Bulk cargo terminal rainwater and sewage early warning system based on weather prediction model

Abstract

The application discloses a rain and sewage early warning system for bulk cargo wharf based on a weather prediction model, which comprises a weather prediction subsystem, a data processing subsystem and a rain and sewage processing subsystem. The weather prediction subsystem obtains a predicted rainfall time and a predicted rainfall amount according to observation information of a national weather observation station and monitoring data of a port operation area. The data processing subsystem processes TSP concentration in port area weather monitoring data, wharf operation ship machine information, wharf operation personnel information, wharf rain and sewage collection equipment state information and wharf rain and sewage processing equipment state information through an isolated forest algorithm. Processed data is input into a rain and sewage early warning subsystem. The rain and sewage early warning subsystem uses DS evidence to deduce a rain and sewage generation probability, and according to the rain and sewage generation probability, sewage amount and pollution concentration, configures wharf rain and sewage collection equipment and processing facilities in advance.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of bulk cargo wharf environment monitoring and early warning management, and particularly relates to a bulk cargo wharf rain and sewage early warning system based on a weather prediction model. BACKGROUND

[0002] Waterway transportation is one of the important modes of comprehensive transportation system; as the hub of waterway transportation, ports play an important role in cargo transportation and transfer. With the acceleration of green port construction process, the monitoring and control of port pollutants have attracted much attention. In the process of cargo loading, transfer and transportation in the port, air or water pollution is easily caused, especially for bulk cargo wharfs; if the pollutants are not timely treated, it is easy to cause port environmental pollution incidents. In rainy days, suspended dust and rainwater are mixed to form mud water; the working cargo of the wharf or the temporarily stored cargo of the yard may be washed by rainwater to produce rain and sewage, and further cause port water pollution incidents. China is a large-scale importer and exporter of coal, ore and grain, and the pollution and control of bulk cargo wharf are crucial for green port construction.

[0003] At present, for the treatment of rain and sewage of bulk cargo wharf, the method of collecting first and then treating is mostly used. For example, the patent for invention with the authorized announcement number CN111039041B discloses a coal-containing sewage treatment system for bulk cargo wharf, which comprises a stacking system, a sewage treatment system and a liftable baffle arranged around the stacking system. The liftable baffle can adjust the bending degree, length and height of the channel according to the amount of rainfall to complete the collection of sewage; the utility model patent with the authorized announcement number CN215828521U discloses a wharf sewage treatment system, which comprises a sedimentation tank, a filter box and a water storage tank. The patent for invention with the publication number CN112279493A discloses a coal wharf sewage treatment equipment, which comprises a sewage collection tank, a mud pumping device and a sludge treatment device.

[0004] The above three patents propose the collection and treatment methods of wharf sewage; when rain and sewage appear, corresponding collection and treatment measures are taken according to the amount of sewage, or the treatment device is innovated and invented, but the rainfall of bulk cargo wharf and the generation of rain and sewage are not predicted. If strong rainfall or continuous rainfall weather occurs, the rain and sewage near the bulk cargo wharf cannot be quickly collected and treated, and a large amount of rain and sewage flows into the sea, which is not conducive to the protection of port water environment. SUMMARY

[0005] The purpose of this invention is to provide a stormwater and sewage early warning system for bulk cargo terminals based on a meteorological forecasting model. The system calculates the probability, range, and pollution concentration of stormwater and sewage in the port area in advance based on the constructed meteorological forecasting model. Port management departments can configure stormwater and sewage collection devices and stormwater and sewage treatment equipment in advance according to the forecast. This solves the problem of emergency response to sudden stormwater and sewage events under the existing technology system, as well as the problem of stormwater and sewage flowing into the sea and causing water pollution when encountering heavy rainfall and continuous rainfall.

[0006] To solve the above-mentioned technical problems, the present invention provides a technical solution: a stormwater and sewage early warning system for bulk cargo terminals based on a meteorological forecasting model, comprising a meteorological forecasting subsystem, a data processing subsystem, and a stormwater and sewage early warning processing subsystem;

[0007] The meteorological forecasting subsystem integrates observation information from national meteorological observation stations and meteorological monitoring data from port areas to generate fused meteorological data. Then, using a GA-BP neural network algorithm, it obtains predicted rainfall time and amount based on the fused meteorological data. Specifically, the meteorological indicators from national meteorological observation station information include rainfall, humidity, temperature, air pressure, wind direction, and wind speed; the meteorological indicators from port area meteorological monitoring data include temperature, humidity, air pressure, wind direction, wind speed, and TSP concentration. The data fusion process involves setting corresponding judgment thresholds for common meteorological indicators from both national meteorological observation station information and port area meteorological monitoring data. Then, the absolute values ​​of the differences between the national meteorological observation station information and the port area meteorological monitoring data for each meteorological indicator are calculated. All obtained absolute values ​​are compared with the corresponding judgment thresholds. If the absolute value of the difference for a certain meteorological indicator is less than its corresponding judgment threshold, then the port area meteorological monitoring data is selected as the input to the GA-BP neural network algorithm for that meteorological indicator; otherwise, the national meteorological observation station information is selected as the input to the GA-BP neural network algorithm.

