Method and device for determining full-cycle maintenance strategy of subsea pipeline system

Abstract

The present invention provides a method and device for determining a full-cycle maintenance strategy for a subsea pipeline system. First, based on nonlinear finite element analysis and the Monte Carlo method, the evolution laws of the ultimate burst pressure and failure probability of the corroded pipeline structure over time are solved, which can effectively describe the influence laws of the interaction between nonlinear factors and corrosion defects on the structural ultimate burst pressure and corresponding failure probability. Subsequently, an economic evaluation system is introduced to calculate the cost of pipeline burst failure and the input cost of the maintenance strategy. Finally, a multi-objective optimization algorithm is coupled with the above two parts of the module to complete the information mapping of the full-cycle failure probability evolution characteristics of different pipe sections and the corresponding economic costs, so as to find the optimal maintenance plan. The maintenance strategy determination method can comprehensively consider the failure risk of the pipeline system under the influence of the interaction between corrosion defects and the corresponding failure economic cost, so as to unify the safety and economy of the pipeline operation project.

Description

Technical Field

[0001] The present invention relates to the technical field of subsea pipeline maintenance, and particularly to a method and device for determining a full-cycle maintenance strategy for a subsea pipeline system. Background Art

[0002] As the main medium for transporting oil and gas resources, subsea pipelines have the advantages of high efficiency, continuity, economy, etc., and are widely used in the offshore oil and gas development industry. Different from onshore pipelines, the working environment of subsea pipelines is harsh, and the corrosion situation is aggravated. Statistical data shows that corrosion is the primary factor threatening the safety of subsea pipelines. As the service time increases, different forms of corrosion defects gradually form on the pipeline surface, and their sizes increase steadily. Therefore, the ultimate burst pressure of the pipeline structure continuously decreases, and when there is an interaction between corrosion defects, the strength reduction is more obvious. When the ultimate burst pressure drops below the design internal pressure, the structure will undergo burst failure, triggering oil spills and even explosion accidents, causing serious losses to life, economy, and the environment. To ensure the safety and operating benefits of the subsea pipeline system, it is necessary to accurately estimate the ultimate burst pressure of corroded pipelines with certain defect forms and the evolution characteristics of their failure probabilities, and combine with an economic evaluation system to finally formulate a full-cycle maintenance strategy for the pipeline system, so as to achieve the unity of the safety and economy of the pipeline project.

[0003] Currently, the existing methods for determining pipeline maintenance strategies mainly have the following two limitations: (1) Most of the solutions for the structural ultimate burst pressure are based on empirical formula methods, which are difficult to describe the influence laws of non-linear factors and the interaction between corrosions on the ultimate burst pressure, introducing large model errors into the failure probability estimation and making the design conservative; (2) Most studies are limited to formulating maintenance strategies within a certain time interval, without considering the evolution law of the full-cycle failure probability of each pipeline segment and its contribution to the failure probability of the pipeline system, making it difficult to plan the full-cycle maintenance strategy of the pipeline system and limiting the application scope. Summary of the Invention

[0004] In view of the above problems, the present invention proposes a method and device for determining a full-cycle maintenance strategy for a subsea pipeline system. First, based on non-linear finite element analysis and the Monte Carlo method, the evolution laws of the ultimate burst pressure and failure probability of the corroded pipeline structure over time are solved, which can effectively describe the influence laws of non-linear factors and the interaction between corrosion defects on the structural ultimate burst pressure and the corresponding failure probability; then an economic evaluation system is introduced to consider the economic consequence cost of pipeline burst failure. Finally, a multi-objective optimization algorithm is coupled with the above two parts of the module to complete the information mapping of the full-cycle failure probability evolution characteristics and the corresponding economic costs of different pipeline segments, so as to find the optimal maintenance plan.

[0005] According to one aspect of the present invention, the present invention provides a method for determining a full-cycle maintenance strategy for a subsea pipeline system, including the following steps:

[0006] S1: Based on the non - linear finite element analysis method and the Monte Carlo method, obtain the evolution laws of the ultimate burst pressure and failure probability of the corroded pipeline structure over time;

[0007] S2: Introduce an economic evaluation system to calculate the economic consequence cost caused by pipeline burst failure;

[0008] S3: Couple the multi - objective optimization algorithm with the results obtained in steps S1 and S2, complete the information mapping of the full - cycle failure probability evolution characteristics of different pipeline segments and the corresponding economic consequence costs, so as to find the optimal maintenance plan to achieve the full - cycle maintenance of the subsea pipeline system.

[0009] Furthermore, in step S1, the non - linear finite element analysis method includes:

[0010] S1.1: According to the corrosion defect size, type, as well as the subsea pipeline size and steel material properties, establish the corresponding solid model pipe d , and divide the grid for it using the structured grid method;

[0011] S1.2: Apply fixed - end constraint conditions to both ends of the solid model pipe d , and constrain the degrees of freedom in six directions; apply a unit pressure load p d to the inner surface of the solid model pipe e = 1 MPa;

[0012] S1.3: Based on the strain failure criterion, solve the ultimate burst pressure p e of the pipeline structure under the unit pressure load p b through the arc - length method iteration;

[0013] S1.4: Based on the non - linear finite element analysis method described in S1.1 - S1.3, use the Latin hypercube sampling method to generate a certain number of training sample sets:

[0014] {x i , p i b,FEM (x i )}(i = 1, 2, …, n) (1)

[0015] In the formula, x i is the independent variable vector of the i - th sample in the training sample set; n is the number of the training sample set; p i b,FEM (x i ) is the finite element simulation value of the ultimate burst pressure corresponding to the independent variable vector of the i - th sample in the training sample set;

[0016] S1.5: Train the preset pipeline ultimate bursting pressure prediction model according to the training sample set, and obtain the surrogate model p of the nonlinear finite element analysis by fitting to realize the prediction of the pipeline ultimate bursting pressure. b,s (x).

[0017] Furthermore, in step S1.1, the corrosion defect sizes include: defect depth d, defect length l, defect width w, defect circumferential spacing s c and defect axial spacing s l ; the corrosion defect types include: single-point corrosion, circumferential distributed corrosion, axial distributed corrosion, and composite corrosion; the subsea pipeline sizes include: pipeline outer diameter D and pipeline wall thickness t; the steel material properties include: elastic modulus E s , Poisson's ratio ν s , yield strength σ y and ultimate tensile strength σ uts .

[0018] Furthermore, step S1.3 includes:

[0019] Plot the pressure-corrosion area von-Mises stress change curve. When the minimum von-Mises stress in the corrosion area exceeds the ultimate tensile strength σ uts , the corresponding pressure is the ultimate bursting pressure p b .

[0020] Furthermore, in step S1.5, the preset pipeline ultimate bursting pressure prediction model includes at least one of: genetic programming model, BP neural network model, radial basis network model, support vector machine model, chaotic quadratic polynomial expansion model, and convolutional neural network model.

