Hydraulic engineering data intelligent monitoring method and system based on digital twinning

Abstract

The invention relates to the technical field of hydraulic engineering, in particular to a hydraulic engineering data intelligent monitoring method and system based on digital twinning. The method comprises the following steps: collecting multiple dimension data at the current moment in a target water area in real time, and forming a current data point; inputting the current data point into the network model, and outputting a corresponding first abnormal probability; acquiring data of multiple dimensions of multiple moments before the current moment, wherein the same dimension comprises data of all sampling moments of the current moment to form a sequence; performing frequency domain analysis on each dimension sequence, determining the abnormal probability of the data at the current moment in the corresponding dimension, and taking the maximum value of the abnormal probability in all dimensions at the current moment as a second abnormal probability; integrating the first abnormal probability and the second abnormal probability to obtain a comprehensive probability; if the comprehensive probability is greater than a set threshold value, determining that the current data point is abnormal, and notifying a worker to take measures; according to the scheme, the hydraulic engineering data can be accurately monitored.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology. More specifically, the present invention relates to a method and system for intelligent monitoring of water conservancy project data based on digital twins. Background Art

[0002] Water conservancy projects (such as dams, reservoirs, levees, and canals) are critical infrastructure for water resource management, flood control, power generation, and irrigation. These projects often reside in complex natural environments and are affected by factors such as water pressure, temperature fluctuations, geological activity, and material aging. These projects may exhibit subtle anomalies, such as cracks, leaks, deformation, and subsidence. If these subtle changes are not identified promptly, they can become major safety hazards, leading to project failure or even catastrophic consequences. Therefore, identifying these subtle anomalies is crucial in the monitoring and maintenance of water conservancy projects.

[0003] At present, with the rapid development of the Internet of Things, cloud computing and big data technologies, digital twin technology has gradually become an important means to improve the management efficiency of water conservancy projects. By creating virtual models of physical water conservancy facilities and monitoring and analyzing them in real time, digital twins can effectively integrate various sensor data and reflect information such as hydrological, meteorological and water resources changes in real time.

[0004] Through high-precision sensors and remote sensing technology, the operation data of water conservancy projects, including key parameters such as water level, flow rate, and pressure, are collected in real time. Using this data to establish a digital twin model, the physical status of water conservancy facilities can be dynamically reflected. The digital twin model can provide real-time monitoring capabilities, allowing managers to check the operating status of water conservancy facilities at any time.

[0005] Through data analysis algorithms, potential faults and risks can be identified, early warning capabilities can be improved, and the safe and stable operation of facilities can be ensured. Environmental data such as meteorology and hydrology can be integrated into the digital twin model to enhance the ability to adapt to changes in external factors. It can simulate engineering performance under different environmental conditions and support decision makers in formulating more scientific management strategies.

[0006] Although existing technologies can perform real-time monitoring of water conservancy project data, they still have the following problems:

[0007] 1) Data collection accuracy limitations: Current sensors and other data collection equipment have limited accuracy, making it difficult to accurately capture some subtle changes in water conservancy projects.

[0008] 2) Data transmission and processing delays: In a water conservancy project's digital twin system, data must be collected from the physical entity and transferred to the digital model for processing and analysis. This process can be subject to certain delays. For example, the speed of data transmission and processing may not keep up with the speed of change in the actual physical process. Consequently, analysis of water conservancy project data may not reflect the current operating status in real time, affecting the timeliness of decision-making. Summary of the Invention

[0009] The purpose of the present invention is to propose a method and system for intelligent monitoring of water conservancy project data based on digital twins, so as to solve the problem in the existing technology that it is difficult to reflect the status changes of water conservancy projects in real time, and thus it is impossible to accurately carry out intelligent monitoring of water conservancy projects; to this end, the present invention provides solutions in the following two aspects.