[0008] The data processing subsystem uses the isolated forest algorithm to clean and remove outliers from the TSP concentration in the meteorological monitoring data of the port area, as well as the information on dock operation vessels and machinery, dock operation personnel, status information of dock rainwater and sewage collection equipment, and status information of dock rainwater and sewage treatment equipment.

[0009] The stormwater and sewage early warning and treatment subsystem obtains the predicted stormwater and sewage concentration at the dock based on the predicted rainfall amount, predicted rainfall time, TSP concentration after data cleaning and outlier removal, and dock operation vessel and machinery information. Based on dock operation personnel information, dock stormwater and sewage collection equipment status information, and dock stormwater and sewage treatment equipment status information after data cleaning and outlier removal, it obtains the current stormwater and sewage collection and treatment capacity. Then, using the DS evidence theory, it determines whether the existing stormwater and sewage equipment meets the demand based on the predicted stormwater and sewage concentration and the current stormwater and sewage collection and treatment capacity. When it is determined that the existing stormwater and sewage equipment does not meet the demand, an early warning is issued.

[0010] According to the above scheme, the GA-BP neural network algorithm is specifically as follows:

[0011] S101. Determine the BP network topology;

[0012] S102. Set the relevant parameter values ​​of the GA algorithm according to the BP network topology. The relevant parameter values ​​of the GA algorithm include the number of individuals in the population, the number of chromosome nodes, the number of iterations, chromosome crossover and mutation probability; where the number of individuals in the population is the number of sampling times of the observation information of the national meteorological observation station and the meteorological monitoring data of the port area, and the number of chromosome nodes is the number of meteorological index types included in the observation information of the national meteorological observation station and the meteorological monitoring data of the port area.

[0013] S103. Initialize the population. Initialize the BP neural network based on the chromosome information contained in each individual in the population. Use the reciprocal of the mean square error of the training results as the fitness value of the individual. Based on the selection algorithm, retain the individual with the highest fitness and perform crossover and mutation calculations on the remaining individuals according to probability to obtain the new generation of population.

[0014] S104. Iteration: Repeatedly calculate the fitness value of the new generation population and perform selection, crossover, and mutation calculations until the convergence condition is met to obtain the optimal chromosome.

[0015] S105. Decode the chromosome of the optimal individual to obtain the optimal weights of each layer of the BP neural network;

[0016] S106. Train the BP neural network and make predictions to obtain the predicted rainfall time and predicted rainfall amount.

[0017] According to the above scheme, the isolated forest algorithm includes a first phase and a second phase.

[0018] The first stage trains t isolated trees to form an isolated forest. The specific steps of the first stage are as follows:

[0019] S201, X = {x1, x2, ..., x} n Let} be a given sample set, in which samples x1 to x2 are defined. nThese are sample points at different times. x i1 ~x id This represents all dimensions of a sample point at a given time, including TSP concentration, information on dockside vessels and machinery, information on dockside personnel, status information of dockside stormwater and sewage collection equipment, and status information of dockside stormwater and sewage treatment equipment; randomly selected from X. The sample points constitute a subset X′ of X and are placed in the root node;

[0020] S202. Randomly specify a dimension q from d dimensions, and randomly generate a cutting point p in the data space formed by the subset X′.

[0021] S203. Generate a hyperplane using this cutting point p to divide the data space formed by the subset X′ into two subspaces. Place sample points with a dimension less than p into the left child node and those with a dimension greater than or equal to p into the right child node.

[0022] S204, recursively S202 and S203, until all leaf nodes have only one sample point or the isolated tree has reached the specified height;

[0023] S205, repeat S201 to S204 until t isolated trees are generated;

[0024] In the second stage, for any sample point x i Let it traverse each isolated tree, based on the sample point x i The height of each isolated tree yields the sample point x. i The average height h(x) in an isolated forest i ), calculate the outlier score s for all sample points,

[0025]

[0026] in, The average height of all standardized sample points in an isolated forest is represented by the following formula.

[0027]

[0028] Where H is the harmonic function, and when the input is k, its output H(k) = ln(k) + ξ, where ξ is Euler's constant;

[0029] After obtaining the outlier scores of all sample points, a mapping from outlier scores to 0-1 is established based on certain threshold screening conditions. 0 represents a sample point as a normal value, and 1 represents a sample point as an outlier. After outlier removal based on the mapping results, the outlier scores of the remaining sample points are recalculated until the mapping results from the outlier scores of all sample points to 0-1 are all 0.

[0030] According to the above scheme, the process by which the stormwater and sewage early warning and treatment subsystem uses the DS evidence theory to determine whether the existing stormwater and sewage equipment meets the requirements is as follows.

