A water conservancy project dam safety monitoring device and monitoring method

Abstract

The present invention discloses a water conservancy project dam safety monitoring device and a monitoring method, belonging to the technical field of water conservancy project dam safety monitoring. By adopting a deep learning algorithm to learn the historical water level impact characteristics and the historical water level corresponding to the historical water level impact characteristics, a water level prediction model that can predict the water level is obtained, and then the real-time water level impact characteristics are identified through the water level prediction model, and finally the water level prediction is realized. Based on actual data, it can be quickly predicted whether there may be danger, so that the staff can respond in time, thereby realizing the safety of the water conservancy project dam.

Description

Technical Field

[0001] The invention belongs to the technical field of water conservancy project dam safety monitoring, and in particular relates to a water conservancy project dam safety monitoring device and a monitoring method. Background Art

[0002] Scheduling water conservancy project dams to ensure their safety is the key to reducing or eliminating flood disasters. Currently, in actual applications, only the water level is often monitored, and the water conservancy project dams are dispatched and safely protected based on the monitored water level. However, the aforementioned existing technologies often have the problem of being untimely, which can cause river overflows, threaten the safety of water conservancy project dams, and may ultimately result in huge property losses. Summary of the invention

[0003] The invention provides a water conservancy project dam safety monitoring device and a monitoring method, which are used to solve the technical problems existing in the prior art.

[0004] On the one hand, the present invention provides a water conservancy project dam safety monitoring device, including: a historical data acquisition module, a deep learning module, a real-time data acquisition module, a water level prediction module and a safety warning module;

[0005] The historical data acquisition module is used to obtain the historical water level impact characteristics of the area where the water conservancy project dam is located and the historical water level corresponding to the historical water level impact characteristics;

[0006] The deep learning module is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics using a deep learning algorithm to obtain a water level prediction model;

[0007] The real-time data acquisition module is used to acquire the real-time water level impact characteristics of the area where the water conservancy project dam is located, and transmit the real-time water level impact characteristics to the water level prediction module;

[0008] The water level prediction module is used to dispatch the water level prediction model to identify the influencing features of the real-time water level and determine the short-term predicted water level;

[0009] The safety warning module is used to perform safety monitoring on the water conservancy project dam according to the short-term predicted water level, obtain the safety monitoring result of the water conservancy project dam, and issue a warning based on the safety monitoring result of the water conservancy project dam.

[0010] Furthermore, the historical water level impact characteristics of the area where the water conservancy project dam is located and the historical water level corresponding to the historical water level impact characteristics are obtained, including:

[0011] Obtain the water level, flow, rainfall and rainfall level at multiple time points in the historical period, and obtain the historical water level impact characteristics of the area where the water conservancy project dam is located;

[0012] For any historical water level impact feature, obtain the water level after a short period of time, and obtain the historical water level corresponding to any historical water level impact feature;

[0013] Among them, short-term is used to represent a fixed time period.

[0014] Furthermore, a deep learning algorithm is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics to obtain a water level prediction model, including:

[0015] The BiGRU-CapsNet algorithm is used to build the initial prediction model;

[0016] According to the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, an improved intelligent optimization algorithm is used to train the initial prediction model to obtain a water level prediction model.

[0017] Furthermore, according to the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, the improved intelligent optimization algorithm is used to train the initial prediction model to obtain a water level prediction model, including:

[0018] For the parameters to be optimized of the initial prediction model, a random initialization method is used to initialize the optimization population;

[0019] The fitness of each individual in the optimal population is obtained according to the historical water level impact characteristics and the historical water level corresponding to the historical water level impact characteristics, and the first target individual, the second target individual, the third target individual and the fourth target individual are determined in descending order of fitness; among which the first target individual is the optimal individual;

[0020] According to the first target individual, each individual in the optimization population is ordered to perform a local search in the optimal direction, and the individual after the local search is determined;

[0021] For the individuals after the local search, a joint search is performed on each individual in the optimal population according to the first target individual, the second target individual, the third target individual, and the fourth target individual to obtain the individuals after the joint search;

[0022] For the individuals after the joint search, each individual in the optimal population is required to perform a volatility search to obtain the individuals after the volatility search;

[0023] For the individuals after the volatility search, each individual in the optimization population is required to perform an escape search around the optimal position to obtain the individuals after the escape search;

[0024] The above local search, joint search, volatility search and escape search are repeated until the maximum number of training times is reached. The final hyperparameters of the initial prediction model are determined according to the individuals after the escape search to obtain the water level prediction model.

