Marine spilled oil emergency recovery optimization method

Abstract

The invention discloses an offshore oil spill emergency recovery optimization method, and belongs to the technical field of petroleum recovery, and the method comprises the following steps: obtaining offshore oil spill accident risk factors; the method comprises the following steps: constructing a three-objective function including maximizing the petroleum recovery amount, minimizing the recovery cost and minimizing the environmental influence, and solving the three-objective function by using a multi-objective genetic-particle swarm hybrid algorithm formed by combining a non-dominated sorting genetic algorithm and a particle swarm optimization algorithm to obtain the maximized petroleum recovery amount. Minimizing the recovery cost and minimizing an optimal solution set of different quantity configuration combinations with respect to different types of oil skimmers under different weights; and by adopting a fuzzy superior and inferior solution distance method, finding out a quantity configuration combination of the optimal oil skimmers under the determined weight from the optimal solution set of different quantity configuration combinations of different types of oil skimmers, namely an emergency recovery scheme for the oil spill at sea. According to the method, environmental and economic losses can be reduced, so that efficient and low-cost light crude oil recovery strategy optimization is realized.

Description

Technical Field

[0001] The present invention relates to the field of oil recovery, and specifically to an optimized method for emergency recovery of offshore oil spills. Background Art

[0002] Offshore oil spill accidents not only pose a serious threat to the marine ecosystem but also may have a profound impact on economic activities such as fisheries and tourism. The diffusion speed of the oil spill, the changes in environmental factors, and the weathering process of the crude oil all make timely and effective emergency response particularly complex. In this context, crude oil recovery operations become an important means to mitigate the impact of oil spills and restore the marine ecosystem. There are significant deficiencies in the current offshore oil spill emergency response technologies. Existing methods mostly focus on single objectives, making it difficult to synergistically optimize multiple conflicting objectives, and generally ignoring the non-linear effects of dynamic environmental parameters such as wind speed and temperature on the oil film diffusion and the efficiency of skimmers; traditional optimization algorithms are prone to falling into local optima or having insufficient diversity of solution sets, and are unable to balance the global search and local development capabilities; most models lack dynamic modeling of the oil weathering process, resulting in the inability to adapt to the attenuation of the oil film thickness in real time during recovery; in addition, existing research mostly uses static equipment efficiency coefficients and does not consider the coupling effect of the environment and equipment performance. Summary of the Invention

[0003] In order to solve the problems of insufficient multi-objective optimization ability, poor adaptability to dynamic environments, and traditional algorithms being prone to falling into local optima and having insufficient diversity of solution sets in the existing offshore oil spill emergency response methods, an optimized method for emergency recovery of offshore oil spills is proposed, including the following steps: Obtain the risk factors of offshore oil spill accidents; Preprocess the risk parameters of offshore oil spill accidents through the Monte Carlo simulation method to obtain a Monte Carlo sample set; Based on the preprocessed risk factors of offshore oil spill accidents, construct a three-objective function including maximizing the amount of oil recovered, minimizing the recovery cost, and minimizing the environmental impact; The expression of the objective function for maximizing the amount of oil recovered is as follows:

[0004] Where: is the volume of the recovered oil, is the operation time, is the upper limit of the operation time for oil recovery using a skimmer, is the number of types of skimmers, is the th type of skimmer, from to The expression of the objective function for minimizing the recovery cost is as follows:

[0005] Wherein: is the total cost, is the type of skimmer, is the upper limit of the skimmer type, is the th type of skimmer's operating cost, related to the volume of the oil spill, is the th type of skimmer's transportation cost from the warehouse to the oil spill point; The objective function expression for minimizing the environmental impact is as follows:

[0006] Wherein, is the base ratio, is the volume, is the geographical location, represents a special management area, is the pollutant characteristic, : represents the environmental impact; Step 4: Construct the constraints of the three-objective function considering the oil weathering process, specifically including: the equality constraint of volume decay and the constraint of recovery equipment resources; The equality constraint of the volume decay is:

[0007]

[0008] Wherein: is the stage's oil film thickness, is the stage's remaining oil spill volume after the oil recovery and natural weathering processes; is the initial volume of the oil spill, A is the oil spill area, is the to time period's oil volume lost through the oil recovery and natural weathering processes, is the to time period's oil volume lost during the oil evaporation weathering process, is the to time period's oil volume lost during the oil dispersion process; Wherein: The oil recovery volume has the following expression:

[0009] Wherein: is the oil recovery volume of the th skimmer in the to stage, , is the empirical coefficient of the skimmer, is the oil film thickness; The equation for the volume of oil lost during the evaporation and weathering process of petroleum is as follows:

[0010]

[0011] Wherein: and are the evaporation equation parameters of a specific oil product, is the evaporation rate in the stage, is the temperature, is the time, represents taking the natural logarithm of , is the to volume (m 3 ) of the evaporated oil in the stage, is the remaining spilled oil volume after the oil recovery and natural weathering processes in the stage; The equation for the volume of oil lost during the oil dispersion process is as follows:

[0012]

[0013]

[0014]

[0015] Wherein: is is the water content rate, and the oil emulsification phenomenon is simulated through the water content rate, is the curve fitting constant varying with the wind speed, is the molar viscosity constant, represents the exponential function, is the wind speed, is the time; is the dispersion rate, is the wind speed, is the dynamic viscosity of the oil, is the oil film thickness in the​ is the interfacial tension between oil and water, is the dynamic viscosity of the oil in stage is the dynamic viscosity of the oil in stage is the evaporation rate in stage is the water cut in stage is the molar viscosity constant, represents the exponential function, is the to volume of dispersed oil in stage is the dispersion rate in stage is the remaining oil spill volume after the oil recovery and natural weathering processes in stage; The resource constraints of the recovery equipment are as follows:

[0016]

[0017] where: is the number of the type of skimmer, is the maximum limit of the skimmer quantity, represents an integer; Combine the preprocessed Monte Carlo sample set, transform the three-objective function into a robust form including the expected value and variance, balance the stability and volatility of the objective function, and obtain the processed objective function; Use the non-dominated sorting genetic algorithm combined with the particle swarm optimization algorithm to form a multi-objective genetic-particle swarm hybrid algorithm to solve the three objective functions, and obtain the optimal solution sets of different weight combinations of maximizing the oil recovery volume, minimizing the recovery cost, and minimizing the environmental impact for different types and quantities of skimmers;

[0018] Furthermore: The risk factors of the offshore oil spill accident include: oil spill parameters, marine environmental factors, and oil spill response conditions; The oil spill parameters include the oil spill type, oil spill area, oil film thickness, oil viscosity, and the degree of weathering of the oil; The marine environmental factors include wind speed, seawater temperature, and interfacial tension; The oil spill response conditions include skimmer parameters, arrival time of the emergency response team, recovery efficiency, and response cost.

