Safety assessment method for new reconstruction and expansion land related to underground concrete storage pool

Abstract

The invention discloses a new reconstruction and expansion land safety assessment method related to an underground concrete storage pool, and relates to the technical field of underground water pollution prevention and control, and the method comprises the following steps: according to a constructed seepage and leakage risk assessment model based on machine learning, assessing the seepage and leakage risk of the underground concrete storage pool, calculating the seepage leakage risk index of each storage pool one by one to obtain an evaluation result with seepage leakage risk; and according to an evaluation result, carrying out underground water pollution risk evaluation by utilizing the constructed underground water pollution risk evaluation model, and evaluating the safety of the newly reconstructed and expanded land of the underground concrete storage tank by calculating pollution risk indexes one by one. According to the method, safe, rapid and qualitative evaluation of the new reconstruction and expansion land of coupling seepage leakage and pollution migration is achieved, the calculation process is simple and practical, the technology is reliable, engineering implementation is easy, reliable reference can be provided for scientific and accurate groundwater pollution prevention and control, and the method has important environmental protection and social significance.

Description

Technical Field

[0001] This invention relates to the field of groundwater pollution prevention and control technology, and more specifically, to a safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks. Background Technology

[0002] Newly built, renovated, or expanded industrial and mining land is currently the main battleground for preventing new groundwater pollution, controlling the risk of pollution spread, and eliminating risks to human health. my country has a large number of newly built, renovated, or expanded industrial and mining enterprises with complex production processes. The raw materials and auxiliary materials used by different enterprises vary greatly, and the types of toxic and hazardous substances are diverse. Leakage from underground concrete storage tanks is highly concealed, and the pollution prevention capabilities of enterprises differ, posing a significant threat to groundwater pollution spread and human health, making it difficult to guarantee the safety of newly built, renovated, or expanded industrial and mining land. Existing methods cannot conduct leakage risk assessments based on the leakage conditions of underground concrete storage tanks, nor can they conduct groundwater pollution risk assessments based on the risk source-material migration pathway-sensitive receptor risk formation factors. Therefore, it is impossible to couple leakage risk assessment and pollution risk assessment to achieve a safety assessment of newly built, renovated, or expanded land. Therefore, it is urgent to establish a safety assessment method for newly built, renovated, or expanded land involving underground concrete storage tanks, in order to comprehensively support enterprises in accurately formulating groundwater pollution prevention and control measures and to facilitate the government's scientific supervision of newly built, renovated, or expanded industrial and mining land. Summary of the Invention

[0003] To address the aforementioned problems, this invention provides a safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks, comprising the following steps: Based on the constructed machine learning-based leakage risk assessment model, the leakage risk of underground concrete storage tanks is assessed. By calculating the leakage risk index of each storage tank, the assessment results with leakage risk are obtained. Based on the assessment results, a groundwater pollution risk assessment model was constructed to conduct a groundwater pollution risk assessment. By calculating the pollution risk index one by one, the safety of the newly built, renovated and expanded land for underground concrete storage tanks was assessed.

[0004] Preferably, when constructing the leakage risk assessment model, PSO and BPNN are coupled to form a PSO-BPNN hybrid machine learning algorithm to construct the leakage risk assessment model.

[0005] Preferably, when constructing the leakage risk assessment model, a leakage risk assessment index system is constructed to form a dataset for training the leakage risk assessment model, wherein the leakage risk assessment index system includes: Properties of stored toxic and hazardous substances: type of substance, quantity of substance, toxicity of substance, corrosiveness of substance; Characteristics of reservoir construction and operation: number of reservoir layers, service life, cathodic protection system, anti-corrosion materials, and anti-seepage materials; Groundwater pollution prevention measures: leakage monitoring devices, remote monitoring systems, routine inspection and maintenance, environmental emergency plans, environmental emergency supplies, daily inspection records, technical training, and surrounding monitoring wells; External environment: groundwater level, groundwater corrosivity, soil corrosivity.

[0006] Preferably, when obtaining the leakage risk index, the leakage risk index is expressed as: LRI=0.045SC+0.035SQ+0.118ST+0.146SCOR+0.036TL+0.044SY+0.065CPS+0.116ACM+0.112ASM+0.053LMD+ 0.035RMS+0.041RM+0.017EEP+0.016EES+0.025DIR+0.022TT+0.019PMW+0.022GWD+0.018GWCOR+0.015SOCOR Wherein, LRI represents the leakage risk index; SC represents the type of substance; SQ represents the quantity of substance; ST represents the toxicity of substance; SCOR represents the corrosivity of substance; TL represents the number of reservoir layers; SY represents the service life; CPS represents the cathodic protection system; ACM represents anti-corrosion materials; ASM represents anti-seepage materials; LMD represents the leakage monitoring device; RMS represents the remote monitoring system; RM represents daily inspection and maintenance; EEP represents the environmental emergency plan; EES represents environmental emergency supplies; DIR represents daily inspection records; TT represents technical training; PMW represents the surrounding monitoring wells; GWD represents the groundwater level; GWCOR represents the groundwater corrosivity; and SOCOR represents the soil corrosivity.