[0031] S301. First, establish an identification framework Θ, which represents the complete set of all possible combinations of predicted stormwater and sewage concentrations at the wharf, current stormwater and sewage collection and treatment capacity, and whether existing stormwater and sewage equipment meets the requirements. Specifically,

[0032] Θ = {θ1, θ2, ..., θ} j , ..., θ N}

[0033] Where, θ j An event or element referred to as the recognition frame Θ;

[0034] The set of all subsets of the recognition frame Θ is called the power set, denoted as 2. Θ The power set is represented as,

[0035] 2 Θ ={Φ,{θ1},{θ2},...,{θ n}, {θ1∪θ2}, {θ1∪θ3},..., Θ}

[0036] S302. Establish the basic trust allocation function m, specifically as follows:

[0037]

[0038] m(A) represents the support of evidence for proposition A. Its value is the basic trust assignment value of the proposition. The basic trust assignment value of the empty set is zero, and the sum of the trust values ​​of all other subsets is equal to 1. A that satisfies the condition that m(A) is greater than 0 is called a focal element.

[0039] S303. If multiple pieces of evidence exist, the Dempster synthesis rule is used to synthesize them. Specifically, assuming there are two pieces of evidence E1 and E2 under the recognition framework Θ, where m1 and m2 are the basic assignment functions corresponding to the two pieces of evidence respectively, A i and B j The synthesis method for the focal elements corresponding to the two pieces of evidence is as follows:

[0040]

[0041] Here, K is the conflict coefficient, and the larger K is, the greater the conflict between different pieces of evidence.

[0042] A method for early warning of stormwater and sewage at bulk cargo terminals based on a meteorological forecasting model, implemented using the aforementioned meteorological forecasting model-based stormwater and sewage early warning system, includes the following steps:

[0043] First, using the GA-BP neural network algorithm, based on observation information from the national meteorological observation station and meteorological monitoring data from the port area, the predicted rainfall time and predicted rainfall amount are obtained;

[0044] Subsequently, the isolated forest algorithm was used to clean and remove outliers from the TSP concentration in the meteorological monitoring data of the port area, as well as the information on dock operation vessels and machinery, dock operation personnel, status information of dock rainwater and sewage collection equipment, and status information of dock rainwater and sewage treatment equipment.

[0045] Finally, based on the predicted rainfall, predicted rainfall time, and TSP concentration after data cleaning and outlier removal, as well as information on vessels and machinery operating at the dock, the predicted values ​​of stormwater and sewage concentration at the dock are obtained. Based on the information of dock workers after data cleaning and outlier removal, the status information of stormwater and sewage collection equipment at the dock, and the status information of stormwater and sewage treatment equipment at the dock, the current stormwater and sewage collection and treatment capacity is obtained. Then, using the DS evidence theory, the existing stormwater and sewage equipment is judged to meet the demand based on the predicted values ​​of stormwater and sewage concentration at the dock and the current stormwater and sewage collection and treatment capacity. When it is determined that the existing stormwater and sewage equipment does not meet the demand, an early warning is issued.

[0046] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for early warning of rainwater and sewage at bulk cargo terminals based on a meteorological forecasting model.

[0047] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method for early warning of rainwater and sewage at bulk cargo terminals based on a meteorological prediction model.

[0048] The beneficial effects of this invention are as follows: This invention integrates and analyzes observation information from national meteorological observation stations and meteorological monitoring data from port operation areas. Utilizing the GA-BP neural network algorithm, it accurately obtains predicted rainfall time and amount. By combining the predicted rainfall time and amount with outlier-removed TSP concentration in the port operation area and information on dockside vessels and machinery after outlier removal using the isolated forest algorithm, the probability, scope, and pollution concentration of rainwater and sewage generation in the port area can be calculated in advance. Port management departments can configure rainwater and sewage collection devices and treatment equipment in advance based on the predicted content. Furthermore, based on outlier-removed dockside personnel information, dockside rainwater and sewage collection equipment status information, and dockside rainwater and sewage treatment equipment status information, the current rainwater and sewage collection and treatment capacity of the port area can be obtained. Early warnings can be issued when the current rainwater and sewage collection and treatment capacity cannot meet the predicted future demand for sewage treatment and collection. This solves the problem of emergency response to sudden rainwater and sewage events under the existing technical system, as well as the water pollution caused by rainwater and sewage flowing into the sea due to equipment shortages during heavy rainfall and continuous rainfall. Attached Figure Description

[0049] Figure 1 This is a block diagram of a stormwater and sewage early warning system for bulk cargo terminals based on a meteorological forecasting model, according to an embodiment of the present invention.

[0050] Figure 2 This is a flowchart of a model based on the GA-BP neural network algorithm according to an embodiment of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the described embodiments of this disclosure without creative effort are within the scope of protection of this disclosure.

[0052] See Figure 1 A stormwater and sewage early warning system for bulk cargo terminals based on a meteorological forecasting model includes a meteorological forecasting subsystem, a data processing subsystem, and a stormwater and sewage early warning and processing subsystem.