[0019] Furthermore: The Monte Carlo sample set is obtained after the combination preprocessing. The three-objective function is transformed into a robust form including the expected value and variance to balance the stability and volatility of the objective function. The process of obtaining the processed objective function is as follows: Randomly draw N groups of independent parameter combinations from the defined probability distribution to generate a sample set including wind speed, temperature, skimmer efficiency coefficient, and recovery cost as the input for the optimization of the three-objective function. The three objective functions are robustified in the following way:

[0020]

[0021]

[0022] Among them: Refers to the expected value of the recovery volume, indicating the average performance of the recovery volume under the parameter distribution. Refers to the standard deviation of the recovery volume, indicating the volatility of the recovery volume. Refers to the expected value of the recovery cost, indicating the average performance of the recovery cost under the parameter distribution. Refers to the standard deviation of the recovery cost, indicating the volatility of the recovery cost. Refers to the expected value of the environmental impact, indicating the average performance of the environmental impact under the parameter distribution. Refers to the standard deviation of the environmental impact, indicating the volatility of the environmental impact. And Are weight parameters used to balance the influence of the expected value and variance, where As a decision variable, represents the combination of the number of skimmers configured.

[0023] Furthermore: The multi-objective genetic-particle swarm hybrid algorithm is constructed by combining the non-dominated sorting genetic algorithm and the particle swarm optimization algorithm to solve the processed objective function. The process of obtaining the optimal solution set regarding different combinations of the number of different types of skimmers under different weights for maximizing the oil recovery volume, minimizing the recovery cost, and minimizing the environmental impact is as follows: Initialize the position and velocity of the particle swarm; For each particle, update the velocity and position according to the particle swarm velocity and position update formula; At the beginning of each generation, dynamically update the inertia weight to balance the global search and local search capabilities; Repair the updated positions of the particles in the particle swarm to ensure that different combinations of the number of different types of skimmers satisfy the constraint conditions of the three-objective function that includes maximizing the oil recovery volume, minimizing the recovery cost, and minimizing the environmental impact; Perform crossover operations on the individuals in the updated population through the simulated binary crossover operator to generate new offspring individuals; Perform mutation operations on the offspring individuals after the crossover operation through the polynomial mutation operator to increase the diversity of solutions; Evaluate the new population after particle swarm update and genetic operations, and calculate the objective function values and constraint values of the new population; Compare the objective function values of the new population with the individual optimal objective function values. If the new solution is better, update the individual optimal position and the corresponding objective function value; Merge the feasible solutions in the new population with the solutions in the archive, re-perform non-dominated sorting and crowding degree calculation, and update the solution set in the archive. The archive stores the current optimal non-dominated solution set; After iterative optimization with the maximum number of iterations, return the non-dominated solution set in the archive and the corresponding objective function to obtain the optimal solution set of different combinations of the number of different types of skimmers under different weights for maximizing the oil recovery volume, minimizing the recovery cost, and minimizing the environmental impact.

[0024] Furthermore: The process of using the fuzzy technique for order preference by similarity to ideal solution (TOPSIS) method to find the optimal combination of the number of skimmers under the determined weights from the optimal solution set regarding different combinations of the number of different types of skimmers, which is the emergency recovery plan for marine oil spills, is as follows: Perform triangular fuzzification on each element in the decision matrix of the optimal solution set; Standardize the fuzzified decision matrix; Weight the standardized decision matrix to obtain the weighted decision matrix; Determine the fuzzy positive ideal solution and the fuzzy negative ideal solution for the weighted decision matrix; Calculate the Euclidean distances from each combination scheme of the number of skimmers to the positive ideal solution and the negative ideal solution respectively; Based on the Euclidean distances from the positive ideal solution and the negative ideal solution, calculate the relative closeness of each combination scheme of the number of skimmers, and sort the combination schemes of the number of skimmers according to the relative closeness. The scheme with the maximum relative closeness is the optimal combination of the number of skimmers.

[0025] The process of using the analytic hierarchy process (AHP) to make pairwise comparisons of the importance of the three objective functions to form a judgment matrix and calculate the relative weights of the importance of the three objective functions is as follows: The decision-making problem is decomposed into three levels: the goal level, the criterion level, and the alternative level; the decision-making problem is: in the scenario of offshore oil spill emergency recovery, based on the three conflicting goals of maximizing oil recovery, minimizing recovery cost, and minimizing environmental impact, select the combination of the optimal number of skimmers with the best comprehensive performance from the Pareto optimal solution set. Compare each factor in the criterion level or alternative level pairwise to determine the relative importance and construct a judgment matrix. Normalize the judgment matrix and calculate the weight vector of each factor. Calculate the consistency index and the random consistency ratio to determine whether the judgment matrix is consistent, and then obtain the relative weights of the importance of the three objective functions.

[0026] An offshore oil spill emergency recovery optimization method proposed by the present invention can provide effective support for the emergency response to offshore oil spill accidents under limited resources, help decision-makers quickly formulate efficient recovery plans, reduce environmental and economic losses, and thus realize the optimization of the efficient and low-cost light crude oil recovery strategy. Secondly, it realizes the balance between recovery efficiency and cost and optimizes the resource allocation. By combining the non-dominated sorting of NSGA-II (Non-dominated Sorting Genetic Algorithm) and the global search ability of PSO (Particle Swarm Optimization), a high-quality Pareto optimal solution set (Pareto optimal solution set) is generated, providing a more comprehensive benefit solution for practical applications. The Pareto solution set generated by the hybrid algorithm is more evenly distributed and closer to the ideal solution than the individual NSGA-II and PSO algorithms. In addition, by dynamically simulating the weathering process of oil, it can more accurately reflect the complex situation in actual oil spill accidents and provide more practical decision-making support for emergency response. Brief Description of the Drawings

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0028] Figure 1 It is the flowchart of the method of the present invention. Figure 2 It is the risk identification framework diagram of the oil spill accident. Figure 3 It is the algorithm flowchart of the NSGA-II-PSO hybrid algorithm (multi-objective genetic-particle swarm hybrid algorithm). Figure 4 It is the Pareto front diagram of the NSGA-II -PSO hybrid algorithm. Figure 5 It is a graph of algorithm performance evaluation indicators; Figure 6 It is a graph of the optimal solution obtained by the NSGA-II-PSO hybrid algorithm on the Pareto front using fuzzy TOPSIS (technique for order preference by similarity to an ideal solution); Figure 7 It is the convergence curve of the NSGA-II-PSO hybrid algorithm for the goal of maximizing oil recovery; Figure 8 It is the convergence curve of the NSGA-II-PSO hybrid algorithm for the goal of minimizing recovery cost; Figure 9 It is the convergence curve of the NSGA-II-PSO hybrid algorithm for the goal of minimizing environmental impact; Figure 10 It is a heat map of the impact of wind speed on oil recovery; Figure 11 It is a heat map of the impact of temperature on oil recovery; Figure 12 It is a Sobol sensitivity analysis graph of the combined impact of wind speed and wind speed on oil recovery. Specific implementation manners

[0029] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the drawings and in combination with the embodiments.