[0025] Further, according to the first target individual, each individual in the optimization population is ordered to perform a local search in an optimal direction, and the individual after the local search is determined, including:

[0026] Determine the current number of training times t, and determine the convergence factor based on the current number of training times t:

[0027]

[0028] in, represents the convergence factor during the t-th training process, represents the starting value of the convergence factor, represents the final value of the convergence factor, represents the maximum number of training times; and =2.5, =0;

[0029] According to the convergence factor And the first target individual, perform local search on each individual in the optimal population, and the individuals after local search are:

[0030]

[0031]

[0032]

[0033] in, Indicates t During the training process m Individuals, Represents the individual after local search , m=1,2,…,M, M represents the total number of individuals, represents the first local search factor, represents the second local search factor, represents the first target individual, | | represents taking the absolute value of the element, represents a random number uniformly distributed between [0,1], Represents a random number between (0,1), Represents a random number between (0,1).

[0034] Further, for the individuals after the local search, a joint search is performed on each individual in the optimal population according to the first target individual, the second target individual, the third target individual, and the fourth target individual to obtain the individuals after the joint search, including:

[0035] For the individuals after the local search, the first search position is determined according to the first target individual:

[0036]

[0037] in, indicates the first search position, represents the individual after the nth local search in the tth training process, n=1,2,…,M;

[0038] The second search position is determined according to the second target individual:

[0039]

[0040] in, indicates the second search position, represents the second target individual;

[0041] The third search position is determined according to the third target individual:

[0042]

[0043] in, indicates the third search position, represents the third target individual;

[0044] The fourth search position is determined according to the fourth target individual:

[0045]

[0046] in, indicates the fourth search position, represents the fourth target individual;

[0047] Identify individuals The Manhattan distances between the first target individual, the second target individual, the third target individual, and the fourth target individual, and the first weighted weight, the second weighted weight, the third weighted weight, and the fourth weighted weight are determined based on the Manhattan distances:

[0048]

[0049]

[0050]

[0051]

[0052] in, represents the first weighted weight, represents the second weighted weight, represents the third weighted weight, represents the fourth weighted weight, Represents an individual The distance to the first target individual, Represents an individual The distance to the second target individual, Represents an individual The distance to the third target individual, Represents an individual The distance to the fourth target individual;

[0053] According to the first search position, the second search position, the third search position, the fourth search position, the first target individual, the second target individual, the third target individual and the fourth target individual, the individual after the joint search is obtained is:

[0054]

[0055] in, Represents the individual after joint search .

[0056] Furthermore, for the individuals after the joint search, each individual in the optimal population is required to perform a volatility search, and the individuals after the volatility search are obtained, including:

[0057] For the individuals after the joint search, each individual is randomly matched with a different other individual to obtain the matching individual corresponding to each individual;

[0058] According to the matching individuals corresponding to each individual, the individuals are searched for volatility, and the individuals after the volatility search are obtained as follows:

[0059]

[0060] in, represents the individual after the kth joint search, represents the individual after the volatility search , k=1,2,…,M; represents the maximum number of training times, sin represents the sine function, cos represents the cosine function, represents pi, Represents a random number between (0,1), Represents a random number between (0,1), Represents a random number between (0,1), Represents an individual The corresponding matching individuals.

[0061] Furthermore, for the individuals after the volatility search, each individual in the optimization population is required to perform an escape search around the optimal position, and the individuals after the escape search are obtained as follows:

[0062]

[0063] in, represents the individual after the qth volatility search in the tth training, Represents the individual after the escape search , represents the escape search coefficient, represents the optimal individual, Represents a random number between (0,1), Represents a random number between (0,1).