[0053] Among them, the meteorological forecasting subsystem uses the GA-BP neural network algorithm to obtain the predicted rainfall time and predicted rainfall amount based on the observation information of the national meteorological observation station and the meteorological monitoring data of the port area;

[0054] The data processing subsystem uses the isolated forest algorithm to clean and remove outliers from the TSP concentration in the meteorological monitoring data of the port area, as well as the information on dock operation vessels and machinery, dock operation personnel, status information of dock rainwater and sewage collection equipment, and status information of dock rainwater and sewage treatment equipment.

[0055] The stormwater and sewage early warning and treatment subsystem obtains the predicted stormwater and sewage concentration at the dock based on the predicted rainfall amount, predicted rainfall time, TSP concentration after data cleaning and outlier removal, and dock operation vessel and machinery information. Based on dock operation personnel information, dock stormwater and sewage collection equipment status information, and dock stormwater and sewage treatment equipment status information after data cleaning and outlier removal, it obtains the current stormwater and sewage collection and treatment capacity. Then, using the DS evidence theory, it determines whether the existing stormwater and sewage equipment meets the demand based on the predicted stormwater and sewage concentration and the current stormwater and sewage collection and treatment capacity. When it is determined that the existing stormwater and sewage equipment does not meet the demand, an early warning is issued.

[0056] Furthermore, the information on dockside vessels and machinery includes the type of cargo, load capacity, type of loading and unloading equipment, and technical parameters of the vessels at different times. In this embodiment, it can be assumed that when the type of cargo is fixed, the type of loading and unloading equipment and technical parameters are also fixed (i.e., each type of cargo corresponds to a fixed type of loading and unloading equipment). In this embodiment, when the types of cargo are grain, coal, ore, and others, the corresponding pollution coefficients are 1.5, 2, and 1, respectively.

[0057] The dock workers' information includes the number of workers and their years of service at different times. The longer the workers' years of service, the stronger their ability to treat stormwater and sewage. The corresponding stormwater and sewage treatment capabilities for years of service are shown in the table below:

[0058] Length of service x x < 1 year 1 year ≤ x < 3 years 3 years ≤ x < 5 years 5 years ≤ x < 10 years x ≥ 10 years Rainwater and sewage treatment capacity 1 2 3 4 5

[0059] The status information of the dock's stormwater and sewage collection equipment includes the types and quantities of stormwater and sewage collection equipment at different times, as well as the operating status of each stormwater and sewage collection equipment, including normal and abnormal operating status.

[0060] The status information of the wharf's stormwater and sewage treatment equipment includes the type and quantity of the equipment at different times, specifically the type and quantity of flocculants, including ferric chloride, polyaluminum chloride, polyacrylamide, and aluminum sulfate.

[0061] Furthermore, the observation information from the national meteorological observation station and the meteorological monitoring data from the port area, which are input into the GA-BP neural network algorithm, are first quantified. The observation information from the national meteorological observation station includes rainfall (mm), humidity (%rh), temperature (°C), air pressure (hPa), wind direction, and wind speed (m / s); the meteorological monitoring data from the port area includes temperature (%rh), humidity (°C), air pressure (hPa), wind direction, wind speed (m / s), and TSP concentration (μg / m³). 3 TSP stands for Total Suspended Particulates, including PM10 and PM2.5.

[0062] The rules for quantifying wind direction data are as follows: starting from due north, rotate clockwise and record the directions of due north, northeast, east, southeast, south, southwest, west, and northwest as 0, 45, 90, 135, 180, 225, 270, and 315 respectively. If the wind speed is 0 at this time, the wind direction value is set to -1.

[0063] In this embodiment, when the difference between the indicator item from the observation information of the national meteorological observation station and the same indicator item in the meteorological monitoring data of the port area is less than or equal to 2 at a certain moment, the meteorological monitoring data of the port area is used as the input of the indicator item to the GA-BP neural network; otherwise, the observation information of the national meteorological observation station is used as the input of the indicator item to the GA-BP neural network. In other embodiments of the present invention, different judgment methods can be used for different indicator items, and the judgment method is not limited to comparing the difference with a preset threshold.

[0064] The following is an example of hourly data from June 1, 2022, used as input to the GA-BP neural network algorithm.

[0065]

[0066]

[0067]

[0068] In this embodiment, the selection time range for the observation information from the national meteorological observation station and the meteorological monitoring data of the port area is a single day, so the meteorological data of each hour within a single day is used as input. In other embodiments, the selection time range for the observation information from the national meteorological observation station and the meteorological monitoring data of the port area is multiple consecutive days, so the meteorological data of each 8-hour period within the previous 7 days is selected as input.

[0069] The GA-BP neural network algorithm is specifically as follows:

[0070] S101. Determine the BP network topology;

[0071] S102. Set the relevant parameter values ​​of the GA algorithm according to the BP network topology. The relevant parameter values ​​of the GA algorithm include the number of individuals in the population, the number of chromosome nodes, the number of iterations, chromosome crossover, and mutation probability. The number of individuals in the population is the number of sampling times of the observation information of the national meteorological observation station and the meteorological monitoring data of the port area. The number of chromosome nodes is the number of meteorological index types included in the observation information of the national meteorological observation station and the meteorological monitoring data of the port area. In this embodiment, the hourly data of June 1, 2022 in the above table is used as input, corresponding to a population of 24 individuals, a chromosome node count of 7, an iteration count of 200, a chromosome crossover of 0.7, and a chromosome mutation probability of 0.01.