[0021] Furthermore, in step S1, the Monte Carlo method includes:

[0022] S1.6: The sizes of the corrosion defects increase steadily with the pipeline service time and are estimated according to the linear growth model. The estimation formulas are as follows:

[0023] d(t) = d0 + v d ·t (2)

[0024] l(t) = l0 + v l ·t (3)

[0025] w(t) = w0 + v w ·t (4)

[0026] Where, d(t) is the depth of the corrosion defect at time t, d0 is the initial depth of the corrosion defect; l(t) is the length of the corrosion defect at time t, l0 is the initial length of the corrosion defect; w(t) is the width of the corrosion defect at time t, w0 is the initial width of the corrosion defect; v d is the annual growth rate of the corrosion depth; v l is the annual growth rate of the corrosion length; v w is the annual growth rate of the corrosion width; t is the service time;

[0027] S1.7: When the ultimate bursting pressure p b of the pipeline is less than the design working internal pressure p op , the pipeline undergoes bursting failure. Based on the reliability theory, the corresponding structural limit state equation is established as follows:

[0028] g(x, t) = p b,s (x, t) - p op (5)

[0029] Where, x is the independent variable vector, t is the service time of the pipeline, p op is the design working internal pressure, p b,s is the ultimate bursting pressure obtained by substituting the model prediction;

[0030] S1.8: The failure probability P f of the pipeline structure is the integral of the joint probability density function f x (x, t) within the failure domain Ω f ∈{x|g(x, t) ≤ 0}. Based on the Monte Carlo method, the unbiased estimate of the failure probability is calculated according to formula (7):

[0031]

[0032]

[0033] Where, f x (x, t) is the joint probability density function; x1, x2,..., x n are the 1st, 2nd,..., nth components of the independent variable vector respectively; I(·) is the indicator function, when y ≤ 0, I(y) = 1, when y > 0, I(y) = 0; N f is the number of failures; N is the total number of Monte Carlo simulations.

[0034] Furthermore, step S2 includes:

[0035] S2.1: Calculate the economic cost consequence C tf after the occurrence of the pipeline failure accident, including the failure loss cost C f and the total maintenance cost C r , and the calculation formula is as follows:

[0036] C tf = C f + C r (8)

[0037] S2.2: Failure loss cost C f includes economic and financial loss cost C eco and environmental and financial loss cost C env , and is calculated according to formula (9); economic and financial loss cost C eco is mainly caused by the loss of crude oil leaked into the sea water and the delay of crude oil production caused by the shutdown and maintenance of the subsea pipeline infrastructure, and is estimated according to formula (10); environmental loss cost C env is estimated according to formula (11), which is mainly used to pay the economic loss compensation fees, environmental loss repair costs, and oil spill cleanup costs of the tourism and fishing industries during the oil spill and cleanup process:

[0038] C f = C env + C eco (9)

[0039] C eco = C p × (Q lp × T lp + Q dp × T dp )(10)

[0040] C env = 54432[0.01(Q h × T lp )] 0.728 (11)

[0041] In the formula, C p is the price of the oil product; Q lp is the output of the lost oil product; Q dp is the output of the oil product with postponed production; T lp is the duration of the oil product loss; T dp is the duration of the postponed production; 0.001(Q h × T lp ) is the oil spill volume;

[0042] S2.3: Total maintenance cost C r is composed of the pipeline inspection cost C inv and the maintenance process cost C rt , and is estimated according to formula (12):

[0043] C r = C inv + C rt(12)

[0044] S2.4: Convert the value of all the above costs to the initial year of project operation. Then, the pipeline life cycle cost LCC is estimated according to formula (13):

[0045]

[0046] In the formula, i is the interest rate; t is the pipeline service time, T is the pipeline service life cycle, C tf is the economic cost consequence; C r is the total maintenance cost; I r (·) is the maintenance indication function. When y ≤ 0, I r (y) = 1. When y > 0, I r (y) = 0.

[0047] Furthermore, step S3 includes:

[0048] S3.1: Define the optimization objectives of the maintenance strategy, which consists of two parts:

[0049] For the pipeline system, its system failure probability P fs is estimated according to formula (14). The optimization objective is to minimize its average failure probability, as shown in formula (15);

[0050] Meanwhile, minimize its pipeline system life cycle cost LCC s to the minimum, as shown in formula (16):

[0051]

[0052]

[0053]

[0054] In the formula, P fs (x,t) is the system failure probability, P f,i (x,t) is the failure probability of the i-th subsystem; P m is the average failure probability; LCC s is the pipeline system life cycle cost; n ps is the number of subsystems; is the maintenance indication function of the j-th subsystem. When y ≤ 0, When y > 0,

[0055] S3.2: Define the constraint conditions: The maximum annual failure probability P max of the pipeline system does not exceed the failure probability criterion FPC, as expressed by formula (17); After taking maintenance measures, the corresponding protection benefit ΔR is positive, as expressed by formula (18):

[0056] P max = max{P fs (x,t),(t = 1,2,…,T)} < FPC(17)

[0057]

[0058]

[0059] wherein, P max is the maximum annual failure probability of the pipeline system; T is the service life cycle of the pipeline; FPC is the failure probability criterion; LCC s ini is the life cycle cost of the pipeline system without maintenance measures; P fs ini (x,t) is the failure probability of the pipeline system without maintenance measures;

[0060] S3.3: According to the optimization objective in step S3.1 and the constraint conditions in S3.2, construct an optimization framework based on the multi-objective genetic algorithm, and couple it with the results of the nonlinear finite element and failure probability analysis in step S1 and the economic evaluation results in step S2 to complete the information mapping of the full-cycle failure probability evolution characteristics and the corresponding economic costs of different pipe segments, so as to find the optimal maintenance plan.