[0030] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and in no way limits the present invention and its application or use. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0031] Figure 1 It is the method flow chart of the present invention; An offshore oil spill emergency recovery optimization method includes the following steps: S1: Obtain the risk factors of offshore oil spill accidents; S2: Preprocess the risk parameters of offshore oil spill accidents through the Monte Carlo simulation method to obtain a Monte Carlo sample set; S3: Based on the preprocessed risk factors of offshore oil spill accidents, construct a three-objective function including maximizing oil recovery, minimizing recovery cost, and minimizing environmental impact; S4: Construct the constraint conditions of the three-objective function considering the oil weathering process, specifically including: the equality constraint of volume attenuation and the constraint of recovery equipment resources; S5: Combine the Monte Carlo sample set obtained after preprocessing, transform the three-objective function into a robust form including expected value and variance, balance the stability and volatility of the objective function, and obtain the processed objective function; S6: Use the NSGA-II algorithm combined with the PSO algorithm to form a multi-objective NSGA-II-PSO hybrid algorithm to solve the three-objective function, and obtain the optimal solution set of different weight combinations of maximizing oil recovery, minimizing recovery cost, and minimizing environmental impact for different types and quantities of skimmers; Integrate the advantages of the global Pareto solution set generation of NSGA-II and the local optimization of the dynamic inertia weight of PSO to solve the problem of imbalance between the convergence and diversity of a single algorithm; S7: Use the analytic hierarchy process to compare the importance of the three objective functions pairwise, form a judgment matrix, calculate the relative weights of the importance of the three objective functions, and use the fuzzy TOPSIS (technique for order preference by similarity to an ideal solution) method to find the optimal offshore oil spill emergency recovery plan under the determined relative weights from the optimal solution set of different types and quantities of skimmers, replacing the traditional linear weighting method.

[0032] This method also includes comparing the hybrid algorithm with the NSGA-II and MOPSO algorithms and the sensitivity analysis of parameters, and outputting the optimal combination of skimmer quantity configuration; In this case, the crude oil recovery operation becomes an important means to reduce the impact of oil spills and restore the marine ecosystem. Figure 2 It shows the risk identification framework of the oil spill accident of the system. This framework classifies the influencing factors into three main categories: oil spill parameters, marine environment, and oil spill response conditions. These factors jointly determine the diffusion mode, influence range, and effectiveness of response measures of the oil spill accident.

[0033] The risk factors of the offshore oil spill accident include: oil spill parameters, marine environmental factors, and oil spill response conditions; The oil spill parameters include oil spill type, oil spill area, oil film thickness, oil viscosity, and the degree of oil weathering, etc. Different types of crude oils have different physical and chemical properties, such as density, viscosity, and volatility. These properties directly affect the diffusion and weathering process of the oil spill. The oil spill area and oil film thickness determine the coverage range and recovery difficulty of the oil spill, while the oil viscosity and weathering degree affect the recovery efficiency of the skimmer. Thus, it deeply affects its behavior performance in the marine environment and the potential harm degree to the ecological environment ultimately.

[0034] Marine environmental factors, such as wind speed, seawater temperature, and interfacial tension, have a significant impact on the diffusion, migration, and various links in the weathering process of oil spills. The magnitude of wind speed can affect the diffusion of oil spills and increase the area covered by the oil spills. Seawater temperature and interfacial tension, on the other hand, affect the evaporation and dissolution processes of oil. High water temperature will accelerate the volatilization of petroleum, reduce its viscosity, and affect the recovery efficiency.

[0035] Oil spill response conditions include skimmer parameters, arrival time of the emergency response team, recovery efficiency, and response costs, etc. Different types of skimmers are suitable for different oil spill conditions and environments. Selecting the appropriate skimmer for oil recovery work can improve the efficiency and effectiveness of oil spill response. Generally, it includes disc skimmers, weir skimmers, belt skimmers, etc. The arrival time is the time when the emergency response team arrives at the accident site. Arriving in time can reduce the diffusion range of the oil spill and improve the recovery efficiency. The recovery efficiency is affected by various factors, such as oil film thickness, sea conditions, and equipment performance. Response costs include equipment costs, transportation costs, labor costs, etc. These costs directly affect the economic feasibility of the emergency response.

[0036] The risk factors of the uncertainty of the offshore oil spill accident are preprocessed, and Monte Carlo simulation is used to quantify the parameter uncertainty.

[0037] First, select the parameters that have a significant impact on oil spill recovery, including wind speed ( ), temperature ( ), skimmer efficiency ( ), and recovery cost ( ). According to historical data or actual observations, define the probability distribution of each parameter. For example, the wind speed ( ) follows a normal distribution with a mean of 10 m / s and a standard deviation of 2 m / s. The temperature ( ) follows a uniform distribution with a range from 275.15K to 281.15 K. Randomly draw a large number of samples from the defined probability distributions to generate multiple sets of parameter combinations, and use this random sample set as the output of data preprocessing.

[0038] These parameter combinations will be used in the subsequent three-objective function optimization and solution process as input data for optimization calculations by algorithms such as the NSGA-II-PSO hybrid algorithm.

[0039] The three-objective function of maximizing the amount of oil recovered, minimizing the recovery cost, and minimizing the environmental impact includes the objective function of maximizing the amount of oil recovered and the objective function of minimizing the recovery cost; The expression of the objective function of maximizing the amount of oil recovered is as follows:

[0040] Where: is the volume of the recovered oil, is the operation time, is the upper limit of the operation time for oil recovery using the skimmer, is the number of skimmer types, is the th type of skimmer's oil recovery volume at to time; The expression of the objective function for minimizing the recovery cost is as follows:

[0041] where: is the total cost, is the type of skimmer, is the upper limit of the skimmer type, is the th skimmer's operation cost, related to the volume of the spilled oil, is the th skimmer's transportation cost from the warehouse to the oil spill point.

[0042] The expression of the objective function for minimizing the environmental impact is as follows:

[0043] where, is the base rate ($1.00), is the volume (m 3 ), is the geographical location (8 = near the coast; 5 = nearshore waters; 1 = offshore), indicates the special management area (2 = yes; 1 = no), is the pollutant characteristic (8 = heavy oil; 4 = medium oil; 1 = light oil); EI represents the environmental impact; The expression of the non - linear equation for the skimmer to recover oil is as follows:

[0044] where: is the th type of skimmer's oil recovery volume at to stage, , is the empirical coefficient of the skimmer, is the oil film thickness;

[0045]

[0046] Wherein: is the oil film thickness at stage is the remaining oil spill volume after the oil recovery and natural weathering processes at stage is the initial volume of the oil spill, A is the oil spill area, is to the oil volume lost during the oil recovery and natural weathering processes from time is to the oil volume lost during the evaporation weathering process of the oil from time is to the oil volume lost during the dispersion process of the oil from time The equation constraints for the volume decay caused by oil weathering include the evaporation equation for the weathering process and the equation for the dispersion process; The evaporation equation for the oil volume lost during the evaporation weathering process of the oil is as follows:

[0047]