[0064] Further, based on the short-term predicted water level, the water conservancy project dam is safety monitored to obtain the water conservancy project dam safety monitoring result, and an early warning is issued based on the water conservancy project dam safety monitoring result, including:

[0065] Determine whether the short-term predicted water level exceeds a preset danger threshold, and if so, determine that the water conservancy project dam safety monitoring result is dangerous, otherwise determine that the water conservancy project dam safety monitoring result is not dangerous;

[0066] When the safety monitoring result of the water conservancy project dam shows that there is danger, an early warning message is generated and the early warning message is transmitted to the equipment designated by the staff.

[0067] In another aspect, the present invention provides a method for monitoring the safety of a dam in a water conservancy project, comprising:

[0068] Obtain the historical water level impact characteristics of the area where the water conservancy project dam is located and the historical water level corresponding to the historical water level impact characteristics;

[0069] A deep learning algorithm is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics to obtain a water level prediction model;

[0070] Acquire the real-time water level impact characteristics of the area where the water conservancy project dam is located, and transmit the real-time water level impact characteristics to the water level prediction module;

[0071] The water level prediction model is dispatched to identify the influencing features of the real-time water level and determine the short-term predicted water level;

[0072] According to the short-term predicted water level, the water conservancy project dam is safety monitored to obtain the water conservancy project dam safety monitoring result, and an early warning is issued according to the water conservancy project dam safety monitoring result.

[0073] The present invention provides a water conservancy project dam safety monitoring device and monitoring method. By adopting a deep learning algorithm to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, a water level prediction model that can predict the water level is obtained. Then, the real-time water level impact characteristics are identified through the water level prediction model, and finally the water level prediction is realized. Based on the actual data, it can quickly predict whether there may be danger, so that the staff can respond in time, thereby realizing the safety of the water conservancy project dam. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0075] Figure 1 The present invention provides a schematic structural diagram of a water conservancy project dam safety monitoring device.

[0076] Figure 2 The present invention provides a flow chart of a method for monitoring the safety of a dam in a water conservancy project.

[0077] Among them, 1-historical data acquisition module, 2-deep learning module, 3-real-time data acquisition module, 4-water level prediction module, and 5-safety warning module.

[0078] The above drawings have shown clear embodiments of the present invention, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present invention in any way, but to illustrate the concept of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0079] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0080] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0081] like Figure 1As shown, the present invention provides a water conservancy project dam safety monitoring device, comprising: a historical data acquisition module 1, a deep learning module 2, a real-time data acquisition module 3, a water level prediction module 4 and a safety warning module 5;

[0082] The historical data acquisition module 1 is used to obtain the historical water level impact characteristics of the area where the water conservancy project dam is located and the historical water level corresponding to the historical water level impact characteristics;

[0083] In an embodiment of the present invention, historical water level influence characteristics of the area where the water conservancy project dam is located and historical water levels corresponding to the historical water level influence characteristics are obtained, including: obtaining water levels, flow rates, rainfall amounts and rainfall levels at multiple time points in a historical period to obtain historical water level influence characteristics of the area where the water conservancy project dam is located; for any historical water level influence characteristic, obtaining the water level after a short period of time to obtain the historical water level corresponding to any historical water level influence characteristic; wherein, short-term is used to represent a fixed time period.

[0084] For example, for any historical moment, the water level, flow, rainfall, and rainfall level at multiple time nodes within the previous hour or several hours can be obtained. The rainfall and rainfall level at each time node can be obtained through the local weather forecast, so that the future water level can be predicted based on these data, and then the water level after one hour or several hours is used as the historical water level corresponding to the historical water level impact feature. However, it is worth noting that the above feature types and times are all distances, and those skilled in the art can adjust them according to actual needs.

[0085] The deep learning module 2 is used to learn the historical water level influencing features and the historical water levels corresponding to the historical water level influencing features by using a deep learning algorithm to obtain a water level prediction model;

[0086] The deep learning algorithm may be a recurrent neural network (RNN), a long short-term memory network (LSTM), a gated recurrent unit (GRU), and a convolutional neural network (CNN). Technicians in this field may set the specific algorithm type according to actual needs to determine the water level prediction model with the best effect.

[0087] A deep learning algorithm is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics to obtain a water level prediction model. This can include: taking the historical water level impact characteristics as the actual input, taking the historical water level corresponding to the historical water level impact characteristics as the output, and using an intelligent optimization algorithm or a gradient descent method to optimize the hyperparameters corresponding to the deep learning algorithm, so as to obtain a water level prediction model.