[0061] Further, step S3.3 includes:

[0062] S3.3.1: Define the framework parameters, including: the population size N pop , the number of generations N evo , the optimization objective, the constraint conditions, and the evolutionary algorithm parameters;

[0063] S3.3.2: Generate the initial population, that is, when the current generation number n evo = 0, randomly generate a chromosome sequence with a size of N pop , and each chromosome sequence corresponds to the full-cycle maintenance plan of the pipeline system;

[0064] S3.3.3: Couple each chromosome sequence with the results of the nonlinear finite element and failure probability analysis in step S1 to obtain its full-cycle system failure probability curve P fs (x,t)-t, and estimate its average system failure probability P m and the maximum annual system failure probability P max ; couple each chromosome sequence and its full-cycle system failure probability curve P fs (x,t)-t with the economic evaluation results in step S2 to obtain its cost C - the number of maintenance times n rCurve and estimate its life cycle cost LCC s and protection benefit ΔR;

[0065] S3.3.4: According to the average system failure probability P corresponding to each chromosome m and life cycle cost LCC s , use the non-dominated sorting algorithm to evaluate the fitness of each chromosome: (1) Non-dominated sorting: According to the dominance relationship between chromosomes, assign them to different fronts and give the corresponding front number i r : Chromosomes in the front with a smaller number dominate chromosomes in the front with a larger number, and chromosomes in the same front are non-dominated with each other; (2) Crowding distance calculation: For chromosomes in the same front, calculate their crowding distance i dis , the larger the crowding distance, the better the chromosome diversity; (3) Sorting operation: According to the front number i of the chromosome r and the crowding distance i dis perform sorting operation according to Equation (19) to obtain the arrangement order of this group of chromosomes: Chromosomes with a smaller front number are prior to chromosomes with a larger front number; when the front numbers of chromosomes are the same, chromosomes with a larger crowding distance are prior to chromosomes with a smaller crowding distance:

[0066]

[0067] where, i r is the front number of the i-th chromosome; j r is the front number of the j-th chromosome; i dis is the crowding distance of the i-th chromosome; j dis is the crowding distance of the j-th chromosome;

[0068] S3.3.5: According to the chromosome arrangement order obtained in S3.3.4 and the constraint conditions, use the tournament selection method based on the feasibility rule to select a group of chromosomes with a quantity of N from the parental chromosome sequence pop and perform crossover operation and mutation operation, and finally obtain the offspring chromosome sequence

[0069] S3.3.6: Based on the elite operation principle, combine the parental chromosome sequence and the offspring chromosome sequence populations into and sort the chromosomes according to the non-dominated sorting algorithm in step S3.3.4, prune the population, and retain the first N pop chromosomes as the new generation of population Update the evolution generation, let n evo = nevo +1;

[0070] S3.3.7: Repeat steps S3.3.4 - S3.3.6 until the number of generations of evolution reaches the set value, i.e., n evo = N evo ;

[0071] S3.3.8: After the evolution ends, the chromosomes located at the first front end constitute the Pareto optimal solution. Each chromosome represents a full - cycle maintenance strategy. At this time, the corresponding average failure probability and life - cycle cost are obtained according to step S3.3.3, that is, the Pareto front.

[0072] According to another aspect of the present invention, the present invention also provides a device for determining the full - cycle maintenance strategy of a subsea pipeline system, including the following modules:

[0073] A non - linear finite - element analysis and failure - probability analysis module, which is used to obtain the evolution law of the ultimate burst pressure and failure probability of the corroded pipeline structure over time based on the non - linear finite - element analysis method and the Monte Carlo method;

[0074] An economic evaluation module, which is used to introduce an economic evaluation system to calculate the economic consequence cost caused by pipeline burst failure;

[0075] A multi - objective optimization module, which is used to couple the multi - objective optimization algorithm with the non - linear finite - element analysis and failure - probability analysis module and the economic evaluation module to complete the information mapping of the full - cycle failure - probability evolution characteristics of different pipe segments and the corresponding economic consequence costs, so as to find the optimal maintenance plan to achieve the full - cycle maintenance of the subsea pipeline system.

[0076] The technical solution provided by the present invention has the following beneficial effects:

[0077] 1. The present invention establishes a non - linear finite - element analysis - failure - probability estimation coupling model for subsea pipelines with interacting corrosion defects; through non - linear finite - element analysis, the ultimate burst pressure of subsea pipelines with interacting corrosion defects is obtained, and it is coupled with the Monte Carlo method through a surrogate model to estimate the variation characteristics of the pipeline failure probability with service time. Through the non - linear finite - element analysis - failure - probability estimation coupling model, the influence of non - linear factors and the interaction of corrosion defects on the bearing capacity of the pipeline structure is mapped into the probability space, and the obtained failure - probability evaluation result is more accurate.

[0078] 2. The present invention proposes a method for determining the full - cycle maintenance strategy of a subsea pipeline system, which can accurately capture the change characteristics of the failure probability of each pipe segment in the subsea pipeline system, comprehensively evaluate the influence of each pipe segment on the failure risk of the pipeline system, and determine the full - cycle maintenance strategy of the pipeline based on a multi - objective optimization algorithm. This method for determining the maintenance strategy can comprehensively consider the failure risk of the pipeline system under the influence of the interaction between corrosion defects and the corresponding failure economic cost, so as to unify the safety and economy of the pipeline operation project. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] The following will further illustrate the specific effects of the present invention in conjunction with the drawings. In the drawings:

[0080] Figure 1 is the implementation flowchart of a method for determining the full - cycle maintenance strategy of a subsea pipeline system in an embodiment of the present invention;

[0081] Figure 2 is the change curve of the pipeline failure probability P f with the service time t;

[0082] Figure 3 is the change curve of the pipeline system failure probability P fs and the pipe segment failure probability P f with the service time t;

[0083] Figure 4 is the full - cycle maintenance plan of the pipeline system determined in an embodiment of the present invention, where (a) shows the change of the pipeline system failure probability P fs with the service time t under different maintenance plans; (b) shows the maintenance plan for pipe segment 1; (c) shows the maintenance plan for pipe segment 2. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0084] In order to have a clearer understanding of the technical features, objectives, and effects of the present invention, the specific embodiments of the present invention will now be described in detail with reference to the drawings.

[0085] Refer to Figure 1 , a method for determining the full - cycle maintenance strategy of a subsea pipeline system in an embodiment of the present invention includes three major steps:

[0086] S1: Based on the non - linear finite element analysis method and the Monte Carlo method, obtain the evolution law of the ultimate burst pressure and failure probability of the corroded pipeline structure over time;

[0087] S1 is specifically divided into the following sub - steps:

[0088] S1.1: According to the corrosion defect size, type, as well as the subsea pipeline size and steel material properties, establish the corresponding solid model pipe d, and a structured grid method is used to divide the grid for it;

[0089] Preferably, the corrosion defect sizes include: defect depth d, defect length l, defect width w, defect circumferential spacing s c and defect axial spacing s l ; the corrosion defect types include: single-point corrosion, circumferential distributed corrosion, axial distributed corrosion, and composite corrosion; the subsea pipeline sizes include: pipeline outer diameter D and pipeline wall thickness t; the steel material properties include: elastic modulus E s , Poisson's ratio ν s , yield strength σ y and ultimate tensile strength σ uts .

