A hydrate blockage risk digital twin early warning method and system for a subsea pipeline coupled with a choke

Abstract

This invention relates to the field of subsea oil and gas gathering and transportation and flow assurance, and discloses a digital twin early warning method and system for hydrate blockage risk in subsea pipelines coupled with nozzles. The method includes: acquiring online monitoring and engineering basic data, preprocessing to form an observation sequence; performing multi-condition continuous state identification based on the observation sequence, establishing or updating the digital twin case structure based on the results of continuous state identification, and performing adaptive time window planning and hierarchical solution scheduling; establishing the coupling boundary between the throttling element and the subsea pipeline and calculating the throttling mass flow rate and throttling temperature drop, using the calculation results as pipeline inlet boundary conditions; executing multiphase transient solutions based on the inlet boundary conditions and hierarchical solution scheduling, inheriting state variables between different solution kernels; calculating the blockage probability and blockage segment events, comparing the blockage probability, severity, or expected blockage time with a threshold and outputting early warning information, and adaptively adjusting subsequent time window planning based on the early warning information to form a closed-loop scheduling.

Description

Technical Field

[0001] This invention relates to the field of subsea oil and gas gathering and transportation and flow assurance technology, and in particular to a digital twin early warning method and system for the risk of hydrate blockage in subsea pipelines coupled with oil nozzles. Background Technology

[0002] When subsea pipelines transport multiphase fluids in deep-water, low-temperature, and high-pressure environments, hydrates are prone to form and deposit, leading to a reduction in the effective flow area within the pipeline and even blockage. Especially under critical operating conditions such as stable production, shutdown, and restart, temperature and pressure change rapidly over time, and the hydrate formation / decomposition and deposition / resuspension processes exhibit significant transient characteristics.

[0003] Existing hydrate risk assessments typically suffer from the following shortcomings: (1) Operating condition identification is mostly based on discrete threshold judgments, making it difficult to stably identify and provide confidence levels for continuous states such as valve closure ramps, initial transients during shutdown, and restart steps, and is easily affected by noise and sampling intervals; (2) Multi-scale solver switching relies on manually preset time periods or fixed windows, which can easily lead to numerical abrupt changes during ramp switching, and it is difficult to balance real-time performance and accuracy; (3) Probabilistic risk representations often lack closed-loop linkage with solver scheduling, and the probability of blockage is low. (3) Failure to effectively drive time window encryption or solver upgrades resulted in insufficient computational resources during high-risk periods; (4) The computational links for stable production, shutdown and restart were fragmented or lacked inter-stage state inheritance, resulting in inconsistent initial restart conditions; (5) Inlet throttling elements (oil nozzles / orifice plates / valve) and subsea pipelines were often calculated separately, ignoring the impact of temperature, pressure and phase changes caused by throttling on hydrate risk. Summary of the Invention

[0004] To address the aforementioned problems, the purpose of this invention is to provide a digital twin early warning method and system for hydrate blockage risk in subsea pipelines coupled with nozzles, employing "continuous state adaptive identification—adaptive window selection hierarchical solution—probabilistic risk" approach. The core closed loop is "Logic and Event-based - Closed-loop Early Warning Output". When there is a throttling element at the inlet, the oil nozzle throttling coupling is used as one of the boundary conditions to realize the continuous evolution prediction and online early warning of risks over time and space under multiple operating conditions such as stable production, shutdown and restart.

[0005] To achieve the above objectives, in a first aspect, the technical solution adopted by the present invention is as follows: a digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with an oil nozzle, comprising: acquiring online monitoring data and engineering basic data, and preprocessing the acquired data to form an observation sequence; performing multi-condition continuous state identification based on the observation sequence to determine the current and predicted operating conditions and outputting the state probability or confidence level and the change point time; establishing or updating the digital twin case structure according to the results of continuous state identification, and performing adaptive time window planning and hierarchical solution scheduling; when a throttling element exists at the inlet, monitoring the oil nozzle and heat transfer... The system updates friction and hydrate dynamic parameters online, establishes the coupling boundary between the throttling element and the subsea pipeline, and calculates the throttling mass flow rate and throttling temperature drop. The calculation results are used as the pipeline inlet boundary conditions. Within the time window, multiphase transient solutions are executed based on the inlet boundary conditions and hierarchical solution scheduling, and state variables are inherited between different solution kernels. The system couples hydrate equilibrium temperature, dynamics, and deposition feedback to calculate the blockage probability and blockage segment events. The system compares the blockage probability, severity, or expected blockage time with the threshold and outputs early warning information. The system adaptively adjusts the subsequent time window planning based on the early warning information to form a closed-loop scheduling.

