A method and system for real-time reliability evaluation of subsea Christmas tree dependent competing failures

Abstract

The present application belongs to the field of offshore oil engineering, and particularly relates to a real-time reliability evaluation method and system for dependent competing failure of a subsea Christmas tree. The real-time reliability evaluation method for dependent competing failure of the subsea Christmas tree comprises five main steps: determining a failure mode, modeling competing failure, modeling a Bayesian network based on competing failure, real-time reliability calculation, and dynamic reliability calculation. The real-time reliability evaluation system for dependent competing failure of the subsea Christmas tree comprises six parts: a subsea Christmas tree mechanical module state monitoring subsystem, a subsea distribution unit state monitoring subsystem, a subsea control unit state monitoring subsystem, a subsea valve group state monitoring subsystem, a subsea Christmas tree wellhead mechanical module state monitoring subsystem, and a data collection and analysis subsystem.

Description

Technical Field

[0001] This invention belongs to the field of marine oil engineering, specifically relating to a real-time reliability assessment method and system for dependent competition failure of subsea production trees. Background Technology

[0002] With social development and technological progress, the country's demand for oil and gas resources is constantly increasing. Due to the dwindling supply of onshore oil and gas resources, development is gradually shifting from land to the ocean, making offshore oil and gas resource extraction particularly important. As a crucial facility for offshore oil extraction, the safe and efficient operation of subsea production trees (DMLs) is paramount. During extraction, any malfunction or leakage in the DML can severely impact the safe production activities of the entire oilfield. To avoid accidents and minimize loss of life and property, rigorous assessments of the DMLs are necessary to ensure their high level of reliability.

[0003] Because subsea production trees operate underwater for extended periods, their reliability is constantly subjected to severe challenges. The internal structure of the subsea production tree system not only experiences oil and gas erosion and corrosion but also endures the complex influences of the external underwater environment, such as high pressure and low temperature. Therefore, higher requirements are placed on the structural reliability of the subsea production tree system. Considering the impact of these factors, conducting a reliability assessment of the subsea production tree system is essential. The various failure factors encountered by the subsea production tree during operation can be categorized into two types: degradation failure caused by internal factors and sudden failure triggered by external impacts. These two types of failure are often not independent but rather interact, with the failure of the subsea production tree system resulting from the competition between these two failure modes. Therefore, the impact of competing failures needs to be considered when conducting a reliability assessment of the subsea production tree system. By establishing a competing failure mechanism for the subsea production tree and using real-time data, real-time reliability assessment of the subsea production tree can be achieved.

[0004] Uncertainty is inherently unavoidable, and Bayesian networks, with their strong uncertainty handling capabilities, have been widely applied in reliability engineering. Currently, research on reliability assessment of competing failures rarely considers the parameter uncertainty of the distribution function. However, due to uncertainties such as the working environment and limited expert experience data, the parameters in the distribution function may also exhibit uncertainty. To date, no comprehensive assessment method has been developed to address the impact of parameter uncertainty in related competing failures. Therefore, proposing a real-time reliability assessment method and system for dependent competing failures of subsea production trees is particularly necessary. Summary of the Invention

[0005] Considering the impact of the interaction between different failure factors on subsea production trees, this invention provides a real-time reliability assessment method and system for dependent competing failures of subsea production trees.

[0006] To achieve the above objectives, according to one aspect of the present invention, a real-time reliability assessment method for dependent competing failure of subsea production trees comprises five main steps: S1: Determine the failure mode. Collect failure data from the established subsea production tree database, classify the failure of the subsea production tree system into degradation failure caused by internal degradation and sudden failure caused by external shock, and select appropriate degradation and shock models.

[0007] S2: Competitive Failure Modeling. This study investigates competitive failure mechanisms, considering the impact of sudden failures on degradation increments and degradation rates. Parameter estimation is performed using a subsea production tree failure database, and a competitive failure model is established.

[0008] S3: Bayesian Network Modeling Based on Competitive Failure. A static Bayesian network is established based on the competitive failure model of the subsea tree. The nodes in the static Bayesian network that change over time during the competitive failure process of the subsea tree are determined. The parameter uncertainty is represented by different probability density functions, and the distribution of the independent variable nodes can be normal distribution, Weibull distribution, etc.

[0009] S4: Real-time reliability calculation. The acquired real-time data is updated using Bayesian methods, and the new parameters are fed into the static Bayesian network of the subsea tree competing failure process to obtain the real-time reliability of the subsea tree system.

[0010] S5: Dynamic Reliability Calculation. The static Bayesian network of the subsea tree competition failure process is extended into a dynamic Bayesian network of the subsea tree competition failure process. By extending the time of the dynamic Bayesian network of the subsea tree competition failure process, with years as the time unit, the reliability of the subsea tree in the next few years starting from the current moment is calculated.

[0011] According to another aspect of the present invention, a real-time reliability assessment system for dependent competition failure of subsea production trees includes six parts: a subsea production tree mechanical unit status monitoring subsystem, a subsea distribution unit status monitoring subsystem, a subsea control module status monitoring subsystem, a subsea valve group status monitoring subsystem, a wellhead mechanical module status monitoring subsystem, and a data collection and analysis subsystem.