[0010] In a first aspect, the present invention provides a method for intelligent monitoring of water conservancy project data based on digital twins, comprising:

[0011] Collect multiple dimensional data at the current moment in the target water area in real time and form the current data point; the multiple dimensional data include dam leakage, ambient temperature, rainfall, water flow, water pressure, and water level;

[0012] Input the current data point into the network model and output the corresponding first abnormality probability;

[0013] Obtain data from multiple dimensions at multiple moments before the current moment. The same dimension contains data from all sampling moments at the current moment to form a sequence. Perform frequency domain analysis on each dimension sequence to determine the abnormal probability of the data at the current moment in the corresponding dimension. The maximum abnormal probability among all dimensions at the current moment is used as the second abnormal probability.

[0014] integrating the first abnormal probability and the second abnormal probability to obtain a comprehensive probability;

[0015] In response to the comprehensive probability being greater than the set threshold, there is an anomaly in the data point at the current moment, and the staff needs to be notified to take measures.

[0016] The above scheme approaches this from two perspectives: one is time-domain analysis, which uses a network model to analyze data to determine the anomaly probability of a data point at each sampling moment; the other is frequency-domain analysis, which captures subtle anomaly changes. Ultimately, the data from these two perspectives is integrated to obtain the final comprehensive probability of the water conservancy project data, thereby determining anomalies at each sampling moment. This approach, by analyzing anomalies from two perspectives, can accurately determine anomalies in water conservancy project data and improve monitoring reliability.

[0017] Optionally, the second weight of the integrated second abnormality probability is:

[0018] Perform STL decomposition on the dimension sequence corresponding to the second anomaly probability to obtain the sum of the seasonal component and residual component under each dimension;

[0019] The second weight is negatively correlated with the sum of the seasonal component and the residual;

[0020] The sum of the first weight and the second weight of the first abnormality probability is 1.

[0021] The above solution adaptively obtains the weights of the first abnormal probability and the second abnormal probability according to the characteristics of the data point at the current moment, which can make the integrated probability after fusion more accurate.

[0022] Optionally, the process of obtaining the abnormality probability is:

[0023] The time-frequency diagram is obtained by processing the sequences of each dimension through short-time Fourier transform;

[0024] According to the absolute value of the difference between the target sampling moment of the target frequency on the time-frequency diagram and the sampling spectrum amplitude of the previous adjacent moment and the mean spectrum amplitude of the target frequency, an abnormality index is obtained, and the abnormality index is normalized to obtain the abnormality probability; the target sampling moment is any sampling moment.

[0025] By analyzing from a frequency domain perspective, it is possible to evaluate abnormal conditions in the target waters and ensure the stability and safety of the system.

[0026] Optionally, the abnormality indicator is the mean of the global deviation and the local deviation;

[0027]

[0028] in, is the global deviation, is the local deviation; is the frequency amplitude mean of the kth frequency on the time-frequency diagram; f k (n) is the spectrum amplitude of the kth frequency at the nth moment on the time-frequency diagram, f k (n-1) is the spectrum amplitude of the kth frequency at the n-1th moment.

[0029] The above solution provides a method for accurately obtaining abnormal indicators.

[0030] Optionally, the network model is a fully connected network model.

[0031] Optionally, the training process of the network model is:

[0032] Obtaining a training set, wherein the training set includes historical data points at multiple moments in a historical period; labels are normal or abnormal conditions of the data points at each moment;

[0033] The network model is trained using the training set to obtain a trained network model.

[0034] Optionally, the method further includes: performing time alignment processing on the dimensional data in the historical data points at each moment to obtain aligned historical data points.

[0035] The above solution provides data support for obtaining an accurate network model by aligning data of various dimensions in time.

[0036] Optionally, the normalization process adopts a softmax function or a maximum-minimum normalization method.

[0037] Optionally, the dam leakage, water flow, water pressure and water level are measured by osmometer, flowmeter, water pressure gauge and water level gauge respectively; the ambient temperature and rainfall are obtained by meteorological data.

[0038] In a second aspect, a water conservancy project data intelligent monitoring system based on digital twins includes:

[0039] processor;

[0040] A memory stores computer instructions for intelligent monitoring of water conservancy project data based on digital twins. When the computer instructions are executed by the processor, the system executes the above-mentioned intelligent monitoring method of water conservancy project data based on digital twins.