[0090] S1.2: Apply fixed-end constraint conditions to both ends of the solid model pipe d , restricting its six degrees of freedom; apply a unit pressure load p d to the inner surface of the solid model pipe e = 1 MPa;

[0091] S1.3: Based on the strain failure criterion, use the arc-length method to iteratively solve the ultimate burst pressure p e of the pipeline structure under the unit pressure load p b ; specifically: draw the curve of the change in von-Mises stress in the pressure-corrosion area, and when the minimum von-Mises stress in the corrosion area exceeds the ultimate tensile strength σ uts , the corresponding pressure is the ultimate burst pressure p b ;

[0092] S1.4: Based on the nonlinear finite element analysis method described in S1.1 - S1.3, use the Latin hypercube sampling method to generate a certain number of training sample sets:

[0093] {x i , p i b,FEM (x i )}(i = 1, 2, …, n) (1)

[0094] In the formula, x is the independent variable vector; x i is the independent variable vector of the i-th sample in the training sample set; n is the number of the training sample set; p i b,FEM (x i ) is the finite element simulation value of the ultimate burst pressure corresponding to the independent variable vector of the i-th sample in the training sample set;

[0095] S1.5: Train the preset pipeline ultimate bursting pressure prediction model according to the training sample set, and obtain the surrogate model p of the nonlinear finite element analysis by fitting to realize the prediction of the pipeline ultimate bursting pressure. b,s (x), to realize the prediction of the pipeline ultimate bursting pressure.

[0096] Among them, the preset pipeline ultimate bursting pressure prediction model is preferably a genetic programming model;

[0097] In some embodiments, it can also be at least one of a BP neural network model, a radial basis network model, a support vector machine model, a chaotic quadratic polynomial expansion model, and a convolutional neural network model.

[0098] S1.6: The size of the corrosion defect grows stably with the pipeline service time and is estimated according to the linear growth model. The estimation formula is as follows:

[0099] d(t) = d0 + v d ·t (2)

[0100] l(t) = l0 + v l ·t (3)

[0101] w(t) = w0 + v w ·t (4)

[0102] In the formula, d(t) is the depth of the corrosion defect at time t, d0 is the initial depth of the corrosion defect; l(t) is the length of the corrosion defect at time t, l0 is the initial length of the corrosion defect; w(t) is the width of the corrosion defect at time t, w0 is the initial width of the corrosion defect; v d is the annual growth rate of the corrosion depth; v l is the annual growth rate of the corrosion length; v w is the annual growth rate of the corrosion width; t is the service time;

[0103] S1.7: When the pipeline ultimate bursting pressure p b is less than the designed working internal pressure p op , the pipeline undergoes bursting failure. Based on the reliability theory, the corresponding structural limit state equation is established as follows:

[0104] g(x, t) = p b,s (x, t) - p op (5)

[0105] In the formula, x is the independent variable vector, t is the pipeline service time, p op is the designed working internal pressure, p b,s is the ultimate bursting pressure obtained by surrogate model prediction;

[0106] S1.8: The failure probability P f (x, t) is the joint probability density function fx (x, t) is within the failure domain Ω f ∈ {x | g(x, t) ≤ 0} for integration. Based on the Monte Carlo method, the unbiased estimate of the failure probability is calculated according to formula (7):

[0107]

[0108]

[0109] Where f x (x, t) is the joint probability density function; x1, x2, …, x n are the 1st, 2nd, …, nth components of the independent variable vector respectively; I(·) is the indicator function, when y ≤ 0, I(y) = 1, when y > 0, I(y) = 0; N f is the number of failures; N is the total number of Monte Carlo simulations.

[0110] S2: Introduce an economic evaluation system to calculate the economic consequence cost caused by pipeline burst failure;

[0111] S2 can be specifically broken down into the following small steps:

[0112] S2.1: Calculate the economic cost consequence C tf after the pipeline failure accident, including the failure loss cost C f and the total maintenance cost C r . The calculation formula is as follows:

[0113] C tf = C f + C r (8)

[0114] S2.2: The failure loss cost C f includes the economic and financial loss cost C eco and the environmental and financial loss cost C env , and is calculated according to formula (9); the economic and financial loss cost C eco is mainly caused by the loss of crude oil leaked into the sea water and the delay in crude oil production caused by the shutdown and maintenance of the subsea pipeline infrastructure, and is estimated according to formula (10); the environmental loss cost C env is estimated according to formula (11), which is mainly used to pay the economic loss compensation fees, environmental loss repair costs, and oil spill cleanup costs in the tourism and fishing industries during the oil spill and cleanup process:

[0115] C f = C env + C eco (9)

[0116] C eco = C p×(Q lp ×T lp +Q dp ×T dp ) (10)

[0117] C env =54432[0.01(Q h ×T lp )] 0.728 (11)

[0118] In the formula, C p is the price of the oil product; Q lp is the output of the lost oil product; Q dp is the output of the oil product with postponed production; T lp is the duration of the oil product loss; T dp is the duration of the postponed production; 0.001(Q h ×T lp ) is the oil spill volume;

[0119] S2.3: The total maintenance cost C r consists of the pipeline inspection cost C inv and the maintenance process cost C rt , and is estimated according to formula (12):

[0120] C r =C inv +C rt (12)

[0121] S2.4: Convert the value of all the above costs to the initial year of the project operation, then the life cycle cost LCC of the pipeline is estimated according to formula (13):

[0122]

[0123] In the formula, i is the interest rate; t is the service time of the pipeline, T is the service life cycle of the pipeline, C tf is the economic cost consequence; C r is the total maintenance cost; I r (·) is the maintenance indication function. When y ≤ 0, I r (y) = 1. When y > 0, I r (y) = 0.

[0124] S3: Couple the multi-objective optimization algorithm with the results obtained in steps S1 and S2 to complete the information mapping of the full-cycle failure probability evolution characteristics of different pipeline segments and the corresponding economic consequence costs, so as to find the optimal maintenance plan to achieve the full-cycle maintenance of the subsea pipeline system.

[0125] Step S3 includes:

[0126] S3.1: Define the optimization objectives for the maintenance strategy, which consists of two parts:

[0127] For the pipeline system, its system failure probability P fs Estimated according to formula (14), the optimization objective is to minimize the average value of its failure probability, as shown in formula (15);

[0128] At the same time, minimize the life cycle cost LCC of its pipeline system s To the minimum, as shown in formula (16):

[0129]

[0130]

[0131]

[0132] In the formula, P fs (x,t) is the system failure probability, and P f,i (x,t) is the failure probability of the i-th subsystem; P m Is the average value of the failure probability; LCC s Is the life cycle cost of the pipeline system; n ps Is the number of subsystems; Is the maintenance indication function of the j-th subsystem. When y ≤ 0, When y > 0,

[0133] S3.2: Define the constraint conditions: The maximum annual failure probability P of the pipeline system max Does not exceed the failure probability criterion FPC, as expressed by formula (17); After taking maintenance measures, the corresponding protection benefit ΔR is positive, as expressed by formula (18):

[0134] P max = max{P fs (x,t),(t = 1,2,…,T)} < FPC (17)

[0135]

[0136]

[0137] In the formula, P max Is the maximum annual failure probability of the pipeline system; T is the pipeline service life; FPC is the failure probability criterion; LCC s ini Is the life cycle cost of the pipeline system under the condition of not taking maintenance measures; P fs ini(x, t) is the failure probability of the pipeline system without maintenance measures;

[0138] S3.3: According to the optimization objectives in step S3.1 and the constraint conditions in S3.2, construct an optimization framework based on the multi-objective genetic algorithm, and couple it with the results of the nonlinear finite element and failure probability analysis in step S1 and the economic evaluation results in step S2 to complete the information mapping of the full-cycle failure probability evolution characteristics of different pipe segments and the corresponding economic costs, so as to find the optimal maintenance plan.