[0007] Preferably, when obtaining assessment results with leakage risk, a leakage risk index score of 6.5 is set as a preset threshold; when the leakage risk index of the reservoir is greater than or equal to 6.5, it is considered to have a high leakage risk; when the leakage risk index of the reservoir is less than 6.5, it is considered to have a low leakage risk.

[0008] Preferably, when constructing a groundwater pollution risk assessment model, the importance score of each indicator is determined by using the 1-9 scale method in the analytic hierarchy process based on the constructed potential index system for groundwater pollution risk. This score is then used to determine the weight of the groundwater pollution risk assessment indicators. Based on the summation method, a groundwater pollution risk assessment model is constructed.

[0009] Preferably, when constructing a potential indicator system for groundwater pollution risk, the potential indicator system for groundwater pollution risk includes: Risk sources: leakage risk index, solubility of substances, volatility of substances; Material migration pathways: wet-dry index, vadose zone lithology, groundwater depth, aquifer lithology; Sensitive receptors: People who come into contact with the reservoir, and water sources within 1 km downstream.

[0010] Preferably, when obtaining the pollution risk index, the pollution risk index is represented as: PRI=0.198LRI+0.079SS+0.085SV+0.088DWI+0.069VZL+0.086GD+0.120AL+0.136CSTP+0.139WS Among them, PRI represents the pollution risk index; LRI represents the leakage risk index; SS represents the solubility of the substance; SV represents the volatility of the substance; DWI represents the dry-wet index; VZL represents the vadose zone lithology; GD represents the groundwater depth; AL represents the aquifer lithology; CSTP represents the population in contact with the reservoir; and WS represents the water source within 1km downstream.

[0011] Preferably, when assessing the safety of newly constructed or expanded underground concrete storage tanks, a pollution risk index score of 6.7 is set as a preset threshold. If the pollution risk index of any storage tank is greater than or equal to 6.7, the newly constructed or expanded land is considered unsafe. If the pollution risk index of all storage tanks is less than 6.7, the newly constructed or expanded land is considered safe.

[0012] The present invention discloses the following technical effects: (1) The safety assessment method for new, renovated and expanded land involving underground concrete storage tanks provided by the present invention addresses the technical requirements for safety assessment of new, renovated and expanded land, and is geared towards underground concrete storage tanks. It constructs a leakage risk assessment index system based on the leakage conditions of the storage tank, and constructs a groundwater pollution risk assessment index system based on the risk formation elements of risk source-material migration pathway-sensitive receptor. This significantly reduces the number of two types of risk assessment indicators and saves the cost of collecting assessment indicator information.

[0013] (2) The safety assessment method for new and expanded land involving underground concrete storage tanks based on machine learning provided by this invention addresses the management needs of safety assessment for new and expanded land. It successively carries out seepage risk assessment and groundwater pollution risk assessment based on the summation method, thereby realizing a rapid qualitative assessment of the safety of new and expanded land coupled with seepage and pollution migration. This greatly reduces the workload of traditional risk assessment based on numerical simulation. The calculation process is simple and practical, the technology is reliable, and it is easy to implement in engineering. It can provide a reliable reference for scientific and accurate groundwater pollution prevention and control, and has important environmental and social significance.

[0014] (3) The safety assessment method for new, renovated and expanded land involving underground concrete storage tanks based on machine learning provided by the present invention selects two indicators: the population in contact with the storage tank and the water source within 1km downstream. When establishing the groundwater pollution risk assessment model, the risk to human health and the risk of pollution spread are considered simultaneously, which makes up for the shortcomings of existing methods that only consider a single risk to human health or the risk of pollution spread. It can be used for groundwater pollution risk assessment of pore water in loose rock.

[0015] (4) The safety assessment method for new, renovated and expanded land involving underground concrete storage tanks provided by the present invention, based on the preliminary determination of the weight of leakage risk assessment indicators by human subjective judgment based on expert judgment, uses the optimal hybrid machine learning algorithm to objectively optimize and determine the weight of leakage risk assessment indicators, realizes the objective assignment of indicator weights, solves the problem of large ambiguity when human subjectively assigns weights, and enhances the scientificity and accuracy of leakage risk assessment model prediction.

[0016] (5) The safety assessment method for new, renovated and expanded land involving underground concrete storage tanks based on machine learning provided by the present invention couples PSO and BPNN to form a PSO-BPNN hybrid machine learning algorithm, which overcomes the defects of existing BPNN such as passive learning, instability during training, and easy getting stuck in local extreme points, and further enhances the scientificity and accuracy of the leakage risk assessment model prediction. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the method described in this invention; Figure 2 This is an optimization of the training function of the PSO-BPNN hybrid machine learning algorithm described in this invention; Figure 3 This invention optimizes the number of hidden layer nodes in the PSO-BPNN hybrid machine learning algorithm. Figure 4 This invention optimizes the learning rate of the PSO-BPNN hybrid machine learning algorithm. Figure 5 This is a comparison chart of the optimal results of the PSO-BPNN hybrid machine learning algorithm described in this invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] like Figures 1-5 As shown, this invention provides a safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks, comprising the following steps: Step 1. Construct a machine learning-based risk assessment model for seepage and leakage in underground concrete storage tanks, including building a case library for seepage and leakage risk assessment, screening seepage and leakage risk evaluation indicators, constructing a seepage and leakage risk assessment indicator system, assigning scores and classifying seepage and leakage risk assessment indicators, optimizing the weights of seepage and leakage risk assessment indicators using machine learning, constructing a seepage and leakage risk assessment model, and determining a preset threshold for seepage and leakage risk (referred to as the first preset threshold).