[0006] Furthermore, multi-condition continuous state identification includes: Construct an observation feature vector within a sliding window. The observation feature vector includes inlet flow rate or production, nozzle or valve opening, the difference between upstream inlet pressure and outlet back pressure or equivalent pressure difference, and the difference terms of each parameter. Change point detection is performed based on feature vectors to identify change points in the observation sequence; Based on the state transition matrix, the state probability is recursively calculated to determine the posterior probability of the current state belonging to each working condition, and the state probability or confidence level is output. The state output is constrained by a combination of dwell time and hysteresis threshold: entering the target state requires the state probability to exceed the entry threshold and to be maintained for a first preset number of sampling periods; exiting the current state requires the state probability to be lower than the exit threshold and to be maintained for a second preset number of sampling periods, so as to suppress state jitter.

[0007] Furthermore, establish or update the digital twin instance structure, and perform adaptive time window planning and hierarchical solution scheduling, including: A unified computational example structure is constructed, which includes pipeline geometry and mesh, environmental boundaries, fluid properties, stable production, shutdown and restart scheduling, throttling elements and model parameters. The operating condition stage information in the computational example structure is updated based on the continuous state identification results. Adaptive time window planning includes: using the stage boundary, the end time of the ramp segment, the change point time, and the risk trigger time as candidate window boundaries; when the candidate boundary falls into the ramp segment, it will be aligned to the end time of the ramp; and the window length will be adjusted segment by segment according to the maximum congestion probability within the window, where the higher the maximum congestion probability, the shorter the window length. The hierarchical solution scheduling includes: using the first solution kernel and the first time step for the gradual change phase; using the second solution kernel and the second time step for the valve closing ramp, restart ramp or risk trigger phase, the second time step is smaller than the first time step, and inheriting the pressure, temperature, liquid holdup and hydrate deposition state along the process when switching solution kernels.

[0008] Furthermore, the coupling boundary between the throttling element and the subsea pipeline is established, and the throttling mass flow rate and throttling temperature drop are calculated. The calculation results are used as the pipeline inlet boundary conditions, including: Obtain the equivalent orifice diameter, flow coefficient, and opening degree of the throttling element; Obtain upstream pressure, upstream temperature, and downstream inlet pressure, and calculate the gas phase volume fraction based on fluid properties; Based on the upstream pressure, downstream inlet pressure, and gas phase volume fraction, the appropriate model is adaptively selected from the single-phase liquid incompressible orifice plate model, the single-phase gas compressible isentropic critical flow model, the two-phase homogeneous HEM model, and the API-14B orifice plate model to calculate the throttling mass flow rate. Among them, for gas or two-phase flow, when the pressure ratio is lower than the critical pressure ratio, the critical flow upper limit is used to limit the mass flow rate. In the two-phase homogeneous HEM model, the Wood sound velocity approximates the mixed phase sound velocity to determine the critical flow upper limit. The throttling temperature drop is calculated based on the isenthalpic or isentropic model to obtain the temperature after throttling; The throttled mass flow rate is used as the boundary of the pipeline inlet mass flow rate, and the temperature after throttling is used as the boundary of the pipeline inlet temperature.