[0139] Step S3.3 further includes:

[0140] S3.3.1: Define the framework parameters, including: the population size N pop 、the number of generations N evo 、the optimization objectives, constraint conditions, and evolutionary algorithm parameters (using binary coding, non-dominated sorting algorithm, tournament selection method);

[0141] S3.3.2: Generate the initial population, that is, when the current generation number n evo = 0, randomly generate a chromosome sequence with a quantity of N pop , and each chromosome sequence corresponds to the full-cycle maintenance plan of the pipeline system;

[0142] S3.3.3: Couple each chromosome sequence with the results of the nonlinear finite element and failure probability analysis in step S1 to obtain its full-cycle system failure probability curve P fs (x, t)-t, and estimate its average system failure probability P m and the maximum annual system failure probability P max ; couple each chromosome sequence and its full-cycle system failure probability curve P fs (x, t)-t with the economic evaluation results in step S2 to obtain its cost C - the number of maintenance times n r curve, and estimate its life cycle cost LCC s and the protection benefit ΔR;

[0143] S3.3.4: According to the average system failure probability P m and the life cycle cost LCC s corresponding to each chromosome, use the non-dominated sorting algorithm to evaluate the fitness of each chromosome: (1) Non-dominated sorting: According to the dominance relationship between chromosomes, assign them to different fronts and give the corresponding front numbers i r : The chromosomes in the front with a smaller number dominate the chromosomes in the front with a larger number, and the chromosomes in the same front are non-dominated with each other; (2) Crowding distance calculation: For the chromosomes in the same front, calculate their crowding distance idis , the larger the crowding distance, the better the chromosome diversity; (3) Sorting operation: According to the front-end serial number i of the chromosome r and the crowding distance i dis Perform sorting operation according to formula (19) to obtain the arrangement order of this group of chromosomes: Chromosomes with smaller front-end serial numbers take precedence over chromosomes with larger front-end serial numbers; when the front-end serial numbers of chromosomes are the same, chromosomes with larger crowding distances take precedence over chromosomes with smaller crowding distances:

[0144]

[0145] In the formula, i r is the front-end serial number of the i-th chromosome; j r is the front-end serial number of the j-th chromosome; i dis is the crowding distance of the i-th chromosome; j dis is the crowding distance of the j-th chromosome;

[0146] S3.3.5: According to the chromosome arrangement order obtained in S3.3.4 and the constraint conditions, use the tournament selection method based on the feasibility rule to select a group of chromosomes with a quantity of N from the parental chromosome sequence pop and perform crossover operation and mutation operation to finally obtain the offspring chromosome sequence

[0147] S3.3.6: Based on the elite operation principle, combine the parental chromosome sequence the offspring chromosome sequence population into and sort the chromosomes according to the non-dominated sorting algorithm in step S3.3.4, prune the population, and retain the first N pop chromosomes as the new generation of population Update the evolutionary generation, let n evo = n evo +1;

[0148] S3.3.7: Repeat steps S3.3.4 - S3.3.6 until the evolutionary generation reaches the set value, that is, n evo = N evo ;

[0149] S3.3.8: After the evolution ends, the chromosomes located at the first front end form the Pareto optimal solution. Each chromosome represents a full-cycle maintenance strategy. At this time, obtain the corresponding average failure probability and life cycle cost according to step S3.3.3, that is, the Pareto front.

[0150] Example calculation:

[0151] According to the above description, the method proposed by the present invention includes non - linear finite element analysis and failure probability estimation. The accuracy of both will directly affect the determination of the maintenance strategy, and it is necessary to verify them first.

[0152] 1. Verification of non - linear finite element analysis

[0153] Based on the non - linear finite element analysis method described in steps S1.1 - S1.3, numerical simulation is carried out for the full - scale pipeline bursting test, and the ultimate bursting pressure p obtained by non - linear finite element analysis b,FEM is compared with the result p in the test b,EXP The material and size parameters of the pipeline in the test and the comparison of calculation results are shown in Table 1. It can be seen that the ultimate bursting pressure obtained by non - linear finite element analysis is in good agreement with the test results, and has high accuracy in the case of single defects and interacting defects (axial distributed defects, circumferential distributed defects, composite defects). The relative error is controlled within the range of 0.039% - 5.31%, meeting the requirements of engineering design. Therefore, non - linear finite element analysis can reasonably describe the failure behavior of the structure under internal pressure load and can be used for subsequent failure probability analysis and determination of maintenance strategies.

[0154] Table 1 Comparison of ultimate bursting pressure results

[0155]

[0156]

[0157] Note: D is the pipeline diameter; t is the pipeline wall thickness; σ y is the yield strength of the pipeline steel material; σ uts is the ultimate tensile strength of the pipeline steel material; d is the depth of the corrosion defect; l is the length of the corrosion defect; w is the width of the corrosion defect; s l is the axial spacing between corrosion defects; s c is the circumferential spacing between corrosion defects.

[0158] 2. Verification of failure probability

[0159] 2.1 Verification of surrogate model accuracy

[0160] To improve the efficiency of failure probability estimation and meet the accuracy requirements, the present invention uses a surrogate model based on non - linear finite element analysis to predict the ultimate bursting pressure of the pipeline. Therefore, when performing failure probability estimation, it is necessary to first verify the prediction accuracy of the surrogate model.

[0161] Use the method described in steps S1.4 - S1.5 to construct a surrogate model, and take 80% of the data in the sample dataset to form the training dataset D train, which is used for the training of the surrogate model, and the remaining 20% of the data constitutes the prediction dataset D test , which is used to verify the prediction accuracy of the surrogate model for unknown data. The statistical results of the prediction errors for the training dataset and the prediction dataset are shown in Table 2. It can be seen that the prediction results of the surrogate model have a very high correlation with the results of the nonlinear finite element analysis, and its average coefficient of determination R 2 is 0.9999. The surrogate model has a stable prediction performance for the ultimate burst pressure of pipelines with single defects and interacting defects, and its average relative root mean square error RAE is 0.0236. At the same time, the average error difference coefficient ΔErr is 0.0166, indicating that when the prediction object changes from the training data to the prediction data, the prediction error of the surrogate model does not decrease significantly, indicating that the surrogate model has good robustness and generalization ability. The above analysis shows that the surrogate model can achieve accurate and rapid prediction of the ultimate burst pressure of pipelines, complete the information mapping of nonlinear factors, and can be used for subsequent failure probability estimation and maintenance strategy determination.