[0021] Step 2. Conduct a leakage risk assessment of underground concrete storage tanks, and use the leakage risk assessment model to calculate the leakage risk index of each storage tank.

[0022] Step 3. Construct a groundwater pollution risk assessment model for underground concrete storage tanks, including screening groundwater pollution risk assessment indicators, constructing a groundwater pollution risk assessment indicator system, assigning scores and classifying groundwater pollution risk assessment indicators, determining the weights of groundwater pollution risk assessment indicators, constructing a groundwater pollution risk assessment model, and determining the preset threshold for pollution risk (referred to as the second preset threshold).

[0023] Step 4. Conduct a groundwater pollution risk assessment of underground concrete storage tanks. Using a pollution risk assessment model, calculate the pollution risk index of each storage tank that is greater than or equal to the first preset threshold.

[0024] Step 5. Determine whether the newly built or expanded land is safe. The safety of the newly built or expanded land is determined by whether there is a reservoir with a pollution risk index greater than or equal to the second preset threshold among all reservoirs with a pollution risk index greater than or equal to the second preset threshold. (That is, if there is any reservoir with a pollution risk index greater than or equal to the second preset threshold, the newly built or expanded land is not safe; if there is no reservoir with a pollution risk index greater than or equal to the second preset threshold, the newly built or expanded land is safe.)

[0025] The leakage risk assessment case library mentioned in this invention contains 48 information items; 20 indicators are selected from 30 potential leakage risk assessment indicators; the leakage risk assessment indicator system includes 20 indicators related to the properties of stored toxic and hazardous substances (substance type, quantity, toxicity, corrosivity), characteristics of reservoir construction and operation (number of reservoir layers, service life, cathodic protection system, anti-corrosion materials, seepage prevention materials), groundwater pollution prevention measures (leakage monitoring devices, remote monitoring systems, daily maintenance, environmental emergency plans, environmental emergency supplies, daily inspection records, technical training, surrounding monitoring wells), and external environment (groundwater level, groundwater corrosivity, soil corrosivity); through performance comparison of basic machine learning algorithms and hyperparameter adjustment of basic machine learning algorithms, an optimal hybrid machine learning algorithm is established to optimize the weights of leakage risk assessment indicators; and a leakage risk assessment model is constructed based on the summation method. Simultaneously, considering the risks of pollution diffusion and human health, nine indicators were selected from 16 potential groundwater pollution risk indicators. The potential groundwater pollution risk indicator system includes nine indicators in three aspects: risk sources (leakage risk index, solubility of substances, volatility of substances), substance migration pathways (dryness-wetness index, vadose zone lithology, groundwater depth, aquifer lithology), and sensitive receptors (people in contact with the reservoir, water sources within 1 km downstream). The 1-9 scale method in the analytic hierarchy process was used to determine the importance score of each indicator, which was used to determine the weight of the groundwater pollution risk assessment indicators. Based on the summation method, a groundwater pollution risk assessment model was constructed.

[0026] This invention uses the fitting coefficient (R) 2 Using mean absolute error (MAE) and root mean square error (RMSE) as performance evaluation metrics, and comparing five basic machine learning algorithms—genetic algorithm (GA), backpropagation neural network (BPNN), optimized particle swarm optimization (PSO), GA-BPNN, and PSO-BPNN—and adjusting hyperparameters such as inertia weights, individual learning factors, and social learning factors in PSO, and training functions, number of hidden layer neurons, and learning rate in BPNN, an optimal hybrid machine learning algorithm is established.

[0027] Example: The present invention discloses a safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks, comprising the following implementation steps: I. Constructing a machine learning-based risk assessment model for seepage and leakage in underground concrete storage tanks: 1. Construction of a case library for leakage risk assessment: By collecting soil pollution hazard investigation reports and enterprise operation and management information from nine key industry enterprises (involving four industries: smelting, coking, chemical and oil and gas extraction), and taking the production equipment in each enterprise as the research object, a case library for seepage and leakage risk assessment was constructed. The case library contains 48 information items for 111 reservoirs, including equipment name, company name, plant location, substance type, substance molecular weight, substance color, substance odor, substance quantity, substance toxicity, substance corrosivity, substance thermal conductivity, number of reservoir layers, service life, installation method, cathodic protection system, anti-corrosion materials, anti-seepage materials, leakage monitoring device, remote monitoring system, daily inspection and maintenance, ground hardening, designated maintenance personnel, environmental emergency plan, environmental emergency supplies, daily inspection records, technical training, surrounding monitoring wells, groundwater level, groundwater hydrochemical type, groundwater flow velocity, groundwater hydraulic gradient, groundwater corrosivity, soil corrosivity, sub-category of land use, environmental pollution accidents, environmental violations, substance solubility, substance volatility, wet / dry index, vadose zone lithology, groundwater flow velocity, microbial activity, groundwater depth, plant root system activity, aquifer lithology, specific surface area of ​​aquifer medium, people in contact with the reservoir, and water sources within 1km downstream.