[0088] The real-time data acquisition module 3 is used to obtain the real-time water level impact characteristics of the area where the water conservancy project dam is located, and transmit the real-time water level impact characteristics to the water level prediction module;

[0089] After optimizing the deep learning algorithm, a water level prediction model that can predict the water level is obtained. The water level prediction model can then be used to identify the characteristics that affect the real-time water level.

[0090] The water level prediction module 4 is used to dispatch the water level prediction model to identify the influencing features of the real-time water level and determine the short-term predicted water level;

[0091] The safety warning module 5 is used to perform safety monitoring on the water conservancy project dam according to the short-term predicted water level, obtain the water conservancy project dam safety monitoring result, and issue a warning according to the water conservancy project dam safety monitoring result.

[0092] The early warning can be an audio-visual warning or an information warning, that is, when the safety monitoring results of the water conservancy project dam are unsafe, the early warning information is transmitted to the staff's equipment.

[0093] The present invention provides a water conservancy project dam safety monitoring device, which uses a deep learning algorithm to learn historical water level impact characteristics and historical water levels corresponding to the historical water level impact characteristics, to obtain a water level prediction model that can predict the water level, and then uses the water level prediction model to identify the real-time water level impact characteristics, and finally realizes the prediction of the water level. It can quickly predict whether there may be danger based on actual data, so that the staff can respond in time, thereby realizing the safety of the water conservancy project dam.

[0094] In an embodiment of the present invention, a deep learning algorithm is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics to obtain a water level prediction model, including:

[0095] The BiGRU-CapsNet (Bidirectional Gated Recurrent Unit-Capsule Network) algorithm is used to build the initial prediction model;

[0096] It is worth noting that the above BiGRU-CapsNet algorithm is only a preferred method of the embodiment of the present invention, and other algorithms can also be used to construct the initial prediction model.

[0097] According to the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, an improved intelligent optimization algorithm is used to train the initial prediction model to obtain a water level prediction model.

[0098] In the prior art, back propagation is generally used to optimize hyperparameters (such as connection weights, biases, etc.), which often results in insufficient training accuracy and easy to fall into local optimality, which ultimately leads to poor water level prediction capabilities. Therefore, the embodiment of the present invention provides an improved intelligent optimization algorithm to train the initial prediction model, thereby solving the technical problems existing in the prior art and improving the accuracy of water level prediction.

[0099] In the embodiment of the present invention, according to the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, an improved intelligent optimization algorithm is used to train the initial prediction model to obtain a water level prediction model, including:

[0100] For the parameters to be optimized of the initial prediction model, a random initialization method is used to initialize the optimization population;

[0101] For example, for any parameter to be optimized, it can be randomly initialized between its upper and lower limits to obtain a single individual, and then repeatedly obtain multiple different individuals to achieve optimal population initialization.

[0102] The fitness of each individual in the optimal population is obtained according to the historical water level impact characteristics and the historical water level corresponding to the historical water level impact characteristics, and the first target individual, the second target individual, the third target individual and the fourth target individual are determined in descending order of fitness; among which the first target individual is the optimal individual;

[0103] According to the first target individual, each individual in the optimization population is ordered to perform a local search in the optimal direction, and the individual after the local search is determined;

[0104] For the individuals after the local search, a joint search is performed on each individual in the optimal population according to the first target individual, the second target individual, the third target individual, and the fourth target individual to obtain the individuals after the joint search;

[0105] For the individuals after the joint search, each individual in the optimal population is required to perform a volatility search to obtain the individuals after the volatility search;

[0106] For the individuals after the volatility search, each individual in the optimization population is required to perform an escape search around the optimal position to obtain the individuals after the escape search;

[0107] The above local search, joint search, volatility search and escape search are repeated until the maximum number of training times is reached. The final hyperparameters of the initial prediction model are determined according to the individuals after the escape search to obtain the water level prediction model.