[0041] The beneficial effects of the present invention are:

[0042] The solution of the present invention can accurately analyze water conservancy project data from the perspectives of time domain and frequency domain, thereby improving the reliability of monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 The following schematically shows a flowchart of the steps of a method for intelligent monitoring of water conservancy project data based on digital twins in this embodiment;

[0044] Figure 2 The structural block diagram of a water conservancy project data intelligent monitoring system based on digital twin in this embodiment is schematically shown. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0046] The present invention is aimed at monitoring the water conservancy project data of the target reservoir, that is, monitoring the abnormal situation of the water conservancy project data based on digital twin technology.

[0047] Among them, digital twin technology is a cutting-edge technology that combines physical entities with virtual digital models. It achieves accurate simulation, optimization and prediction of physical entities through real-time data interaction and advanced analysis.

[0048] The technical architecture of digital twins includes the physical layer, data layer, analysis layer, model layer, and application layer. As the above technical architecture is existing technology, it will not be described in detail here.

[0049] In the following, in combination with digital twin technology, a digital twin-based intelligent monitoring method for water conservancy project data of the present invention is specifically introduced.

[0050] The digital twin system in this embodiment includes:

[0051] Physical layer: IoT sensors are deployed at monitoring points within the target reservoir area, such as water level gauges, piezometers, accelerometers, and flow meters, to collect real-time data from the corresponding monitoring points, including water level, pressure, flow, vibration, etc.

[0052] Data layer: Pre-process sensor data (such as denoising) through edge computing nodes and upload it to the cloud-based digital twin platform.

[0053] Analysis layer: Process and analyze the collected data to detect and predict data anomalies.

[0054] Model layer: Build a digital twin model of the water conservancy project and use computer simulation technology to simulate and predict the behavior and performance of physical entities.

[0055] The digital twin model in this embodiment includes a geometric model (3D modeling) and a physical model (water level change model, hydrodynamics model, and dam structure model) to simulate water conservancy facilities and realize dynamic simulation of actual facilities.

[0056] Application layer: Provides visual interface, early warning notifications and optimization suggestions (such as scheduling strategies and maintenance plans).

[0057] It should be noted that the solution of this embodiment focuses on the process of processing and analyzing the collected data in the analysis layer. Therefore, the following takes the water conservancy data monitoring of the target water area as an example to introduce the processing and analysis of data in the analysis layer.

[0058] Specifically, if Figure 1 As shown, a method for intelligent monitoring of water conservancy project data based on digital twins in this embodiment includes the following steps:

[0059] Step S1, obtaining a plurality of historical data points in a historical period and a plurality of data points in a current period in an area where a target reservoir is located.

[0060] Specifically, the data points at any sampling moment include multi-dimensional data, and the multi-dimensional data include dam leakage, ambient temperature, rainfall, water flow, water pressure, and water level.

[0061] The above historical period may be half a year, one year or longer, and the current period may be a number of consecutive days, specifically 3 days, 5 days or more days.

[0062] Data from the same dimension at different sampling times constitutes a sequence. The multiple dimensions collected during the historical period are the same as the dimensions in the current period.

[0063] The above ambient temperature and rainfall are obtained from meteorological data sources.

[0064] Furthermore, the sequences consisting of historical data and current data in multiple dimensions are also standardized.

[0065] Step S2: Determine the first abnormality probability of the current data point.

[0066] In this embodiment, the data in the current period is analyzed to determine the first abnormal probability of the data of the current water conservancy project.

[0067] The process of obtaining the first abnormality probability is as follows:

[0068] Step S21: construct a network model.

[0069] The network model in this embodiment is a fully connected network model (such as an FCN model).

[0070] Step S22: Obtain a training set, and use the training set to train the network model to obtain a trained network model.

[0071] The training set in this embodiment is a plurality of historical data points and corresponding labels within a historical period in the region where the target reservoir is located. The label is normal or abnormal for each historical data point, where normal is 1 and abnormal is 0.

[0072] The training process is as follows:

[0073] The training set is input into the network model for training, and the loss function is used to calculate the loss value. The parameters of the network model are adjusted using the gradient descent algorithm until the loss value between the output theoretical value and the label is less than the threshold or the number of training times reaches the set number. The training is stopped and a trained network model is obtained.