[0028] 2. Screening of leakage risk assessment indicators: Based on the characteristics of underground concrete storage tanks and the conditions under which leakage occurs, and using expert consultation and scoring methods, 16 industry experts were required to select 20 indicators from 30 potential indicators and rank them by importance. This aimed to reduce individual expert bias towards potential indicators, enhance the authority and rationality of the selected indicators, reduce the workload of indicator information collection, and increase the efficiency of leakage risk assessment. Specifically, the most important indicator was assigned 20 points, the second most important indicator 19 points, the third most important indicator 18 points, the fourth most important indicator 17 points, the fifth most important indicator 16 points, the sixth most important indicator 15 points, the seventh most important indicator 14 points, the eighth most important indicator 13 points, the ninth most important indicator 12 points, the tenth most important indicator 11 points, the eleventh most important indicator 10 points, the twelfth most important indicator 9 points, the thirteenth most important indicator 8 points, the fourteenth most important indicator 7 points, the fifteenth most important indicator 6 points, the sixteenth most important indicator 5 points, the seventeenth most important indicator 4 points, the eighteenth most important indicator 3 points, the nineteenth most important indicator 2 points, and the twentieth most important indicator 1 point. As shown in Table 1, 20 potential indicators were selected as leakage risk assessment indicators, including substance type, substance quantity, substance toxicity, substance corrosivity, number of reservoir layers, service life, cathodic protection system, anti-corrosion materials, anti-seepage materials, leakage monitoring devices, remote monitoring system, daily maintenance, environmental emergency plan, environmental emergency supplies, daily inspection records, technical training, surrounding monitoring wells, groundwater level, groundwater corrosivity, and soil corrosivity (Table 1).

[0029] Table 1 Importance of Potential Indicators for Leakage Risk Assessment

[0030] Note: Importance = S i / [(1+2+3+4+5+6+7+8+9+10+11+12+13+14+15+16+17+18+19+20)×16], where Si is the sum of the scores of the 16 experts for the i-th potential indicator.

[0031] 3. Construction of Leakage Risk Assessment Index System Based on four aspects—the nature of the stored toxic and hazardous substances, the characteristics of the reservoir's construction and operation, groundwater pollution prevention measures, and the external environment—the aforementioned 20 leakage risk assessment indicators are classified to form a reservoir leakage risk assessment indicator system (Table 2).

[0032] Table 2 Risk Assessment Index System for Reservoir Leakage

[0033] (1) The properties of stored toxic and harmful substances include four indicators: substance type, substance quantity, substance toxicity, and substance corrosivity.

[0034] ① Substance type: refers to the category to which the stored toxic and harmful substances belong, such as organic matter or inorganic matter.

[0035] ② Quantity of substances: refers to the quantity of toxic and harmful substances stored, usually measured in units of mass (kilograms) or volume (cubic meters).

[0036] ③ Toxicity of substances: refers to the degree of toxicity of a substance to an organism (such as a human, animal or plant), usually measured by indicators such as toxicity level, median lethal dose (LD50) or median lethal concentration (LC50).

[0037] ④ Corrosiveness of substances: refers to the ability of toxic and harmful substances to cause chemical damage and erosion from the inside of the storage tank.

[0038] (2) The characteristics of the construction and operation of the reservoir include five indicators: number of reservoir layers, service life, cathodic protection system, anti-corrosion materials, and anti-seepage materials.

[0039] ① Number of reservoir layers: refers to the number of structural layers of the reservoir designed and constructed in the vertical direction, such as single-layer reservoir or double-layer reservoir.

[0040] ② Service life: refers to the actual time when the reservoir begins to be used, that is, the duration of operation or use, in years.

[0041] ③ Cathodic protection system: refers to a protection system established through electrochemical principles (such as sacrificial anode method and impressed current method) to mitigate corrosion of the reservoir.

[0042] ④ Anti-corrosion materials: These are materials that are coated or lined on the inner and outer walls of the storage tank to protect the surface of the tank from corrosion by the medium and environmental erosion, such as anti-corrosion coatings and fiberglass linings.

[0043] ⑤ Impermeable materials: These are materials laid at the bottom of the reservoir foundation and inside the impermeable pool to prevent toxic and harmful substances leaking from the reservoir from seeping into the soil or groundwater, such as high-density polyethylene film and bentonite waterproof blanket.

[0044] (3) Groundwater pollution prevention measures include eight indicators such as leakage monitoring devices, remote monitoring systems, daily inspection and maintenance, environmental emergency plans, environmental emergency supplies, daily inspection records, technical training, and surrounding monitoring wells.

[0045] ① Leakage monitoring device: refers to equipment installed at the bottom of the reservoir, the seepage prevention layer or the surrounding area, used to detect the leakage of toxic and harmful substances in real time or periodically, such as leakage sensors and detection wells.

[0046] ② Remote monitoring system: refers to a system that uses the Internet of Things and communication technology to remotely monitor and transmit data on the operating status of the reservoir, leakage monitoring data, and surrounding environmental parameters in real time.