[0162] Table 2 Statistical results of the prediction errors of the surrogate model

[0163]

[0164]

[0165] Note: R 2 is the coefficient of determination (Equation 20); RAE is the relative absolute error (Equation 21); RRSE is the root relative square error (Equation 22); ΔErrors is the error difference coefficient, indicating the degree of decrease in the prediction accuracy of the surrogate model for unknown data and known data (Equation 23)

[0166]

[0167]

[0168]

[0169]

[0170] In the formula, p i b,FEM is the value of the nonlinear finite element analysis of the ultimate burst pressure of the i-th pipeline; p i b,s is the predicted value of the ultimate burst pressure of the i-th pipeline by the surrogate model; n is the sample size of the dataset.

[0171] 2.2 Verification of Failure Probability Estimation

[0172] Using the method described in steps S1.6 - S1.8, the full - cycle failure probability of pipelines with single defects and interacting defects is estimated. For comparative analysis, for pipelines with single defects, the Monte Carlo - code method is additionally used to estimate their full - cycle failure probability, and the result can be approximately regarded as the true value.

[0173] The calculation results are as Figure 2 shown. It can be seen that for pipelines with single defects, the results of the estimation method of the present invention and the Monte Carlo - code method are in good agreement, and the failure probability - service time change curves of the two almost coincide, proving the correctness of the failure probability estimation method of the present invention. Further, it can be seen that the failure probability change characteristics of pipelines with different defects are relatively consistent: when the service period t < 10 y, due to the small size of the corrosion defects, the ultimate bearing capacity of the structure is not significantly reduced, so the failure probability change curves of pipelines with different corrosion defects approximately coincide and all approach 0. After that, the increase in service time is accompanied by the stable growth of the corrosion defect size, the structural bearing capacity gradually decreases, and the failure probability gradually increases. After the service time t > 10 y, each curve shows an obvious upward trend. At the same time, the interaction effect between corrosion defects on the decline of the structural bearing capacity gradually appears, and the failure probability curve - service time curve of pipelines under different corrosion defect conditions separates. The failure probability of pipelines with interacting defects is significantly greater than that of pipelines with single defects, and the difference increases with the increase of service time. The above analysis shows that the failure probability method of the present invention can accurately describe the full - cycle change characteristics of the pipeline failure probability, and at the same time can fully consider the influence of non - linear factors and the interaction between corrosion defects on the structural bearing capacity and failure risk, which is more in line with the actual situation and lays a foundation for the determination of subsequent full - cycle maintenance strategies.

[0174] 3 Calculation and Analysis

[0175] Taking a certain offshore pipeline system as the research object, the applicability of the algorithm proposed by the present invention is verified, and its full - cycle maintenance strategy is determined. The offshore pipeline system consists of two sub - pipe segments, and their corrosion types are single defect (pipe segment 1) and composite defect (pipe segment 2) respectively. The pipeline geometric dimensions and material strength are shown in Table 3, and the corrosion defect - related parameters are shown in Table 4. The random variables considered in this example and their statistical characteristics are shown in Table 5. The pipeline service period is T = 30 y, and the failure probability criterion is FPC = 0.01.

[0176] Table 3 Pipeline Size and Material Strength Parameters

[0177]

[0178] Table 4 Corrosion Defect Types and Size Parameters

[0179]

[0180] Table 5 Statistical Characteristics of Random Variables

[0181]

[0182] Figure 3 The evolution law of the system failure probability of the offshore pipeline system over time is given. It can be seen that due to the interaction of corrosion defects, the failure probability of the pipe section with composite defects is significantly greater than that of the pipeline with a single defect, and the former dominates the failure risk of the offshore pipeline system. Figure 4 The corresponding full-cycle maintenance strategy is given, which includes two schemes: Scheme 1: For pipe section 1 and pipe section 2, maintenance is carried out in their 11th year and 21st year respectively. This result exactly coincides with Figure 3 the variation law of the failure probability reflected in. In the first 10 years, the corrosion defects have not fully developed, and their failure probability does not increase significantly. Therefore, the first maintenance year is selected as the 11th year to inhibit the subsequent rapid growth of corrosion defects. At the same time, the second maintenance year is selected as the 21st year, keeping the maintenance interval within 10 years, so that the growth law of pipeline corrosion defects is controlled within the first 10 years, and the failure probability of the pipeline is kept at a low level. The life cycle cost LCC s = 1930.7×10 4 $, the average failure probability P m = 1.71×10 -3 . The maximum annual failure probability P max = 9.18×10 -3 < FPC, meeting the constraint condition of the failure probability, and the protection benefit ΔR = 2015.6×10 4 $, meeting the constraint condition of the protection benefit; Scheme 2: As mentioned above, Figure 3 the pipe section 2 with composite defects in dominates the failure risk of the entire offshore pipeline system, so more maintenance resources are tilted towards it. For pipe section 1, the maintenance plan is still in the 11th year and 21st year, with a total of 2 maintenance times. For pipe section 2, maintenance is carried out in the 11th year, 18th year, and 24th year respectively, with a total of 3 maintenance times, shortening its maintenance interval to 6 - 7 years to further reduce its failure risk, thereby reducing the overall system failure probability. The life cycle cost LCC s = 2373.6×10 4 $, the average failure probability P m = 8.9×10 -4 . The maximum annual failure probability P max = 9.18×10 -3 < FPC, meeting the constraint condition of the failure probability, and the protection benefit ΔR = 1584×10 4$, satisfying the constraint conditions of the protection benefit. From the above analysis, it can be seen that the method for determining the full-cycle maintenance strategy of the submarine pipeline system proposed by the present invention can accurately describe the change characteristics of the failure probability of the pipeline system, effectively identify the pipe segments with dominant failure risks and give a certain inclination of maintenance resources, reasonably reduce the failure probability of the entire pipeline system, and at the same time ensure the protection benefit.

[0183] In some embodiments, there is also provided a device for determining the full-cycle maintenance strategy of a submarine pipeline system, including the following modules:

[0184] A non-linear finite element analysis and failure probability analysis module, which is used to obtain the evolution law of the ultimate burst pressure and failure probability of the corroded pipeline structure over time based on the non-linear finite element analysis method and the Monte Carlo method;

[0185] An economic evaluation module, which is used to introduce an economic evaluation system to calculate the economic consequence cost caused by the pipeline burst failure;

[0186] A multi-objective optimization module, which is used to couple the multi-objective optimization algorithm with the non-linear finite element analysis and failure probability analysis module and the economic evaluation module to complete the information mapping of the full-cycle failure probability evolution characteristics of different pipe segments and the corresponding economic consequence costs, so as to find the optimal maintenance plan to achieve the full-cycle maintenance of the submarine pipeline system.