[0108] In the embodiment of the present invention, according to the first target individual, each individual in the optimization population is made to perform a local search in the optimal direction, and the individual after the local search is determined, including:

[0109] Determine the current number of training times t, and determine the convergence factor based on the current number of training times t:

[0110]

[0111] in, represents the convergence factor during the t-th training process, represents the starting value of the convergence factor, represents the final value of the convergence factor, represents the maximum number of training times; and =2.5, =0;

[0112] The nonlinear convergence factor provided in the embodiment of the present invention decays slowly in the early stage of iteration and maintains a large value to enhance the global exploration capability; it decays rapidly in the later stage of iteration and decreases to a smaller value to improve the local search capability and increase the probability of the algorithm jumping out of the local extreme value.

[0113] According to the convergence factor And the first target individual, perform local search on each individual in the optimal population, and the individuals after local search are:

[0114]

[0115]

[0116]

[0117] in, Indicates t During the training process m Individuals, Represents the individual after local search , m=1,2,…,M, M represents the total number of individuals, represents the first local search factor, represents the second local search factor, represents the first target individual, | | represents taking the absolute value of the element, represents a random number uniformly distributed between [0,1], Represents a random number between (0,1), Represents a random number between (0,1).

[0118] The local search provided by the embodiment of the present invention can enable the algorithm to continuously search for the optimal position based on its own position and the position corresponding to the optimal individual, thereby preliminarily realizing the search of the local area.

[0119] In the embodiment of the present invention, for the individual after the local search, a joint search is performed on each individual in the optimal population according to the first target individual, the second target individual, the third target individual, and the fourth target individual to obtain the individual after the joint search, including:

[0120] For the individuals after the local search, the first search position is determined according to the first target individual:

[0121]

[0122] in, indicates the first search position, represents the individual after the nth local search in the tth training process, n=1,2,…,M;

[0123] The second search position is determined according to the second target individual:

[0124]

[0125] in, indicates the second search position, represents the second target individual;

[0126] The third search position is determined according to the third target individual:

[0127]

[0128] in, indicates the third search position, represents the third target individual;

[0129] The fourth search position is determined according to the fourth target individual:

[0130]

[0131] in, indicates the fourth search position, represents the fourth target individual;

[0132] Identify individuals The Manhattan distances between the first target individual, the second target individual, the third target individual, and the fourth target individual, and the first weighted weight, the second weighted weight, the third weighted weight, and the fourth weighted weight are determined based on the Manhattan distances:

[0133]

[0134]

[0135]

[0136]

[0137] in, represents the first weighted weight, represents the second weighted weight, represents the third weighted weight, represents the fourth weighted weight, Represents an individual The distance to the first target individual, Represents an individual The distance to the second target individual, Represents an individual The distance to the third target individual, Represents an individual The distance to the fourth target individual;

[0138] According to the first search position, the second search position, the third search position, the fourth search position, the first target individual, the second target individual, the third target individual and the fourth target individual, the individual after the joint search is obtained is:

[0139]

[0140] in, Represents the individual after joint search .

[0141] The joint search provided by the embodiment of the present invention can not only perform learning in the optimal area, but also make each individual more inclined to learn target individuals with larger differences and exclude target individuals with smaller differences. It can effectively alleviate the serious assimilation of the algorithm in the later stage and improve the possibility of the algorithm searching for the optimal position in the optimal area.

[0142] In the embodiment of the present invention, for the individual after the joint search, each individual in the optimal population is required to perform a volatility search to obtain the individual after the volatility search, including:

[0143] For the individuals after the joint search, each individual is randomly matched with a different other individual to obtain the matching individual corresponding to each individual;

[0144] According to the matching individuals corresponding to each individual, the individuals are searched for volatility, and the individuals after the volatility search are obtained as follows:

[0145]

[0146] in, represents the individual after the kth joint search, represents the individual after the volatility search , k=1,2,…,M; represents the maximum number of training times, sin represents the sine function, cos represents the cosine function, represents pi, Represents a random number between (0,1), Represents a random number between (0,1), Represents a random number between (0,1), Represents an individual The corresponding matching individuals.

[0147] Generally, algorithms are divided into two stages when searching for the best solution, namely global search and local search. These two stages promote and contradict each other. Global search is used to quickly locate the range of the optimal solution, and local search is used to find the optimal solution. When there are too many global searches, the operation will be very slow, and when there are too many local searches, it is easy to fall into the local optimum. Therefore, the volatility search provided by the embodiment of the present invention can enable individuals to perform volatility searches towards the optimal position, and can explore different areas more widely, balancing the local and global development capabilities of the algorithm.