[0009] Furthermore, based on the inlet boundary conditions and hierarchical solution scheduling, multiphase transient solutions are executed, including: Within the time window, the corresponding solution kernel is selected based on the hierarchical solution scheduling to execute the transient solution: When in the gradual change phase, the first solution kernel is used for solving. The first solution kernel uses quasi-steady-state hydraulics as the framework and couples a radial RC thermal resistance-heat capacity network to solve the transient energy equation to simulate the thermal inertia of the pipe wall, insulation layer and external medium. The radial RC thermal resistance-heat capacity network abstracts the pipe wall, insulation layer and external medium into series and parallel thermal resistance and distributed heat capacity, and updates the heat transfer flux between the pipe wall and the fluid through time stepping. When in the valve-closing ramp, restart ramp, or risk-triggered phase, a second solution kernel is used for solving. The second solution kernel uses a fully implicit two-fluid conservation equation discretization and solves the mass, momentum, and energy conservation equations through backward Euler discretization and Newton iteration to capture compressibility and pressure wave propagation. When switching between different solution kernels and time windows, the pressure, temperature, liquid holdup, and hydrate deposition state parameters along the process are inherited to ensure numerical continuity.

[0010] Furthermore, the probability of clogging and clogging events are calculated by coupling hydrate equilibrium temperature, kinetics, and deposition feedback, including: Based on the current pressure and gas composition, and combined with salinity correction and inhibitor concentration correction, the hydrate equilibrium temperature is calculated. The degree of supercooling or superheating is determined based on the difference between the fluid temperature and the hydrate equilibrium temperature. The mass transfer coefficient at the gas-liquid interface is the main control factor. The hydrate formation term is represented by the empirical mapping of the degree of supercooling. The heat transfer limitation at the solid-liquid interface is the main control factor. The hydrate decomposition term is represented by the empirical mapping of the degree of superheating. The rate of change of hydrate volume fraction is output. A two-pool exchange model of slurry pool and sediment pool is used to update the hydrate deposition state: sediment pool growth is controlled by the sedimentation coefficient and velocity inhibition scale, and resuspension is controlled by the resuspension coefficient and velocity enhancement scale; the effective inner diameter is calculated based on the sediment volume fraction in the sediment pool, and the effective inner diameter is fed back into the pressure drop calculation and heat transfer calculation; Based on the uncertainty of the sediment volume fraction, the basic clogging probability is calculated using either the transcendental probability calculation or the Monte Carlo uncertainty propagation. The clogging probability is then obtained by adaptive correction based on the operating conditions, such as flow velocity, liquid holdup, tilt angle, and effective inner diameter. Using the deposition volume fraction exceeding the first threshold and its growth rate exceeding the second threshold as the criteria for blockage formation, adjacent trigger grids are merged, and the location, length, formation time, and severity of the blockage are output.

[0011] Furthermore, it also includes the step of restarting the feasibility assessment: based on production baseline pressure reduction. For reference, the pressure drop required for restart is estimated based on the cross-sectional area ratio. And predict pressure drop With available pressure differential In comparison, if the predicted pressure drop Exceeding available pressure differential If so, a restart risk warning will be output.

[0012] Secondly, the technical solution adopted by this invention is: a digital twin early warning system for the risk of hydrate blockage in subsea pipelines coupled with an oil nozzle, comprising: The data preprocessing module acquires online monitoring data and basic engineering data, and preprocesses the acquired data to form an observation sequence. The continuous state identification module performs multi-condition continuous state identification based on the observation sequence to determine the current and predicted working condition stages and output the state probability or confidence level and the time of change. The time window planning and hierarchical solution scheduling module establishes or updates the digital twin case structure based on the results of continuous state identification, and performs adaptive time window planning and hierarchical solution scheduling. The nozzle throttling coupling module updates the dynamic parameters of the nozzle, heat transfer, friction and hydrate online when a throttling element is present at the inlet. It also establishes the coupling boundary between the throttling element and the subsea pipeline and calculates the throttling mass flow rate and throttling temperature drop. The calculation results are used as the pipeline inlet boundary conditions. The multiphase transient solution module executes multiphase transient solutions within a time window based on the entry boundary conditions and hierarchical solution scheduling, and inherits state variables between different solution kernels; The blockage probability assessment and event-based module couples hydrate equilibrium temperature, kinetics, and deposition feedback to calculate blockage probability and blockage segment events. The early warning output module compares the congestion probability, severity, or expected congestion time with a threshold and outputs early warning information. It also adaptively adjusts subsequent time window planning based on the early warning information to form a closed-loop scheduling.

