data and structural design parameters of the sluice project, and builds a prediction model based on machine learning, so that a large amount of data that was originally stored in isolation can be effectively integrated and deeply mined, providing a comprehensive and accurate information basis for safety assessment, and significantly improving the efficiency and value of data utilization. Using machine learning technology, this system can predict the trend of changes in the safety status of the sluice project in the future. Through predictive analysis, managers can understand possible safety risks in advance, so as to take necessary preventive measures, effectively avoid safety accidents caused by lack of foresight, and improve the safety management level of the sluice project.

[0051] Through the status comparison module and risk assessment module, the system compares and analyzes the real-time monitoring data with the predicted data, scientifically calculates the difference between the two sets of curves, and makes risk judgments based on the preset risk threshold. Through the systematic assessment method, not only the accuracy and reliability of risk assessment are improved, but also potential safety hazards can be accurately identified, providing strong support for timely elimination of risks.

[0052] The early warning notification module can quickly send the identified risk points and comprehensive assessment results to managers in a variety of ways, ensuring the timely transmission of early warning information. This efficient early warning mechanism enables managers to grasp the safety status at the first time, respond quickly and take treatment measures, effectively avoiding losses caused by untimely early warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a working principle diagram of a sluice engineering safety risk assessment and early warning system according to the present invention;

[0054] Figure 2 A flowchart for building a safety status prediction model;

[0055] Figure 3 This is a workflow diagram for the risk assessment and early warning notification described in the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] See also Figure 1-3 The present invention provides a technical solution: a sluice engineering safety risk assessment and early warning system, the system comprising:

[0058] Data analysis module, the data analysis module is the core part of this system, it is responsible for using machine learning models to predict the trend of safety status changes of sluice projects. This module contains the following units:

[0059] Data Integration Unit: This unit is responsible for collecting various relevant data such as historical operation records, environmental monitoring reports, structural design documents, and maintenance records of the sluice project, and constructing a comprehensive database. These data will serve as the basis for subsequent analysis.

[0060] Data preprocessing unit: cleans the data in the comprehensive database to remove invalid and erroneous data records; performs format conversion to ensure that all data is stored in a unified format; and performs normalization to eliminate dimensional differences between data, providing a standardized data set for subsequent feature extraction and model training.

[0061] Feature mining unit: Based on the preprocessed data, this unit uses statistical methods and data mining techniques to extract key features that affect the safety of the sluice, such as water level change rate, structural stress peak, seepage rate, etc., and constructs a feature set as model input.

[0062] Model building unit: Using machine learning algorithms, a sluice project safety status prediction model is built and trained based on the extracted feature set. Through model training, the complex relationship between the safety status of the sluice project and various influencing factors is learned, and the changing trend of the future safety status is predicted.

[0063] Real-time monitoring module, which is responsible for collecting real-time safety parameters of the sluice project, including water level, flow, gate opening, structural stress, seepage conditions and surrounding geological changes. This module pre-processes the original monitoring data, including data verification, error data elimination and noise filtering, to ensure the accuracy and reliability of the data.

[0064] The status comparison module receives the real-time safety parameter data output by the real-time monitoring module and the predicted safety status data output by the data analysis module. For each set of safety parameters, the module draws the real-time curve and the predicted curve over time, and performs time synchronization to ensure that the two sets of curves are aligned on the same time axis. Subsequently, the two sets of curves are displayed in the same coordinate system, which makes it easier for managers to intuitively compare the difference between the real-time status and the predicted status.

[0065] Risk assessment module, the risk assessment module calculates the difference between the real-time curve and the predicted curve generated by each set of safety parameter data. The difference can be calculated using a variety of mathematical methods, such as the Euclidean distance between curves, slope difference or area difference, etc. According to the preset risk threshold, the module determines whether the difference exceeds the acceptable range. If it exceeds, the group of data is marked as a risk point, and the detailed information of the risk point is recorded, including the difference value, the corresponding safety parameter, the timestamp, etc.

[0066] Early warning notification module, based on the risk points identified by the risk assessment module and the preset safety assessment rules, the early warning notification module conducts a comprehensive assessment of the safety status of the sluice project. The assessment results include whether there are potential safety risks and the severity of the risks. This module sends the assessment results and early warning information to designated managers via email, SMS, system messages or other communication methods to ensure that they can receive them in time and take appropriate measures.