[0047] ③ Routine inspection and maintenance: refers to the daily work of inspecting, repairing, cleaning and maintaining facilities such as the storage tank body, valves and monitoring devices in accordance with the prescribed cycle to ensure their normal operation.

[0048] ④ Environmental emergency response plan: refers to a pre-formulated plan that includes emergency organization, response procedures, and disposal measures to deal with sudden environmental incidents such as reservoir leaks and fires that may cause soil or groundwater pollution.

[0049] ⑤ Environmental emergency supplies: refers to supplies that are stockpiled in advance to deal with sudden environmental incidents such as reservoir leaks, such as adsorbents, protective equipment, and emergency monitoring instruments.

[0050] ⑥ Daily inspection records: These are written or electronic documents recorded by staff during daily inspections, documenting the status of the storage tank, the operation of monitoring devices, and the surrounding environment.

[0051] ⑦ Technical training: refers to training conducted to improve the professional capabilities of reservoir operators, managers, and emergency personnel on topics such as the characteristics of toxic and hazardous substances, reservoir operation and maintenance, pollution prevention and control, and emergency response.

[0052] ⑧ Surrounding monitoring wells: These are well facilities excavated or constructed in the area surrounding the reservoir according to certain layout principles, used to periodically collect soil or groundwater samples to monitor whether they are contaminated.

[0053] (4) The external environment includes three indicators: groundwater level, groundwater corrosivity, and soil corrosivity.

[0054] ① Groundwater level: refers to the elevation of the groundwater level in the area where the reservoir is located relative to the local ground reference level, reflecting the depth of the groundwater.

[0055] ② Groundwater corrosivity: refers to the ability of groundwater surrounding the reservoir to chemically corrode the concrete and other components of the reservoir due to the different chemical compositions it contains.

[0056] ③ Soil corrosivity: refers to the ability of the soil in the area where the storage tank is located to corrode the concrete of the storage tank due to the characteristics of pH value, salt content, and water content.

[0057] 4. Scoring and grading of leakage risk assessment indicators Primary indicator: Leakage risk (200 points); Secondary indicators: nature of stored toxic and hazardous substances (40 points), characteristics of storage pool construction and operation (50 points), groundwater pollution prevention measures (80 points), external environment (30 points); Level 3 indicators: Level 3 indicators are divided into categories and assigned corresponding scores (Table 3).

[0058] Table 3 Scoring and Classification of Leakage Risk Assessment Indicators

[0059] 5. Determination of initial weights for leakage risk assessment indicators Based on the analytic hierarchy process (AHP), pairwise comparisons of each indicator are performed using the 1-9 scale to determine the importance scores of each indicator, thereby constructing a judgment matrix. The consistency of this judgment matrix is ​​then tested. Consistency indicators are determined (…). The value was 0.0779, and the random consistency index ( The consistency ratio was 1.630. The value of 0.0394 < 0.1, indicating acceptable consistency. Finally, the weights of the leakage risk assessment indicators were obtained through normalization (Table 4).

[0060] Table 4 Initial Weights of Leakage Risk Assessment Indicators

[0061] 6. Performance Comparison of Different Basic Machine Learning Algorithms Using the fitting coefficient (R) 2The root mean square error (RMSE) and root mean square error (RMSE) were used as performance evaluation metrics to compare the accuracy of five basic machine learning algorithms—Genetic Algorithm (GA), Backpropagation Neural Network (BPNN), Optimized Particle Swarm Optimization (PSO), GA-BPNN, and PSO-BPNN—in predicting leakage risk in the test centralized reservoir (Table 5). Typically, R... 2 The larger the value and the smaller the RMSE, the higher the prediction accuracy of the basic machine learning algorithm. Table 5 shows that, compared with the other four basic algorithms, PSO-BPNN has the highest prediction accuracy (R²). 2 The value was 0.967, and the RMSE was 0.184.

[0062] Table 5 Performance of different basic machine learning algorithms

[0063] PSO is a swarm intelligence-based stochastic search method (Equation 1). It overcomes the drawback of BPNN, which is prone to getting trapped in local optima using gradient descent. Through cooperation and information sharing among particles, it efficiently handles complex, multivariable, and nonlinear problems, and has the advantages of simple computation, fast convergence, and good robustness. PSO uses the weights and thresholds of the BPNN neural network as particles. After initializing the population, calculating fitness values, and updating particle velocity and position, it outputs the optimal solution of each particle, which serves as the optimal individual weight threshold for the BPNN, thereby obtaining the output result of the BPNN (i.e., the weights of the leakage risk assessment index).

[0064] (1) In the formula: w is the weight matrix; δ is the threshold of each neuron in the layer; f is called the activation function.

[0065] 7. Optimization of PSO-BPNN Hybrid Machine Learning Algorithm The PSO-BPNN algorithm was optimized by adjusting three hyperparameters: training functions (trainlm, trainingd, trainingdm, trainingda, trainingdx, trainingrp, trainingcgf, trainingcgp, trainingcgb, trainingscg, trainingbfg, trainingoss), the number of hidden layer neurons (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16), and the learning rate (0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.11, 0.12, 0.13, 0.14, 0.15). The data was divided into a training set:test set ratio of 7:3, with an inertia weight of 0.6 and individual learning factors and social learning factors of 2. R0 was used as the training set. 2The prediction accuracy of GA-BPNN is evaluated using mean absolute error (MAE) and RMSE as evaluation metrics to determine the optimal hyperparameters of the GA-BPNN algorithm.

