Early warning analysis method and system for disaster damage of mariculture net cage

Abstract

The invention discloses a disaster damage early warning analysis method and system for a mariculture net cage, and relates to the technical field of marine aquaculture engineering. The method comprises the following steps: sampling wave characteristic parameters at a central point O and each monitoring point Pi of an area S where a to-be-monitored aquaculture net cage is located, constructing a first neural network which takes a normalized value of a wave characteristic difference value between the monitoring point Pi and the central point O as an input quantity and takes a relative position parameter as an output quantity, and determining a disaster damage level at the central point O; constructing a second neural network which takes the wave characteristic parameter at the central point O as an input quantity and the disaster damage level as an output quantity, taking the wave characteristic parameter at the central point O at the current moment as the input quantity, obtaining a wave characteristic predicted value parameter at any point p according to the first neural network, and taking the wave characteristic predicted value parameter as the input quantity to obtain a wave characteristic predicted value parameter at any point p; and acquiring a disaster damage level at the point p through the constructed second neural network, and drawing a disaster damage level distribution diagram of the region S.

Description

Technical Field

[0001] The present invention relates to the technical field of marine aquaculture engineering, and in particular to a method and system for early warning analysis of disaster damage of marine aquaculture cages. Background Art

[0002] Aquaculture cages in my country's offshore and deep seas are under serious threat from severe marine disasters such as storm surges. Every year, marine aquaculture suffers huge economic losses due to disasters such as coastal storm surges. Therefore, there is a need for a disaster damage early warning system and method for offshore aquaculture cages to warn of the damage level of aquaculture cages and assess the degree of damage. Existing offshore aquaculture cages are mainly composed of a floating frame system, a net system, an anchor system and a counterweight system. At present, there are two main methods for studying the dynamic characteristics of aquaculture cages under the influence of marine environmental elements: physical model testing and numerical simulation. The data of the physical model test method is more real and reliable, but the test cycle is long and it is impossible to quickly judge the dynamic characteristics of the cage. With the development of computer technology, numerical simulation has gradually become the main method for studying the dynamic characteristics of marine engineering structures.

[0003] Generally, offshore aquaculture cages do not exist alone, but are composed of several small cages. At present, the relevant research on offshore aquaculture cages mainly focuses on the dynamic characteristics under the action of marine environmental loads. In the simulation process, the cages in the aquaculture area are usually regarded as a whole for finite element analysis, ignoring the impact of marine environmental loads on the internal components of the cages, resulting in a deviation between the analysis results and the actual situation, making the early warning effect unsatisfactory. To this end, we propose a disaster damage early warning analysis method and system for offshore aquaculture cages. Summary of the invention

[0004] The main purpose of the present invention is to provide a method and system for early warning analysis of damage to offshore aquaculture cages, which can effectively solve the problems in the background technology.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] A method for early warning analysis of damage to offshore aquaculture cages, comprising:

[0007] Set the area where the aquaculture cages to be monitored are located as S, select the center point O of area S and several monitoring points P i , where P i It is represented as the i-th monitoring point. Within the monitoring period λ, the center point O of the area S and each monitoring point P i Wave characteristic parameter WO j , W ij Sampling was performed, where WO jIt is represented by the jth wave characteristic parameter sampling value at the center point O; W ij It is represented as the jth wave characteristic parameter sampling value of the i-th monitoring point, wherein the wave characteristic parameter includes at least one of the horizontal drag force, horizontal inertia force, wave height, period, and flow velocity of the wave; the relative position parameter includes azimuth and straight-line distance; the monitoring point P i Located at the boundary of area S and evenly distributed outside the center point O;

[0008] Calculate and obtain monitoring point P i The wave characteristic difference and relative position parameter between the monitoring point P and the center point O are used to construct a first neural network model with the normalized value of the wave characteristic difference as the input and the relative position parameter as the output. The number of neurons in the intermediate layer of the first neural network model is adjusted according to the prediction result until its prediction accuracy is not less than the first expected value. i The normalized formula for the difference in wave characteristics from the center point O is:

[0009]

[0010] In the formula, is the normalized value of the j-th wave characteristic difference; y jmin ,y jmax are the minimum and maximum sampling values ​​of the jth wave characteristic difference within the monitoring period λ, respectively;

[0011] Determine the wave characteristic parameter WO at the center point O at the kth sampling kj Damage level L k , where the disaster damage levels are divided into first-level disasters, second-level disasters, and third-level disasters in order of increasing severity;

