Search

CN-121614877-B - Marine disaster-causing biological monitoring system and method for coastal nuclear power station

CN121614877BCN 121614877 BCN121614877 BCN 121614877BCN-121614877-B

Abstract

The invention discloses a marine disaster-causing biological monitoring system and a method for a coastal nuclear power station, in particular to the technical field of marine monitoring, which acquire multi-source heterogeneous data in the marine environment of the water intake area of the coastal nuclear power station through an optical imaging subunit, an acoustic detection subunit and an environmental parameter sensing subunit which are deployed in the water intake area, perform fusion analysis and feature extraction on the acquired multi-source heterogeneous data, and outputting a species identification result, the number of individuals and individual length data, constructing a biological disaster quantity quantitative curve, carrying out coupling analysis on the biological disaster quantity quantitative curve and key operation parameters of a cold source system of the nuclear power station, calculating a comprehensive risk index of the current biological situation on the cold source system, automatically generating grading early warning information according to the comprehensive risk index, and providing an automatic control instruction for operators of the nuclear power station based on the generated early warning information and the comprehensive risk index.

Inventors

  • JI JIANDA
  • LIN JING
  • ZHANG JINZHAO

Assignees

  • 自然资源部第三海洋研究所

Dates

Publication Date
20260508
Application Date
20260203

Claims (8)

