Disclosure of Invention
The invention aims to provide a dam deformation risk prediction method and system based on an LSTM model, which are used for solving the defects in the background technology.
In order to achieve the above object, the present invention provides the following technical solutions: dam deformation risk prediction method based on LSTM model, the prediction method includes the following steps:
Acquiring the number of the dams after the management area, marking all the dams so that each dam has a unique identifier, and acquiring multisource factors related to the operation of the dams at regular time through a multispeculiarity data acquisition system;
Preprocessing various data in the multi-source factors, extracting multiple characteristic data in the multi-source factors, substituting the multiple data into a joint analysis model for analysis, and outputting deformation assignment of the dam;
Predicting deformation risk of the dam according to the comparison result of the dam deformation assignment and the level threshold value, and dividing the deformation risk of the dam into no risk, primary risk and secondary risk according to the deformation risk prediction result;
collecting hydrological data of an area where the dam is located, substituting the hydrological data into a trained LSTM model for analysis, and predicting the water flow change of a river where the dam is located for a period of time;
After a water flow change prediction result of a river where a dam is positioned is combined with a dam deformation risk division result for analysis, generating corresponding management strategies for all the dams in a management area, and sending the management strategies to management staff;
judging whether all the dams reach the maintenance time point, if so, analyzing the historical deformation assignment of the dams, generating management assignment for the dams, sorting all the dams from large to small according to the management assignment, generating a management list, and sending the management list to management staff.
In a preferred embodiment, extracting multiple characteristic data in multiple source factors, substituting the multiple characteristic data into a joint analysis model for analysis, and outputting deformation assignment of a dam, wherein the method comprises the following steps:
extracting multiple characteristic data in multiple source factors, wherein the multiple data comprise dam data and geological data;
Dam data comprise flood discharge flow deviation of flood discharge holes, equipment vibration indexes and displacement rates, and geological data comprise stratum compression indexes;
And substituting the water flow deviation of the flood discharge hole, the equipment vibration index, the displacement rate and the rock stratum compression resistance index into a joint analysis model for analysis, and outputting deformation assignment of the dam.
In a preferred embodiment, the deformation risk prediction is performed on the dam according to the comparison result of the deformation assignment of the dam and the level threshold, and the method comprises the following steps:
According to the calculation expression of deformation assignment of the dam, the larger the deformation assignment is, the larger the risk of deformation of the dam in the future is indicated, so that after the deformation assignment is obtained, the deformation assignment is compared with a preset level threshold;
The hierarchy threshold comprises a first deformation threshold and a second deformation threshold, wherein the first deformation threshold is used for judging whether deformation risks exist in future operation of the dam, and the second deformation threshold is used for judging the severity of the deformation risks of the dam;
if the deformation assignment is smaller than or equal to the first deformation threshold value, judging that no deformation risk exists in future operation of the dam;
if the deformation assignment is larger than the first deformation threshold, judging that the deformation risk exists in future operation of the dam;
if the deformation assignment is larger than the first deformation threshold value and smaller than or equal to the second deformation threshold value, judging that the deformation risk of the dam is small;
And if the deformation assignment is larger than the second deformation threshold value, judging that the deformation risk of the dam is large.
In a preferred embodiment, the deformation risk of the dam is classified into no risk, primary risk and secondary risk according to the deformation risk prediction result, and the method comprises the following steps:
If the deformation risk does not exist in the future operation of the dam, dividing the deformation risk of the dam into risk-free grades;
If the deformation risk of the predicted dam is small, dividing the deformation risk of the dam into first-level risk grades;
and if the deformation risk of the dam is predicted to be large, classifying the deformation risk of the dam into a secondary risk level.
In a preferred embodiment, after the water flow change prediction result of the river where the dam is located and the dam deformation risk division result are combined and analyzed, a corresponding management strategy is generated for all the dams in the management area, and the method comprises the following steps:
acquiring a water flow change prediction result of a river where a dam is positioned and a dam deformation risk division result;
If the dam deformation risk is divided into secondary risk grades, automatically generating a management strategy for the dam, wherein the management strategy needs to be managed in time;
If the dam deformation risk is divided into first-level risk grades, and the water flow of the river where the dam is located is predicted to be small in the next few days through the LSTM model, a management strategy which needs to be alleviated and managed is automatically generated for the dam;
If the deformation risk of the dam is divided into first-level risk grades, and the water flow of the river where the dam is located is predicted to be large in the next few days through the LSTM model, a management strategy which needs to be managed in time is automatically generated for the dam.