[0012] Construct the wave characteristic parameters WO at the center point O kj is the input amount, with the damage level L k A second neural network model for the output quantity, adjusting the number of neurons in the middle layer of the second neural network model according to the prediction result until its prediction accuracy is not less than a second expected value;

[0013] Get the wave characteristic parameter WO at the center point O of area S at the current moment j * , with wave characteristic parameter WO j * As the input, the wave characteristic prediction value parameter W at any point p in the area S is obtained according to the constructed first neural network model. ij * ;

[0014] The wave characteristic prediction value parameter Wij * The disaster damage level at point p in area S is obtained through the constructed second neural network model, and the disaster damage level distribution map of area S is drawn according to the obtained disaster damage level at any point p.

[0015] The calculation formulas of the first expected value and the second expected value are:

[0016]

[0017] Among them, E(Y) r represents the rth expected value; N r represents the number of input samples of the rth neural network; f(X tr ) r represents the output function of the rth neural network; X tr represents the tth output sample of the rth neural network, where r = 1, 2;

[0018] The first neural network and the second neural network are both BP neural networks, wherein the number of neurons in the input layer of the first neural network is equal to the feature type of the input wave feature difference, and the number of neurons in the output layer is equal to the type of the output relative position parameter; the number of neurons in the input layer of the second neural network is equal to the feature type of the input wave feature parameter.

[0019] A disaster damage early warning analysis system for offshore aquaculture cages, comprising a data acquisition module, a data processing module, a first neural network module, a disaster damage level determination module, a second neural network module, and a disaster damage level distribution drawing module;

[0020] The data acquisition module is used to set the area where the aquaculture cages to be monitored are located as S, select the center point O of the area S and several monitoring points P i , where P i It is represented as the i-th monitoring point. Within the monitoring period λ, the center point O of the area S and each monitoring point P i Wave characteristic parameter WO j , W ij Sampling was performed, where WO j It is represented by the jth wave characteristic parameter sampling value at the center point O; W ij It is represented by the sampling value of the j-th wave characteristic parameter at the ith monitoring point;

[0021] The data processing module is used to calculate and obtain the monitoring point P i The wave characteristic difference and relative position parameters between the monitoring point P and the center point O are i The difference of wave characteristics between the center point O is normalized;

[0022] The first neural network module is used to construct a first neural network model with the normalized value of the wave feature difference as input and the relative position parameter as output, and the number of neurons in the middle layer of the first neural network model is adjusted according to the prediction result so that its prediction accuracy is not less than the first expected value, wherein the first neural network is a BP neural network, and the number of neurons in the input layer of the first neural network is equal to the feature type of the input wave feature difference, and the number of neurons in the output layer is equal to the type of the output relative position parameter;

[0023] The disaster damage level determination module is used to determine the wave characteristic parameter WO of the center point O at the kth sampling time. kj The damage level is L k , where the disaster damage levels are divided into first-level disasters, second-level disasters, and third-level disasters in order of increasing severity;

[0024] The second neural network module is used to construct a wave characteristic parameter WO at the center point O. kj is the input amount, with the damage level L k A second neural network model for output, adjusting the number of neurons in the middle layer of the second neural network model according to the prediction result until its prediction accuracy is not less than a second expected value, wherein the second neural network is a BP neural network, and the number of neurons in the input layer of the second neural network is equal to the feature type of the input wave feature parameter;

[0025] The damage level distribution drawing module is used to calculate the wave characteristic parameter WO at the center point O of the area S at the current moment. j * As the input, the wave characteristic prediction value parameter W at any point p in the area S is obtained according to the constructed first neural network model ij * , and the wave characteristic prediction value parameter W ij * As input, the disaster damage level at point p in area S is obtained through the constructed second neural network model, and the disaster damage level distribution map of area S is drawn according to the obtained disaster damage level at any point p;

[0026] The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor.