[0072] S103. Initialize the population. Initialize the BP neural network based on the chromosome information contained in each individual in the population. Use the reciprocal of the mean square error of the training results as the fitness value of the individual. Based on the selection algorithm, retain the individual with the highest fitness and perform crossover and mutation calculations on the remaining individuals according to probability to obtain the new generation of population.

[0073] S104. Iteration: Repeatedly calculate the fitness value of the new generation population and perform selection, crossover, and mutation calculations until the convergence condition is met to obtain the optimal chromosome.

[0074] S105. Decode the chromosome of the optimal individual to obtain the optimal weights of each layer of the BP neural network;

[0075] S106. Train the BP neural network and make predictions to obtain the best rainfall prediction results.

[0076] Furthermore, the isolated forest algorithm includes a first phase and a second phase.

[0077] The first stage trains t isolated trees to form an isolated forest. The specific steps of the first stage are as follows:

[0078] S201, X = {x1, x2, ..., x} n Let} be a given sample set, in which samples x1 to x2 are defined. n These are sample points at different times. x i1 ~x id This represents all dimensions of a sample point at a given time, including TSP concentration, information on dockside vessels and machinery, information on dockside personnel, status information of dockside stormwater and sewage collection equipment, and status information of dockside stormwater and sewage treatment equipment; randomly selected from X. The sample points constitute a subset X′ of X and are placed in the root node;

[0079] S202. Randomly specify a dimension q from d dimensions, and randomly generate a cutting point p in the data space formed by the subset X′.

[0080] S203. Generate a hyperplane using this cutting point p to divide the data space formed by the subset X′ into two subspaces. Place sample points with a dimension less than p into the left child node and those with a dimension greater than or equal to p into the right child node.

[0081] S204, recursively S202 and S203, until all leaf nodes have only one sample point or the isolated tree has reached the specified height;

[0082] S205, repeat S201 to S204 until t isolated trees are generated;

[0083] In the second stage, for any sample point x i Let it traverse each isolated tree, based on the sample point x i The height of each isolated tree yields the sample point x. i The average height h(x) in an isolated forest i ), calculate the outlier score s for all sample points,

[0084]

[0085] in, The average height of all standardized sample points in an isolated forest is represented by the following formula.

[0086]

[0087] Where H is the harmonic function, and when the input is k, its output H(k) = ln(k) + ξ, where ξ is Euler's constant. In this embodiment, ξ is 0.5772156649.

[0088] After obtaining the outlier scores of all sample points, a mapping from outlier scores to 0-1 is established based on certain threshold screening conditions. 0 represents a sample point as a normal value, and 1 represents a sample point as an outlier. After outlier removal based on the mapping results, the outlier scores of the remaining sample points are recalculated until the mapping results from the outlier scores of all sample points to 0-1 are all 0.

[0089] In this embodiment, hourly meteorological data from 00:00 to 10:00 on June 2, 2022 are used as sample points. The process of establishing the isolated forest is as follows:

[0090] Train 10 isolated trees, each with a specified height of 50, to form an isolated forest;

[0091] S201, X = {x1, x2, ..., x} 10} is a given sample set, x i (i∈[1, 10], and i is an integer) is a seven-dimensional sample point, x i =(x i1 x i2 , ..., x i7 ), x i1 Indicates the predicted rainfall time, x i2 Indicates predicted rainfall, x i3 Indicates TSP concentration, x i4 Indicates information about dockside vessels and machinery, x i5 Indicates information about dock workers, x i6 Indicates the status information of the dock's rainwater and sewage collection equipment, x i7 This indicates the status information of the wharf's stormwater and sewage treatment equipment; randomly selected from X. The sample points constitute a subset X′ of X and are placed in the root node;

[0092] S202. Randomly specify one dimension q from the seven dimensions, and randomly generate a cutting point p in the data space formed by the subset X′.

[0093] S203. Generate a hyperplane using this cutting point p to divide the data space formed by the subset X′ into two subspaces. Place sample points with a dimension less than p into the left child node and those with a dimension greater than or equal to p into the right child node.

[0094] S204, recursively S202 and S203, until all leaf nodes have only one sample point or the isolated tree has reached the specified height;

[0095] S205, repeat S201 to S204 until 10 isolated trees are generated.