[0013] Thirdly, the technical solution adopted by the present invention is: a computer-readable storage medium for storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any of the methods described above.

[0014] Fourthly, the technical solution adopted by the present invention is: a computing device comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for performing any of the methods described above.

[0015] The present invention has the following advantages due to the adoption of the above technical solutions: 1. This invention directly drives time window planning and solver hierarchical scheduling with continuous state recognition results, rather than using fixed windows or manual segmentation. This invention employs multi-condition continuous state adaptive recognition: it stably identifies continuous states such as valve closing ramp, initial transient state during shutdown, and step-by-step restart maintenance, and outputs state probabilities / confidence levels, reducing misjudgments caused by noise and sampling intervals.

[0016] 2. This invention explicitly couples the throttling mass flow rate and throttling temperature drop at the inlet nozzle as the subsea pipeline inlet boundary and updates it online; it automatically plans and schedules time windows based on the identification results and risk triggers. The solver aligns the window boundaries with the end of the ramp to avoid numerical abrupt changes caused by switching in the ramp segment, balancing real-time performance and accuracy, and achieving adaptive window selection and layered solving.

[0017] 3. This invention directly uses the Pb obtained from explicit uncertainty quantization for window refinement and solver upgrade, forming a probability-driven closed loop; it employs explicit uncertainty quantization and adaptive working condition correction to construct the blockage probability. and use Triggering time window encryption and solver upgrades enables adaptive matching of computing resources and risk sensitivity.

[0018] 4. This invention supports executable early warning decisions by providing event-based output of congestion segments (location, length, formation time, severity, TTB). It outputs a list of blocked sections, severity, and estimated blockage time in conjunction with indicators such as deposition growth rate and pressure drop gradient, and supports restart feasibility assessment, which facilitates on-site prioritization and decision-making.

[0019] 5. This invention inherits P, T, HL and hydrate deposition states during layered solution and working condition switching, avoiding inconsistencies in initial values ​​caused by broken links, and making risk evolution traceable.

[0020] 6. This invention explicitly calculates the throttling mass flow rate and temperature drop when there is an oil nozzle / orifice plate / valve at the inlet, and continuously updates the oil nozzle, heat transfer, friction and dynamic parameters by combining online sensor residuals, thereby improving adaptability. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall architecture of the digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with oil nozzles in this embodiment of the invention; Figure 2 This is a continuous state recognition—adaptive window selection and hierarchical solution in the embodiments of the present invention. A flowchart illustrating the event-driven digital twin early warning method; Figure 3 This is a schematic diagram of multi-condition continuous state recognition and state machine (including sub-states) in an embodiment of the present invention; Figure 4 This is a schematic diagram of the throttling coupling boundary of the inlet nozzle in an embodiment of the present invention (including the calculation of critical flow and throttling temperature drop). Figure 5 This is a schematic diagram of the adaptive window selection layered solution strategy in an embodiment of the present invention (continuous state recognition + (Triggering, window alignment, and state inheritance). Figure 6 This is a schematic diagram illustrating the evolution and effective inner diameter feedback of the two hydrate reservoirs (sediment reservoir / slurry reservoir) in an embodiment of the present invention; Figure 7This is the blocking probability in the embodiments of the present invention. Schematic diagram of spatiotemporal distribution and early warning threshold; Figure 8 This is a schematic diagram showing the output of the blockage event list, severity, and restart feasibility assessment in an embodiment of the present invention; Figure 9 This is the explicit uncertainty quantification and blocking probability in the embodiments of the present invention. Mapping diagram ( (Monte Carlo and Operating Condition Correction); Figure 10 This is a schematic diagram of the blockage segment detection, event merging, and output fields in an embodiment of the present invention; Figure 11 This is a schematic diagram of the radial RC heat transfer network and thermal resistance-heat capacity modeling in an embodiment of the present invention; Figure 12 The input working conditions in Embodiment 1 of this invention are , Schematic diagram of the evolution curve over time; Figure 13 This is a schematic diagram of the risk comparison curve between Example 2 and Example 1 under the inhibitor-off condition.