[0067] The present invention will be further described below in conjunction with Examples 1 to 3:

[0068] Embodiment 1:

[0069] The data preprocessing unit includes normalizing the numerical data in the comprehensive database to ensure that data with different characteristics have similar scales, thereby eliminating the impact of dimensional differences on subsequent analysis. The specific implementation method is as follows:

[0070] For each numerical data item in the comprehensive database, the following formula is used for normalization:

[0071]

[0072] Where x is the original data, min(x) and max(x) are the minimum and maximum values ​​of the parameter data, respectively, and x' is the normalized data.

[0073] For example, suppose the historical operation data of a sluice project includes the data item of water level, and its minimum value in a certain period of time is 1 meter and the maximum value is 10 meters. If the water level at a certain moment is 5 meters, then after normalization according to the above formula, the water level value obtained is:

[0074]

[0075] Through such processing, all water level data will be mapped to the interval [0,1], achieving data normalization.

[0076] The feature mining unit is responsible for extracting key features that affect the safety of the sluice gate from the preprocessed data. To achieve this goal, the unit uses the Pearson correlation coefficient method to calculate the linear correlation coefficient between each data item in the comprehensive database and the sluice gate safety parameters. The specific implementation method is as follows:

[0077] For each pair of data items (X, Y) in the comprehensive database, where X is the data item and Y is the safety parameter of the sluice (such as structural stress, seepage rate, etc.), the correlation analysis subunit calculates the Pearson correlation coefficient r between them, and the calculation formula is:

[0078]

[0079] Where n represents the number of data points, X i and Y i are the X and Y values ​​of the ith data point, and are the means of X and Y respectively.

[0080] For example, suppose you want to analyze the correlation between water level and structural stress. First, calculate their respective average values. and Then, the Pearson correlation coefficient r is calculated according to the above formula. If the calculated r value is close to 1 or -1, it means that there is a strong linear correlation between the water level and the structural stress; if the r value is close to 0, it means that there is almost no linear correlation between them.

[0081] Based on the calculated Pearson correlation coefficient, a correlation threshold (such as 0.5 or -0.5) is set, and data items with correlation coefficients with sluice safety parameters higher than the threshold are selected as key features to construct a feature set. These key features will be used as inputs to the machine learning model to predict the trend of safety status changes in sluice projects.

[0082] Embodiment 2:

[0083] In the sluice engineering safety risk assessment and early warning system of the present invention, the safety state prediction model adopts a long short-term memory network (LSTM) structure to capture and memorize the long-term dependencies and complex patterns of time series data. The construction process of the model includes:

[0084] ①Model structure design:

[0085] Input layer: input data x i With the weight matrix W x and the bias vector b x Multiply to get the output of the input layer The input layer is responsible for receiving multivariate time series data features of the sluice project, including but not limited to water level changes, flow data, gate operation records, structural stress monitoring values, and environmental factors (such as temperature, rainfall, etc.). These data features are organized into a suitable format in the input layer and are ready to be sent to the LSTM model for processing.

[0086] LSTM hidden layer: according to the output of the input layer The hidden state h at the previous moment t-1 and the memory cell state c t-1 , calculate the hidden state h at the current moment through the forget gate, input gate and output gate of LSTM t and the memory cell state c t ; The LSTM hidden layer consists of one or more layers of LSTM units, and each layer of LSTM units contains three key components: forget gate, input gate, and output gate. The forget gate determines how much of the memory unit state at the previous moment needs to be forgotten; the input gate determines how much of the input at the current moment needs to be written into the memory unit; the output gate generates the output at the current moment based on the current memory unit state and hidden state. These gate structures work together on the memory unit, enabling it to capture and remember long-term dependencies in time series data.

[0087] Output layer: The output h of the LSTM hidden layer t With the weight matrix W y and the bias vector b y Multiply them together to get the predicted output sequence According to the output of the LSTM hidden layer, it is further processed through the fully connected layer (usually one or more fully connected neural network layers) to finally predict the trend of the safety status of the sluice project in the future. The output layer can be a continuous value (such as the score or probability of the safety status) or a discrete category label (such as "safe", "warning", "danger", etc.).

[0088] ②Model training process:

[0089] Data partitioning: First, the feature set is divided into a training set and a validation set. The training set is used for model training, and the validation set is used to evaluate model performance and adjust model parameters. The partition ratio can be determined according to the actual situation. Usually, the training set accounts for a larger proportion (such as 70%-80%), and the validation set accounts for a smaller proportion (such as 20%-30%).

[0090] Model initialization: Initialize the weights and biases of the LSTM model, usually using random initialization or pre-training initialization. At the same time, set training parameters such as learning rate, number of iterations, batch size, etc.

[0091] Iterative training:

[0092] Step 3.1: Take a batch of data from the training set as input for the current iteration.

[0093] Step 3.2: Input layer processing: Multiply the input data with the weight matrix and bias vector to get the output of the input layer.