[0012] The underwater wellhead mechanical unit condition monitoring subsystem includes a pressure sensor group, a temperature sensor group, a vibration sensor group, a corrosion monitoring sensor group, and an underwater wellhead mechanical system degradation information acquisition module, which is used to understand the degradation status of the mechanical system in the underwater wellhead mechanical unit during the operation of the underwater wellhead.

[0013] The underwater distribution unit condition monitoring subsystem includes a pressure sensor group, a flow sensor group, a temperature sensor group, a vibration sensor group, an underwater distribution unit electronic system information collection module, and an underwater distribution unit hydraulic system information collection module. It is used to understand the degradation status of the underwater distribution unit hydraulic system and electronic system during underwater wellhead operation.

[0014] The underwater control unit condition monitoring subsystem includes a pressure sensor group, a flow sensor group, a temperature sensor group, a vibration sensor group, an underwater control unit electronic system information collection module, and an underwater control unit hydraulic system information collection module. It is used to understand the degradation status of the underwater control unit hydraulic system and electronic system during underwater wellhead operation.

[0015] The subsea valve assembly condition monitoring subsystem includes a pressure sensor group, a flow sensor group, a temperature sensor group, a vibration sensor group, and a subsea valve assembly hydraulic system information acquisition module, which is used to understand the degradation status of the subsea valve assembly hydraulic system during subsea wellhead operation.

[0016] The subsea wellhead mechanical module condition monitoring subsystem includes a pressure sensor group, a temperature sensor group, a vibration sensor group, a corrosion monitoring sensor group, and a subsea wellhead mechanical system degradation information collection module, which is used to understand the degradation status of the wellhead mechanical system during subsea well operation.

[0017] The data collection and analysis subsystem includes a communication module, a signal conversion module, a probability information generation and update module, a competing failure probability algorithm module, and a reliability assessment module. It can calculate the failure probability of the subsea production tree system and perform real-time reliability assessments of the system. Compared with the prior art, the beneficial result of the present invention is: a real-time reliability assessment method and system for dependent competing failure of subsea production trees, which analyzes the reliability of subsea production trees under competing failure modes, studies the competing failure mechanism, considers the parameter uncertainty problem of the model, and realizes real-time assessment and effective prediction of system reliability. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a real-time reliability assessment method for dependent competing failures of subsea oil production trees.

[0019] Figure 2 This is a schematic diagram of a static Bayesian network for reliability in underwater oil production tree interdependence competition failure.

[0020] Figure 3 This is a schematic diagram of a dynamic Bayesian network for reliability in underwater oil production tree interdependence competition failure.

[0021] Figure 4 This is a schematic diagram of an underwater wellhead system.

[0022] Figure 5 This is a schematic diagram of a real-time reliability assessment system for dependent competition failure of subsea oil production trees.

[0023] In the diagram, 101 is the underwater production tree mechanical unit, 102 is the production tree cap, 103 is the production tree trunk, 104 is the underwater distribution unit, 105 is the hydraulic distribution module, 106 is the electronic distribution module, 107 is the underwater control unit, 108 is the control pilot valve, 109 is the first solenoid valve, 110 is the second solenoid valve, 111 is the third solenoid valve, 112 is the underwater electronic module, 113 is the fourth solenoid valve, and 114 is the fifth solenoid valve. 115. Sixth solenoid valve; 116. Subsea valve assembly; 117. Annular wing valve; 118. Chemical injection valve; 119. Annular main valve; 120. Production main valve; 121. Switching valve; 122. Production wing valve; 123. Subsea Christmas tree wellhead mechanical unit; 124. Christmas tree connector; 125. Wellhead; 126. Downhole safety valve; 201. Subsea Christmas tree mechanical unit condition monitoring subsystem; 202. Pressure sensor assembly. 203. Temperature sensor group; 204. Vibration sensor group; 205. Corrosion monitoring sensor group; 206. Subsea production tree mechanical unit degradation information acquisition module; 207. Subsea distribution unit status monitoring subsystem; 208. Pressure sensor group; 209. Flow sensor group; 210. Temperature sensor group; 211. Vibration sensor group; 212. Subsea distribution unit electronic system information acquisition module; 213. Subsea distribution unit hydraulic system information acquisition module; 214. Subsea control unit status monitoring subsystem; 215. Pressure sensor group 216. Flow sensor group; 217. Temperature sensor group; 218. Vibration sensor group; 219. Subsea control unit electronic system information acquisition module; 220. Subsea control unit hydraulic system information acquisition module; 221. Subsea valve group status monitoring subsystem; 222. Pressure sensor group; 223. Flow sensor group; 224. Temperature sensor group; 225. Vibration sensor group; 226. Subsea valve group hydraulic system information acquisition module; 227. Subsea wellhead mechanical module status monitoring subsystem; 228. Pressure sensor group; 229. Temperature sensor group; 230. Vibration sensor group; 231. Corrosion monitoring sensor group; 232. Wellhead mechanical system degradation information acquisition module; 233. Data collection and analysis subsystem; 234. Communication module; 235. Signal conversion module; 236. Probabilistic information generation and update module; 237. Competition failure probability algorithm module; 238. Reliability assessment module. Detailed Implementation

[0024] like Figure 1 As shown, a real-time reliability assessment method for dependent competing failures of subsea production trees comprises five main steps: S1: Determine the failure mode. Collect failure data from the established subsea production tree database, classify the failure of the subsea production tree system into degradation failure caused by internal degradation and sudden failure caused by external shock, and select appropriate degradation and shock models.