[0022] Figure label: 1- Online data source; 2- Data interface and preprocessing module; 3- Continuous state identification module; 4- Parameter assimilation module; 5- Nozzle throttling coupling module; 6- Time window planning and hierarchical solution scheduling module; 7- Hydrate model module; 8- Logic and event-driven modules; 9-Early warning output module. Detailed Implementation

[0023] To address the shortcomings of existing hydrate risk assessment methods, this invention provides a digital twin early warning method and system for hydrate blockage risk in subsea pipelines coupled with nozzles. This invention focuses on online monitoring data, firstly performing time alignment and quality control on inlet / outlet pressure, temperature, flow rate or production, nozzle opening, inhibitor injection volume, and environmental parameters. Then, based on multi-feature sequences, it adaptively identifies continuous states under multiple operating conditions, including stable production, shutdown, valve closure ramp, initial transient shutdown, restart ramp, and step-by-step restart maintenance, outputting state probabilities / confidence levels and change point times. Based on this, it automatically plans sliding calculation time windows and performs hierarchical solution scheduling: long-term windows employ rapid... The kernel and short-time sensitive windows use full implicit mode. The kernel follows the constraint of "aligning window boundaries with the end of the ramp and avoiding switching within the ramp," inheriting P, T, HL, and other parameters between different kernels. , Isostatic quantities; further coupled with hydrate equilibrium temperature Dynamics and two-reservoir sedimentation feedback were used to obtain state evolution; a blockage probability based on explicit uncertainty quantification and adaptive correction of operating conditions was constructed. and will Indicators such as deposition rate and pressure drop gradient are converted into a list of blocked segments, severity, and expected blockage time; when Alternatively, when the expected congestion time triggers a threshold, an early warning is generated, which in turn drives the encryption of the time window and the upgrade of the solver, realizing a closed-loop digital twin early warning system for continuous evolution across multiple operating conditions.

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0025] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0026] In one embodiment of the present invention, a digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with nozzles is provided. This method is designed for online monitoring and early warning of hydrate blockage risk in subsea pipelines and encompasses multiphase flow and heat transfer transient simulation, inlet nozzle throttling coupling, hydrate phase equilibrium and dynamics, sedimentation feedback, probabilistic risk assessment, and early warning decision-making. Inputs include: online monitoring data (inlet / outlet pressure P, temperature T, flow rate Q or production, nozzle opening, inhibitor injection amount, etc.), engineering basic data (pipeline length L, inner and outer diameter D, insulation layer parameters, centerline profile / burial depth, seawater depth and seabed temperature, etc.), and fluid property parameters (Black-Oil correlation or tabular PVT). In this embodiment, a unified computational structure c is used to describe the engineering object, including at least: c.pipeline (geometry and mesh), c.environment (environmental boundary), c.fluid and c.blackoil (physical properties and phases), c.operation (boundary conditions), c.schedule (stable production-stop-restart scheduling), c.components (inlet nozzles / pipeline throttling elements), and c.model (model switches and parameters). The system output includes: , , , , , , , And congestion events and early warning information.

[0027] Specifically, in this embodiment, the operating condition identification is upgraded from "single-point threshold discrimination" to "continuous state adaptive identification of multiple feature sequences." Feature vectors are constructed for inlet / outlet pressure, temperature, flow rate or output, opening degree and its rate of change, etc., and combined with change point detection (e.g., CUSUM / sliding difference / Bayesian change point) and residence time / hysteresis constraints, outputting discrete state labels and state probabilities (or confidence levels) to suppress frequent switching caused by noise. Figures 1 to 11 As shown, the method includes the following steps: 1) Acquire online monitoring data and basic engineering data, and preprocess the acquired data to form an observation sequence.

[0028] 2) Perform continuous state identification for multiple operating conditions based on the observation sequence to determine the current and predicted operating conditions and output the state probability or confidence level and the time of change.