[0074] The above loss function is the mean square error loss. Since the specific training process is an existing technology, it will not be described here.

[0075] Furthermore, in order to obtain a more accurate trained network model, it also includes: performing time alignment processing on the sequences of data in each dimension.

[0076] When monitoring water conservancy project data, when collecting data from different dimensions at the same time, there will be a certain delay between the data. For example, changes in water level usually precede changes in flow, changes in rainfall precede changes in reservoir water level (after rainfall, rainwater needs to enter the reservoir through processes such as surface runoff and underground infiltration), and changes in ambient temperature precede changes in water level. Therefore, when analyzing water conservancy project data, it is necessary to first calculate the delay time between different dimensional parameters.

[0077] In this embodiment, the cross-correlation function is used to calculate the delay time between parameters of different dimensions. Since the cross-correlation function describes the degree of correlation between the values ​​of different random signals at any two different moments, the delay time between parameters of different dimensions in the historical parameter data is determined by the cross-correlation function.

[0078] Specifically, taking the sequence of data of any dimension as the target sequence, the mutual correlation coefficients between each target data in the target sequence and the data in the sequence of data of any other dimension are calculated respectively, and the moment corresponding to the data when the mutual correlation coefficient is the largest is selected as the moment after delay. At this point, the difference between the moment after delay and the moment corresponding to the target data can be obtained, and this difference is the delay time.

[0079] At this point, a sequence of multi-dimensional data after time alignment can be obtained, and then the historical data points and current data points after time alignment can be obtained.

[0080] The purpose of obtaining the above delay time is to obtain data of different dimensions with high consistency of changes, so as to facilitate the subsequent accurate analysis of data of different dimensions.

[0081] Step S23: Use the trained network model to perform anomaly detection on the current data point to obtain a first anomaly probability.

[0082] Furthermore, due to the problem of model accuracy, the product of the model accuracy and the first abnormality probability can also be used as the corrected first abnormality probability to avoid obvious errors caused by inaccurate model accuracy.

[0083] Step S3: Determine the second abnormality probability of the current data point.

[0084] In this embodiment, considering that data with abnormal frequency components may still appear normal in the time domain (without obvious abnormalities), it is not comprehensive to only consider the abnormalities of the time domain amplitude of each dimensional data in the data point. Therefore, in order to detect data with abnormal frequency components whose time domain abnormalities are not obvious, it is also necessary to analyze the frequency domain of each dimensional data.

[0085] Specifically, the process of the second abnormal probability is:

[0086] First, the sequence of each dimension data in each data point is processed by short-time Fourier transform to obtain a time-frequency diagram.

[0087] Specifically, the time-frequency diagram of the sequence of data in each dimension is obtained by short-time Fourier transform. The time-frequency diagram uses frequency as the vertical axis and time as the horizontal axis, reflecting the frequency amplitude of data at different frequencies at different sampling moments.

[0088] Secondly, the anomaly index is obtained based on the absolute value of the difference between the target sampling moment of the target frequency on the time-frequency diagram and the sampling spectrum amplitude of the previous moment and the mean spectrum amplitude of the target frequency. The anomaly index is normalized to obtain the anomaly probability.

[0089] The above target sampling time is any sampling time.

[0090] In one embodiment, the abnormality indicator is the average of the global deviation and the local deviation.

[0091]

[0092] in, is the global deviation, is the local deviation; is the frequency amplitude mean of the kth frequency on the time-frequency diagram; f k (n) is the spectrum amplitude of the kth frequency at the nth moment on the time-frequency diagram, f k (n-1) is the spectrum amplitude of the kth frequency at the n-1th moment.

[0093] when It reflects the fluctuation of the spectrum amplitude of the kth frequency in all time frames, that is, the overall deviation degree. The larger the value, the greater the impact of the abnormality at the kth frequency at the nth moment. It reflects the local fluctuation or mutation of the kth frequency in the adjacent time frame. The larger the mutation, the more abnormal the kth frequency at the nth moment is.