In a preferred embodiment, determining whether all dams reach a maintenance time point, if so, analyzing historical deformation assignment of the dams, generating management assignment for the dams, and sorting all dams from large to small according to the management assignment to generate a management list, wherein the method comprises the following steps:
Acquiring all historical deformation assignments of the dam, and establishing an assignment set of all the historical deformation assignments;
calculating average deformation assignment and assignment standard deviation in an assignment set, wherein the expression is as follows:
Wherein Bavg is an average deformation assignment, BQ is an assignment standard deviation, i=1, 2,3, & gt, n is the number of deformation assignments in the assignment set, and Bi is an ith deformation assignment in the assignment set;
After the average deformation assignment and the assignment standard deviation are obtained, management assignment is generated for each dam, and the function expression is as follows:
Wherein, Gpz is a management assignment, Bavg is an average deformation assignment, BQ is an assignment standard deviation, BZ is a standard deviation threshold, eb is a second deformation threshold, and the larger the management assignment is, the worse the historical overall operation state of the dam is;
after the management assignment is acquired, all dams are ordered from big to small according to the management assignment, and a management list is generated.
In a preferred embodiment, collecting hydrological data of an area where a dam is located, substituting the hydrological data into a trained LSTM model for analysis, and predicting a water flow change of a river where the dam is located for a period of time, comprising the steps of:
Acquiring hydrological data in a period of time through a weather bureau and a water bureau, including water level, rainfall and flow rate data in the future days, cleaning the data, processing missing values and abnormal values, converting the data into a time sequence format of an LSTM model, and carrying out data normalization processing;
loading an LSTM model trained on historical hydrologic data, and inputting the preprocessed hydrologic data into the LSTM model, wherein the method comprises the steps of predicting water flow in the next days by using water level, rainfall and flow rate data in the previous days;
The LSTM model utilizes input historical data to learn a water flow mode and predicts the water flow change of the dam river in the next days;
finally, the prediction results of the LSTM model are presented using a chart or visualization tool.
The dam deformation risk prediction system based on the LSTM model comprises a marking module, a data acquisition module, a feature extraction module, an analysis module, a deformation prediction module, a risk division module, an LSTM module, a strategy generation module and a sequencing management module;
And a marking module: acquiring the number of the dams after the management area, and marking all the dams so that each dam has a unique identifier;
and a data acquisition module: the method comprises the steps of acquiring multisource factors related to dam operation at regular time through a multispeculiarity data acquisition system;
And the feature extraction module is used for: preprocessing various data in the multi-source factors, and extracting multiple characteristic data in the multi-source factors;
and an analysis module: substituting a plurality of data into the joint analysis model for analysis, and outputting deformation assignment of the dam;
deformation prediction module: predicting the deformation risk of the dam according to the comparison result of the dam deformation assignment and the level threshold;
Risk division module: dividing the deformation risk of the dam into no risk, primary risk and secondary risk according to the deformation risk prediction result;
LSTM module: collecting hydrological data of an area where the dam is located, substituting the hydrological data into a trained LSTM model for analysis, and predicting the water flow change of a river where the dam is located for a period of time;
The strategy generation module: after a water flow change prediction result of a river where a dam is positioned is combined with a dam deformation risk division result for analysis, generating corresponding management strategies for all the dams in a management area, and sending the management strategies to management staff;
The ordering management module: judging whether all the dams reach the maintenance time point, if so, analyzing the historical deformation assignment of the dams, generating management assignment for the dams, sorting all the dams from large to small according to the management assignment, generating a management list, and sending the management list to management staff.
In the technical scheme, the invention has the technical effects and advantages that:
1. According to the method, multisource factors related to dam operation are obtained regularly through a multisource data acquisition system, after preprocessing is carried out on various data in the multisource factors, multiple characteristic data in the multisource factors are extracted, after the multiple characteristic data are substituted into a joint analysis model for analysis, deformation assignment of the dam is output, deformation risk prediction is carried out on the dam according to a comparison result of the dam deformation assignment and a level threshold, deformation risk of the dam is divided into risk-free, primary risk and secondary risk according to the deformation risk prediction result, hydrological data of an area where the dam is located are collected, the hydrological data are substituted into a trained LSTM model for analysis, water flow change of a river where the dam is located for a period of time is predicted, and after the water flow change prediction result of the river where the dam is located and the dam deformation risk division result are combined for analysis, corresponding management strategies are generated for all the dams in a management area. The prediction method can effectively predict the deformation risk of the dam, and predicts the water flow change of the river where the dam is positioned in a period of time through the LSTM model, so that a corresponding management strategy can be automatically generated, and the safety and stability of the operation of the dam are ensured.