[0066] R of the training function "trainlm" 2 R0 (0.973) is higher than other training functions, while MAE (0.129) and RMSE (0.248) are lower than other training functions; R0 for a hidden layer with 10 nodes is higher. 2 The R-value (0.986) is higher than other values, while the MAE (0.135) and RMSE (0.181) are lower; the R-value with a learning rate of 0.11 is higher. 2 (0.987) is higher than other learning rates, while MAE (0.097) and RMSE (0.173) are lower than other learning rates. Figure 2-4 Therefore, the training function was ultimately determined to be trainlm, the number of hidden layer nodes was 10, and the learning rate was 0.11.

[0067] In MATLAB R2019b, run the PSO-BPNN hybrid machine learning algorithm with its hyperparameters adjusted until the optimal result is obtained. Figure 5 The optimal training results of the algorithm show that R0 is the optimal value for the test set data. 2 The value was 0.998, the MAE was 0.039, and the RMSE was 0.075.

[0068] 8. Optimization of weights for leakage risk assessment indicators Extract the weight coefficients between hidden layer neurons and input layer neurons from the aforementioned optimal operating results. and the weight coefficients between output layer neurons and hidden layer neurons The significance coefficient, correlation index and absolute influence coefficient are calculated (Formula 2-4). The absolute influence coefficient is the weight of each indicator after optimization (Table 6).

[0069] (2) (3) (4) In the formula, denoted as the significance coefficient; i represents the input unit of the neural network, i=1,…m; j represents the output unit of the neural network, j=1,…n; k represents the hidden layer unit of the neural network, k=1,…p; These are the weighting coefficients between hidden layer neurons and input layer neurons; These are the weighting coefficients between output layer neurons and hidden layer neurons; For related indices; This represents the absolute influence coefficient.

[0070] Table 6. Optimized weights of leakage risk assessment indicators

[0071] 9. Construction of Leakage Risk Assessment Model Based on the aforementioned assessment indicators and their weights, a leakage risk assessment model (Formula 5) is constructed, and a leakage risk index score of 6.5 is set as a preset threshold. When the leakage risk index of the reservoir is greater than or equal to 6.5, it is considered to have a high leakage risk; when the leakage risk index of the reservoir is less than 6.5, it is considered to have a low leakage risk.

[0072] LRI=0.045SC+0.035SQ+0.118ST+0.146SCOR+0.036TL+0.044SY+0.065CPS+0.116ACM+0.112ASM+0.053LMD+ 0.035RMS+0.041RM+0.017EEP+0.016EES+0.025DIR+0.022TT+0.019PMW+0.022GWD+0.018GWCOR+0.015SOCOR (5) Wherein, LRI represents the leakage risk index; SC represents the type of substance; SQ represents the quantity of substance; ST represents the toxicity of substance; SCOR represents the corrosivity of substance; TL represents the number of reservoir layers; SY represents the service life; CPS represents the cathodic protection system; ACM represents anti-corrosion materials; ASM represents anti-seepage materials; LMD represents the leakage monitoring device; RMS represents the remote monitoring system; RM represents daily inspection and maintenance; EEP represents the environmental emergency plan; EES represents environmental emergency supplies; DIR represents daily inspection records; TT represents technical training; PMW represents the surrounding monitoring wells; GWD represents the groundwater level; GWCOR represents the groundwater corrosivity; and SOCOR represents the soil corrosivity.

[0073] II. Conduct a risk assessment of seepage and leakage in underground concrete storage tanks: For 20 underground concrete storage tanks in a newly constructed or expanded land area, a leakage risk assessment model was used to calculate the leakage risk index for each tank (Table 7). According to Table 7, the leakage risk index for tanks #2 (sulfuric acid), #1 (hydrochloric acid), #2 (ethanol), #1 (benzene), #2 (benzene), and #3 (benzene) is greater than or equal to 6.5. Therefore, these tanks pose a leakage risk, and the company should strengthen its supervision of them.

[0074] Table 7 Risk Assessment Scores for Leakage in the Tested Reservoir

[0075] III. Constructing a groundwater pollution risk assessment model for underground concrete storage tanks: 1. Screening of groundwater pollution risk assessment indicators Based on the characteristics of underground concrete storage tanks, and using expert consultation and scoring methods, 22 industry experts were asked to select 9 indicators from 16 potential indicators and rank them by importance (Table 8). This aimed to reduce individual expert biases regarding potential indicators, enhance the authority and rationality of the selected indicators, reduce the workload of indicator information collection, and increase the efficiency of pollution risk assessment. The most important indicator was assigned 9 points, the second most important indicator 8 points, the third most important indicator 7 points, the fourth most important indicator 6 points, the fifth most important indicator 5 points, the sixth most important indicator 4 points, the seventh most important indicator 3 points, the eighth most important indicator 2 points, and the ninth most important indicator 1 point. Table 8 shows that the following 9 potential indicators were selected as pollution risk assessment indicators: leakage risk index, substance solubility, substance volatility, wet-dry index, vadose zone lithology, groundwater depth, aquifer lithology, population in contact with the storage tank, and water sources within 1 km downstream.