  1. 1. The marine disaster-causing biological monitoring system for the coastal nuclear power station is characterized by comprising a data acquisition module, an edge intelligent processing module, a risk quantification and early warning module and an intelligent decision and control module; The data acquisition module comprises an optical imaging subunit, an acoustic detection subunit and an environmental parameter sensing subunit, and is used for acquiring multi-source heterogeneous data in the marine environment of the water intake area of the coastal nuclear power station; The edge intelligent processing module is in communication connection with the data acquisition module and is used for carrying out fusion analysis and feature extraction on the multi-source heterogeneous data and outputting species identification results, individual quantity and individual length data; the risk quantification and early warning module is connected with the edge intelligent processing module, generates a biological disaster quantity quantitative curve based on data output by the edge intelligent processing module, performs coupling analysis on the biological disaster quantity quantitative curve and key operation parameters of a cold source system of the nuclear power station, calculates a comprehensive risk index of the cold source system caused by the current biological situation, and automatically generates grading early warning information according to the comprehensive risk index, wherein the grading early warning information specifically comprises the following contents: Based on species identification results, individual number and individual length data output by the edge intelligent processing module, disaster-causing quantitative models of different disaster-causing organisms are built, and disaster-causing intensity functions are built for single-species disaster-causing organisms according to physical characteristics of the single-species disaster-causing organisms For quantifying individual species of the species at different body lengths Disaster potential of the lower part; for a particular disaster causing organism class j, the total disaster causing amount at time point t The sum of all individual disaster-causing contribution degrees of the category is defined as: , wherein, Is the number of individuals of the disaster causing organism class j identified at time t, Is the body length of the ith individual of the category at time t; the total disaster-causing quantity of the marine organisms at the current moment is obtained by accumulating all the identified disaster-causing organism types Wherein J is the number of species of the disaster causing organism, Is a weight coefficient and will Drawing a biological disaster-causing quantitative curve according to the change of time; the key operation parameters of the cold source system of the nuclear power station are obtained, including the pressure difference of a filter screen Flow rate of circulating water End difference of condenser Power of circulating water pump Quantitatively evaluating the comprehensive risk of the cold source system caused by the current biological disaster situation according to the total disaster causing amount of the living beings and key operation parameters of the cold source system to obtain a comprehensive risk index , wherein, Is the first Normalized risk values for the individual critical operating parameters at time t, Is the number of key operating parameters that are to be used, Is the first Risk weights for the individual critical operating parameters, The risk weight of the biological total disaster causing quantity represents the direct risk of the biological total disaster causing quantity on the system; The intelligent decision and control module is respectively connected with the risk quantification and early warning module and the control interface of the cold source system of the nuclear power station, and provides an automatic control instruction for the operator of the nuclear power station based on the early warning information and the comprehensive risk index generated by the risk quantification and early warning module.
  2. 2. The marine disaster-causing biological monitoring system for the coastal nuclear power plant of claim 1 wherein the optical imaging subunit is deployed in an underwater area of a water intake and is used for continuously collecting high-definition video streams and static images containing disaster-causing biological targets; The acoustic detection subunit is deployed in an underwater area of the water intake, and is used for transmitting and receiving sound waves to detect organisms in a water body by utilizing multi-band sonar, so as to obtain the position, density and motion trail information of a biological group, and specifically comprises the following contents: Acquiring the acoustic target intensity, volume scattering intensity and radial movement velocity relative to sonar of a living being by transmitting broadband acoustic pulses of different center frequencies and analyzing echo signals received from a body of water, the acoustic target intensity being obtained by measuring the transmitted sound source level Sound pressure level of reception Propagation loss And the system gain G, the formula is , wherein, , Is the acoustic target intensity, R is the target distance, Is the absorption attenuation coefficient; the volume scattering intensity is obtained by summing the power of the acoustic target intensities of all N detection targets in the sampling volume and taking the logarithm, and reflects the density of organisms in the unit volume, and the specific formula is that , wherein, Is the acoustic target intensity of the a-th probe target; the radial movement speed is obtained by measuring the frequency change of the echo signal relative to the transmitting signal Calculated, the specific formula is Wherein, c is the sound velocity in water, In order to emit the center frequency of the sound wave, Is the radial velocity component of the target along the direction of propagation of the acoustic wave; Detecting the same target at different spatial positions by using a multi-sonar equipment triangulation method, acquiring three-dimensional coordinates of the target, performing time sequence analysis on sonar echo data, and matching and linking the same target detected at different moments by using a target association and tracking algorithm to construct a continuous motion path of the same target within a period of time; the environment parameter sensing subunit is deployed in the water intake water body, integrates a multi-parameter water quality sensor array, comprises a high-precision digital water temperature sensor, a conductivity sensor, a photoelectric scattering turbidity sensor, a three-dimensional ADCP flow rate and a fluorescence chlorophyll a sensor, and is used for monitoring and transmitting marine environment parameters including water temperature, salinity, turbidity, flow rate and chlorophyll concentration data in real time.
  3. 3. The marine disaster-causing biological monitoring system for a coastal nuclear power station of claim 2 wherein the method is characterized in that the method comprises the steps of detecting the same target at different spatial positions by a multi-sonar equipment triangulation method, obtaining three-dimensional coordinates of the target, performing time sequence analysis on sonar echo data, matching and linking the same target detected at different moments by a target association and tracking algorithm, and constructing a continuous motion path of the same target in a period of time, and further comprises the steps of: deploying at least three multi-frequency sonar devices with known spatial positions and non-collinear with each other When a target is detected by a plurality of sonars, the time delay from the transmission of sound waves to the reception is measured Calculating target to sonar Distance of (2) Will sonar Expressed as coordinates of (a) Solving the equation set by a least square method to obtain the three-dimensional coordinates of the target The system of equations consists of the measured geometry of each sonar: Wherein M is the sonar quantity of the detected target, b is the sonar equipment index, and c is the underwater sound speed; obtaining known target sets of time step k by sonar equipment respectively And a new set of probe points for time step k+1 The method comprises the steps of processing a Kalman filter algorithm with known target states, obtaining states of the targets in different time steps, generating state estimation points, and connecting the state estimation points in continuous time steps for the same target to generate a real-time motion track of the target.