[0027] The present invention has the following beneficial effects:

[0028] Compared with the prior art, the method sets the area where the aquaculture cages to be monitored as S, selects the center point O of the area S and several monitoring points P i , within the monitoring period λ, the center point O of the area S and each monitoring point Pi Wave characteristic parameter WO j , W ij Sampling is performed and the monitoring point P is calculated. i The wave characteristic difference and relative position parameter between the center point O and the center point O are used to construct a first neural network model with the normalized value of the wave characteristic difference as the input and the relative position parameter as the output. The number of neurons in the intermediate layer of the first neural network model is adjusted according to the prediction result so that its prediction accuracy is not lower than the first expected value, and the wave characteristic parameter WO of the center point O at the kth sampling is determined. kj The damage level is L k , construct the wave characteristic parameter WO at the center point O kj is the input amount, with the damage level L k The second neural network model with the output quantity is used to adjust the number of neurons in the middle layer of the second neural network model according to the prediction result until its prediction accuracy is not lower than the second expected value, and obtain the wave characteristic parameter WO at the center point O of the area S at the current moment. j * , with wave characteristic parameter WO j * As the input, the wave characteristic prediction value parameter W at any point p in the area S is obtained according to the constructed first neural network model ij * , using the wave characteristic prediction value parameter W ij * The second neural network model is used as input to obtain the disaster damage grade at point p in area S. According to the obtained disaster damage grade at any point p, the disaster damage grade distribution map of area S is drawn. Through the constructed neural network, the influence of marine environmental loads on the internal components of the cage is comprehensively considered, and the disaster damage situation at each point inside the cage under the action of marine environmental loads can be truly reflected. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a flow chart of a method for early warning analysis of disaster damage to offshore aquaculture cages according to the present invention;

[0030] Figure 2 This is a structural block diagram of a disaster damage early warning analysis system for offshore aquaculture cages according to the present invention;

[0031] Figure 3 A structural diagram of a first neural network constructed for the solution of the present invention;

[0032] Figure 4 A structural diagram of a second neural network constructed for the solution of the present invention;

[0033] Figure 5It is a schematic diagram of one distribution of monitoring points when the area S is a rectangle in the solution of the present invention;

[0034] Figure 6 It is a schematic diagram of one distribution of monitoring points when the area S is circular in the solution of the present invention. DETAILED DESCRIPTION

[0035] The present invention will be further described below in conjunction with specific implementation methods, wherein the accompanying drawings are only used for exemplary descriptions and represent only schematic diagrams rather than actual drawings, and should not be understood as limiting the present invention. In order to better illustrate the specific implementation methods of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product.

[0036] The implementation process of the technical solution of the present invention includes the following steps:

[0037] Step 1: Set the area where the aquaculture cages to be monitored are located as S, select the center point O of area S and several monitoring points P i , where P i It is represented as the i-th monitoring point. Within the monitoring period λ, the center point O of the area S and each monitoring point P i Wave characteristic parameter WO j , W ij Sampling was performed, where WO j It is represented by the jth wave characteristic parameter sampling value at the center point O; W ij It is expressed as the jth wave characteristic parameter sampling value of the i-th monitoring point, where the wave characteristic parameters include at least one of the horizontal drag force, horizontal inertia force, wave height, period, and flow velocity of the wave; the relative position parameters include azimuth and straight-line distance; the monitoring point P i Located at the boundary of area S and evenly distributed outside the center point O;

[0038] It should be noted that the monitoring point P i Located at the boundary of region S and evenly distributed outside the center point O, such as Figure 5 and Figure 6 As shown in the figure, when the number of monitoring points is ten groups and the area S is rectangular and circular respectively, the monitoring point P i The distribution of can be distributed according to the position in the diagram;

[0039] Step 2: Calculate and obtain monitoring point P i The wave characteristic difference and relative position parameters between the monitoring point P and the center point O, where the wave characteristic difference = monitoring point P i The wave characteristic parameters at - the wave characteristic parameters at the center point O;

[0040] Step 3: Construct a first neural network model with the normalized value of the wave feature difference as input and the relative position parameter as output. According to the prediction results, adjust the number of neurons in the middle layer of the first neural network model until its prediction accuracy is not less than the first expected value. i The normalized formula for the difference in wave characteristics from the center point O is:

[0041]

[0042] In the formula, is the normalized value of the j-th wave characteristic difference; y jmin ,y jmax are the minimum and maximum sampling values ​​of the jth wave characteristic difference within the monitoring period λ respectively; the calculation formula of the first expected value is:

[0043]

[0044] Where E(Y)1 represents the first expected value; N1 represents the number of input samples of the first neural network; f(X t1 )1 represents the output function of the first neural network; X t1 represents the t-th output sample of the first neural network; wherein, the number of neurons in the input layer of the first neural network is equal to the feature type of the input wave feature difference, and the number of neurons in the output layer is equal to the output relative position parameter type; taking the input wave features including the horizontal drag force, horizontal inertia force, wave height, period, and flow velocity of the wave as an example for explanation, in this embodiment, the number of neurons in the input layer of the first neural network is 5, the number of neurons in the output layer is 2, and the initial value of the number of neurons in the intermediate layer can be preliminarily determined according to an empirical formula, and the empirical formula is: Among them, s is the initial value of the number of neurons in the middle layer, p and q are the number of neurons in the input layer and output layer respectively. In this formula, p = 5, q = 2, For The value of is rounded up, a is a positive integer from 1 to 9, and the structure diagram of the first neural network is as follows Figure 3 As shown;