[0076] Table 8 Potential Indicators for Groundwater Pollution Risk Assessment

[0077] Note: Importance = S i / [(1+2+3+4+5+6+7+8+9)×22], where S i The sum of scores from 22 experts for the i-th potential indicator 2. Construction of Groundwater Pollution Risk Assessment Index System The nine pollution risk assessment indicators selected above are classified into three aspects: risk sources, material migration pathways, and sensitive receptors, and a groundwater pollution risk assessment indicator system for the reservoir is constructed (Table 9).

[0078] Table 9 Groundwater Pollution Risk Assessment Index System

[0079] (1) Risk sources include three indicators: leakage risk index, substance solubility, and substance volatility.

[0080] ①Leakage Risk Index: This refers to the magnitude of leakage risk in the storage tank.

[0081] ② Solubility of a substance: refers to the mass of a substance that can be dissolved per unit volume of solvent under certain temperature and pressure. It is usually expressed as the ratio of the maximum mass of a substance that can be dissolved to the mass of the solvent.

[0082] ③ Volatility of a substance: refers to the ability of a liquid substance to transform into a gaseous state at temperatures below its boiling point. It is used to characterize whether a substance is easily volatile and whether it is easily inhaled or comes into contact with the skin.

[0083] (2) The material migration pathways include four indicators: wet-dry index, vadose zone lithology, groundwater depth, and aquifer lithology.

[0084] ①Dryness index: refers to the ratio of rainfall to evaporation (Formula 6).

[0085] (6) Wherein, DWI represents the dryness index; This represents the average annual precipitation over many years. This represents the average annual evaporation over many years.

[0086] ② Vadose zone lithology: refers to the lithology with the highest permeability coefficient in the vadose zone.

[0087] ③Groundwater depth: refers to the distance from the ground surface to the water table.

[0088] ④ Aquifer lithology: refers to the lithology with the highest permeability coefficient in loose rock pore aquifers.

[0089] (3) Sensitive receptors include two indicators: the population that comes into contact with the reservoir and the water source within 1 km downstream.

[0090] ① People who come into contact with the reservoir: refers to the number of employees and visitors who may come into contact with the reservoir within the newly built, renovated or expanded land area.

[0091] ②Water sources within 1km downstream: refers to the number of surface water or groundwater sources within 1km of newly built, renovated or expanded land.

[0092] 3. Scoring and grading of groundwater pollution risk assessment indicators Primary indicator: Groundwater pollution risk (90 points); Secondary indicators: Risk source (30 points), substance migration pathway (40 points), sensitive receptor (20 points); Level 3 indicators: Level 3 indicators are categorized and assigned corresponding scores (Table 10).

[0093] Table 10 Scoring and Classification of Groundwater Pollution Risk Assessment Indicators

[0094] 4. Determination of weights for groundwater pollution risk assessment indicators Based on the analytic hierarchy process (AHP), pairwise comparisons of each indicator are performed using the 1-9 scale to determine the importance scores of each indicator, thereby constructing a judgment matrix. The consistency of this judgment matrix is ​​then tested. Consistency indicators are determined (…). The value was 0.014, and the random consistency index ( The consistency ratio was 1.575. The value of 0.0089 < 0.1, indicating acceptable consistency. Finally, the weights of the pollution risk assessment indicators were obtained through normalization (Table 11).

[0095] Table 11 Weights of Groundwater Pollution Risk Assessment Indicators

[0096] 5. Construction of groundwater pollution risk assessment model: Based on the aforementioned assessment indicators and their weights, a pollution risk assessment model (Formula 7) is constructed, and a pollution risk index score of 6.7 is set as a preset threshold. If the pollution risk index of any reservoir is greater than or equal to 6.7, the newly constructed or expanded land is considered unsafe; if the pollution risk index of all reservoirs is less than 6.7, the newly constructed or expanded land is considered safe.

[0097] PRI=0.198LRI+0.079SS+0.085SV+0.088DWI+0.069VZL+0.086GD+0.120AL+0.136CSTP+0.139WS (7) Among them, PRI represents the pollution risk index; LRI represents the leakage risk index; SS represents the solubility of the substance; SV represents the volatility of the substance; DWI represents the dry-wet index; VZL represents the vadose zone lithology; GD represents the groundwater depth; AL represents the aquifer lithology; CSTP represents the population in contact with the reservoir; and WS represents the water source within 1km downstream.

[0098] IV. Conduct a groundwater pollution risk assessment for underground concrete storage tanks: For six underground concrete storage tanks in a newly constructed or expanded land area, a pollution risk assessment model was used to calculate the pollution risk index of each tank (Table 12). According to Table 12, the pollution risk index of the tanks ranges from 3.5 to 7.7.

[0099] Table 12 Safety Assessment Scores for Newly Built, Renovated, and Expanded Land Use

[0100] V. Determining the safety of newly built, renovated, or expanded land: The pollution risk index of the No. 2 sulfuric acid storage tank, the No. 1 hydrochloric acid storage tank, and the No. 2 ethanol storage tank is greater than or equal to 6.7 points (Table 12). Therefore, the newly built, renovated, and expanded land to be tested is unsafe, and the land user should investigate the causes of pollution risks to avoid risks to human health and downstream water sources.