[0148] In the embodiment of the present invention, for the individuals after the volatility search, each individual in the optimization population is required to perform an escape search around the optimal position, and the individuals after the escape search are obtained as follows:

[0149]

[0150] in, represents the individual after the qth volatility search in the tth training, Represents the individual after the escape search , represents the escape search coefficient, represents the optimal individual, Represents a random number between (0,1), Represents a random number between (0,1).

[0151] The escape search provided in the embodiment of the present invention can enable the algorithm to have the ability to escape the optimal area, thereby effectively avoiding the algorithm from falling into the local optimum. However, it is worth noting that the escape search process can be controlled by a greedy algorithm or an annealing simulation algorithm, thereby effectively ensuring the convergence speed of the algorithm.

[0152] Optionally, after each search, out-of-bounds processing can be performed on the parameters in the individual to ensure the search effectiveness of the algorithm.

[0153] In an embodiment of the present invention, the water conservancy project dam is safety monitored according to the short-term predicted water level to obtain a water conservancy project dam safety monitoring result, and an early warning is issued according to the water conservancy project dam safety monitoring result, including:

[0154] Determine whether the short-term predicted water level exceeds a preset danger threshold, and if so, determine that the water conservancy project dam safety monitoring result is dangerous, otherwise determine that the water conservancy project dam safety monitoring result is not dangerous;

[0155] When the safety monitoring result of the water conservancy project dam shows that there is danger, an early warning message is generated and the early warning message is transmitted to the equipment designated by the staff.

[0156] like Figure 2 As shown, the present invention provides a method for monitoring the safety of a dam in a water conservancy project, comprising:

[0157] S1. Obtain the historical water level impact characteristics of the area where the water conservancy project dam is located and the historical water level corresponding to the historical water level impact characteristics;

[0158] S2. Using a deep learning algorithm to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, so as to obtain a water level prediction model;

[0159] S3, obtaining the real-time water level impact characteristics of the area where the water conservancy project dam is located, and transmitting the real-time water level impact characteristics to the water level prediction module;

[0160] S4, dispatching the water level prediction model to identify the influencing features of the real-time water level and determine the short-term predicted water level;

[0161] S5. Conduct safety monitoring of the water conservancy project dam according to the short-term predicted water level, obtain safety monitoring results of the water conservancy project dam, and issue an early warning based on the safety monitoring results of the water conservancy project dam.

[0162] A method for monitoring the safety of a dam in a water conservancy project provided in an embodiment of the present invention is applied to the above-mentioned device. Its principles and beneficial effects are similar and will not be described in detail here.

[0163] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0164] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0165] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0166] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0167] A person of ordinary skill in the art can understand that all or part of the steps in realizing the above-mentioned facts and methods can be completed by instructing the relevant hardware through a program, and the program involved or the program described can be stored in a computer-readable storage medium. When the program is executed, it includes the following steps: At this time, the corresponding method steps are derived, and the storage medium can be ROM / RAM, a disk, an optical disk, etc.