[0025] S101: Determine the degradation failure model.

[0026] use The process describes the internal degradation process of the subsea wellhead, and the amount of internal degradation can be expressed as: ; In the formula, This represents the initial value of the degradation. and These represent the drift coefficient and the diffusion coefficient, respectively. This is standard Brownian motion. Drift parameters. It follows a normal distribution, that is... diffusion coefficient These are fixed parameters. (By...) The nature of the process is known. Follows a normal distribution: ; S102: Determine the sudden failure model.

[0027] Under external shocks, the number of shocks experienced by a subsea production tree system has considerable uncertainty. Assume... This indicates the time until the subsea production tree system is completed. The number of times the impact has occurred. Obedience strength is of The process. Then, at the designated time, the subsea production tree system... It just so happens The probability density function of the secondary impact can be expressed as: ; make For the first subsea production tree system Damage caused by secondary impact, impact damage It is a sequence of independent and identically distributed random variables that follow a normal distribution. .

[0028] ; make Indicates the subsea wellhead system The sum of the damage caused by each impact is then The expression is: ; Then before The sum of damage caused by the secondary impact The distribution can be represented as: ; S2: Establish a competing failure model. Study the competing failure mechanism, consider the impact of sudden failure on degradation increment and degradation rate, use the subsea production tree failure database to estimate parameters and establish a competing failure model.

[0029] S201: Degradation increment caused by external shock.

[0030] Under the influence of external shocks, subsea production tree systems experience additional degradation. This invention employs a cumulative shock model to describe the impact of shocks on internal degradation. Assumptions The additional degradation caused by constant impact is as follows ,but The cumulative additional degradation of the subsea production tree system at any given time can be expressed as: ; Additional degradation caused by each impact on the subsea drilling tree It is a series of independent, identically distributed random variables. If it follows a normal distribution, then the cumulative additional degradation amount The distribution is as follows: ; S202: External shocks cause changes in degradation rate.

[0031] When the cumulative impact damage exceeds the set initial sudden failure threshold At that time, the initial degradation rate will change. If the first... If a shock initially causes a change in the degradation rate, then this shock is defined as a trigger shock. The time at which the trigger shock occurs is called the transition time, denoted by . This indicates that no triggering impact occurred when the subsea production tree system was operating, i.e. At that time, the degradation rate was When a triggering impact occurs during the operation of the subsea production tree system, i.e. At that time, the degradation rate from , increased to Assuming the degradation rate of the subsea production tree system changes only once during operation, the degradation amount due to external shock causing a change in the degradation rate can be expressed as: ; in, The time of change of degradation rate (i.e., the first) The time it takes for the second impact to arrive is called the transition time, and the transition time of the degradation rate. Following a gamma distribution, the scale parameter is The shape parameters are . The initial degradation rate, To change the degradation rate of the system; and They all follow a normal distribution, that is Assuming , In order to make It is necessary to ensure .

[0032] Total degradation of subsea wellhead systems Degradation caused by internal factors Additional degradation caused by impact The sum of It can be represented as: ; When the total degradation within the subsea production tree system exceeds the degradation failure threshold... At time [time], the system fails. The probability that a subsea production tree system will not degrade or fail can be expressed as: ; When the impact is triggered At that time, the total degradation of the underwater oil production tree It can be represented as: ; When the impact is triggered At that time, the total degradation of the underwater oil production tree It can be represented as: ; S203: Establish a competitive failure model.

[0033] By combining the sudden failure model caused by shock with the degradation failure model caused by internal degradation, when the total internal degradation of the subsea production tree system exceeds the degradation failure threshold... Or, the total degradation caused by a sudden failure exceeds the sudden failure threshold. Subsea Christmas trees can fail. When considering the degradation rate and degradation increment caused by shocks, the reliability calculation of subsea Christmas trees is divided into the following cases: When time No impact occurred, or At that time, the reliability of the subsea production tree system can be expressed as: ; When time At least one impact occurred, or At this time, there are two scenarios in which the subsea production tree will not fail. The probabilities of these two mutually exclusive scenarios must be added together.

[0034] (1) When the cumulative degradation caused by external shock is less than Sometimes, or The reliability of the subsea production tree system is: ; (2) When the cumulative degradation caused by external shock is greater than ,or The reliability of the subsea production tree system is: ; The reliability function of the subsea production tree system is the sum of the three formulas above. Therefore, the reliability function of the subsea production tree can be expressed as: ; S204: By estimating the parameters of known data in the subsea production tree database and combining them with the field expert database, the probability distribution is corrected to obtain the values ​​of the parameters corresponding to degradation failure and sudden failure.

[0035] The maximum likelihood estimation method is used to estimate parameters in degradation failure and sudden failure: ; in, These are the parameters being estimated in the model; This refers to status monitoring data, i.e., samples.