[0096] Furthermore, the process by which the stormwater and sewage early warning and treatment subsystem uses the DS evidence theory to determine whether existing stormwater and sewage equipment meets the requirements is as follows:

[0097] S301. First, establish an identification framework Θ, which represents the complete set of all possible combinations of predicted stormwater and sewage concentrations at the wharf, current stormwater and sewage collection and treatment capacity, and whether existing stormwater and sewage equipment meets the requirements. Specifically,

[0098] Θ = {θ1, θ2, ..., θ} j , ..., θ N}

[0099] Where, θ j An event or element referred to as the recognition frame Θ;

[0100] The set of all subsets of the recognition frame Θ is called the power set, denoted as 2. Θ The power set is represented as,

[0101] 2 Θ ={Φ,{θ1},{θ2},...,{θ n}, {θ1∪θ2}, {θ1∪θ3},..., Θ}

[0102] S302. Establish the basic trust allocation function m, specifically as follows:

[0103]

[0104] m(A) represents the support of evidence for proposition A. Its value is the basic trust assignment value of the proposition. The basic trust assignment value of the empty set is zero, and the sum of the trust values ​​of all other subsets is equal to 1. A that satisfies the condition that m(A) is greater than 0 is called a focal element.

[0105] S303. If multiple pieces of evidence exist, the Dempster synthesis rule is used to synthesize them. Specifically, assuming there are two pieces of evidence E1 and E2 under the recognition framework Θ, where m1 and m2 are the basic assignment functions corresponding to the two pieces of evidence respectively, A i and Bj The synthesis method for the focal elements corresponding to the two pieces of evidence is as follows:

[0106]

[0107] Here, K is the conflict coefficient, and the larger K is, the greater the conflict between different pieces of evidence.

[0108] In this embodiment, the predicted value of the wharf stormwater and sewage concentration, the current stormwater and sewage collection and treatment capacity, and whether the existing stormwater and sewage equipment meets the requirements are first processed into data in the event identification framework Θ. The specific data processing process is as follows: the predicted value of the wharf stormwater and sewage concentration (mg / L), the current stormwater and sewage collection and treatment capacity include the number of available stormwater and sewage treatment equipment (pieces) and the amount of flocculant in reserve (kg), the existing stormwater and sewage equipment meeting the requirements is represented as 1, and the existing stormwater and sewage equipment not meeting the requirements is represented as 0;

[0109] For example, let's illustrate the events in the recognition framework Θ. θ1 represents a stormwater / sewage concentration of 0, a number of stormwater / sewage treatment devices of 0, and a stormwater / sewage flocculant reserve of 0, where existing equipment meets requirement 1; θ2 represents a stormwater / sewage concentration of 0, a number of stormwater / sewage treatment devices of 2, and a stormwater / sewage flocculant reserve of 10, where existing equipment meets requirement 1; θ3 represents a stormwater / sewage concentration of 0.3, a number of stormwater / sewage treatment devices of 0, and a stormwater / sewage flocculant reserve of 0, where existing equipment does not meet requirement 0; θ4 represents a stormwater / sewage concentration of 0.05, a number of stormwater / sewage treatment devices of 1, and a stormwater / sewage flocculant reserve of 10, where existing equipment meets requirement 1; θ5 represents a stormwater / sewage concentration of 0.02, a number of stormwater / sewage treatment devices of 0, and a stormwater / sewage flocculant reserve of 5, where existing equipment meets requirement 1; θ6 ​​represents a stormwater / sewage concentration of 0.07, a number of stormwater / sewage treatment devices of 1, and a stormwater / sewage flocculant reserve of 1, where existing equipment does not meet requirement 0...

[0110] It should be noted that the identification framework Θ described in this invention is not limited to the combination of the above-mentioned events.

[0111] To illustrate the premise of synthesizing multiple pieces of evidence, for the identification framework Θ = {A, B}, if focal element A represents the predicted concentration of rainwater and sewage at the dock as 0.3, the number of rainwater and sewage treatment equipment as 0, the storage of rainwater and sewage flocculants as 0, and the existing equipment not meeting requirement 0; focal element B represents the predicted concentration of rainwater and sewage at the dock as 0, the number of rainwater and sewage treatment equipment as 0, the storage of rainwater and sewage flocculants as 0, and the existing equipment meeting requirement 1; and there are 2 pieces of evidence, evidence 1: m1(A) = 0.7, m2(B) = 1; evidence 2: m1(A) = 0.4, m2(B) = 0.7, then the evidence needs to be synthesized.

[0112] A method for early warning of stormwater and sewage at bulk cargo terminals based on a meteorological forecasting model, implemented using the aforementioned meteorological forecasting model-based stormwater and sewage early warning system, includes the following steps:

[0113] First, using the GA-BP neural network algorithm, based on observation information from the national meteorological observation station and meteorological monitoring data from the port area, the predicted rainfall time and predicted rainfall amount are obtained;

[0114] Subsequently, the isolated forest algorithm was used to clean and remove outliers from the TSP concentration, dock operation vessel and machinery information, dock operation personnel information, dock stormwater and sewage collection equipment status information, and dock stormwater and sewage treatment equipment status information in the port area meteorological monitoring data.