[0168] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A water conservancy project dam safety monitoring device, characterized in that: include: Historical data acquisition module, deep learning module, real-time data acquisition module, water level prediction module and safety warning module; The historical data acquisition module is used to obtain the historical water level impact characteristics of the area where the water conservancy project dam is located and the historical water level corresponding to the historical water level impact characteristics; The deep learning module is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics using a deep learning algorithm to obtain a water level prediction model; The real-time data acquisition module is used to acquire the real-time water level impact characteristics of the area where the water conservancy project dam is located, and transmit the real-time water level impact characteristics to the water level prediction module; The water level prediction module is used to dispatch the water level prediction model to identify the influencing features of the real-time water level and determine the short-term predicted water level; The safety early warning module is used to perform safety monitoring on the water conservancy project dam according to the short-term predicted water level, obtain the safety monitoring result of the water conservancy project dam, and issue an early warning according to the safety monitoring result of the water conservancy project dam; A deep learning algorithm is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics to obtain a water level prediction model, including: The BiGRU-CapsNet algorithm is used to build the initial prediction model; According to the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, the improved intelligent optimization algorithm is used to train the initial prediction model to obtain the water level prediction model; According to the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, the improved intelligent optimization algorithm is used to train the initial prediction model to obtain the water level prediction model, including: For the parameters to be optimized of the initial prediction model, a random initialization method is used to initialize the optimization population; The fitness of each individual in the optimal population is obtained according to the historical water level impact characteristics and the historical water level corresponding to the historical water level impact characteristics, and the first target individual, the second target individual, the third target individual and the fourth target individual are determined in descending order of fitness; among which the first target individual is the optimal individual; According to the first target individual, each individual in the optimization population is ordered to perform a local search in the optimal direction, and the individual after the local search is determined; For the individuals after the local search, a joint search is performed on each individual in the optimal population according to the first target individual, the second target individual, the third target individual, and the fourth target individual to obtain the individuals after the joint search; For the individuals after the joint search, each individual in the optimal population is required to perform a volatility search to obtain the individuals after the volatility search; For the individuals after the volatility search, each individual in the optimization population is required to perform an escape search around the optimal position to obtain the individuals after the escape search; The above local search, joint search, volatility search and escape search are repeated until the maximum number of training times is reached. The final hyperparameters of the initial prediction model are determined according to the individuals after the escape search to obtain the water level prediction model.

2. The water conservancy project dam safety monitoring device according to claim 1 is characterized in that: Obtain the historical water level impact characteristics of the area where the water conservancy project dam is located and the historical water level corresponding to the historical water level impact characteristics, including: Obtain the water level, flow, rainfall and rainfall level at multiple time points in the historical period, and obtain the historical water level impact characteristics of the area where the water conservancy project dam is located; For any historical water level impact feature, obtain the water level after a short period of time, and obtain the historical water level corresponding to any historical water level impact feature; Among them, short-term is used to represent a fixed time period.

3. The water conservancy project dam safety monitoring device according to claim 2 is characterized in that: According to the first target individual, each individual in the optimization population is made to perform a local search in the optimal direction, and the individual after the local search is determined, including: Determine the current number of training times t, and determine the convergence factor based on the current number of training times t: in, represents the convergence factor during the t-th training process, represents the starting value of the convergence factor, represents the final value of the convergence factor, represents the maximum number of training times; and =2.5, =0; According to the convergence factor And the first target individual, perform local search on each individual in the optimal population, and the individuals after local search are: in, Indicates t During the training m Individuals, Represents the individual after local search , m=1,2,…,M, M represents the total number of individuals, represents the first local search factor, represents the second local search factor, represents the first target individual, | | represents taking the absolute value of the element, represents a random number uniformly distributed between [0,1], Represents a random number between (0,1), Represents a random number between (0,1).

4. The water conservancy project dam safety monitoring device according to claim 3 is characterized in that: For the individuals after the local search, a joint search is performed on each individual in the optimal population according to the first target individual, the second target individual, the third target individual, and the fourth target individual to obtain the individuals after the joint search, including: For the individuals after the local search, the first search position is determined according to the first target individual: in, indicates the first search position, represents the individual after the nth local search in the tth training process, n=1,2,…,M; The second search position is determined according to the second target individual: in, indicates the second search position, represents the second target individual; The third search position is determined according to the third target individual: in, indicates the third search position, represents the third target individual; The fourth search position is determined according to the fourth target individual: in, indicates the fourth search position, represents the fourth target individual; Identify individuals The Manhattan distances between the first target individual, the second target individual, the third target individual, and the fourth target individual, and the first weighted weight, the second weighted weight, the third weighted weight, and the fourth weighted weight are determined based on the Manhattan distances: in, represents the first weighted weight, represents the second weighted weight, represents the third weighted weight, represents the fourth weighted weight, Represents an individual The distance to the first target individual, Represents an individual The distance to the second target individual, Represents an individual The distance to the third target individual, Represents an individual The distance to the fourth target individual; According to the first search position, the second search position, the third search position, the fourth search position, the first target individual, the second target individual, the third target individual and the fourth target individual, the individual after the joint search is obtained is: in, Represents the individual after joint search .