[0029] 3) Based on the results of continuous state identification, establish or update the digital twin case structure, and perform adaptive time window planning and hierarchical solution scheduling.

[0030] 4) When a throttling element is present at the inlet, the dynamic parameters of the nozzle, heat transfer, friction and hydrate are updated online, and the coupling boundary between the throttling element and the subsea pipeline is established. The throttling mass flow rate and throttling temperature drop are calculated, and the calculation results are used as the pipeline inlet boundary conditions.

[0031] 5) Within the time window, multiphase transient solutions are executed based on the entry boundary conditions and hierarchical solution scheduling, and state variables are inherited between different solution kernels.

[0032] 6) Couple hydrate equilibrium temperature, kinetics and deposition feedback to calculate clogging probability and clogging segment events.

[0033] 7) After comparing the congestion probability, severity, or expected congestion time with the threshold, output early warning information, and adaptively adjust the subsequent time window planning based on the early warning information to form a closed-loop scheduling.

[0034] In step 1) above, the data is time-aligned, unit-converted, outlier-handled, and missing data filled to form an observation sequence. Outlier-handling includes at least: setting physical upper and lower limits for pressure, temperature, and flow rate; setting slope thresholds for the difference between adjacent sampling points; and using median filtering or robust regression to suppress spike noise.

[0035] In step 2) above, the multi-condition continuous state identification includes the following steps: 2.1) In the sliding window Internal structure observation eigenvectors The observed feature vectors include inlet flow rate or output Q, nozzle or valve opening. The difference between the upstream pressure at the inlet and the back pressure at the outlet, or the equivalent pressure difference. and the difference terms of the aforementioned parameters. , , Where the subscript i represents the discrete sampling time (or sampling sequence number); This represents the average inlet flow rate or output Q within the sliding window; This represents the variance of Q within that window; The slope of the trend of Q changing with time within this window should be obtained by least squares linear fitting.

[0036] In this embodiment, the differential term employs robust differential and performs interpolation to complete the missing measurement points, thereby enhancing its adaptability to changes in noise and sampling interval.

[0037] 2.2) Detect change points based on feature vectors to identify change points in the observation sequence.

[0038] 2.3) Based on the state transition matrix A, perform state probability recursion, calculate the posterior probability of belonging to each working condition at the current time, and output the state probability or confidence level.

[0039] In this embodiment, to obtain an interpretable confidence level, the continuous state is denoted as... And based on the transition matrix A, the state probabilities are recursively calculated (example expression): ; in, Let be the posterior probability that time i belongs to state k; Let be the likelihood function of state k; S represents a predefined set of continuous operating conditions. This represents the observation sequence from the first sampling time to the i-th sampling time; Let represent the posterior probability of being in state j at time i-1.

[0040] 2.4) Constrain the state output by combining dwell time and hysteresis threshold: entering the target state requires the state probability to exceed the entry threshold and to be maintained for a first preset number of sampling periods; exiting the current state requires the state probability to be lower than the exit threshold and to be maintained for a second preset number of sampling periods, so as to suppress state jitter.

[0041] In this embodiment, both dwell and hysteresis constraints are introduced: entry condition And continuously satisfy Number of (first preset number) sampling periods; Exit condition And continuously satisfy If there are 1 (second preset number) sampling periods, the output state probability is: (1) In the formula, Let be the posterior probability that time i belongs to state k. Let A be the likelihood function under state k, and A be the state transition matrix. This is the normalization coefficient.

[0042] In step 3) above, if Figure 2 As shown, establishing or updating the digital twin instance structure and performing adaptive time window planning and hierarchical solution scheduling includes the following steps: 3.1) Construct a unified computational example structure that includes pipeline geometry and mesh, environmental boundaries, fluid properties, stable production, shutdown, restart scheduling, throttling elements, and model parameters. Update the operating condition stage information in the computational example structure based on the continuous state identification results, such as... Figure 3 As shown.

[0043] In this embodiment, the continuous status tag includes at least: stable production. ; Stop the service (Valve closing ramp) (Short-term transient state after valve closure) (Stop transmission and slow change); Restart (Step-by-step slope and maintenance) and (End-of-phase retention). The identified state boundaries are used as the phase transition times of the work condition scheduling schedule, and are updated continuously within a sliding time window.