[0045] Step 4: Determine the wave characteristic parameter WO at the center point O at the kth sampling kj Damage level L k , where the disaster damage levels are divided into first-level disasters, second-level disasters, and third-level disasters in order of increasing severity;

[0046] Step 5: Construct the wave characteristic parameters WO at the center point O kj is the input amount, with the damage level L kThe second neural network model with the output quantity is used to adjust the number of neurons in the middle layer of the second neural network model according to the prediction result until its prediction accuracy is not less than the second expected value, wherein the calculation formula of the second expected value is:

[0047]

[0048] Where E(Y)2 represents the second expected value; N2 represents the number of input samples of the second neural network; f(X t2 )2 represents the output function of the second neural network; X t2 represents the t-th output sample of the second neural network; the number of neurons in the input layer of the second neural network is equal to the feature type of the input wave feature parameter;

[0049] In this embodiment, the number of neurons in the input layer of the second neural network is 5, the number of neurons in the output layer is 1, and the initial value of the number of neurons in the intermediate layer can be preliminarily determined according to an empirical formula, which is: Then the structure diagram of the first neural network constructed is as follows Figure 4 As shown;

[0050] Step 6: Get the wave characteristic parameter WO at the center point O of area S at the current moment j * , with wave characteristic parameter WO j * As the input, the wave characteristic prediction value parameter W at any point p in the area S is obtained according to the constructed first neural network model. ij * ;

[0051] Step 7: Use wave characteristics to predict the value parameter W ij * The disaster damage level at point p in area S is obtained through the constructed second neural network model, and the disaster damage level distribution map of area S is drawn according to the obtained disaster damage level at any point p.

[0052] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for early warning analysis of damage to marine aquaculture cages, characterized in that: include: Set the area where the aquaculture cages to be monitored are located as S, select the center point O of area S and several monitoring points P i , where P i It is represented as the i-th monitoring point. Within the monitoring period λ, the center point O of the area S and each monitoring point P i Wave characteristic parameter WO j , W ij Sampling was performed, where WO j It is represented by the jth wave characteristic parameter sampling value at the center point O; W ij It is represented by the sampling value of the j-th wave characteristic parameter at the ith monitoring point; Calculate and obtain monitoring point P i The wave characteristic difference and relative position parameter between the center point O and the center point O are used to construct a first neural network model with the normalized value of the wave characteristic difference as the input and the relative position parameter as the output, and the number of neurons in the middle layer of the first neural network model is adjusted according to the prediction result so that its prediction accuracy is not less than the first expected value; Determine the wave characteristic parameter WO at the center point O at the kth sampling kj Damage level L k , where the disaster damage levels are divided into first-level disasters, second-level disasters, and third-level disasters in order of increasing severity; Construct the wave characteristic parameters WO at the center point O kj is the input amount, with the damage level L k A second neural network model for the output quantity, adjusting the number of neurons in the middle layer of the second neural network model according to the prediction result until its prediction accuracy is not less than a second expected value; Get the wave characteristic parameter WO at the center point O of area S at the current moment j * , with wave characteristic parameter WO j * As the input, the wave characteristic prediction value parameter W at any point p in the area S is obtained according to the constructed first neural network model ij * ; The wave characteristic prediction value parameter W ij * The disaster damage level at point p in area S is obtained through the constructed second neural network model, and the disaster damage level distribution map of area S is drawn according to the obtained disaster damage level at any point p.

2. The method for early warning analysis of damage to marine aquaculture cages according to claim 1, characterized in that: The wave characteristic parameters include at least one of the horizontal drag force, horizontal inertia force, wave height, period, and flow velocity of the wave; and the relative position parameters include azimuth and straight-line distance.

3. The method for early warning analysis of damage to marine aquaculture cages according to claim 1, characterized in that: The calculation formulas of the first expected value and the second expected value are: Among them, E(Y) r represents the rth expected value; N r represents the number of input samples of the rth neural network; f(X tr ) r represents the output function of the rth neural network; X tr represents the tth output sample of the rth neural network, where r=1,2.