[0048] Wherein: and are the evaporation equation parameters for a specific oil product, is the evaporation rate (%) at stage is the temperature (K), is the time (h), represents taking the natural logarithm of is the to volume of the evaporated oil at stage 3 ), is the remaining oil spill volume (m 3 ) after the oil recovery and natural weathering processes at stage

[0049] The emulsification phenomenon was simulated using the embedding rate of water in the oil film:

[0050] Wherein, is the water content, is the curve fitting constant (2×10 -6 ) varying with the wind speed, is the molar viscosity constant (0.7, applicable to crude oil and heavy fuel oil), represents the exponential function, is the wind speed (m / s), is the time (h); The volume of oil lost during the oil dispersion process has the following equation:

[0051]

[0052]

[0053]

[0054] Where: is the water cut at the stage, and the oil emulsification phenomenon is simulated through the water cut, is the curve fitting constant varying with the wind speed, is the molar viscosity constant, represents the exponential function, is the wind speed, is the time; is the dispersion rate, is the wind speed, is the dynamic viscosity of the oil, is the oil film thickness at the stage, is the oil-water interface tension, is the dynamic viscosity of the oil at the stage, is the dynamic viscosity of the oil at the stage, is the evaporation rate at the stage, is the water cut at the stage, is the molar viscosity constant, represents the exponential function, is the to volume of the dispersed oil at the stage, is the dispersion rate at the stage, is the remaining oil spill volume after the oil recovery and natural weathering processes at the stage; The resource constraints of the recovery equipment are as follows:

[0055]

[0056] Where: is the The number of skimmers, is the maximum limit of the skimmer quantity, which represents an integer.

[0057] The process of obtaining the Monte Carlo sample set after the combination preprocessing, transforming the three-objective function into a robust form including the expected value and variance, and balancing the stability and volatility of the objective function to obtain the processed objective function is as follows: From the sample set obtained by the Monte Carlo simulation, N groups of independent parameter combinations are randomly drawn from the defined probability distribution to generate a sample set including wind speed, temperature, skimmer efficiency coefficient, and recovery cost, which is used as the input for the optimization of the three-objective function. The three-objective function is robustified in the following way:

[0058]

[0059]

[0060] Where: refers to the expected value of the recovery quantity, indicating the average performance of the recovery quantity under the parameter distribution. refers to the standard deviation of the recovery quantity, indicating the volatility of the recovery quantity. refers to the expected value of the recovery cost, indicating the average performance of the recovery cost under the parameter distribution. refers to the standard deviation of the recovery cost, indicating the volatility of the recovery cost. refers to the expected value of the environmental impact, indicating the average performance of the environmental impact under the parameter distribution. refers to the standard deviation of the environmental impact, indicating the volatility of the environmental impact. and are weight parameters used to balance the influence of the expected value and variance.

[0061] The process of using the non-dominated sorting genetic algorithm combined with the particle swarm optimization algorithm to form a multi-objective genetic-particle swarm hybrid algorithm to solve the three-objective function and obtain the optimal solution set of different weight combinations of different types of skimmers with different quantities for maximizing the oil recovery quantity, minimizing the recovery cost, and minimizing the environmental impact is as follows: Initialize the position and velocity of the particle swarm; For each particle, update the velocity and position according to the particle swarm velocity and position update formula; At the beginning of each generation, dynamically update the inertia weight to balance the global search and local search capabilities; Repair the positions of the particles in the particle swarm after updating to ensure that different combinations of the number of different types of skimmers satisfy the constraints of the three-objective function that includes maximizing the oil recovery amount, minimizing the recovery cost, and minimizing the environmental impact; Perform crossover operations on the individuals in the updated population through the simulated binary crossover operator to generate new offspring individuals; Perform mutation operations on the offspring individuals after the crossover operation through the polynomial mutation operator to increase the diversity of solutions; Evaluate the new population after particle swarm update and genetic operations, and calculate the objective function values and constraint values of the new population; Compare the objective function values of the new population with the individual optimal objective function values. If the new solution is better, update the individual optimal position and the corresponding objective function value; Merge the feasible solutions in the new population with the solutions in the archive, re-perform non-dominated sorting and crowding degree calculation, and update the solution set in the archive. The solutions saved in the archive are the current optimal non-dominated solution set; After iterative optimization with the maximum number of iterations, return the non-dominated solution set in the archive and the corresponding objective function to obtain the optimal solution set of different combinations of the number of different types of skimmers under different weights for maximizing the oil recovery amount, minimizing the recovery cost, and minimizing the environmental impact.

[0062] The process of using the fuzzy TOPSIS method to find the optimal offshore oil spill emergency recovery plan under the determined weights from the optimal solution set is as follows: Perform triangular fuzzification on each element in the decision matrix of the optimal solution set; Standardize the fuzzified decision matrix; Weight the standardized decision matrix to obtain the weighted decision matrix; Determine the fuzzy positive ideal solution and the fuzzy negative ideal solution for the weighted decision matrix; Calculate the Euclidean distances from each combination scheme of the number of skimmers to the positive ideal solution and the negative ideal solution respectively; Calculate the relative closeness of each combination scheme of the number of skimmers based on the Euclidean distances from the positive ideal solution and the negative ideal solution. According to the relative closeness Sort all the schemes: the combination schemes of the number of skimmers; The scheme with the largest value is the optimal combination of the number of skimmers.

[0063] The process of using the analytic hierarchy process to compare the importance of the three objective functions pairwise to form a judgment matrix and calculate the relative weights of the importance of the three objective functions is as follows: Decompose the decision-making problem into three levels: the goal level, the criterion level, and the alternative level; the decision-making problem is: in the scenario of offshore oil spill emergency recovery, based on the three conflicting goals of maximizing oil recovery, minimizing recovery cost, and minimizing environmental impact, screen out the combination of the optimal number of skimmers with the best comprehensive performance from the Pareto optimal solution set. Make pairwise comparisons of the factors at the criterion level or alternative level to determine the relative importance and construct a judgment matrix. Normalize the judgment matrix and calculate the weight vector of each factor. Calculate the consistency index and the random consistency ratio to determine whether the judgment matrix is consistent, and then obtain the relative weights of the importance of the three objective functions.

[0064] Example 1: Assume that a light crude oil spill accident occurred in a certain sea area. S1: Obtain the risk factors of the offshore oil spill accident. The initial volume of the oil spill is 4000 cubic meters, the oil film thickness is 40 millimeters. During the "golden 24-hour" window period of emergency response, under the combined action of wind field shear force and turbulent diffusion, the oil film reaches a quasi-steady state, and the area expansion rate drops below 5% of the initial stage. Assume that the oil spill area remains relatively stable to improve the calculation efficiency. To respond to this accident, the emergency response team plans to use three different types of skimmers for oil recovery. Assume that due to the load limit of the transportation vehicle, there are ten sets of each type of skimmer, and the maximum capacity of the container for oil recovery operation is 20 sets of skimmers. Due to the sudden occurrence of the accident and transportation conditions, no more skimmers can be supplied within 48 hours. Therefore, the core goal at the current stage is to determine the best combination of the three skimmers within 24 hours to achieve the maximum oil spill recovery and minimize the recovery cost. Table 1 lists the empirical coefficients and transportation costs of the three skimmers.