5. The water conservancy project dam safety monitoring device according to claim 4 is characterized in that: For the individuals after the joint search, each individual in the optimal population is required to perform a volatility search to obtain the individuals after the volatility search, including: For the individuals after the joint search, each individual is randomly matched with a different other individual to obtain the matching individual corresponding to each individual; According to the matching individuals corresponding to each individual, the individuals are searched for volatility, and the individuals after the volatility search are obtained as follows: in, represents the individual after the kth joint search, represents the individual after the volatility search , k=1,2,…,M; represents the maximum number of training times, sin represents the sine function, cos represents the cosine function, represents pi, Represents a random number between (0,1), Represents a random number between (0,1), Represents a random number between (0,1), Represents an individual The corresponding matching individuals.

6. The water conservancy project dam safety monitoring device according to claim 5, characterized in that: For the individuals after the volatility search, each individual in the optimization population is required to perform an escape search around the optimal position, and the individuals after the escape search are obtained as follows: in, represents the individual after the qth volatility search in the tth training, Represents the individual after the escape search , represents the escape search coefficient, represents the optimal individual, Represents a random number between (0,1), Represents a random number between (0,1).

7. The water conservancy project dam safety monitoring device according to claim 6, characterized in that: According to the short-term predicted water level, the water conservancy project dam is safety monitored to obtain the water conservancy project dam safety monitoring result, and an early warning is issued according to the water conservancy project dam safety monitoring result, including: Determine whether the short-term predicted water level exceeds a preset danger threshold, and if so, determine that the water conservancy project dam safety monitoring result is dangerous, otherwise determine that the water conservancy project dam safety monitoring result is not dangerous; When the safety monitoring result of the water conservancy project dam shows that there is danger, an early warning message is generated and the early warning message is transmitted to the equipment designated by the staff.

8. A method for monitoring the safety of a water conservancy project dam, the method for monitoring the safety of a water conservancy project dam being applied to the water conservancy project dam safety monitoring device according to any one of claims 1 to 7, characterized in that: include: Obtain the historical water level impact characteristics of the area where the water conservancy project dam is located and the historical water level corresponding to the historical water level impact characteristics; A deep learning algorithm is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics to obtain a water level prediction model; Acquire the real-time water level impact characteristics of the area where the water conservancy project dam is located, and transmit the real-time water level impact characteristics to the water level prediction module; The water level prediction model is dispatched to identify the influencing features of the real-time water level and determine the short-term predicted water level; Conduct safety monitoring of the water conservancy project dam according to the short-term predicted water level, obtain the water conservancy project dam safety monitoring result, and issue an early warning according to the water conservancy project dam safety monitoring result; A deep learning algorithm is used to learn the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics to obtain a water level prediction model, including: The BiGRU-CapsNet algorithm is used to build the initial prediction model; According to the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, the improved intelligent optimization algorithm is used to train the initial prediction model to obtain the water level prediction model; According to the historical water level impact characteristics and the historical water levels corresponding to the historical water level impact characteristics, the improved intelligent optimization algorithm is used to train the initial prediction model to obtain the water level prediction model, including: For the parameters to be optimized of the initial prediction model, a random initialization method is used to initialize the optimization population; The fitness of each individual in the optimal population is obtained according to the historical water level impact characteristics and the historical water level corresponding to the historical water level impact characteristics, and the first target individual, the second target individual, the third target individual and the fourth target individual are determined in descending order of fitness; among which the first target individual is the optimal individual; According to the first target individual, each individual in the optimization population is ordered to perform a local search in the optimal direction, and the individual after the local search is determined; For the individuals after the local search, a joint search is performed on each individual in the optimal population according to the first target individual, the second target individual, the third target individual, and the fourth target individual to obtain the individuals after the joint search; For the individuals after the joint search, each individual in the optimal population is required to perform a volatility search to obtain the individuals after the volatility search; For the individuals after the volatility search, each individual in the optimization population is required to perform an escape search around the optimal position to obtain the individuals after the escape search; The above local search, joint search, volatility search and escape search are repeated until the maximum number of training times is reached. The final hyperparameters of the initial prediction model are determined according to the individuals after the escape search to obtain the water level prediction model.

Images