[0101] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0102] In the description of this invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0103] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A safety assessment method for new construction, renovation, and expansion projects involving underground concrete storage tanks, characterized in that, Includes the following steps: Based on the constructed machine learning-based leakage risk assessment model, the leakage risk of underground concrete storage tanks is assessed. By calculating the leakage risk index of each storage tank, the assessment results with leakage risk are obtained. Based on the assessment results, a groundwater pollution risk assessment was conducted using the constructed groundwater pollution risk assessment model. By calculating the pollution risk index one by one, the safety of the newly built, renovated and expanded land for underground concrete storage tanks was assessed.

2. The safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks according to claim 1, characterized in that: When constructing the leakage risk assessment model, PSO and BPNN are coupled to form a PSO-BPNN hybrid machine learning algorithm to construct the leakage risk assessment model.

3. The safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks according to claim 2, characterized in that: When constructing a leakage risk assessment model, a leakage risk assessment index system is built to form a dataset for training the leakage risk assessment model. The leakage risk assessment index system includes: Properties of stored toxic and hazardous substances: type of substance, quantity of substance, toxicity of substance, corrosiveness of substance; Characteristics of reservoir construction and operation: number of reservoir layers, service life, cathodic protection system, anti-corrosion materials, and anti-seepage materials; Groundwater pollution prevention measures: leakage monitoring devices, remote monitoring systems, routine inspection and maintenance, environmental emergency plans, environmental emergency supplies, daily inspection records, technical training, and surrounding monitoring wells; External environment: groundwater level, groundwater corrosivity, soil corrosivity.

4. The safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks according to claim 3, characterized in that: When obtaining the leakage risk index, the leakage risk index is expressed as: LRI=0.045SC+0.035SQ+0.118ST+0.146SCOR+0.036TL+0.044SY+0.065CPS+0.116ACM+0.112ASM+0.053LMD+ 0.035RMS+0.041RM+0.017EEP+0.016EES+0.025DIR+0.022TT+0.019PMW+0.022GWD+0.018GWCOR+0.015SOCOR Wherein, LRI represents the leakage risk index; SC represents the type of substance; SQ represents the quantity of substance; ST represents the toxicity of substance; SCOR represents the corrosivity of substance; TL represents the number of reservoir layers; SY represents the service life; CPS represents the cathodic protection system; ACM represents anti-corrosion materials; ASM represents anti-seepage materials; LMD represents the leakage monitoring device; RMS represents the remote monitoring system; RM represents daily inspection and maintenance; EEP represents the environmental emergency plan; EES represents environmental emergency supplies; DIR represents daily inspection records; TT represents technical training; PMW represents the surrounding monitoring wells; GWD represents the groundwater level; GWCOR represents the groundwater corrosivity; and SOCOR represents the soil corrosivity.

5. The safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks according to claim 4, characterized in that: When obtaining assessment results with leakage risk, a leakage risk index score of 6.5 is set as a preset threshold; when the leakage risk index of the reservoir is greater than or equal to 6.5, it is considered to have a high leakage risk; when the leakage risk index of the reservoir is less than 6.5, it is considered to have a low leakage risk.

6. The safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks according to claim 5, characterized in that: When constructing the groundwater pollution risk assessment model, based on the constructed potential index system for groundwater pollution risk, the 1-9 scale method in the analytic hierarchy process is used to determine the importance score of each index, which is used to determine the weight of the groundwater pollution risk assessment index. Based on the summation method, the groundwater pollution risk assessment model is constructed.

7. The safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks according to claim 6, characterized in that: When constructing a potential indicator system for groundwater pollution risk, the potential indicator system for groundwater pollution risk includes: Risk sources: leakage risk index, solubility of substances, volatility of substances; Material migration pathways: wet-dry index, vadose zone lithology, groundwater depth, aquifer lithology; Sensitive receptors: People who come into contact with the reservoir, and water sources within 1 km downstream.

8. The safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks according to claim 7, characterized in that: When obtaining the pollution risk index, the pollution risk index represents: PRI=0.198LRI+0.079SS+0.085SV+0.088DWI+0.069VZL+0.086GD+0.120AL+0.136CSTP+0.139WS Among them, PRI represents the pollution risk index; LRI represents the leakage risk index; SS represents the solubility of the substance; SV represents the volatility of the substance; DWI represents the dry-wet index; VZL represents the vadose zone lithology; GD represents the groundwater depth; AL represents the aquifer lithology; CSTP represents the population in contact with the reservoir; and WS represents the water source within 1km downstream.

9. The safety assessment method for new, renovated, or expanded land use involving underground concrete storage tanks according to claim 8, characterized in that: When assessing the safety of newly constructed or expanded underground concrete storage tanks, a pollution risk index score of 6.7 is set as a preset threshold. If the pollution risk index of any storage tank is greater than or equal to 6.7, the newly constructed or expanded land is considered unsafe. If the pollution risk index of all storage tanks is less than 6.7, the newly constructed or expanded land is considered safe.

Images