[0187] As Figure 1 shown, each module of the device respectively implements each step of the above method and can achieve the same technical effects.

[0188] It should be noted that in this article, the terms "including", "comprising" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or system. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or system including that element.

[0189] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. Among the several unit claims of the device, several of these devices may be specifically embodied by the same hardware item. The use of the words first, second, and third, etc. does not represent any order and these words can be interpreted as identifiers.

[0190] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be similarly included in the patent protection scope of the present invention.

Claims

1. A method for determining the full-cycle maintenance strategy of a subsea pipeline system, characterized in that, It includes the following steps: S1: Based on the nonlinear finite element analysis method and the Monte Carlo method, obtain the evolution laws of the ultimate burst pressure and failure probability of the corroded pipeline structure over time; S2: Introduce an economic evaluation system to calculate the economic consequence cost caused by pipeline burst failure; S3: Couple the multi-objective optimization algorithm with the results obtained in steps S1 and S2, complete the information mapping of the full-cycle failure probability evolution characteristics and the corresponding economic consequence cost of different pipe segments, and thus search for the optimal maintenance plan to achieve the full-cycle maintenance of the subsea pipeline system; Step S3 includes: S3.1: Define the optimization objectives of the maintenance strategy, which consist of two parts: For the pipeline system, the optimization objective is to minimize the average failure probability; meanwhile, minimize the life cycle cost of the pipeline system The formula is as follows: Wherein, is the system failure probability, is the failure probability of the i-th subsystem; is the average value of the failure probability; is the life cycle cost of the pipeline system; is the number of subsystems; is the maintenance indication function of the j-th subsystem. When y ≤ 0, = 1. When y > 0, = 0; S3.2: Define the constraint conditions, and the formula is as follows: In the formula, is the maximum annual failure probability of the pipeline system; T is the service life cycle of the pipeline; FPC is the failure probability criterion; is the life cycle cost of the pipeline system without taking maintenance measures; is the failure probability of the pipeline system without taking maintenance measures; S3.3: According to the optimization objectives in step S3.1 and the constraint conditions in S3.2, construct an optimization framework based on the multi-objective genetic algorithm, and couple it with the nonlinear finite element and failure probability analysis results in step S1 and the economic evaluation results in step S2, complete the information mapping of the full-cycle failure probability evolution characteristics and the corresponding economic cost of different pipe segments, and thus search for the optimal maintenance plan.

2. The method for determining the full-cycle maintenance strategy of the subsea pipeline system according to claim 1, wherein In step S1, the nonlinear finite element analysis method includes: S1.1: Establish the corresponding solid model pipe according to the corrosion defect size, type, as well as the subsea pipeline size and steel material properties d , and use the structured grid method to divide the grid for it; S1.2: Create a solid model of pipe d Apply fixed-end boundary conditions at both ends to constrain its six degrees of freedom; for the solid model of pipe d Apply a unit pressure load to the inner surface p e = 1 MPa; S1.3: Based on the strain failure criterion, the limit bursting pressure of the pipeline structure under unit pressure load is iteratively solved by the arc-length method p e p b ;​ S1.4: Based on the nonlinear finite element analysis method described in S1.1 - S1.3, use the Latin hypercube sampling method to generate a preset number of training sample sets: (1) In the formula, x i is the independent variable vector of the i th sample in the training sample set; n is the number of the training sample set; p i b,FEM ( x i ) is the finite element simulation value of the ultimate bursting pressure corresponding to the independent variable vector of the i th sample in the training sample set; S1.5: Train a preset pipeline ultimate bursting pressure prediction model according to the training sample set, and fit to obtain a surrogate model for nonlinear finite element analysis p b,s ( x ), to achieve the prediction of the pipeline ultimate bursting pressure.

3. The method for determining the full-cycle maintenance strategy of the subsea pipeline system according to claim 2, characterized in that In step S1.1, the corrosion defect sizes include: defect depth d , defect length l , defect width w , defect circumferential spacing s c and defect axial spacing s l ; the corrosion defect types include: single-point corrosion, circumferential distributed corrosion, axial distributed corrosion, and composite corrosion; the subsea pipeline sizes include: pipeline outer diameter D and pipeline wall thickness t ; the steel material properties include: elastic modulus E s , Poisson's ratio ν s , yield strength σ y and ultimate tensile strength σ uts .

4. The method for determining the full-cycle maintenance strategy of the submarine pipeline system according to claim 2, wherein Step S1.3 includes: Plot the von-Mises stress variation curve of the pressure-corrosion area. When the minimum von-Mises stress in the corrosion area exceeds the ultimate tensile strength σ uts the corresponding pressure is the ultimate bursting pressure p b .

5. The method for determining the full-cycle maintenance strategy of the subsea pipeline system according to claim 2, characterized in that, In step S1.5, the preset pipeline ultimate burst pressure prediction model includes at least one of a genetic programming model, a BP neural network model, a radial basis network model, a support vector machine model, a chaotic quadratic polynomial expansion model, and a convolutional neural network model.

6. The method for determining the full-cycle maintenance strategy of the submarine pipeline system according to claim 1, characterized in that In step S1, the Monte Carlo method includes: S1.6: The size of the corrosion defect grows steadily with the service time of the pipeline, and is estimated according to the linear growth model, and the estimation formula is as follows: (2) (3) (4) Wherein, is the depth of the corrosion defect at t ; d 0 is the initial depth of the corrosion defect; is the length of the corrosion defect at t ; l 0 is the initial length of the corrosion defect; is the width of the corrosion defect at time t, w 0 is the initial width of the corrosion defect; v d is the annual growth rate of the corrosion depth; v l is the annual growth rate of the corrosion length; v w is the annual growth rate of the corrosion width; t is the service time; S1.7: When the ultimate bursting pressure of the pipeline p b is less than the designed working internal pressure p op the pipeline bursts and fails. Based on the reliability theory, the corresponding structural limit state equation is established as follows: (5) In the formula, x is the independent variable vector, t is the pipeline service time, p op is the designed internal working pressure, p b,s is the ultimate bursting pressure obtained by surrogate model prediction; S1.8: Failure probability of pipeline structure P f ( x , t ) is the joint probability density function f x ( x , t ) within the failure domain Integrate. Based on the Monte Carlo method, the unbiased estimate of the failure probability is calculated according to formula (7): (6) (7) In the formula, f x ( x , t ) is the joint probability density function; x 1, x 2, …, x n are the 1st, 2nd, …, n th components of the independent variable vector respectively; I ( ) is the indicator function. When y ≤0, I ( y ) = 1. When y >0, I ( y ) = 0; N f is the number of failure times; N is the total number of Monte Carlo simulations.