[0115] Finally, based on the predicted rainfall, predicted rainfall time, and TSP concentration after data cleaning and outlier removal, as well as information on vessels and machinery operating at the dock, the predicted values ​​of stormwater and sewage concentration at the dock are obtained. Based on the information of dock workers after data cleaning and outlier removal, the status information of stormwater and sewage collection equipment at the dock, and the status information of stormwater and sewage treatment equipment at the dock, the current stormwater and sewage collection and treatment capacity is obtained. Then, using the DS evidence theory, the existing stormwater and sewage equipment is judged to meet the demand based on the predicted values ​​of stormwater and sewage concentration at the dock and the current stormwater and sewage collection and treatment capacity. When it is determined that the existing stormwater and sewage equipment does not meet the demand, an early warning is issued.

[0116] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for early warning of rainwater and sewage at bulk cargo terminals based on a meteorological forecasting model.

[0117] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method for early warning of rainwater and sewage at bulk cargo terminals based on a meteorological prediction model.

[0118] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A stormwater and sewage early warning system for bulk cargo terminals based on a meteorological forecasting model, characterized in that: It includes a weather forecasting subsystem, a data processing subsystem, and a rainwater and sewage early warning and processing subsystem; The meteorological forecasting subsystem integrates observation information from national meteorological observation stations and meteorological monitoring data from port areas to generate fused meteorological data. Then, using a GA-BP neural network algorithm, it obtains predicted rainfall time and amount based on the fused meteorological data. Specifically, the meteorological indicators from national meteorological observation station information include rainfall, humidity, temperature, air pressure, wind direction, and wind speed; the meteorological indicators from port area meteorological monitoring data include temperature, humidity, air pressure, wind direction, wind speed, and TSP concentration. The data fusion process involves setting corresponding judgment thresholds for common meteorological indicators from both national meteorological observation station information and port area meteorological monitoring data. Then, the absolute values ​​of the differences between the national meteorological observation station information and the port area meteorological monitoring data for each meteorological indicator are calculated. All obtained absolute values ​​are compared with the corresponding judgment thresholds. If the absolute value of the difference for a certain meteorological indicator is less than its corresponding judgment threshold, then the port area meteorological monitoring data is selected as the input to the GA-BP neural network algorithm for that meteorological indicator; otherwise, the national meteorological observation station information is selected as the input to the GA-BP neural network algorithm. The data processing subsystem uses the isolated forest algorithm to clean and remove outliers from the TSP concentration in the meteorological monitoring data of the port area, as well as the information on dock operation vessels and machinery, dock operation personnel, status information of dock rainwater and sewage collection equipment, and status information of dock rainwater and sewage treatment equipment. The stormwater and sewage early warning and treatment subsystem obtains the predicted stormwater and sewage concentration at the dock based on the predicted rainfall amount, predicted rainfall time, TSP concentration after data cleaning and outlier removal, and dock operation vessel and machinery information. Based on dock operation personnel information, dock stormwater and sewage collection equipment status information, and dock stormwater and sewage treatment equipment status information after data cleaning and outlier removal, it obtains the current stormwater and sewage collection and treatment capacity. Then, using the DS evidence theory, it combines the predicted stormwater and sewage concentration at the dock with the current stormwater and sewage collection and treatment capacity to determine whether the existing stormwater and sewage equipment meets the demand. When it is determined that the existing stormwater and sewage equipment does not meet the demand, an early warning is issued.

2. The stormwater and sewage early warning system for bulk cargo terminals based on a meteorological forecasting model according to claim 1, characterized in that: The GA-BP neural network algorithm is specifically as follows: S101. Determine the BP network topology; S102. Set the relevant parameter values ​​of the GA algorithm according to the BP network topology. The relevant parameter values ​​of the GA algorithm include the number of individuals in the population, the number of chromosome nodes, the number of iterations, chromosome crossover and mutation probability; where the number of individuals in the population is the number of sampling times of the observation information of the national meteorological observation station and the meteorological monitoring data of the port area, and the number of chromosome nodes is the number of meteorological index types included in the observation information of the national meteorological observation station and the meteorological monitoring data of the port area. S103. Initialize the population. Initialize the BP neural network based on the chromosome information contained in each individual in the population. Use the reciprocal of the mean square error of the training results as the fitness value of the individual. Based on the selection algorithm, retain the individual with the highest fitness and perform crossover and mutation calculations on the remaining individuals according to probability to obtain the new generation of population. S104. Iteration: Repeatedly calculate the fitness value of the new generation population and perform selection, crossover, and mutation calculations until the convergence condition is met to obtain the optimal chromosome. S105. Decode the chromosome of the optimal individual to obtain the optimal weights of each layer of the BP neural network; S106. Train the BP neural network and make predictions to obtain the predicted rainfall time and predicted rainfall amount.