2. According to the method, whether all the dams reach the maintenance time point is judged, if yes, the historical deformation assignment of the dams is analyzed, then the management assignment is generated for the dams, after all the dams are ordered from big to small according to the management assignment, a management list is generated, the management list is sent to management staff, the management staff selects the management sequence of the dams according to the positive sequence of the management list, and the management efficiency of the dams is effectively improved.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1: referring to fig. 1, the dam deformation risk prediction method based on the LSTM model according to the present embodiment includes the following steps:
the method comprises the steps of obtaining the number of the dams after the management area, marking all the dams so that each dam has a unique identifier, wherein the identifier comprises information such as the number and the position of the dam, and the like, and comprises the following steps:
Boundary coordinates or other location information for the management area is obtained using a Geographic Information System (GIS) or other related tool. This may be accomplished by way of a query or API call for geospatial data, or the like. Dam data located within the management area is obtained from a relational database, dataset, or API. The data should contain the location, number, basic information, etc. of each dam. For the acquired dam data, the number of dams in the management area is calculated. This can be obtained by simple statistics of the data or by spatial analysis based on location information.
A unique identifier is generated for each dam, and numbering, GUID (globally unique identifier) or the like can be adopted. The identification is ensured to be unique, and subsequent management and inquiry are convenient. Information such as unique identification, dam number, location, etc. is associated with the data for each dam, creating a data set or database table containing complete information. The marked dam information is visualized on the map using a map tool or other data visualization tool. This can help to intuitively understand the distribution and basic information of the dams within the management area. If the dam data is from a database or dataset, it is ensured that the generated unique identification and other information is updated back to the original data source for use in subsequent queries and management.
The method comprises the steps of acquiring multisource factors related to dam operation at regular time through a multisource data acquisition system, preprocessing various data in the multisource factors, and extracting multiple characteristic data in the multisource factors, wherein the steps are as follows:
a plurality of data acquisition systems are arranged, including sensor networks, weather stations, hydrologic stations, geological monitoring equipment and the like, so as to collect various factors related to dam operation. Ensuring that different types of data sources can provide comprehensive and accurate information. And cleaning the acquired data, and processing missing values, abnormal values and error data. And (5) performing data quality inspection to ensure the reliability and consistency of the data. The time stamps of the different data sources are aligned and synchronized to ensure consistency of the data over time. This facilitates subsequent time series analysis.
The data is standardized to ensure that different types of data have the same scale and unit, and features are extracted from various types of data, which may include statistical features, frequency domain features, time domain features, and the like. Depending on the nature of the operation of the dam, suitable features are selected to describe the state of the system. If the feature dimensions are high, dimension reduction techniques such as Principal Component Analysis (PCA) or t-distribution neighborhood embedding (t-SNE) may be considered to reduce the number of features while retaining key information. The collected time series data is analyzed, including periodic analysis, trend analysis, and the like. This helps to find regularity and trend in the data. The processed data is stored in a reliable database and is backed up periodically. The security and usability of the data are ensured.
After substituting a plurality of data into a joint analysis model for analysis, outputting deformation assignment of the dam, predicting deformation risk of the dam according to a comparison result of the deformation assignment of the dam and a hierarchical threshold value, dividing the deformation risk of the dam into risk-free, primary risk and secondary risk according to the deformation risk prediction result, collecting hydrological data of an area where the dam is located, substituting the hydrological data into a trained LSTM model for analysis, predicting water flow change of a river where the dam is located for a period of time, combining and analyzing the water flow change prediction result of the river where the dam is located with the dam deformation risk division result, generating corresponding management strategies for all the dams in a management area, transmitting the management strategies to management staff, managing the dams according to the management strategies by the management staff, judging whether all the dams reach maintenance time points or not, if so, generating management assignment for the dams after analyzing the history deformation assignment of the dams, ordering all the dams from large to small, generating a management list, transmitting the management list to the management staff, and selecting a management order for the dams according to the management list;
According to the method, multisource factors related to dam operation are obtained regularly through a multisource data acquisition system, after preprocessing is carried out on various data in the multisource factors, multiple characteristic data in the multisource factors are extracted, after the multiple characteristic data are substituted into a joint analysis model for analysis, deformation assignment of the dam is output, deformation risk prediction is carried out on the dam according to a comparison result of the dam deformation assignment and a level threshold, deformation risk of the dam is divided into risk-free, primary risk and secondary risk according to the deformation risk prediction result, hydrological data of an area where the dam is located are collected, the hydrological data are substituted into a trained LSTM model for analysis, water flow change of a river where the dam is located for a period of time is predicted, and after the water flow change prediction result of the river where the dam is located and the dam deformation risk division result are combined for analysis, corresponding management strategies are generated for all the dams in a management area. The prediction method can effectively predict the deformation risk of the dam, and predicts the water flow change of the river where the dam is positioned in a period of time through the LSTM model, so that a corresponding management strategy can be automatically generated, and the safety and stability of the operation of the dam are ensured.