[0065] Table 1 Skimmer coefficients

[0066] The weathering parameters of Arab light crude oil involved in the case are shown in Table 2: Table 2 Weathering parameters of Arab light crude oil

[0067] S2: Use Monte Carlo simulation to randomly sample parameters such as wind speed and temperature to quantify the impact of parameter uncertainty on the recovery strategy.

[0068] Randomly sample the parameters involved in the weathering process as shown in the following table: Table 3 Definition of parameter probability distribution

[0069] S3: Based on the pre-processed risk factors of offshore oil spill accidents, construct a three-objective function that includes maximizing the amount of oil recovered, minimizing the recovery cost, and minimizing the environmental impact, which is expressed as:

[0070]

[0071]

[0072] Step 4: Construct the constraints of the three-objective function considering the oil weathering process, specifically including: the equality constraint of volume decay and the recovery equipment resource constraint; The equality constraint of volume decay and the recovery equipment resource constraint are:

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084] Where: is to maximize the amount of oil recovered; is to minimize the recovery cost; is to minimize the environmental impact; is the decision variable of the objective function, that is, the number of different types of skimmers, is the type of skimmer, is the th type of skimmer at the stage of the oil recovery rate; is the th type of skimmer's operating cost, is the The transportation cost of an oil skimmer; For the initial volume of the oil spill (m 3 ); For The remaining oil spill volume (m 3 ) after the oil recovery and natural weathering processes in the , , The volumes of oil recovered, evaporated, and dispersed during the to h time periods are respectively (m 3 ); For the oil film thickness (mm) in the stage, For the oil spill area (m 2 ); For the evaporation rate (%) in the stage, And Are the parameters of the oil evaporation equation, For the temperature (K), For the time (h), For the water content (%) in the stage, For the oil viscosity (cP) in the stage, For the dispersion rate (%) in the stage, For the wind speed (m / s), For the interfacial tension (dyne / m).

[0085] Step 5: Combine the preprocessed Monte Carlo sample set, transform the three-objective function into a robust form including the expected value and variance, balance the stability and volatility of the objective function, and obtain the processed objective function; Couple the random samples generated by Monte Carlo simulation in data preprocessing with the three-objective optimization to construct a robust decision-making framework. Monte Carlo simulation is used to generate random samples of parameters, and the samples are used to construct the robust decision-making framework. The steps of generating samples can be regarded as data preprocessing. Robust decision-making optimization is an optimization method that modifies the three-objective function based on the parameter combinations generated by Monte Carlo simulation to find the optimal solution under uncertain conditions, which can enhance the robustness of the objective function, that is, stably output the solution set under different conditions. This step belongs to the optimization process. The robust decision-making framework is the objective function below, in order to consider the randomness and volatility of parameters.

[0086] Among the parameters generated by Monte Carlo simulation, each parameter has its own probability distribution, and the joint distribution of these parameters describes the probability of their simultaneous occurrence. From this joint distribution, N independent samples are randomly drawn. Each sample (s = 1, 2, ..., N) represents a possible combination of parameters, such as it could be a combination of parameters with a wind speed of 10 m / s and a temperature of 278 K, or it could be another combination of parameters with a wind speed of 12 m / s and a temperature of 280 K. Substituting different combinations into the following robust multi-objective problem can ensure that the optimization results remain stable under various uncertainty conditions. At the same time, sensitivity analysis can be performed on the two parameters of wind speed and temperature by calculating values; For each set of parameters , robustify the multi-objective problem:

[0087]

[0088]

[0089] Where: refers to the expected value of the recovery amount, representing the average performance of the recovery amount under the parameter distribution. refers to the standard deviation of the recovery amount, representing the volatility of the recovery amount. refers to the expected value of the recovery cost, representing the average performance of the recovery cost under the parameter distribution. refers to the standard deviation of the recovery cost, representing the volatility of the recovery cost. refers to the expected value of the environmental impact, representing the average performance of the environmental impact under the parameter distribution. refers to the standard deviation of the environmental impact, representing the volatility of the environmental impact. and are weight parameters used to balance the influence of the expected value and variance.

[0090] Combine each set of parameters (s = 1, 2, …, N) in the Monte Carlo sample set with the position vector of the particle swarm, and calculate the corresponding recovery amount , ) and cost ( , ) for each set of ( ( , ) for each particle For each particle , based on N sets of ( ( , ), ( , ), ( , )) Calculate the expected value and standard deviation to generate a robust three-objective value ([[]] , ); Quantify the contribution of parameters to the objective through the Sobol method (Sobol method) and conduct sensitivity analysis:

[0091] Among them: Represents the parameter 's first-order sensitivity index. ∈[0,1], the larger the value, the more significant the influence of the parameter on the objective function. Represents the average value of the objective function RV when the parameter is determined, Represents the variance of the average value of the objective function RV when the parameter is determined, Represents the variance of the recovery amount; S6: Combine the non-dominated sorting genetic algorithm and the particle swarm optimization algorithm to form a multi-objective genetic-particle swarm hybrid algorithm to solve the processed objective function, and obtain the optimal solution set of different weight combinations of different types of skimmers with different numbers for maximizing oil recovery, minimizing recovery cost, and minimizing environmental impact; Fuse the advantages of the global Pareto solution set generation of NSGA-II and the dynamic inertia weight local optimization of PSO to solve the problem of imbalance between the convergence and diversity of a single algorithm, Figure 3 Figure S61: Initialize the position and velocity ;

[0092]

[0093] Among them: Represents the position vector of the th particle, respectively represent the th particle in the th dimension of the position component; represents the velocity vector of the th particle, respectively represents the th particle in the dimensional velocity component, is the total number of particles in the particle swarm.

[0094] S62: Iterative optimization: For each particle, update according to the velocity and position update formula of the PSO particle swarm algorithm. The formula is as follows:

[0095]

[0096] Among them: is the velocity of particle in the th iteration and the dimensional velocity, is the current iteration number, is the velocity of particle in the th iteration and the dimensional velocity, is the inertia parameter, is the individual learning coefficient, is the global learning coefficient, and are random numbers uniformly distributed in the range [0, 1], used to introduce randomness and help the algorithm jump out of local optima. is the individual extreme value, is the global extreme value, is the position of particle in the th iteration and the dimensional position. is the current position of the particle, is the position of particle after update, is the velocity of particle after update.

[0097] At the beginning of each generation, dynamically update the inertia weight , to balance the global search and local search capabilities.

[0098]

[0099] Among them: is the inertia weight of the th iteration, is the maximum value of the inertia weight, is the minimum value of the inertia weight, is the current iteration number, is the maximum number of iterations.

[0100] Solution repair: Repair the updated position to ensure that the solution satisfies the constraint conditions.

[0101] Genetic operations: Crossover operation: Crossover the individuals in the population through the simulated binary SBX crossover operator to generate new offspring individuals.