[0036] S3: Bayesian Network Modeling Based on Competitive Failure. A static Bayesian network is established based on the competitive failure model of the subsea tree. The nodes in the static Bayesian network that change over time during the competitive failure process of the subsea tree are determined. The parameter uncertainty is represented by different probability density functions, and the distribution of the independent variable nodes can be normal distribution, Weibull distribution, etc.

[0037] S301: The conditional probability table is determined by the physical relationships between variables in the competing failure model and the number of samplings. The relationships between all independent and dependent variable nodes are determined based on the physical formulas provided by the competing failure model. Then, appropriate methods such as Monte Carlo simulation, particle filtering, and random filtering are used to sample the parent nodes. Finally, the conditional probability table is obtained through physical relationships with sufficient sampled values ​​and corresponding probabilities.

[0038] S302: The prior probabilities of variables are determined by assuming that the relevant variables follow a certain probability distribution. The prior probabilities are obtained from the distribution of the performance variables corresponding to the parent node at the initial time step. Depending on the actual physical conditions, the distribution can be a normal distribution, a Weibull distribution, etc. Subsequently, discretization methods are used to discretize these corresponding nodes to form discrete prior probabilities.

[0039] S303: Establish a static Bayesian network based on competition failure. Map each parameter in the formulas of S2 to a parameter node in the Bayesian network. Input the conditional probability table and prior probabilities obtained in S302 into the static Bayesian network to establish the static Bayesian network of the competition failure reliability assessment model, as follows. Figure 2 As shown, where, , and Used for calculation , and The probability value at time, the independent variable node , , , , , , , For uncertain parameters, consider the parameter uncertainty problem in the competitive failure model, where all independent variable nodes follow a probability distribution.

[0040] S4: Real-time reliability calculation. The acquired real-time data is processed and fed into the static Bayesian network of the subsea tree competition failure process to obtain the real-time reliability of the subsea tree system.

[0041] S401: Real-time failure data processing, which denoises sensor signals and performs time-frequency domain analysis.

[0042] (1) Noise Reduction Processing: The signals from sensors in the subsea production tree system are easily affected by various external factors during transmission. This patent uses wavelet transform for noise reduction. Threshold denoising of noisy signals is... Perform wavelet transform, compare the threshold and coefficients, and process the coefficients that are greater than and less than the threshold separately.

[0043] ; in It is a wavelet function. yes The wavelet transform is used, and a general thresholding rule is employed to select the threshold for the wavelet transform thresholding denoising method. As shown below: ; in, The standard deviation of the added noise signal is given, where N is the actual measured signal. The sum of wavelet coefficients obtained after wavelet transform decomposition. A soft thresholding function is used to perform nonlinear thresholding on the wavelet transform coefficients of the measured signal. ; (2) Time-Frequency Domain Analysis: Given the denoised sensor signals from the subsea production tree system, in addition to obtaining the frequency domain characteristics of the signals, it is also necessary to process the time history changes of the signals to obtain transient information. This invention uses short-term Fourier transform for time-frequency domain analysis: ; in, This indicates a truncation window, which allows for step-by-step analysis of each segment of the signal after truncation. By comparing the local spectrum of each time period, the characteristics of the signal changing over time can be obtained.

[0044] S402: Perform Bayesian updates on real-time data. According to Bayesian theory, the posterior probability density function of the relevant parameters of the subsea production tree system is: ; In the formula, Indicates parameters The prior probability density function; Indicates parameters The posterior probability density function, where , The number of random variables; It is a proportionality constant. ; Let be the likelihood function.

[0045] S403: Input the real-time parameters of the competing failure model of the subsea production tree system calculated in S402. Figure 2 Real-time reliability assessment is achieved in a static Bayesian network.

[0046] S5: Dynamic Reliability Calculation. The static Bayesian network of the subsea tree competition failure process is extended into a dynamic Bayesian network of the subsea tree competition failure process. By extending the time of the dynamic Bayesian network of the subsea tree competition failure process, with years as the time unit, the reliability change of the subsea tree system in the next few years starting from the current moment is calculated.

[0047] S501: The static Bayesian network based on the competing failure physical model is extended into a dynamic Bayesian network on an annual basis. The dynamic Bayesian network is used for real-time reliability analysis of subsea production trees based on competing failures. Figure 3As shown. Real-time reliability assessment of subsea production trees based on competition failures. All nodes in the dynamic Bayesian network and... Figure 2 The competitive failure static Bayesian network in the model reflects the reliability change of the subsea production tree system from one time point to the next.