According to the method, whether all the dams reach the maintenance time point is judged, if yes, the historical deformation assignment of the dams is analyzed, then the management assignment is generated for the dams, after all the dams are ordered from big to small according to the management assignment, a management list is generated, the management list is sent to management staff, the management staff selects the management sequence of the dams according to the positive sequence of the management list, and the management efficiency of the dams is effectively improved.
Example 2: after preprocessing various data in the multi-source factors, extracting multiple characteristic data in the multi-source factors, wherein the method comprises the following steps of:
extracting multiple characteristic data in multiple source factors, wherein the multiple data comprise dam data and geological data;
dam data includes flood discharge flow deviation, equipment vibration index and displacement rate, and geological data includes rock stratum compression index.
After substituting a plurality of data into the joint analysis model for analysis, outputting deformation assignment of the dam, wherein the method comprises the following steps of:
substituting the water flow deviation of the flood discharge hole, the equipment vibration index, the displacement rate and the rock stratum compression index into a joint analysis model for analysis, and then outputting deformation assignment of the dam;
the functional expression of the joint analysis model is:
Wherein bxz is a deformation assignment, XHS is a flood discharge hole water flow deviation, SBZ is an equipment vibration index, WYS is a displacement rate, KYC is a rock stratum compression resistance index, and alpha, beta, gamma and delta are proportional coefficients of the flood discharge hole water flow deviation, the equipment vibration index, the displacement rate and the rock stratum compression resistance index respectively, and alpha, beta, gamma and delta are all larger than 0.
The logic for obtaining the water flow deviation of the flood discharge hole is as follows: the water flow sensor that sets up through flood discharge hole water inlet end department obtains first discharge, obtains the second discharge through the water flow sensor that sets up through flood discharge hole water outlet end department, subtracts first discharge with the second discharge and obtains flood discharge hole water flow deviation, and the flood discharge hole water flow deviation is bigger, and the jam situation that indicates the flood discharge hole is more serious, can lead to the dam future operation to have deformation risk, specifically does:
The water pressure rises: blockage of the flood discharge holes can cause blockage of the flood discharge channels, so that water flow of the flood discharge channels is blocked. This may lead to an increase in water pressure upstream of the flood discharge channel, increasing the water pressure load of the dam structure. And (3) increasing the pressure of the dam body: blocking the flood discharge hole means that the dam cannot rapidly relieve the water level pressure through the flood discharge path. This may result in increased water pressure at the bottom of the dam and within the dam, creating additional forces on the dam structure. Uneven water pressure distribution: flood discharge hole blockage can cause uneven flow through other channels, resulting in uneven local water pressure distribution. This may lead to uneven stressing of the dam structure, increasing the risk of deformation of the structure. Deformation and settlement of the dam body: blocking the flood discharge holes may cause excessive saturation and softening of the soil body at the bottom of the dam, resulting in settlement and deformation of the dam. Sedimentation non-uniformity may lead to tilting and asymmetric deformation of the dam. Flood transit difficulty: in the case that the flood discharge hole is severely blocked, the dam may not efficiently process the flood, increasing the water pressure at the top and bottom of the dam. This may have a negative impact on the overall stability of the dam.