[0102] Mutation operation: Mutate the offspring individuals after crossover through the polynomial mutation operator to increase the diversity of the solution.

[0103] New population evaluation: Evaluate the new population after particle swarm update and genetic operations, and calculate its objective function value and constraint value.

[0104] Individual best update: Compare the objective function value of the new population with the individual best objective function value. If the new solution is better, update the individual best position and the corresponding objective function value.

[0105] Archive update: Merge the feasible solutions in the new population with the solutions in the archive, re-perform non-dominated sorting and crowding degree calculation, and update the solution set in the archive. The archive stores the current optimal non-dominated solution set.

[0106] Return results After iterative optimization for the maximum number of iterations, return the non-dominated solution set in the archive and the corresponding objective function. These solutions represent the Pareto optimal trade-off between the recovery volume and the cost.

[0107] Figure 4 is the Pareto front graph of the NSGA-II -PSO hybrid algorithm; Figure 5 is the comparison between the NSGA-II-PSO hybrid algorithm and other multi-objective intelligent optimization algorithms, and quantifies the differences in convergence and distribution of each algorithm with Hypervolume and Spacing values. The Hypervolume metric refers to the volume of the region in the objective space enclosed by the non-dominated solution set obtained by the algorithm and the reference point. The larger the HV value, the better the comprehensive performance of the algorithm. Its calculation formula is:

[0108] where represents the volume calculation function, which is used to calculate the volume of a certain region in the objective space. represents taking the union of all solutions v i in the region defined in the objective space, represents the number of solutions in the solution set S.

[0109] The smaller the Spacing value, the more uniform the distribution of the solution set. Its calculation formula is:

[0110] where N represents the number of all solutions, d i represents the distance from solution i to its nearest neighbor solution, and represents the average value of the distances from all solutions to their nearest neighbor solutions.

[0111] Generally speaking, by integrating the non-dominated sorting mechanism of NSGA-II and the global search ability of PSO, NSGA-II-PSO demonstrates stronger comprehensive optimization performance in terms of the Hypervolume index. At the same time, it maintains a high distribution quality in terms of the Spacing index, effectively balancing convergence, diversity, and uniformity, providing a better and more comprehensive Pareto solution set for the decision-making of offshore oil spill recovery, and verifying its superiority in complex multi-objective optimization problems.

[0112] S7: Use the analytic hierarchy process to make pairwise comparisons of the importance of the three objective functions to form a judgment matrix, calculate the relative weights of the importance of the three objective functions, and adopt the fuzzy superior and inferior solution distance method to find the optimal skimmer quantity configuration combination under the determined relative weights from the optimal solution set regarding different quantities of different types of skimmers, which is the offshore oil spill emergency recovery plan.

[0113] Adopt triangular fuzzy weight and closeness calculation to quickly screen the comprehensive optimal plan from the Pareto solution set, replacing the traditional linear weighting method.

[0114] S71: Fuzzify each element in the decision matrix of the optimal solution set;

[0115] where each element is a triangular fuzzy number, representing the fuzzy evaluation value of the th plan on the th index, represents the fuzzy evaluation value of the 1st plan on the 1st index, represents the fuzzy evaluation value of the 1st plan on the mth index, represents the fuzzy evaluation value of the nth plan on the 1st index, represents the fuzzy evaluation value of the nth plan on the mth index; is the minimum possible value of this fuzzy evaluation value, is the most likely value of the fuzzy evaluation value, is the maximum possible value of the fuzzy evaluation value; S72: Standardize the fuzzified decision matrix; For benefit-type indicators (the larger the better), in the invention, it refers to the oil recovery volume:

[0116] For cost-type indicators (the smaller the value, the better), in the invention, it refers to the recovery cost:

[0117] Among them, is the fuzzy evaluation value after standardization; represents the maximum value of the th indicator, minimum value of the is the minimum possible value of the fuzzy evaluation value, is the most likely value of the fuzzy evaluation value, is the maximum possible value of the fuzzy evaluation value; S73: Weight the standardized decision matrix to obtain the weighted decision matrix ;

[0118] Among them, is the weighted fuzzy evaluation value; represents the weight assigned to each indicator, satisfying , represents the number of rows in the decision matrix, representing the number of decision-making schemes; represents the number of columns in the decision matrix, representing the number of evaluation indicators.

[0119] S74: Determine the fuzzy positive ideal solution (FPIS) and fuzzy negative ideal solution (FNIS) for the weighted decision matrix; Fuzzy positive ideal solution (FPIS):

[0120] For benefit-type attributes:

[0121] For cost-type attributes:

[0122] Fuzzy negative ideal solution (FNIS):

[0123] For benefit-type attributes:

[0124] For cost-type attributes:

[0125] Among them, represents the ideal optimal solution vector obtained from the weighted decision matrix, is its component; represents the ideal worst solution vector obtained from the weighted decision matrix, is its component, is the number of components of the solution vector. represents the fuzzy positive ideal solution of the th component, represents the fuzzy positive ideal solution of the th component.

[0126] S75: Calculate the Euclidean distances from different quantity configuration combination schemes of different types of skimmers to the positive ideal solution and the negative ideal solution respectively:

[0127]

[0128] Among them, represents the Euclidean distance from the th scheme to the fuzzy positive ideal solution, represents the Euclidean distance from the th scheme to the fuzzy negative ideal solution; is the weighted fuzzy evaluation value, represents the fuzzy positive ideal solution of the th component, represents the fuzzy positive ideal solution of the th component.

[0129] S76: Calculate the relative closeness of different schemes based on the Euclidean distances of the positive ideal solution and the negative ideal solution of each scheme :

[0130] Among them represents the relative closeness of the th scheme, , and the larger the value, the closer the scheme is to the ideal solution.

[0131] S77: Calculate the relative closeness of each scheme , according to the relative proximity Sort all different configuration schemes of different types of oil skimmers with different quantities, The largest scheme is the optimal solution.

[0132] The process of using the analytic hierarchy process to compare the importance of the three objective functions pairwise, form a judgment matrix, and calculate the relative weights of the importance of the three objective functions is as follows: Step1: Construct a hierarchical structure model Decompose the decision-making problem into three levels: the goal level, the criterion level, and the scheme level.

[0133] Goal level: Select the optimal solution on the Pareto front.

[0134] Criterion level: Three goals (maximizing oil recovery, minimizing recovery cost, minimizing environmental pollution).

[0135] Scheme level: Candidate solutions on the Pareto front.

[0136] Step2: Construct a judgment matrix; By pairwise comparison, compare each factor at the criterion level or the scheme level pairwise to determine their relative importance to the goal level (or the upper level). The comparison results are represented by the 1-9 scale method proposed by Saaty, as shown in the following table, and based on the values in the scale table, construct a judgment matrix A=(a ij ) 3*3 .