[0048] like Figure 4 As shown, the subsea production tree includes a subsea production tree mechanical unit 101, a subsea distribution unit 104, a subsea control unit 107, a subsea valve assembly 116, and a subsea production tree wellhead mechanical unit 123. The subsea production tree mechanical unit 101 includes a production tree cap 102 and a production tree body 103. The production tree cap sits atop the production tree body 103 for corrosion protection and to prevent foreign objects from falling into the wellhead. The subsea distribution unit 104 includes a hydraulic distribution module 105 and an electronic distribution module 106, installed on the outside of the subsea production tree. The hydraulic distribution module 105 is connected to a control pilot valve 108 and provides it with hydraulic power, while the electronic distribution module 106 is connected to a subsea electronic module 112 and provides it with electrical power. The subsea control unit 107 is located between the subsea distribution unit and the subsea production tree and includes a control pilot valve 108 and a first solenoid valve 109. The system includes a second solenoid valve 110, a third solenoid valve 111, a subsea electronic module 112, a fourth solenoid valve 113, a fifth solenoid valve 114, and a sixth solenoid valve 115. A pilot valve 108 is connected to the six solenoid valves to supply hydraulic power to each. The subsea electronic module 112 is also connected to the six solenoid valves to control them. The first solenoid valve 109, second solenoid valve 110, third solenoid valve 111, fourth solenoid valve 113, fifth solenoid valve 114, and sixth solenoid valve 115 are respectively connected to annular wing valve 117, annular main valve 119, chemical injection valve 118, production main valve 120, switching valve 121, and production wing valve 122 to control the valve opening and closing of the valve group. A subsea Christmas tree wellhead mechanical unit 123 is installed at the bottom of the subsea Christmas tree and includes a Christmas tree connector 124, a wellhead 125, and a downhole safety valve 126.

[0049] like Figure 5 As shown, the real-time reliability analysis system for dependent competition failure of subsea production trees based on dynamic Bayesian networks consists of six parts: subsea production tree mechanical unit status monitoring subsystem 201, subsea distribution unit status monitoring subsystem 207, subsea control module status monitoring subsystem 214, subsea valve group status monitoring subsystem 221, wellhead mechanical module status monitoring subsystem 227, and data collection and analysis subsystem 233.

[0050] The subsea production tree mechanical unit status monitoring subsystem 201 includes a pressure sensor group 202, a temperature sensor group 203, a vibration sensor group 204, a corrosion monitoring sensor group 205, and a subsea production tree mechanical system degradation information collection module 206. The pressure sensor group 202, temperature sensor group 203, and corrosion monitoring sensor group 205 are respectively attached to the production tree cap 102 and the production tree body 103 to collect pressure status, temperature changes, and corrosion amount. The vibration sensor group 204 is respectively attached to the outside of the production tree cap 102 and the production tree body 103 to detect external impact information of the subsea production tree mechanical unit. The subsea production tree mechanical system degradation information collection module 206 is used to integrate the degradation information of the subsea production tree mechanical unit.

[0051] The underwater distribution unit status monitoring subsystem 207 includes a pressure sensor group 208, a flow sensor group 209, a temperature sensor group 210, a vibration sensor group 211, an underwater distribution unit electronic system information collection module 212, and an underwater distribution unit hydraulic system information collection module 213. The pressure sensor group 208, flow sensor group 209, and temperature sensor group 210 are used to collect the pressure status, flow rate changes, and temperature changes of the underwater distribution unit hydraulic distribution module 105 and electronic distribution module 106. The vibration sensor group 211 is used to detect external impact information of the underwater distribution unit hydraulic distribution module 105 and electronic distribution module 106. The underwater distribution unit electronic system information collection module 212 is used to collect degradation information of the underwater distribution unit electronic system, and the underwater distribution unit hydraulic system information collection module 213 is used to collect degradation information of the underwater distribution unit hydraulic system.

[0052] The underwater control unit status monitoring subsystem 214 includes a pressure sensor group 215, a flow sensor group 216, a temperature sensor group 217, a vibration sensor group 218, an underwater control unit electronic system information collection module 219, and an underwater control unit hydraulic system information collection module 220. The pressure sensor group 215, flow sensor group 216, and temperature sensor group are used to collect the pressure status, flow rate changes, and temperature changes of each component in the underwater control unit 107. The vibration sensor group 211 is used to detect external impact information of the underwater control module. The underwater control unit electronic system information collection module 219 is used to collect the total degradation information of the underwater control unit electronic system, and the underwater control unit hydraulic system information collection module 220 is used to collect the total degradation information of the underwater control unit hydraulic system.

[0053] The underwater valve group status monitoring subsystem 221 includes a pressure sensor group 222, a flow sensor group 223, a temperature sensor group 224, a vibration sensor group 225, and an underwater valve group hydraulic system information acquisition module 226. The pressure sensor group 222, the flow sensor group 223, and the temperature sensor group 224 are used to collect the pressure status, flow rate changes, and temperature changes of the annular wing valve 117, the chemical injection valve 118, the annular main valve 119, the production main valve 120, the switching valve 121, and the production wing valve 122. The vibration sensor group 225 is used to detect external impact information of the underwater valve group. The underwater valve group hydraulic system information collection module 226 is used to collect the total degradation information of the underwater valve group hydraulic system.

[0054] The subsea wellhead mechanical module status monitoring subsystem 227 includes a pressure sensor group 228, a temperature sensor group 229, a vibration sensor group 230, a corrosion monitoring sensor group 231, and a subsea wellhead mechanical system degradation information acquisition module 232. The pressure sensor group 228, temperature sensor group 229, and corrosion monitoring sensor group 231 are used to collect the pressure status, temperature changes, and corrosion amount of the wellhead connector 124, wellhead 125, and downhole safety valve 126. The vibration sensor group 230 is used to detect external impact information of the wellhead mechanical system. The wellhead mechanical system degradation information collection module 232 is used to collect the total degradation information of the wellhead mechanical system.