  4. 4. The marine disaster-causing biological monitoring system for the coastal nuclear power plant of claim 1, wherein the edge intelligent processing module is in communication connection with the data acquisition module and is used for carrying out fusion analysis and feature extraction on multi-source heterogeneous data and outputting species identification results, individual number and individual length data, and the marine disaster-causing biological monitoring system specifically comprises the following contents: receiving video stream and static image from the optical imaging subunit, sonar echo data of the acoustic detection subunit and environment parameter data of the environment parameter sensing subunit, performing space-time registration and association on the optical, acoustic and environment data, and extracting visual features, acoustic features and environment features, wherein the method comprises the following steps: Establishing a global coordinate system with the center of a water intake as an origin, and for each optical image pixel point, reversely projecting rays under the global coordinate system through a camera calibration matrix and a pose matrix, wherein for each acoustic detection point, the three-dimensional coordinates of the three-dimensional coordinates are as follows Directly in the global coordinate system by calculating the vertical distance from the acoustic point to the optical ray and setting the threshold value Establishing an association pair of an optical pixel area and an acoustic detection point , wherein, Representing the first in the optical image A plurality of optical pixel points are arranged on the substrate, Representing the first in the optical image The number of acoustic detection points is chosen, Represent the first Rays corresponding to the optical pixel points and the first Vertical distance between the acoustic detection points; Preprocessing the obtained video stream and static image, pre-training on a large marine organism image data set by using a lightweight convolutional neural network model, fine-tuning the pre-training model by using a labeled target disaster-causing organism image data set aiming at target disaster-causing organism types, inputting the preprocessed image into a fine-tuned CNN model, and extracting the output of the last convolutional layer as a high-dimensional visual characteristic diagram And converting the feature map into one-dimensional feature vectors by global averaging pooling , wherein, Representing the first feature vector The elements H, W are the height and the width of the feature map, p represents the position index of the feature map in the height direction, and q represents the position index of the feature map in the width direction; Processing the obtained sonar echo data, including filtering and time-frequency analysis, and constructing a multidimensional feature vector for each detected and tracked acoustic target , wherein, Is the intensity of the acoustic target and, Is the difference in TS of the multiple frequencies, Is the radial velocity component of the velocity, Is the volume scattering intensity; The environmental parameter data provided by the environmental parameter sensing subunit is subjected to standardized processing and spliced into an environmental state vector ; And a double-flow neural network architecture is adopted, visual features, acoustics and environmental features of an optical image are fused, a fine-granularity classification task is executed, disaster-causing organism types are distinguished with high accuracy, and a species identification result, the number of individuals and individual length data are output.
  5. 5. The marine disaster-causing biological monitoring system for a coastal nuclear power plant of claim 4, wherein the system is characterized in that the system adopts a double-flow neural network architecture, combines visual features and acoustic and environmental features of an optical image, performs fine-grained classification tasks, distinguishes disaster-causing biological species with high accuracy, outputs species identification results, individual number and individual length data, and further comprises: to acoustic feature vectors And an environmental state vector Input to the respective full connection layers for dimension alignment and preliminary feature transformation, including acoustic feature transformation Transformation with environmental characteristics And performing splicing operation through splicing operation to form early fusion feature vectors It is passed through several fully connected layers to obtain the fusion characteristic representation of acoustic-environment mode , wherein, Respectively a learning weight matrix of acoustic and environmental features, The bias vectors, reLU as an activation function, 、 The weight matrix and the bias vector of the full connection layer are respectively; fusion features of visual features and acoustic-environmental modalities Weighted fusion is carried out to obtain a final fusion feature vector And fusing the final feature vectors Inputting into a classifier, using Softmax activation function as output layer, predicting the probability distribution of disaster-causing organism, and for the e-th organism category, outputting probability of the output layer Is that , wherein, 、 Is the weight and bias of the classifier output layer, Representing visual features in final fused feature vectors The weight of (a); End-to-end training on a large number of labeled multi-modal datasets by minimizing cross entropy loss functions that optimize parameters of the entire network Where U is the number of samples, E is the number of categories, Is the true class label of sample u, The probability that the model predicted sample u belongs to the class e; By selecting the class with the highest probability as the final species identification result Carrying out boundary box detection and classification on disaster-causing biological individuals in the optical image by using a lightweight target detection model, carrying out target tracking on detection results in continuous video frames, and outputting the number of individuals of different species detected in the visual field; For the successfully-associated optical pixel area-acoustic detection point association pair, according to the acoustic detection point And (3) converting the pixel size of the biological individual in the optical image into the actual physical body length by combining the boundary box information obtained by target detection.
  6. 6. The marine disaster-causing biological monitoring system for the coastal nuclear power plant of claim 1 wherein the comprehensive risk index automatically generates hierarchical early warning information, and specifically comprises the following contents: according to the calculated comprehensive risk index Dividing the early warning information into a plurality of grades, automatically generating corresponding early warning measures and reports, and setting four incremental risk thresholds as To divide the early warning level: When (when) Triggering a first-level blue early warning to indicate that the biological disaster is increased; When (when) Triggering a secondary yellow early warning to indicate that the biological disaster is continuously increased; When (when) Triggering three-level orange early warning to indicate that the biological disaster amount reaches a high level; When (when) And triggering four-level red early warning to indicate that the biological disaster amount reaches a critical level.
  7. 7. The marine disaster-causing biological monitoring system for the coastal nuclear power plant of claim 1, wherein the intelligent decision and control module is connected with the risk quantification and early warning module and a control interface of a cold source system of the nuclear power plant, and specifically comprises the following contents: when the primary blue early warning is triggered, the monitoring frequency of the data acquisition module is automatically adjusted, continuous tracking and observation are carried out on specific biological types and areas, and an early warning and cleaning mode of the grid cleaning equipment is started; When triggering the secondary yellow early warning, adjusting the operation mode of the grating trash remover to an enhanced trash removal mode, and checking the standby circulating water pump to enable the standby circulating water pump to be in a standby state which can be started at any time; When triggering three-level orange early warning, starting all the grid trash removers and the circulating water pumps, running at the maximum frequency and intensity, sending out forced advice for reducing the running load of the nuclear power unit to operators, and sending out loud audible and visual alarm; When the four-level red early warning is triggered, all the cleaning and emergency clearing measures are started, the operation is carried out at the maximum power and speed, a mandatory instruction is immediately sent to an operator, an emergency shutdown program is started, the safety of a nuclear reactor is protected, the highest-level audible and visual alarm is sent, and all relevant personnel and an external emergency mechanism are notified through multiple channels.
  8. 8. A marine disaster-causing biological monitoring method for a coastal nuclear power station, which is applied to the marine disaster-causing biological monitoring system for the coastal nuclear power station according to any one of claims 1 to 7, and is characterized by comprising the following steps: The multi-source heterogeneous data in the marine environment of the water intake area of the coastal nuclear power station is obtained through an optical imaging subunit, an acoustic detection subunit and an environmental parameter sensing subunit which are deployed in the water intake area; Performing fusion analysis and feature extraction on the obtained multi-source heterogeneous data, and outputting species identification results, individual number and individual length data; Based on the output species identification result, the number of individuals and the individual length data, constructing a biological disaster-causing quantitative curve, performing coupling analysis on the biological disaster-causing quantitative curve and key operation parameters of a cold source system of the nuclear power station, calculating a comprehensive risk index of the current biological situation on the cold source system, and automatically generating grading early warning information according to the comprehensive risk index; And providing an automatic control instruction for the operator of the nuclear power station based on the generated early warning information and the comprehensive risk index.