[0137] Table 4 Judgment matrix scale value and meaning

[0138] Step3: Calculate the weight vector By normalizing the judgment matrix, calculate the weight vector of each factor. In the judgment matrix A, its element is a ij , the normalization matrix R The calculation formula is:

[0139] Among them, rij is the element in the i-th row and j-th column of the normalization matrix R, is the sum of the j-th column of the judgment matrix A, and n is the order of the judgment matrix. After normalization, calculate the weight vector W, and W is the average value of each row of the normalization matrix R. The calculation formula is:

[0140] And verify that the sum of all elements of the weight vector W is 1:

[0141] Step4: Consistency Check Check whether the judgment matrix has consistency by calculating the consistency index (CI) and the random consistency ratio (CR) to ensure no logical contradictions. The formula for the consistency index CI is:

[0142] is the largest eigenvalue of the judgment matrix A; The formula for the consistency ratio is:

[0143] where RI is the average random consistency index. For n = 1 to 9, the RI values are shown in the following table: Table 5 Average Random Consistency Index (RI) Table

[0144] When CR < 0.1, it is considered that the consistency of the matrix is acceptable.

[0145] S8: Dynamic Policy Output and Verification Output the optimal equipment combination, compare the hybrid algorithm with NSGA-II and MOPSO algorithms, and conduct a sensitivity analysis of the parameters.

[0146] Figure 6 is the optimal solution graph obtained by using the fuzzy TOPSIS (technique for order preference by similarity to an ideal solution) method on the Pareto front for the NSGA-II-PSO hybrid algorithm; in a marine oil spill accident, quickly recovering the oil spill is the core environmental protection goal. The spread of oil spills will directly threaten the marine ecosystem, fishery resources, and coastal communities, and its environmental and social costs are much higher than the economic investment in emergency operations. Therefore, it is considered that the maximum oil recovery volume (G1) is strongly more important than minimizing the recovery cost (G2), slightly more important than minimizing the environmental impact (G3), and minimizing the environmental impact (G3) is significantly more important than minimizing the recovery cost (G2). Construct a judgment matrix, and calculate the weights of the three objective functions through the AHP method as:

[0147] The largest eigenvalue , Substituting the RI value into the consistency test formula gives CR = 0.026 < 1, indicating that the matrix has satisfactory consistency. Therefore, the weights of the three objective functions are set as maximizing the oil recovery (0.643), total cost (0.074), and minimizing the environmental impact (0.283), and then substituted into the fuzzy TOPSIS method. This reflects the emergency management principle of "pollution control takes precedence over short-term costs", making the optimal solution tend to choose the plan with a higher recovery rate and less environmental impact, even if the cost is relatively high. The obtained optimal solution is as follows Figure 6 shown. At this time, the distribution of skimmers is SK1 = 0, SK2 = 6, SK3 = 9. The oil recovery volume is 2777.1 m3, the total cost is 28,632,421.2 yuan, and the environmental impact is 47,558,862.9 yuan.

[0148] Figure 7 is the convergence curve of the NSGA-II-PSO hybrid algorithm for the objective of maximizing the oil recovery volume; Figure 8 is the convergence curve of the NSGA-II-PSO hybrid algorithm for the objective of minimizing the recovery cost; Figure 9 is the convergence curve of the NSGA-II-PSO hybrid algorithm for the objective of minimizing the environmental impact; The results show that the NSGA-II-PSO algorithm can effectively balance the recovery efficiency and cost during the optimization process. In the first 20 generations of iteration, the changes in the objective function are relatively large, indicating that the algorithm can quickly find a better solution in the initial stage. Thanks to the global search ability of the PSO algorithm, it can quickly explore the solution space in the early stage. After 20 generations of iteration, the decline gradually slows down, indicating that the algorithm enters the local search stage. The local refinement ability of the NSGA-II algorithm plays an important role in this stage. Through non-dominated sorting and crowding degree calculation, the distribution and quality of the solution set are further optimized.

[0149] Figure 10 is the three-dimensional surface plot of the influence of wind speed on oil recovery; Figure 11 is the three-dimensional surface plot of the influence of temperature on oil recovery; Figure 12 is the Sobol index of the influence of wind speed and temperature on oil recovery; The results show that with the increase of wind speed and temperature, the weathering rate of oil accelerates, the amount of evaporated and dispersed oil increases, resulting in a decrease in the recovery amount. The increase in temperature accelerates the evaporation process of oil, but has a relatively small impact on the recovery amount. When the temperature rises from 274 K to 281 K, the recovery amount decreases by about 11.8%. When the wind speed increases from 5 m / s to 15 m / s, the volume of oil lost due to dispersion increases, and the recovery amount decreases by about 5.4%. Among the first-order Sobol indices, the wind speed (U) is 0.2334 and the temperature (T) is 0.7711, indicating that the temperature has a greater impact on the oil spill recovery amount, which is the same as the changing trend shown in the heat map.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An optimization method for emergency recovery of offshore oil spills, characterized in that: It includes the following steps: Step 1: Obtain the risk factors of the marine oil spill accident; Step 2: Preprocess the risk parameters of the marine oil spill accident by the Monte Carlo simulation method to obtain a Monte Carlo sample set; Step 3: Based on the preprocessed risk factors of the marine oil spill accident, construct a three-objective function including maximizing the oil recovery volume, minimizing the recovery cost, and minimizing the environmental impact; The expression of the objective function for maximizing the oil recovery volume is as follows: Wherein: is the volume of the recovered oil, is the operation time, is the upper limit of the operation time for oil recovery using a skimmer, is the number of skimmers, is the th type of skimmer's oil recovery volume in the to stage; The expression of the objective function for minimizing the recovery cost is as follows: Wherein: is the total cost, is the type of skimmer, is the upper limit of the skimmer type, is the operating cost of the - th skimmer, related to the volume of the oil spill, is the transportation cost of the - th skimmer from the warehouse to the oil spill point; The expression of the objective function for minimizing the environmental impact is as follows: Among them, is the base rate, is the volume, is the geographical location, represents the special management area, is the pollutant characteristic, : represents the environmental impact; Step 4: Construct the constraints of the three-objective function considering the oil weathering process, specifically including: the equality constraint of volume decay and the constraint of recovery equipment resources; The equality constraint of the volume decay is: Wherein: is the oil film thickness at stage is the remaining oil spill volume after the oil recovery and natural weathering processes at stage is the initial volume of the oil spill, A is the oil spill area, is from to is to the oil volume lost during the oil evaporation and weathering processes at stage is to the oil volume lost during the oil dispersion process at stage Wherein: the amount of oil recovered The expression is as follows: Wherein: is the oil recovery volume of the th skimmer in the to stage, , is the empirical coefficient of the skimmer, is the oil film thickness; The volume of oil lost during the evaporation and weathering process of oil The equation is as follows: Wherein: and are the evaporation equation parameters of a specific oil product, is the evaporation rate at stage the temperature, the time, denotes taking the natural logarithm of is the to volume of the evaporated oil at stage is the remaining oil spill volume after the oil recovery and natural weathering processes at stage The volume of oil lost during the oil dispersion process The equation is as follows: Wherein: is the water content at stage, simulating the phenomenon of petroleum emulsification through the water content, is the curve fitting constant varying with wind speed, is the molar viscosity constant, represents the exponential function, is the wind speed, is the time; is the dispersion rate, is the wind speed, is the dynamic viscosity of the oil, is the oil film thickness at stage, is the interfacial tension between oil and water, is the dynamic viscosity of the oil at stage, is the dynamic viscosity of the oil at stage, is the evaporation rate at stage, is the water content at stage, is the molar viscosity constant, represents the exponential function, is the to volume of the dispersed oil at stage, is the dispersion rate at stage, is the remaining oil spill volume at stage after the processes of petroleum recovery and natural weathering; The constraint of the recovery equipment resources is as follows: Wherein: is the number of the th oil skimmer, is the maximum limit of the number of oil skimmers, represents an integer; Step 5: Combine the obtained Monte Carlo sample set after preprocessing, transform the three-objective function into a robust form including the expected value and variance, balance the stability and volatility of the objective function, and obtain the processed objective function; Step 6: Use the multi-objective genetic-particle swarm hybrid algorithm composed of the non-dominated sorting genetic algorithm and the particle swarm optimization algorithm to solve the processed objective function, and obtain the optimal solution set of different weight combinations of maximizing the oil recovery volume, minimizing the recovery cost, and minimizing the environmental impact for different numbers of different types of skimmers; Step 7: Use the analytic hierarchy process to make pairwise comparisons of the importance of the three objective functions to form a judgment matrix, calculate the relative weights of the importance of the three objective functions, and use the fuzzy TOPSIS method to find the optimal number configuration combination of skimmers under the determined relative weights from the optimal solution set of different numbers of different types of skimmers, which is the marine oil spill emergency recovery plan.