[0055] The data collection and analysis subsystem 233 includes a communication module 234, a signal conversion module 235, a probability information generation and update module 236, a competition failure probability algorithm module 237, and a reliability assessment module 238. The communication module 234 is used to receive information from the following modules: underwater production tree mechanical system degradation information collection module 206, underwater distribution unit electronic system information collection module 212, underwater distribution unit hydraulic system information collection module 213, underwater control unit electronic system information collection module 219, underwater control unit hydraulic information collection module 220, underwater valve group hydraulic information collection module 226, and water... The data information is sent by the subsea production tree wellhead mechanical system degradation information collection module 232; the signal conversion module 235 is used to convert the received signal into an electrical signal; the probability information generation and update module 236 is used to read the electrical signal converted by the signal conversion module 235 and use the read electrical signal for probability information generation, tracking and updating; the competition failure probability algorithm module 237 is used to summarize the probability information of each part and generate the probability of failure of the subsea production tree system; the reliability assessment module 238 reads the probability information generated by the subsea production tree competition failure probability algorithm module 237 and performs a reliability assessment.

Claims

1. A real-time reliability assessment method for dependent competing failure of subsea production trees, characterized in that... It includes five main steps: S1: Determine the failure mode; collect failure data from the established subsea tree database, classify the failure of the subsea tree system into degradation failure caused by internal degradation and sudden failure caused by external shock, and select degradation and shock models; S101: Determine the degradation failure model; use The process describes the internal degradation process of the subsea production tree, and the amount of internal degradation failure is expressed as: ； In the formula, Indicates the amount of internal degradation and failure. and These represent the drift coefficient and the diffusion coefficient, respectively. Standard Brownian motion; drift parameters It follows a normal distribution, that is... diffusion coefficient For fixed parameters; by The nature of the process is known. Follows a normal distribution: ； S102: Determine the sudden failure model; Under the influence of external shocks, the number of shocks experienced by the subsea production tree system has considerable uncertainty; assuming This indicates the time until the subsea production tree system is completed. The number of times the impact has occurred so far. Obedience strength is of The process; then, at the time when the subsea production tree system reaches its designated point. It just so happens The probability density function of the second impact is expressed as: make For the first subsea production tree system Damage caused by secondary impact, impact damage It is a sequence of independent and identically distributed random variables that follow a normal distribution. ; ； make Indicates the subsea wellhead system The sum of the damage caused by the secondary impacts, The expression is: ； Then before The sum of damage caused by the secondary impact The distribution is represented as: ； S2: Establish a competitive failure model; This study investigates the competitive failure mechanism, considering the impact of sudden failure on degradation increment and degradation rate, and uses the subsea production tree failure database to estimate parameters and establish a competitive failure model. S201: Changes in degradation increment caused by external shocks; Under the influence of external shocks, subsea production tree systems experience additional degradation. A cumulative shock model is used to describe the impact of shocks on internal degradation; assuming... The additional degradation caused by constant impact is as follows ,but The cumulative additional degradation of the subsea production tree system at any given time is expressed as: ； Additional degradation caused by each impact on the subsea drilling tree It is a series of independent, identically distributed random variables; If it follows a normal distribution, then the cumulative additional degradation amount is... The distribution is as follows: ； S202: External shocks cause changes in degradation rate; When the cumulative impact damage exceeds the set initial sudden failure threshold At that time, the initial degradation rate will change; if the first If a shock initially causes a change in the degradation rate, then this shock is defined as the triggering shock; the time at which the triggering shock occurs is called the transition time, denoted by . This indicates that no triggering impact occurred when the subsea Christmas tree system was operating, i.e. At that time, the degradation rate was When a triggering impact occurs during the operation of the subsea production tree system, i.e. At that time, the degradation rate from , increased to Assuming the subsea production tree system experiences only one change in degradation rate during operation, the degradation amount due to external degradation failure when the degradation rate changes due to external shock is expressed as: ； in, The time of change of degradation rate (i.e., the first) The time it takes for the second impact to arrive is called the transition time, and the transition time of the degradation rate. Following a gamma distribution, the scaling parameter is The shape parameters are The initial degradation rate, To change the degradation rate of the system; and They all follow a normal distribution, that is Assuming , In order to make It is necessary to ensure ; Total degradation of subsea wellhead systems External degradation failure degradation amount Additional degradation caused by impact The sum of Represented as: When the total degradation within the subsea production tree system exceeds the degradation failure threshold... At time 1, the system fails; at time 2, the system fails. The probability that a subsea production tree system will not degrade or fail is expressed as: ； When the impact is triggered At that time, the total degradation of the underwater oil production tree Represented as: ; When the impact is triggered At that time, the total degradation of the underwater oil production tree Represented as: ； S203: Establish a competitive failure model; By combining the sudden failure model caused by shock with the degradation failure model caused by internal degradation, when the total internal degradation of the subsea production tree system exceeds the degradation failure threshold... Or, the total degradation caused by a sudden failure exceeds the sudden failure threshold. Subsea Christmas systems can fail under certain conditions. When considering the degradation rate and degradation increment caused by shocks, the reliability calculation of subsea Christmas systems can be divided into the following cases: When time No impact occurred, or At that time, the reliability of the subsea production tree system is expressed as: ； When time At least one impact occurred, or At this time, there are two scenarios in which the subsea production tree will not fail; the probabilities of these two mutually exclusive scenarios must be added together. (1) When the cumulative degradation caused by external shock is less than Sometimes, or The reliability of the subsea production tree system is: ； (2) When the cumulative degradation caused by external shock is greater than ,or The reliability of the subsea production tree system is: ； The reliability function of the subsea production tree system is the sum of the three formulas above. Therefore, the reliability function of the subsea production tree is expressed as: ； S204: By estimating the parameters of known data in the subsea production tree database and combining them with the field expert database, the probability distribution