3. The stormwater and sewage early warning system for bulk cargo terminals based on a meteorological forecasting model according to claim 1, characterized in that: The isolated forest algorithm consists of a first phase and a second phase. The first stage trains t isolated trees to form an isolated forest. The specific steps of the first stage are as follows: S201, X = {x1, x2, ..., x} n Let} be a given sample set, in which samples x1 to x2 are defined. n These are sample points at different times. x i1 ~x id This represents all dimensions of a sample point at a given time, including TSP concentration, information on dockside vessels and machinery, information on dockside personnel, status information of dockside stormwater and sewage collection equipment, and status information of dockside stormwater and sewage treatment equipment; randomly selected from X. A subset X' of sample points is placed in the root node; S202. Randomly specify a dimension q from d dimensions, and randomly generate a cutting point p in the data space formed by the subset X'. S203. Generate a hyperplane using this cutting point p to divide the data space formed by the subset X' into two subspaces. Place sample points with a dimension less than p into the left child node and those with a dimension greater than or equal to p into the right child node. S204, recursively S202 and S203, until all leaf nodes have only one sample point or the isolated tree has reached the specified height; S205, repeat S201 to S204 until t isolated trees are generated; In the second stage, for any sample point x i Let it traverse each isolated tree, based on the sample point x i The height of each isolated tree yields the sample point x. i The average height h(x) in an isolated forest i ), calculate the outlier score s for all sample points, in, The average height of all standardized sample points in an isolated forest is represented by the following formula. Where H is the harmonic function, and when the input is k, its output H(k) = ln(k) + ξ, where ξ is Euler's constant; After obtaining the outlier scores of all sample points, a mapping from outlier scores to 0-1 is established based on certain threshold screening conditions. 0 represents a sample point as a normal value, and 1 represents a sample point as an outlier. After outlier removal based on the mapping results, the outlier scores of the remaining sample points are recalculated until the mapping results from the outlier scores of all sample points to 0-1 are all 0.

4. The stormwater and sewage early warning system for bulk cargo terminals based on a meteorological forecasting model according to claim 1, characterized in that: The process by which the stormwater and sewage early warning and treatment subsystem uses the DS evidence theory to determine whether existing stormwater and sewage equipment meets the requirements is as follows. S301. First, establish an identification framework Θ, which represents the complete set of all possible combinations of predicted stormwater and sewage concentrations at the wharf, current stormwater and sewage collection and treatment capacity, and whether existing stormwater and sewage equipment meets the requirements. Specifically, Θ={θ1,θ2,…,θ j ,…,θ N } Where, θ j An event or element referred to as the recognition frame Θ; The set of all subsets of the recognition frame Θ is called the power set, denoted as 2. Θ The power set is represented as, 2 Θ ={Φ,{θ1},{θ2},…,{θ n },{θ1∪θ2},{θ1∪θ3},…,θ} S302. Establish the basic trust allocation function m, specifically as follows: m(A) represents the support of evidence for proposition A. Its value is the basic trust assignment value of the proposition. The basic trust assignment value of the empty set is zero, and the sum of the trust values ​​of all other subsets is equal to 1. A that satisfies the condition that m(A) is greater than 0 is called a focal element. S303. If multiple pieces of evidence exist, the Dempster synthesis rule is used to synthesize them. Specifically, assuming there are two pieces of evidence E1 and E2 under the recognition framework Θ, where m1 and m2 are the basic assignment functions corresponding to the two pieces of evidence respectively, A i and B j The synthesis method for the focal elements corresponding to the two pieces of evidence is as follows: Here, K is the conflict coefficient, and the larger K is, the greater the conflict between different pieces of evidence.

5. A method for early warning of rainwater and sewage at a bulk cargo terminal based on a meteorological prediction model, implemented using the early warning system for rainwater and sewage at a bulk cargo terminal based on a meteorological prediction model as described in any one of claims 1-4, characterized in that: Includes the following steps, First, the observation information from the national meteorological observation station and the meteorological monitoring data from the port area are fused to generate fused meteorological data. Then, the GA-BP neural network algorithm is used to obtain the predicted rainfall time and predicted rainfall amount based on the fused meteorological data. Subsequently, the isolated forest algorithm was used to clean and remove outliers from the TSP concentration in the meteorological monitoring data of the port area, as well as the information on dock operation vessels and machinery, dock operation personnel, status information of dock rainwater and sewage collection equipment, and status information of dock rainwater and sewage treatment equipment. Finally, based on the predicted rainfall, predicted rainfall time, and TSP concentration after data cleaning and outlier removal, as well as information on vessels and machinery operating at the dock, the predicted values ​​of stormwater and sewage concentration at the dock are obtained. Based on the information of dock workers after data cleaning and outlier removal, the status information of stormwater and sewage collection equipment at the dock, and the status information of stormwater and sewage treatment equipment at the dock, the current stormwater and sewage collection and treatment capacity is obtained. Then, using the DS evidence theory, the existing stormwater and sewage equipment is judged to meet the demand based on the predicted values ​​of stormwater and sewage concentration at the dock and the current stormwater and sewage collection and treatment capacity. When it is determined that the existing stormwater and sewage equipment does not meet the demand, an early warning is issued.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the steps of the stormwater and sewage early warning method for bulk cargo terminals based on a meteorological prediction model as described in claim 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the stormwater and sewage early warning method for bulk cargo terminals based on a meteorological prediction model as described in claim 5.

Images