[0094] Step 3.3: LSTM hidden layer processing: Based on the output of the input layer, the hidden state and memory cell state at the previous moment, the current hidden state and memory cell state are calculated through the LSTM forget gate, input gate, and output gate. This process is performed inside each LSTM unit and passed layer by layer.

[0095] Step 3.4: Output layer processing: Multiply the output of the LSTM hidden layer with the weight matrix and bias vector to obtain the predicted output sequence.

[0096] Step 3.5: Calculate the loss function: Calculate the loss function value (such as mean square error) based on the predicted output sequence and the actual safety status data.

[0097] Step 3.6: Back propagation and parameter update: Calculate the gradient through the back propagation algorithm, and update the weights and bias of the model according to the gradient.

[0098] Performance evaluation and adjustment: After each iteration or a certain number of iterations, use the validation set to evaluate the performance of the model (such as accuracy, F1 score, etc.). According to the performance evaluation results, adjust the learning rate, increase or decrease the number of LSTM hidden layers, change the batch size and other parameters to optimize the model performance.

[0099] Training termination and model saving: When the model performance reaches the expected level or no longer improves significantly, stop training. Save the weights and biases of the trained LSTM model for subsequent deployment and prediction.

[0100] Taking water level changes and structural stress monitoring values ​​as input features as an example, assume that there are 1,000 historical data records, each of which contains water level changes and structural stress monitoring values ​​for a consecutive week. First, divide these data into a training set (800 records) and a validation set (200 records). Then, build a model with two LSTM hidden layers, each with 128 LSTM units. Set the learning rate to 0.001, the number of iterations to 1,000, and the batch size to 64. During the training process, regularly use the validation set to evaluate the model performance and adjust the learning rate as needed. Finally, after the training is completed, save the weights and biases of the model.

[0101] During the training process of the safety status prediction model, the mean squared error (MSE) algorithm is used to evaluate the model performance to ensure the prediction accuracy and stability of the model. The following is a detailed description of the implementation of the evaluation method:

[0102] Step 1: Validation set preparation: Before model training begins, the dataset is divided into a training set and a validation set. The validation set is used to evaluate the performance of the model after each training iteration to ensure that the model not only performs well on the training data, but also maintains generalization capabilities on unseen data.

[0103] Step 2: Iteration cycle and performance evaluation:

[0104] After each training iteration (usually including multiple batches of data processing), the training process is paused and the performance of the LSTM model is evaluated using the validation set data. This step is intended to monitor the performance changes of the model during training and prevent overfitting or underfitting.

[0105] Step 3: Calculate the mean square error (MSE): For each sample in the validation set, calculate the square of the difference between the predicted output of the LSTM model and the actual output. Then, sum these squared differences and divide them by the number of samples in the validation set, N, to get the mean square error (MSE). The calculation formula for MSE is:

[0106]

[0107] Among them, N is the number of samples in the validation set, y i is the actual output value of the i-th sample in the validation set, is the predicted output value of the LSTM model for the i-th sample.

[0108] Step 4: Performance evaluation and decision making:

[0109] Evaluate the performance of the LSTM model based on the calculated MSE value. The smaller the MSE value, the higher the prediction accuracy of the model and the better the performance. Monitor the trend of the MSE value. If the MSE value continues to decrease, it means that the model is gradually being optimized; if the MSE value begins to increase, it may indicate that the model is beginning to overfit. Set a preset performance threshold. When the MSE value reaches or falls below the threshold, it can be considered that the model performance meets the requirements, and consider stopping training or making further fine-tuning. If the MSE value is higher than expected, or the performance improvement is not obvious, you need to adjust the model parameters (such as increasing the number of LSTM hidden layers, changing the learning rate, etc.), or increase the number of training iterations to further improve the model performance.

[0110] Embodiment 3:

[0111] In the sluice project safety risk assessment and early warning system, the risk assessment module and early warning notification module are used to realize real-time analysis of safety parameter data and risk early warning. The following is a detailed description of the specific implementation methods of these two modules:

[0112] ①Risk Assessment Module:

[0113] Difference calculation: For each set of safety parameter data (such as water level, flow, structural stress, etc.), the risk assessment module first calculates the difference between the two curves (actual monitoring curve and expected or historical curve). The difference calculation takes into account the distance between the curves and the slope change rate.

[0114] The distance between curves can be calculated using common distance measurement methods such as Euclidean distance and Manhattan distance, which are used to measure the degree of deviation between two curves in terms of values.

[0115] The calculation of the slope change rate focuses on the similarity of the trends of the two curves, and evaluates whether the changing trends of the curves are consistent by comparing the slope differences at each point.