4. The method for early warning analysis of damage to marine aquaculture cages according to claim 1, characterized in that: The monitoring point P i Located at the boundary of area S and evenly distributed outside the center point O.

5. The method for early warning analysis of damage to marine aquaculture cages according to claim 1, characterized in that: Monitoring point P i The normalized formula for the difference in wave characteristics from the center point O is: In the formula, is the normalized value of the j-th wave characteristic difference; y jmin ,y jmax are respectively the minimum and maximum sampling values ​​of the j-th wave characteristic difference within the monitoring period λ.

6. The method for early warning analysis of damage to marine aquaculture cages according to claim 1, characterized in that: The first neural network and the second neural network are both BP neural networks, wherein the number of neurons in the input layer of the first neural network is equal to the feature type of the input wave feature difference, and the number of neurons in the output layer is equal to the type of the output relative position parameter; the number of neurons in the input layer of the second neural network is equal to the feature type of the input wave feature parameter.

7. A disaster damage early warning analysis system for offshore aquaculture cages, characterized in that: It includes a data acquisition module, a data processing module, a first neural network module, a disaster damage level determination module, a second neural network module, and a disaster damage level distribution drawing module; The data acquisition module is used to set the area where the aquaculture cages to be monitored are located as S, select the center point O of the area S and several monitoring points P i , where P i It is represented as the i-th monitoring point. Within the monitoring period λ, the center point O of the area S and each monitoring point P i Wave characteristic parameter WO j , W ij Sampling was performed, where WO j It is represented by the jth wave characteristic parameter sampling value at the center point O; W ij It is represented as the j-th wave characteristic parameter sampling value of the i-th monitoring point, wherein the wave characteristic parameter includes at least one of the horizontal drag force, horizontal inertia force, wave height, period, and flow velocity of the wave; the relative position parameter includes azimuth and straight-line distance; The data processing module is used to calculate and obtain the monitoring point P i The wave characteristic difference and relative position parameters between the monitoring point P and the center point O are i The wave characteristic difference between the center point O is normalized, where the normalization formula is: In the formula, is the normalized value of the j-th wave characteristic difference; y jmin ,y jmax are the minimum and maximum sampling values ​​of the jth wave characteristic difference within the monitoring period λ, respectively; The first neural network module is used to construct a first neural network model with the normalized value of the wave feature difference as input and the relative position parameter as output, and the number of neurons in the intermediate layer of the first neural network model is adjusted according to the prediction result so that its prediction accuracy is not less than the first expected value, wherein the first neural network is a BP neural network, and the number of neurons in the input layer of the first neural network is equal to the feature type of the input wave feature difference, and the number of neurons in the output layer is equal to the type of the output relative position parameter; the calculation formula of the first expected value is: Where E(Y)1 represents the first expected value; N1 represents the number of input samples of the first neural network; f(X t1 )1 represents the output function of the first neural network; X t1 represents the tth output sample of the first neural network; The disaster damage level determination module is used to determine the wave characteristic parameter WO of the center point O at the kth sampling time. kj Damage level L k , where the disaster damage levels are divided into first-level disasters, second-level disasters, and third-level disasters in order of increasing severity; The second neural network module is used to construct a wave characteristic parameter WO at the center point O. kj is the input amount, with the damage level L k The second neural network model is an output quantity, and the number of neurons in the middle layer of the second neural network model is adjusted according to the prediction result to make its prediction accuracy not lower than the second expected value, wherein the second neural network is a BP neural network, and the number of neurons in the input layer of the second neural network is equal to the feature type of the input wave feature parameter; the calculation formula of the second expected value is: Where E(Y)2 represents the second expected value; N2 represents the number of input samples of the second neural network; f(X t2 )2 represents the output function of the second neural network; X t2 represents the tth output sample of the second neural network; The damage level distribution drawing module is used to calculate the wave characteristic parameter WO at the center point O of the area S at the current moment. j * As the input, the wave characteristic prediction value parameter W at any point p in the area S is obtained according to the constructed first neural network model ij * , and the wave characteristic prediction value parameter W ij * The disaster damage level at point p in area S is obtained through the constructed second neural network model, and the disaster damage level distribution map of area S is drawn according to the obtained disaster damage level at any point p.

8. The disaster damage early warning analysis system for offshore aquaculture cages according to claim 7, characterized in that: The system comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the steps of the method according to any one of claims 1 to 6 are implemented when the processor executes the program.

Images