[0094] In another embodiment, the anomaly index can also be obtained by measuring the similarity between the time-frequency diagram of the sequence of each dimensional data and the standard time-frequency diagram of the sequence of normal dimensional data. The anomaly index is negatively correlated with the similarity, for example, the anomaly index can be 1 minus the similarity.

[0095] At this time, the abnormality indicator can be directly used as the second abnormality probability without normalization.

[0096] The normal dimensional data sequence described above refers to dimensional data without abnormalities, and can be specifically obtained from historical data using big data technology. It should be noted that because the sequence of each dimensional data is also related to seasonality, the normal dimensional data sequence obtained above needs to be identical or similar to the current dimensional data sequence.

[0097] Among them, the maximum abnormal probability in all dimensions at the current moment is selected as the second abnormal probability.

[0098] In the above embodiment, the short-time Fourier transform can link time and frequency, thereby analyzing anomalies that occur at each frequency over time. Furthermore, since the energy distribution at each frequency varies, anomalies at high-energy frequencies have a greater impact, and therefore, anomalies at high-energy frequencies attract greater attention.

[0099] The above normalization function can adopt the softmax function or the maximum and minimum value normalization method.

[0100] Step S4: Integrate the first abnormal probability and the second abnormal probability to obtain a comprehensive probability; in response to the comprehensive probability being greater than a set threshold, it indicates that the current data point is abnormal and the staff needs to be notified to take measures.

[0101] In this embodiment, the first abnormal probability and the second abnormal probability of each data point are integrated together to obtain a comprehensive probability.

[0102] The process of obtaining the first weight of the first abnormal probability and the second weight of the second abnormal probability during integration is as follows:

[0103] First, the dimension sequence corresponding to the second anomaly probability is decomposed by STL to obtain the sum of the seasonal component and residual component under each dimension.

[0104] Among them, STL decomposition is a commonly used method in time series decomposition algorithms, which decomposes the time series into trend component, seasonal component and remainder based on LOESS (locally weighted regression).

[0105] Next, the second weight is calculated, and the sum of the first weight and the second weight is 1.

[0106] The second weight in this embodiment is negatively correlated with the sum of the seasonal component and the remainder.

[0107] Specifically, the second weight is T k is the absolute value of the sum of the seasonal component and the remainder of the k-th dimension, and exp() is an exponential function with the natural constant e as the base.

[0108] Since in time series analysis, the trend component represents the continuous and long-term change of the sequence mean, if the absolute value of the sum of the seasonal component and the remainder is larger, it proves that the trend component is smaller, the second anomaly probability obtained from the time series is less important, and its weight should be smaller; otherwise, the second weight should be larger.

[0109] The threshold value in this embodiment may be set to 0.8. Of course, in other implementations, it may also be set based on experience.

[0110] Furthermore, in this embodiment, water conservancy project data at a future time can also be predicted to determine whether the predicted data point is abnormal data.

[0111] In one embodiment, an LSTM model can also be used for prediction, where the input is the data point at the current moment and the data points before the current moment, to obtain the data point at the next moment, and determine the comprehensive probability of the data point at the next moment to determine whether the data point at the next moment is abnormal.

[0112] Before using the LSTM model for prediction, the LSTM model must be trained. Since the training process is an existing technology, it will not be described in detail here.

[0113] The solution of the present invention can accurately analyze the abnormal situation of hydraulic engineering data at the current moment, thereby improving the safety of water conservancy facility monitoring.

[0114] The present invention also provides a water conservancy project data intelligent monitoring system based on digital twins. Figure 2 As shown, the system includes a processor and a memory, and the memory stores computer program instructions. When the computer program instructions are executed by the processor, the above-mentioned intelligent monitoring method of water conservancy project data based on digital twins according to the present invention is implemented.

[0115] The system further includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and thus will not be described in detail here.

[0116] In the present invention, the aforementioned memory can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, the computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by applications, modules, or both. Any such computer storage medium can be part of, accessible to, or connectable to a device. Any application or module described in the present invention can be implemented by computer-readable / executable instructions stored or otherwise retained by such a computer-readable medium.