[0116] Risk threshold setting and comparison: The risk assessment module presets one or more risk thresholds, which are derived from the safety standards of the sluice project and the statistical analysis of historical data to define the acceptable range of differences. For each set of safety parameter data, the risk assessment module compares the calculated difference with the preset risk threshold. If the difference exceeds any preset threshold, the group of data is automatically marked as a risk point.

[0117] Risk point information record: For data marked as risk points, the risk assessment module records the difference value, corresponding safety parameters (such as water level, flow, etc.) and timestamp in detail for subsequent analysis and tracing.

[0118] Assume that in a certain monitoring, the difference between the water level monitoring curve and the expected curve is calculated to be 0.15, and the preset risk threshold is 0.1. Since 0.15>0.1, this group of water level data is marked as a risk point, and the difference value 0.15, water level parameters and current timestamp are recorded.

[0119] ② Early warning notification module:

[0120] Receiving risk point information: The early warning notification module receives the risk point information marked by the risk assessment module in real time, including difference value, safety parameters and timestamp.

[0121] Application of safety assessment rules: The early warning notification module has a set of built-in safety assessment rules. These rules are based on the safety standards and historical experience of the sluice project and are used to analyze and evaluate the received risk point information. The number of risk points of a single safety parameter reaches a certain threshold, risk points appear at the same time in multiple safety parameters, and the difference value of risk points continues to increase. According to the safety assessment rules, the early warning notification module conducts a comprehensive analysis of the received risk point information and evaluates the overall safety status of the sluice project.

[0122] If the assessment results show that there are potential safety risks, the early warning notification module will generate an assessment result report, which includes detailed information on risk points, overall safety status assessment, recommended measures, etc. After the assessment result report is generated, the early warning notification module will send the report to relevant managers or security monitoring systems through preset communication channels (such as SMS, email, system notification, etc.) so that timely response measures can be taken to ensure the safe operation of the sluice project.

[0123] Continuing with the risk point example of water level monitoring above, after the early warning notification module receives the risk point information, it determines that the water level difference exceeds the standard according to the safety assessment rules, which may indicate a potential water level control problem. Combined with the monitoring of other safety parameters (such as flow, structural stress, etc.), the early warning notification module assesses that there is currently a risk of abnormal water level increase in the sluice project and generates an assessment result report. The report points out the specific information of the risk point and recommends strengthening water level monitoring and control to prevent safety hazards of the sluice project caused by excessively high water levels. Subsequently, the early warning notification module sends the report to the managers of the sluice project via email so that they can take timely measures to deal with potential risks.

[0124] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0125] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A sluice project safety risk assessment and early warning system, characterized in that: The system comprises: The data analysis module is used to use machine learning models to combine the historical operation data, environmental monitoring data and structural design parameters of the sluice project to predict the trend of the safety status of the sluice project in the future; Real-time monitoring module, used to monitor the real-time safety parameters of the sluice project, including data on water level, flow, gate opening, structural stress, seepage conditions and surrounding geological changes, and pre-process the original monitoring data to eliminate erroneous data and noise interference; The status comparison module obtains the real-time safety parameter data output by the real-time monitoring module and the predicted safety status data output by the data analysis module, draws the real-time and predicted curves of each set of safety parameter data over time, and performs time synchronization processing on the two sets of curves and displays them in the same coordinate system; The risk assessment module is used to calculate the difference between the two curves generated by each set of safety parameter data, and determine whether the difference exceeds the acceptable range based on the preset risk threshold. If it exceeds, it is marked as a risk point; The early warning notification module is used to conduct a comprehensive assessment of the safety status of the sluice project based on the identified risk points and preset safety assessment rules, determine whether there are potential safety risks, and send the assessment results and early warning information to management personnel.

2. A sluice engineering safety risk assessment and early warning system according to claim 1, characterized in that: The data analysis module includes: Data integration unit, which is used to collect historical operation records, environmental monitoring reports, structural design documents and maintenance records of the sluice project to build a comprehensive database; The data preprocessing unit is responsible for cleaning, format conversion and normalization of the data in the comprehensive database; The feature mining unit extracts key data that affects the safety of the sluice gate from the preprocessed data and constructs a feature set; The model building unit, based on the feature set and using machine learning algorithms, builds and trains a sluice project safety status prediction model to predict the changing trend of the future safety status.

3. A sluice engineering safety risk assessment and early warning system according to claim 2, characterized in that: The data preprocessing unit uses a normalization algorithm to normalize the data so that data with different characteristics have similar scales; specifically, for each numerical data, the following formula is used for normalization: Where x is the original data, min(x) and max(x) are the minimum and maximum values ​​of the parameter data, respectively, and x' is the normalized data.