is corrected to obtain the values ​​of the parameters corresponding to degradation failure and sudden failure. The maximum likelihood estimation method is used to estimate parameters in degradation failure and sudden failure: ； in, These are the parameters being estimated in the model; This refers to condition monitoring data, i.e., samples; S3: Bayesian network modeling based on competitive failure; a static Bayesian network is established based on the competitive failure model of the subsea tree, and the nodes in the static Bayesian network that change over time in the competitive failure process of the subsea tree are determined. The parameter uncertainty is represented by different probability density functions, and the distribution of the independent variable nodes is normal distribution and Weibull distribution. S301: The conditional probability table is determined by the physical relationships between variables in the competing failure model and the number of samplings; the relationships between all independent and dependent variable nodes are determined according to the physical formulas provided by the competing failure model; then Monte Carlo simulation, particle filtering, and random filtering are used to sample the parent nodes; finally, the conditional probability table is obtained through physical relationships with sufficient sampled values ​​and corresponding probabilities. S302: The prior probability of the variable is determined by assuming that the relevant variable follows a probability distribution; the prior probability is obtained from the distribution of the performance variable corresponding to the parent node at the initial time; according to the actual physical situation, the distribution is a normal distribution or a Weibull distribution; then the discretization method is used to discretize these corresponding nodes to form discrete prior probabilities; S303: Establish a static Bayesian network based on competition failure; map each parameter in the formulas of S2 to a parameter node in the Bayesian network; input the conditional probability table and prior probability obtained in S302 into the static Bayesian network to establish the static Bayesian network for the competition failure reliability assessment model, wherein... , and Used for calculation , and The probability value at time, the independent variable node , , , , , , , For uncertain parameters, consider the parameter uncertainty problem in the competitive failure model, where all independent variable nodes follow a probability distribution; S4: Real-time reliability calculation; The acquired real-time data is processed and fed into the static Bayesian network of the subsea tree competition failure process to obtain the real-time reliability of the subsea tree system; S401: Real-time failure data processing, denoising sensor signals and performing time-frequency domain analysis; (1) Noise Reduction: The signals from sensors in the subsea production tree system are easily affected by various external factors during transmission. Wavelet transform is used for noise reduction; threshold denoising of noisy signals is... Perform wavelet transform, compare the threshold and the coefficients, and process the coefficients that are greater than and less than the threshold separately; ； in, It is a wavelet function. yes The wavelet transform is used, and a general thresholding rule is employed to select the threshold for the wavelet transform thresholding denoising method. As shown below: ； in, The standard deviation of the added noise signal is given, where N is the actual measured signal. The sum of wavelet coefficients obtained after wavelet transform decomposition; a soft thresholding function is used to perform nonlinear thresholding on the wavelet transform coefficients of the measured signal: ； (2) Time-frequency domain analysis: Given the denoised sensor signals from the subsea production tree system, in addition to obtaining the frequency domain characteristics of the signals, it is also necessary to process the time history changes of the signals to obtain transient information; short-term Fourier transform is used for time-frequency domain analysis: ； in, This indicates a truncation window, which analyzes each segment of the signal step by step, and uses the comparison of the local spectrum of each time period to obtain the characteristics of the signal changing over time. S402: Perform Bayesian updates on real-time data. According to Bayesian theory, the posterior probability density function of the relevant parameters of the subsea production tree system is: In the formula, Indicates parameters The prior probability density function; Indicates parameters The posterior probability density function, where , The number of random variables; It is a proportionality constant. ; It is the likelihood function; S403: Input the real-time parameters of the competing failure model of the subsea production tree system calculated in S402 into the static Bayesian network to achieve real-time reliability assessment. S5: Dynamic reliability calculation; The static Bayesian network of the underwater tree competition failure process is extended into a dynamic Bayesian network of the underwater tree competition failure process. By extending the time of the dynamic Bayesian network of the underwater tree competition failure process, the reliability change of the underwater tree system in the next few years from the current moment is calculated in years. S501: The static Bayesian network based on the competing failure physical model is extended into a dynamic Bayesian network on an annual basis; the dynamic Bayesian network for real-time reliability assessment of subsea production trees based on competing failure is consistent with the static Bayesian network for competing failure, reflecting the reliability change of the subsea production tree system from one moment to the next. A real-time reliability assessment method for dependent competition failure of subsea production trees is applied to a real-time reliability assessment system for dependent competition failure of subsea production trees. The system consists of six parts: subsea production tree mechanical unit status monitoring subsystem, subsea distribution unit status monitoring subsystem, subsea control module status monitoring subsystem, subsea valve group status monitoring subsystem, subsea production tree wellhead mechanical module status monitoring subsystem, and data collection and analysis subsystem. The underwater production tree mechanical unit condition monitoring subsystem includes a pressure sensor group, a temperature sensor group, a vibration sensor group, a corrosion monitoring sensor group, and an underwater production tree mechanical system degradation information acquisition module; The underwater distribution unit status monitoring subsystem includes a pressure sensor group, a flow sensor group, a temperature sensor group, a vibration sensor group, an underwater distribution unit electronic system information collection module, and an underwater distribution unit hydraulic system information collection module. The underwater control unit condition monitoring subsystem includes a pressure sensor group, a flow sensor group, a temperature sensor group, a vibration sensor group, an underwater control unit electronic system information collection module, and an underwater control unit hydraulic system information collection module; The underwater valve group status monitoring subsystem includes a pressure sensor group, a flow sensor group, a temperature sensor group, a vibration sensor group, and an underwater valve group hydraulic system information acquisition module; The subsea wellhead mechanical module condition monitoring subsystem includes a pressure sensor group, a temperature sensor group, a vibration sensor group, a corrosion monitoring sensor group, and a subsea wellhead mechanical system degradation information collection module; The data collection and analysis subsystem includes a communication module, a signal conversion module, a probability information generation and update module, a competing failure probability algorithm module, and a reliability assessment module; it can calculate the failure probability of the subsea production tree system and perform real-time reliability assessment of the system.