The acquisition logic of the vibration index of the equipment is as follows: vibration amplitude of each device is acquired in real time through vibration sensors arranged on a plurality of devices of the dam, the vibration amplitudes of the plurality of devices are added to obtain a device vibration index, the device vibration index is larger, the condition that the vibration amplitude of a single device is overlarge or the plurality of devices resonate simultaneously is indicated, deformation is easily caused in future operation of the dam, and the method comprises the following steps:
Structural fatigue and damage: excessive vibration amplitude of a single device or simultaneous resonance of multiple devices may lead to fatigue and damage of the dam structure. The prolonged vibration action may lead to cracking, deformation and instability of the structural element. Soil body response: the equipment vibration can be transmitted into the soil body of the dam, and the vibration of the soil body is caused. If the seismic capacity of the soil is limited, compaction and deformation of the soil can be caused, thereby affecting the stability of the dam. Soil-structure interaction: soil vibration caused by equipment vibration can interact with a dam structure, so that the soil and the structure are unevenly deformed, and the instability of the structure is increased. Foundation settlement: the strong vibrations may cause sedimentation of the foundation soil mass. Uneven foundation settlement may lead to uneven deformation and tilting of the dam. Vibration induced earth and rock flow: in the case of unstable geological conditions, equipment vibration may induce earth and rock flow, which has a damaging effect on the terrain surrounding the dam.
The displacement rate is calculated as: Where WZ Horizontal level represents the final horizontal position of the dam, WC Horizontal level represents the initial horizontal position of the dam, WZ Vertical direction represents the final vertical position of the dam, WC Vertical direction represents the initial vertical position of the dam, jsc is the monitoring time period, and the greater the displacement rate is, the greater the horizontal, vertical or horizontal vertical displacement of the dam in the monitoring time period is, the easier the dam deformation is caused, specifically:
dam body sliding risk: a significant increase in horizontal displacement may be indicative of the overall horizontal slip or tendency of the dam to slip. This may be due to instability of the earth at the bottom of the dam or other factors.
Dam settlement or swelling: a significant increase in vertical displacement may be indicative of settlement or bulging of the dam body. This may be due to compaction of the foundation soil, water level changes, or other geological factors.
And (3) integral deformation: if the displacement in both the horizontal and vertical directions increases significantly, this may indicate complex deformation of the dam as a whole, possibly related to overall instability of the structure or soil-structure interactions.
The formation compressive index is calculated as: kyc= YKT/DYL, where YKT represents the compressive strength of the dam bearing formation, DYL represents the ground stress of the dam bearing formation, and the smaller the formation compressive index, the worse the compressive property of the dam formation is, the easier the dam deformation is caused, specifically:
Basic stability problem: poor formation compression resistance may result in poor dam foundation stability. If the formation does not provide sufficient support, it may cause settling and deformation of the dam foundation. Formation cracking and breaking: the formation with poor compression resistance is more prone to cracking and breaking, which may lead to instability of the geological structure around the dam, increasing the foundation problem of the dam. Soil erosion and sideslip: formations with poor compression resistance may be susceptible to erosion by water currents and weathering, resulting in sideslip and loss of soil. This may increase the risk of deformation of the dam. Groundwater infiltration: poor formation compression may mean increased permeability of the formation, resulting in groundwater penetration through the formation to the bottom of the dam, which may cause softening of the soil and settling of the dam. Rock formation displacement and inclination: rock formations with poor compression resistance may be more prone to displacement and tilting, resulting in uneven stress of the dam structure, increasing the risk of deformation of the dam.
Referring to fig. 2, according to the comparison result of the deformation assignment and the level threshold of the dam, the deformation risk prediction is performed on the dam, including the following steps:
According to the calculation expression of deformation assignment of the dam, the larger the deformation assignment is, the larger the risk of deformation of the dam in the future is indicated, so that after the deformation assignment is obtained, the deformation assignment is compared with a preset level threshold;
The hierarchy threshold comprises a first deformation threshold and a second deformation threshold, wherein the first deformation threshold is used for predicting whether deformation risks exist in future operation of the dam, and the second deformation threshold is used for predicting the severity of the deformation risks of the dam;
If the deformation assignment is smaller than or equal to the first deformation threshold value, predicting that no deformation risk exists in future operation of the dam;
If the deformation assignment is larger than the first deformation threshold value, predicting that the future operation of the dam has deformation risk;
If the deformation assignment is larger than the first deformation threshold value and smaller than or equal to the second deformation threshold value, the deformation risk of the predicted dam is small;
And if the deformation assignment is larger than the second deformation threshold, predicting that the deformation risk of the dam is large.