4. A sluice engineering safety risk assessment and early warning system according to claim 3, characterized in that: The feature mining unit uses the Pearson correlation coefficient method to calculate the linear correlation coefficient between each data item in the comprehensive database and the sluice safety parameter; specifically, for each pair of data (X, Y), where X is the data item and Y is the sluice safety parameter, the correlation analysis subunit calculates the Pearson correlation coefficient r between them, and the calculation formula is: Where n represents the number of data points, X i and Y i are the X and Y values ​​of the ith data point, and are the average values ​​of X and Y respectively; Based on the calculated Pearson correlation coefficient, a correlation threshold is set, and data items with correlation coefficients with sluice safety parameters higher than the threshold are selected as key features to construct a feature set.

5. A sluice engineering safety risk assessment and early warning system according to claim 4, characterized in that: The safety status prediction model adopts a long short-term memory network LSTM structure, which is composed of an input layer, multiple LSTM hidden layers and an output layer connected in sequence; The input layer is responsible for receiving the multivariate time series data characteristics of the sluice project, including water level changes, flow data, gate operation records, structural stress monitoring values, and environmental factors; The LSTM hidden layer consists of one or more layers of LSTM units. Each layer of LSTM units contains a forget gate, an input gate, and an output gate to capture and remember long-term dependencies and complex patterns in time series data. The output layer predicts the changing trend of the safety status of the sluice project in the future through the fully connected layer based on the output of the LSTM hidden layer.

6. A sluice engineering safety risk assessment and early warning system according to claim 5, characterized in that: The step of training the safety status prediction model by the model building unit includes: Step 1: Divide the feature set into a training set and a validation set for model training and performance evaluation; Step 2: Use the training set data to iteratively train the LSTM model. In each iteration, a batch of input data is sent to the LSTM model, and the values ​​of the input layer, LSTM hidden layer, and output layer are calculated using the following formula, and the memory cells and hidden states are updated to finally obtain the predicted output sequence: Input layer: input data x i With the weight matrix W x and the bias vector b x Multiply to get the output of the input layer LSTM hidden layer: according to the output of the input layer The hidden state h at the previous moment t-1 and the memory cell state c t-1 , calculate the hidden state h at the current moment through the forget gate, input gate and output gate of LSTM t and the memory cell state c t ; Output layer: The output h of the LSTM hidden layer t With the weight matrix W y and the bias vector b y Multiply them together to get the predicted output sequence During training, regularly use the validation set to evaluate model performance and adjust the learning rate or stop training early as needed to prevent overfitting. Step 4: After training is completed, save the weights and biases of the LSTM model and deploy the model.

7. A sluice engineering safety risk assessment and early warning system according to claim 6, characterized in that: During the training process of the safety status prediction model, the mean square error (MSE) algorithm is used to evaluate the model performance. The specific implementation steps include: Step 1: After each training iteration, use the validation set data to evaluate the performance of the LSTM model. Step 2: Calculate the mean square error (MSE) between the predicted output of the LSTM model and the actual output of the validation set. The calculation formula for MSE is: Among them, N is the number of samples in the validation set, y i is the actual output value of the i-th sample in the validation set, is the predicted output value of the LSTM model for the i-th sample; Step 3: Evaluate the performance of the LSTM model based on the calculated MSE value. The smaller the MSE value, the higher the prediction accuracy of the model and the better the performance. Step 4: Based on the changing trend of the MSE value and the preset performance threshold, decide whether to continue training the model, adjust the model parameters, or stop training early.

8. A sluice engineering safety risk assessment and early warning system according to claim 1, characterized in that: The risk assessment module calculates the difference between two curves in each set of safety parameter data, wherein the calculation of the difference includes the distance between the curves and the slope change rate.

9. A sluice engineering safety risk assessment and early warning system according to claim 8, characterized in that: The risk assessment module presets one or more risk thresholds, and compares the calculated difference with the preset risk threshold to determine whether the difference exceeds an acceptable range; if the difference exceeds the preset risk threshold, the group of data is automatically marked as a risk point, and the detailed information of the risk point is recorded, including the difference value, the corresponding security parameters and the timestamp.

10. A sluice engineering safety risk assessment and early warning system according to claim 9, characterized in that: The early warning notification module receives the risk point information marked by the risk assessment module; According to the preset safety assessment rules, the received risk point information is analyzed, the overall safety status of the sluice project is evaluated, whether there are potential safety risks, and an assessment result report is generated.

Images