Description

Marine disaster-causing biological monitoring system and method for coastal nuclear power station Technical Field The invention relates to the technical field of ocean monitoring, in particular to an ocean disaster-causing biological monitoring system and method for a coastal nuclear power station. Background The coastal nuclear power station is used as an important clean energy base, and the operation safety of the coastal nuclear power station is closely related to the marine environment condition. Nuclear power plants typically employ seawater as a cooling medium that is introduced into a cooling system through a water intake structure. However, there are a variety of disaster-causing organisms in the marine environment that pose a potential threat to the safe operation of nuclear power plants, with the explosive aggregation of plankton, especially jellyfish and shrimp, being the most prominent. At present, the monitoring of the coastal nuclear power cold source disaster-causing organisms mainly depends on manual inspection, underwater video monitoring or simple acoustic detection equipment. The manual inspection has poor timeliness and limited coverage, can not be performed in severe weather, is easy to be influenced by turbidity of water bodies and illumination conditions, is difficult to accurately count and measure body length when organisms are dense, and is difficult to accurately identify species although the conventional acoustic equipment can detect biomass, and particularly has weak discrimination capability on molluscs such as jellyfish and shrimp. Disclosure of Invention In order to overcome the above-mentioned drawbacks of the prior art, embodiments of the present invention provide a marine disaster-causing biological monitoring system and method for a coastal nuclear power plant, so as to solve the problems set forth in the above-mentioned background art. In order to achieve the above purpose, the invention provides a technical scheme that the marine disaster-causing biological monitoring system for the coastal nuclear power station comprises a data acquisition module, an edge intelligent processing module, a risk quantification and early warning module and an intelligent decision and control module; The data acquisition module comprises an optical imaging subunit, an acoustic detection subunit and an environmental parameter sensing subunit, and is used for acquiring multi-source heterogeneous data in the marine environment of the water intake area of the coastal nuclear power station; The edge intelligent processing module is in communication connection with the data acquisition module and is used for carrying out fusion analysis and feature extraction on the multi-source heterogeneous data and outputting species identification results, individual quantity and individual length data; The risk quantification and early warning module is connected with the edge intelligent processing module, generates a biological disaster quantity quantitative curve based on data output by the edge intelligent processing module, performs coupling analysis on the biological disaster quantity quantitative curve and key operation parameters of a cold source system of the nuclear power station, calculates a comprehensive risk index of the cold source system caused by the current biological situation, and automatically generates grading early warning information according to the comprehensive risk index; The intelligent decision and control module is respectively connected with the risk quantification and early warning module and the control interface of the cold source system of the nuclear power station, and provides an automatic control instruction for the operator of the nuclear power station based on the early warning information and the comprehensive risk index generated by the risk quantification and early warning module. Preferably, as a preferred scheme of the marine disaster-causing biological monitoring system for the coastal nuclear power station, the marine disaster-causing biological monitoring system for the coastal nuclear power station comprises a data acquisition module, a data acquisition module and a data processing module, wherein the data acquisition module comprises an optical imaging subunit, an acoustic detection subunit and an environmental parameter sensing subunit, and is used for acquiring multi-source heterogeneous data in a marine environment in a water intake area of the coastal nuclear power station; The optical imaging subunit comprises at least one high-definition underwater camera, and is provided with a self-adaptive spectrum light supplementing light source and an ultrasonic self-cleaning device so as to ensure the stability of image quality under different illumination and water conditions, and is deployed in an underwater area of a water intake for continuously collecting high-definition video stream and static images containing disaster-causing biological targets; the acou