Dividing the deformation risk of the dam into no risk, primary risk and secondary risk according to the deformation risk prediction result, and comprising the following steps:
If the deformation risk does not exist in the future operation of the dam, dividing the deformation risk of the dam into risk-free grades;
If the deformation risk of the predicted dam is small, dividing the deformation risk of the dam into first-level risk grades;
and if the deformation risk of the dam is predicted to be large, classifying the deformation risk of the dam into a secondary risk level.
Referring to fig. 3, collecting hydrological data of a dam area, substituting the hydrological data into a trained LSTM model for analysis, and predicting a water flow change of a river where the dam is located for a period of time, including the following steps:
the method comprises the steps of acquiring hydrologic data in a period of time from a weather bureau and a water bureau, including water level, rainfall and flow rate data in the future days, cleaning the data, processing missing values and abnormal values, converting the data into a time sequence format suitable for an LSTM model, carrying out data normalization, loading an LSTM model trained on historical hydrologic data, and inputting the preprocessed hydrologic data into the LSTM model. For example, water level, rainfall and flow rate data for the first days are used to predict water flow for the next days.
The LSTM model learns the pattern of water flow using the input historical data and attempts to predict the water flow change for several days in the future. The prediction results of the LSTM model are shown using a chart or visualization tool. And comparing the predicted result with actual observed data. And evaluating the performance of the model, and comparing the predicted result with actual observed data. And according to the evaluation result, adjusting and optimizing the model.
The training procedure for the LSTM model is as follows:
Historical water level, rainfall and flow rate data are collected. Ensure that the data contains the appropriate time stamps and covers a long enough historical period for training the model.
Cleaning and preprocessing the collected data:
processing the missing values: interpolation or other methods are used to fill in missing values.
Processing outliers: and detecting and processing abnormal values to ensure the data quality.
Smoothing data: the data is smoothed using a filtering technique.
The water level, rainfall and flow rate data are organized into a time series format. Each time step contains observations of water level, rainfall and flow rate.
The time series data set is divided into a training set and a testing set, so that the model is ensured to have enough data for training and verification.
Characteristic engineering:
Hysteresis characteristics: relationship between current time water level, rainfall and flow rate and past time.
Seasonal features: periodic changes each day or week.
External features: other factors that may affect the water flow, such as temperature, etc.
And normalizing the extracted features, and scaling the extracted features to the same scale range so as to improve the training efficiency of the model.
Establishing an LSTM model:
The number of layers and the number of neurons of the model are determined.
The structure of the input layer, the hidden layer and the output layer is defined.
Appropriate loss functions, optimizers, and evaluation criteria are selected.
The LSTM model is trained using the training set, iteratively optimizing model parameters to minimize the loss function.
And verifying the performance of the model by using the test set, and evaluating the generalization capability of the model to unseen data. An index such as Root Mean Square Error (RMSE) or the like is used to evaluate the accuracy of the model.
The trained LSTM model is used to predict water flow for several days in the future, and inputs include historical water level, rainfall and flow rate data, and external features for several days in the future.
And analyzing the prediction result to understand the trend of the model on the future water flow change. The prediction results are visualized using a chart or visualization tool to better understand the behavior of the model.
Depending on the actual predicted effect, the model may need to be tuned, for example, to adjust super parameters, to add training data, or to improve feature engineering.
After combining and analyzing a water flow change prediction result of a river where a dam is positioned and a dam deformation risk division result, generating corresponding management strategies for all dams in a management area, sending the management strategies to management staff, and managing the dams by the management staff according to the management strategies, wherein the method comprises the following steps of:
acquiring a water flow change prediction result of a river where a dam is positioned and a dam deformation risk division result;
If the dam deformation risk is classified into a secondary risk level, indicating that the dam deformation is serious, and risks such as collapse can occur, automatically generating a management strategy for the dam, wherein the management strategy needs to be managed in time;
If the dam deformation risk is classified into a first-level risk grade, and the water flow of the river where the dam is located for several days in the future is predicted by the LSTM model, when the dam deformation risk is classified into the first-level risk grade, the future deformation change speed of the dam is slow, and the influence of the water flow of the river on the dam is small due to the predicted water flow of the river for several days in the future, so that a management strategy needing to be alleviated and managed is automatically generated for the dam;
If the dam deformation risk is classified into a first-level risk grade, and the water flow rate of the river where the dam is located is predicted to be large for a few days in the future through the LSTM model, when the dam deformation risk is classified into the first-level risk grade, the future deformation change speed of the dam is slow, but the influence of the river water flow rate on the dam is large due to the predicted water flow rate of the river for a few days in the future, and the deformation change speed of the dam is aggravated in the river water flow rate for a few days in the future, so that a management strategy which needs timely management is automatically generated for the dam.