2. The real-time reliability assessment method for dependent competing failure of subsea production trees according to claim 1, characterized in that, The pressure sensor group, temperature sensor group, and corrosion monitoring sensor group of the subsea production tree mechanical unit condition monitoring subsystem are respectively attached to the production tree cap and the production tree body to collect pressure status, temperature changes, and corrosion amount; the vibration sensor group is respectively attached to the outside of the production tree cap and the production tree body to detect external impact information of the subsea production tree mechanical system; the subsea production tree mechanical system degradation information collection module is used to integrate the degradation information of the subsea production tree mechanical unit.

3. The real-time reliability assessment method for dependent competing failure of subsea production trees according to claim 1, characterized in that, The pressure sensor group, flow sensor group, and temperature sensor group of the underwater distribution unit status monitoring subsystem are used to collect the pressure status, flow rate changes, and temperature changes of the hydraulic distribution module and electronic distribution module of the underwater distribution unit; the vibration sensor group is used to detect external impact information of the hydraulic distribution module and electronic distribution module of the underwater distribution unit; the underwater distribution unit electronic system information collection module is used to collect degradation information of the underwater distribution unit electronic system, and the underwater distribution unit hydraulic system information collection module is used to collect degradation information of the underwater distribution unit hydraulic system.

4. The real-time reliability assessment method for dependent competing failure of subsea production trees according to claim 1, characterized in that, The pressure sensor group, flow sensor group, and temperature sensor group of the underwater control unit status monitoring subsystem are used to collect the pressure status, flow rate changes, and temperature changes of each component of the underwater control unit; the vibration sensor group is used to detect external impact information of the underwater control module; the underwater control unit electronic system information collection module is used to collect the total degradation information of the underwater control unit electronic system, and the underwater control unit hydraulic system information collection module is used to collect the total degradation information of the underwater control unit hydraulic system.

5. The real-time reliability assessment method for dependent competing failure of subsea production trees according to claim 1, characterized in that, The pressure sensor group, flow sensor group, and temperature sensor group of the underwater valve group status monitoring subsystem are used to collect the pressure status, flow rate changes, and temperature changes of the annular wing valve, chemical injection valve, annular main valve, production main valve, switching valve, and production wing valve; the vibration sensor group is used to detect external impact information of the underwater valve group; and the underwater valve group hydraulic system information collection module is used to collect the total degradation information of the underwater valve group hydraulic system.

6. The real-time reliability assessment method for dependent competing failure of subsea production trees according to claim 1, characterized in that, The pressure sensor group, temperature sensor group, and corrosion monitoring sensor group of the subsea wellhead mechanical module status monitoring subsystem are used to collect the pressure status, temperature changes, and corrosion amount of the wellhead connector, wellhead, and downhole safety valve; the vibration sensor group is used to detect external impact information of the wellhead mechanical system; and the wellhead mechanical system degradation information collection module is used to collect the total degradation information of the wellhead mechanical system.

7. The real-time reliability assessment method for dependent competing failure of subsea production trees according to claim 1, characterized in that, During operation, the communication module of the data collection and analysis subsystem receives data from the following modules: the underwater tree mechanical system degradation information collection module, the underwater distribution unit electronic system information collection module, the underwater distribution unit hydraulic system information collection module, the underwater control unit electronic system information collection module, the underwater control unit hydraulic information collection module, the underwater valve group hydraulic information collection module, and the underwater tree wellhead mechanical system degradation information collection module. The signal conversion module converts the received signals into electrical signals. The probability information generation and update module reads the electrical signals converted by the signal conversion module and uses these signals for probability information generation, tracking, and updating. The competitive failure probability algorithm module aggregates the probability information from each component to generate the probability of failure for the underwater tree system. The reliability assessment module reads the probability information generated by the underwater production tree competition failure probability algorithm module and performs a reliability assessment.

Images