7. The method for determining the full-cycle maintenance strategy of the subsea pipeline system according to claim 1, characterized in that, Step S2 includes: S2.1: Calculate the economic cost consequences after a pipeline failure accident C tf , including failure loss costs C f and total maintenance costs C r , and the calculation formula is as follows: (8) S2.2: Failure loss cost C f Including economic and financial loss costs C eco and environmental and financial loss costs C env , which are calculated according to formula (9); the economic and financial loss costs C eco are mainly caused by the loss of crude oil leaked into the sea water and the delay of crude oil production caused by the shutdown and maintenance of the subsea pipeline infrastructure, and are estimated according to formula (10); the environmental loss costs C env are estimated according to formula (11), which are mainly used to pay the economic loss compensation fees, environmental loss repair costs, and oil spill cleanup costs in the process of oil spill and cleanup: (9) (10) (11) Wherein, C p is the price of the oil product; Q lp is the output of the lost oil product; Q dp is the output of the oil product with production postponed; T lp is the duration of the loss of the oil product; T dp is the duration of the postponed production; 0.001( Q h × T lp ) is the oil spill volume; S2.3: Total maintenance cost C r Composed of pipeline inspection cost C inv and maintenance process cost C rt It is estimated according to formula (12): (12) S2.4: Convert the value of all the above costs to the initial year of project operation, then the pipeline life cycle cost LCC is estimated according to formula (13): (13) In the formula, i is the interest rate; t is the pipeline service time, T is the pipeline service cycle, C tf is the economic cost consequence; C r is the total maintenance cost; I r (·) is the maintenance indication function. When , I r ( y ) = 1. When , I r ( y ) = 0.

8. The method for determining the full-cycle maintenance strategy of the subsea pipeline system according to claim 1, wherein Step S3.3 includes: S3.3.1: Define the framework parameters, including: the population size N pop , the number of generations N evo , the optimization objective, the constraint conditions, and the evolutionary algorithm parameters; S3.3.2: Generate the initial population, i.e., the current generation number of evolution n evo When = 0, randomly generate a set of chromosome sequences with the number of N pop Each chromosome sequence corresponds to the full-cycle maintenance plan of the pipeline system; S3.3.3: Couple each chromosome sequence with the results of the nonlinear finite element and failure probability analysis in step S1 to obtain its full-cycle system failure probability curve P fs ( x ,t)- t , and estimate its average system failure probability P m and the maximum annual system failure probability P max ; Couple each chromosome sequence and its full-cycle system failure probability curve P fs ( x , t )- t with the economic evaluation results in step S2 to obtain its cost C -maintenance times n r curve, and estimate its life cycle cost LCC s and the protection benefit Δ R ; S3.3.4: According to the average system failure probability corresponding to each chromosome P m and the life cycle cost LCC s , use the non-dominated sorting algorithm to evaluate the fitness of each chromosome: (1) Non-dominated sorting: According to the dominance relationship between chromosomes, assign them to different fronts and assign corresponding front numbers i r : Chromosomes in the front with a smaller number dominate chromosomes in the front with a larger number, and chromosomes in the same front are non-dominated with each other; (2) Crowding distance calculation: For chromosomes in the same front, calculate their crowding distance i dis , the larger the crowding distance, the better the chromosome diversity; (3) Sorting operation: According to the front number of the chromosome i r and the crowding distance i dis perform a sorting operation according to Equation (19) to obtain the arrangement order of this group of chromosomes: Chromosomes with a smaller front number take precedence over chromosomes with a larger front number; when the front numbers of chromosomes are the same, chromosomes with a larger crowding distance take precedence over chromosomes with a smaller crowding distance: (19) In the formula, i r is the front-end serial number of the i th chromosome; j r is the front-end serial number of the j th chromosome; i dis is the crowding distance of the i th chromosome; j dis is the crowding distance of the j th chromosome; S3.3.5: According to the chromosome arrangement order and constraint conditions obtained in S3.3.4, use the tournament selection method based on the feasibility rule to select a group of P n chromosome sequences with a quantity of N pop from the parental chromosome sequences Q n evo, and perform crossover operation and mutation operation to finally obtain the offspring chromosome sequences Q n evo; S3.3.6: Based on the elite operation principle, combine the parental chromosome sequences P n evo and the offspring chromosome sequences Q n evo populations into , and sort the chromosomes according to the non-dominated sorting algorithm in step S3.3.4, prune the population, and retain the first N pop chromosomes as the new generation of population Q n evo +1 ; Update the evolution generation, let n evo = n evo + 1; S3.3.7: Repeat steps S3.3.4 - S3.3.6 until the number of generations of evolution reaches the set value, i.e., n evo = N evo ; S3.3.8: After the evolution ends, the chromosomes located at the first front form the Pareto optimal solution. Each chromosome represents a full-cycle maintenance strategy. At this time, obtain the corresponding average failure probability and life cycle cost according to step S3.3.3, that is, the Pareto front.

9. An apparatus for determining a full-cycle maintenance strategy for a subsea pipeline system, characterized in that, It includes the following modules: Nonlinear finite element analysis and failure probability analysis module, which is used to obtain the evolution laws of the ultimate burst pressure and failure probability of the corroded pipeline structure over time based on the nonlinear finite element analysis method and the Monte Carlo method; Economic evaluation module, which is used to introduce an economic evaluation system to calculate the economic consequence cost caused by pipeline burst failure; Multi-objective optimization module, which is used to couple the multi-objective optimization algorithm with the nonlinear finite element analysis and failure probability analysis module and the economic evaluation module, complete the information mapping of the full-cycle failure probability evolution characteristics and the corresponding economic consequence cost of different pipe segments, and thus search for the optimal maintenance plan to achieve the full-cycle maintenance of the subsea pipeline system; Specifically include: Define the optimization objectives of the maintenance strategy, which consist of two parts: For the pipeline system, the optimization goal is to minimize the average failure probability; meanwhile, minimize the life cycle cost of the pipeline system The formula is as follows: Wherein, is the system failure probability, is the failure probability of the i-th subsystem; is the average value of the failure probability; is the life cycle cost of the pipeline system; is the number of subsystems; is the maintenance indication function of the j-th subsystem. When y ≤ 0, = 1. When y > 0, = 0; Define the constraint conditions, and the formula is as follows: In the formula, is the maximum annual failure probability of the pipeline system; T is the service life cycle of the pipeline; FPC is the failure probability criterion; is the life cycle cost of the pipeline system without taking maintenance measures; is the failure probability of the pipeline system without taking maintenance measures; According to the optimization objectives and constraint conditions, construct an optimization framework based on the multi-objective genetic algorithm.

Images