Judging whether all the dams reach a maintenance time point, if so, analyzing historical deformation assignment of the dams, generating management assignment for the dams, sorting all the dams from large to small according to the management assignment, generating a management list, sending the management list to a manager, and selecting a management sequence of the dams by the manager according to the positive sequence of the management list, wherein the method comprises the following steps of:
Acquiring all historical deformation assignments of the dam, and establishing an assignment set of all the historical deformation assignments;
calculating average deformation assignment and assignment standard deviation in an assignment set, wherein the expression is as follows:
Wherein Bavg is an average deformation assignment, BQ is an assignment standard deviation, i=1, 2,3, & gt, n is the number of deformation assignments in the assignment set, and Bi is an ith deformation assignment in the assignment set;
After the average deformation assignment and the assignment standard deviation are obtained, management assignment is generated for each dam, and the function expression is as follows:
wherein, Gpz is a management assignment, Bavg is an average deformation assignment, BQ is an assignment standard deviation, BZ is a standard deviation threshold, eb is a second deformation threshold, and the larger the management assignment is, the worse the historical overall operation state of the dam is, and the more the dam needs to be maintained preferentially;
if the value of the standard deviation of the dam is smaller, the value of the standard deviation is smaller, and the value of the standard deviation is used for distinguishing whether the value of the deformation in the value of the dam is larger or smaller;
Therefore, after the management assignment is obtained, the management list is generated after all the dams are ordered from big to small according to the management assignment.
Example 3: the dam deformation risk prediction system based on the LSTM model comprises a marking module, a data acquisition module, a feature extraction module, an analysis module, a deformation prediction module, a risk division module, an LSTM module, a strategy generation module and a sequencing management module;
And a marking module: acquiring the number of the dams after the management area, marking all the dams so that each dam has a unique identifier, wherein the identifier comprises information such as the number and the position of the dam, and the dam identifier information is sent to a data acquisition module, a strategy generation module and a sequencing management module;
And a data acquisition module: the method comprises the steps of acquiring multisource factors related to dam operation through a multisystem data acquisition system at regular time, and sending the multisource factors to a feature extraction module;
And the feature extraction module is used for: after preprocessing various data in the multi-source factors, extracting multiple characteristic data in the multi-source factors, and sending the multiple characteristic data to an analysis module;
And an analysis module: after substituting a plurality of items of data into the joint analysis model for analysis, outputting deformation assignment of the dam, and transmitting the deformation assignment to the deformation prediction module and the ordering management module;
Deformation prediction module: according to the comparison result of the dam deformation assignment and the level threshold value, performing deformation risk prediction on the dam, and transmitting the deformation risk prediction result to a risk division module;
risk division module: dividing the deformation risk of the dam into no risk, primary risk and secondary risk according to the deformation risk prediction result, and transmitting the deformation risk division result to the strategy generation module;
LSTM module: collecting hydrological data of an area where the dam is located, substituting the hydrological data into a trained LSTM model for analysis, predicting water flow change of a river where the dam is located for a period of time, and sending a water flow prediction result to a strategy generation module;
the strategy generation module: after a water flow change prediction result of a river where a dam is positioned is combined with a dam deformation risk division result for analysis, generating corresponding management strategies for all dams in a management area, wherein the management strategies comprise management suggestions and dam identifications, sending the management strategies to management staff, and managing the dams by the management staff according to the management strategies;
The ordering management module: judging whether all the dams reach the maintenance time point, if so, analyzing the historical deformation assignment of the dams, generating management assignment for the dams, sorting all the dams from large to small according to the management assignment, generating a management list, sending the management list to management staff, and selecting the management sequence of the dams by the management staff according to the positive sequence of the management list.
The above formulas are all formulas with dimensions removed and numerical values calculated, the formulas are formulas with a large amount of data collected for software simulation to obtain the latest real situation, and preset parameters in the formulas are set by those skilled in the art according to the actual situation.
In the description of the present specification, the descriptions of the terms "one embodiment," "example," "specific example," and the like, mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The preferred embodiments of the invention disclosed above are intended only to assist in the explanation of the invention. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention. The invention is limited only by the claims and the full scope and equivalents thereof.