2. The optimization method for emergency recovery of offshore oil spills according to claim 1, characterized in that: The risk factors of the marine oil spill accident include: oil spill parameters, marine environmental factors, and oil spill response conditions; The oil spill parameters include the oil spill type, oil spill area, oil film thickness, oil viscosity, and degree of oil weathering; The marine environmental factors include wind speed, seawater temperature, and interfacial tension; The oil spill response conditions include skimmer parameters, arrival time of the emergency response team, recovery efficiency, and response cost.

3. The optimized method for emergency recovery of offshore oil spills according to claim 1, characterized in that: The process of combining the obtained Monte Carlo sample set after preprocessing, transforming the three-objective function into a robust form including the expected value and variance, balancing the stability and volatility of the objective function, and obtaining the processed objective function is as follows: Randomly draw N groups of independent parameter combinations from the defined probability distribution to generate a sample set including wind speed, temperature, skimmer efficiency coefficient, and recovery cost as the input for the optimization of the three-objective function, and robustify the three-objective function in the following way: Wherein: Refers to the expected value of the recovery volume, representing the average performance of the recovery volume under the parameter distribution, Refers to the standard deviation of the recovery volume, representing the volatility of the recovery volume, Refers to the expected value of the recovery cost, representing the average performance of the recovery cost under the parameter distribution, Refers to the standard deviation of the recovery cost, representing the volatility of the recovery cost, Refers to the expected value of the environmental impact, representing the average performance of the environmental impact under the parameter distribution, Refers to the standard deviation of the environmental impact, representing the volatility of the environmental impact, And Is a weight parameter used to balance the influence of the expected value and the variance, where As a decision variable, represents the combination of the number configuration of the skimmers.

4. The optimized method for emergency recovery of offshore oil spills according to claim 1, characterized in that: The process of using the multi-objective genetic-particle swarm hybrid algorithm composed of the non-dominated sorting genetic algorithm and the particle swarm optimization algorithm to solve the processed objective function, and obtaining the optimal solution set of different weight combinations of maximizing the oil recovery volume, minimizing the recovery cost, and minimizing the environmental impact for different numbers of different types of skimmers is as follows: Initialize the position and velocity of the particle swarm; For each particle, update its velocity and position according to the particle swarm velocity and position update formula; At the beginning of each generation, dynamically update the inertia weight to balance the global search and local search capabilities; Repair the updated positions of the particles in the particle swarm to ensure that different combinations of the number of different types of skimmers satisfy the constraints of the three-objective function that includes maximizing the oil recovery amount, minimizing the recovery cost, and minimizing the environmental impact; Perform crossover operations on the individuals in the updated population through the simulated binary crossover operator to generate new offspring individuals; Perform mutation operations on the offspring individuals after the crossover operation through the polynomial mutation operator to increase the diversity of solutions; Evaluate the new population after the particle swarm update and genetic operations, and calculate the objective function values and constraint values of the new population; Compare the objective function values of the new population with the individual optimal objective function values. If the new solution is better, update the individual optimal position and the corresponding objective function value; Merge the feasible solutions in the new population with the solutions in the archive, re-perform non-dominated sorting and crowding degree calculation, and update the solution set in the archive. The archive stores the current optimal non-dominated solution set; After iterative optimization with the maximum number of iterations, return the non-dominated solution set and the corresponding objective function in the archive to obtain the optimal solution set of different combinations of the number of different types of skimmers under different weights for maximizing the oil recovery amount, minimizing the recovery cost, and minimizing the environmental impact.

5. The optimized method for emergency recovery of offshore oil spills according to claim 1, characterized in that: The process of using the fuzzy technique for order preference by similarity to ideal solution (TOPSIS) method to find the optimal combination of the number of skimmers under the determined relative weights from the optimal solution set of different combinations of the number of different types of skimmers, which is the offshore oil spill emergency recovery plan, is as follows: Perform triangular fuzzification on each element in the decision matrix of the optimal solution set; Standardize the fuzzified decision matrix; Weight the standardized decision matrix to obtain the weighted decision matrix; Determine the fuzzy positive ideal solution and fuzzy negative ideal solution for the weighted decision matrix; Calculate the Euclidean distances from each combination scheme of the number of skimmers to the positive ideal solution and the negative ideal solution respectively; Based on the Euclidean distances from the positive ideal solution and the negative ideal solution, calculate the relative closeness of each combination scheme of the number of skimmers. Sort the combination schemes of the number of skimmers according to the relative closeness, and the scheme with the largest relative closeness is the optimal combination of the number of skimmers.

6. The optimized method for emergency recovery of marine oil spills according to claim 1, characterized in that: The process of using the analytic hierarchy process (AHP) to pairwise compare the importance of the three objective functions to form a judgment matrix and calculate the relative weights of the importance of the three objective functions is as follows: Decompose the decision problem into three levels: the goal level, the criterion level, and the scheme level; the decision problem is: in the offshore oil spill emergency recovery scenario, based on the three conflicting goals of maximizing the oil recovery amount, minimizing the recovery cost, and minimizing the environmental impact, screen out the optimal combination of the number of skimmers with the best comprehensive performance from the Pareto optimal solution set; Pairwise compare the factors at the criterion level or the scheme level to determine the relative importance and construct a judgment matrix; Normalize the judgment matrix and calculate the weight vector of each factor; Calculate the consistency index and the random consistency ratio to determine whether the matrix has consistency, and then obtain the relative weights of the importance of the three objective functions.

Images