[0117] In the description of this specification, “a plurality of” means at least two, for example, two, three or more, etc., unless otherwise clearly defined.

[0118] Although this specification has shown and described several embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and substitutions without departing from the idea and spirit of the present invention.

Claims

1. A method for intelligent monitoring of water conservancy project data based on digital twins, characterized in that: include: Collect multiple dimensional data of the target water area at the current moment in real time and form the current data point; The multiple dimensional data include dam leakage, ambient temperature, rainfall, water flow, water pressure, and water level; Input the current data point into the network model and output the corresponding first abnormality probability; Obtain data from multiple dimensions at multiple moments before the current moment. The same dimension contains data from all sampling moments at the current moment to form a sequence. Perform frequency domain analysis on each dimension sequence to determine the abnormal probability of the data at the current moment in the corresponding dimension. The maximum abnormal probability among all dimensions at the current moment is used as the second abnormal probability. integrating the first abnormal probability and the second abnormal probability to obtain a comprehensive probability; In response to the comprehensive probability being greater than the set threshold, there is an anomaly in the data point at the current moment, and the staff needs to be notified to take measures.

2. The method for intelligent monitoring of water conservancy project data based on digital twins according to claim 1 is characterized in that: The second weight of the integrated second abnormal probability is: Perform STL decomposition on the dimensional sequence corresponding to the second anomaly probability to obtain the sum of the seasonal component and the residual component; The second weight is negatively correlated with the sum of the seasonal component and the residual; The sum of the first weight and the second weight of the first abnormality probability is 1.

3. The method for intelligent monitoring of water conservancy project data based on digital twin according to claim 1 is characterized in that: The process of obtaining the abnormal probability is as follows: The time-frequency diagram is obtained by processing the sequences of each dimension through short-time Fourier transform; According to the absolute value of the difference between the sampling spectrum amplitude of any frequency at the target sampling moment and its adjacent previous moment on the time-frequency diagram and the mean spectrum amplitude of any frequency, an abnormality index is obtained, and the abnormality index is normalized to obtain the abnormality probability; the target sampling moment is any sampling moment.

4. The method for intelligent monitoring of water conservancy project data based on digital twins according to claim 3 is characterized in that: The abnormality index is the average of the global deviation and the local deviation; in, is the global deviation, is the local deviation; is the frequency amplitude mean of the kth frequency on the time-frequency diagram; f k (n) is the spectrum amplitude of the kth frequency at the nth moment on the time-frequency diagram, f k (n-1) is the spectrum amplitude of the kth frequency at the n-1th moment.

5. The method for intelligent monitoring of water conservancy project data based on digital twin according to claim 1 is characterized in that: The network model is a fully connected network model.

6. The method for intelligent monitoring of water conservancy project data based on digital twins according to claim 5 is characterized in that: The training process of the network model is: Obtaining a training set, wherein the training set includes historical data points at multiple moments in a historical period; labels are normal or abnormal conditions of the data points at each moment; The network model is trained using the training set to obtain a trained network model.

7. The method for intelligent monitoring of water conservancy project data based on digital twins according to claim 6 is characterized in that: Also includes: Time alignment is performed on the dimensional data in the historical data points at each moment to obtain aligned historical data points.

8. The method for intelligent monitoring of water conservancy project data based on digital twins according to claim 3 is characterized in that: The normalization process adopts a softmax function or a maximum-minimum value normalization method.

9. The method for intelligent monitoring of water conservancy project data based on digital twin according to claim 1, characterized in that: The dam leakage, water flow, water pressure and water level are measured respectively by means of an osmometer, a flow meter, a water pressure gauge and a water level gauge; the ambient temperature and rainfall are obtained by means of meteorological data.

10. A water conservancy project data intelligent monitoring system based on digital twins, characterized by: include: processor; A memory storing computer instructions for intelligent monitoring of water conservancy project data based on digital twins. When the computer instructions are executed by the processor, the system executes a method for intelligent monitoring of water conservancy project data based on digital twins according to any one of claims 1 to 9.

Images