[0044] 3.2) Adaptive time window planning includes: using stage boundaries, ramp end times, change point times, and risk trigger times as candidate window boundaries, and automatically selecting long windows (for gradual change stages such as shutdown and cooling) and short windows (for sensitive stages such as valve closure / restart). When a candidate boundary falls into a ramp segment, it will be aligned to the ramp end time, and the window length will be adjusted segment by segment according to the maximum congestion probability within the window, with a shorter window length for higher maximum congestion probability.

[0045] In this embodiment, when the candidate window boundary falls into the middle of the slope segment, the window boundary is aligned to the end time of the slope segment to avoid numerical abrupt changes caused by switching the solver or switching the time step during the slope process.

[0046] The time window length can be adaptively determined by risk: maximizing the probability of congestion within the window. = Then the window length Set according to piecewise function: when Time to take , This indicates the maximum time window length used in a low-risk state. This represents the threshold value under the probability of congestion; when Time to take , This indicates the minimum time window length used for high-risk states. This represents a threshold indicating the probability of congestion, and Linear interpolation for the remaining intervals, combined with continuous-state confidence. By weighting and shortening the window for sensitive states, we can achieve adaptive resource allocation that enables "high risk / high uncertainty → shorter window → stronger transient solution".

[0047] Window length for: (2) in, This represents the probability of blockage at axial position x at time t, where x represents the axial position coordinate of the pipe and t represents time.

[0048] The clip function in equation (2) restricts the linear interpolation result to the interval [0,1], which is equivalent to a piecewise definition and facilitates engineering implementation.

[0049] Window boundary alignment rules can be formalized as: if candidate boundaries satisfy And the i-th stage is ( This represents a slope stage where the boundary conditions change continuously according to a preset slope (e.g., the stage where the nozzle opening, flow rate, or pressure gradually transitions from one steady-state value to another). Let... This ensures that the solver switching and time step switching occur at the end of the ramp segment, avoiding forced switching that introduces numerical pseudo-oscillations while the boundary conditions are still changing.

[0050] (3) To ensure continuous evolution across multiple operating conditions, the system inherits key state variables (P, T, liquid holdup HL, and hydrate deposition state) during window switching and solver switching. With effective inner diameter (etc.), to avoid errors caused by simplifying the initial conditions for restart to uniform initial values, and to make the evolution of risks traceable over time.

[0051] 3.3) Layered solution scheduling includes: using the first solution kernel and the first time step for the gradual change phase; using the second solution kernel and the second time step for the valve closing ramp, restart ramp or risk trigger phase, the second time step is smaller than the first time step, and inheriting the pressure, temperature, liquid holdup and hydrate deposition state along the process when switching solution kernels.

[0052] In this embodiment, as Figure 5 As shown, within each sliding time window, a hierarchical solution is performed based on the continuous state identification results and risk triggering conditions: for gradually changing stages such as shutdown and cooling, a hierarchical solution is selected. A kernel (first solver kernel) is used to improve computational efficiency; for valve shut-off / restart ramps, pressure fluctuations, or... Selecting sensitive stages such as trigger thresholds A kernel (second solver kernel) is used to improve transient accuracy; and state inheritance is performed between the two types of kernels to maintain numerical continuity.

[0053] In step 4) above, the coupling boundary between the throttling element and the subsea pipeline is established, and the throttling mass flow rate and throttling temperature drop are calculated. The calculation results are used as the pipeline inlet boundary conditions, such as... Figure 4 As shown, it includes the following steps: 4.1) Obtain the equivalent orifice diameter, flow coefficient, and opening degree of the throttling element.

[0054] In this embodiment, the inlet nozzle is described by components, which includes at least: equivalent orifice diameter d, flow coefficient. (or ), opening degree.

[0055] 4.2) Obtaining upstream pressure upstream temperature and downstream inlet pressure And calculate the gas phase volume fraction based on fluid properties. ; 4.3) Based on upstream pressure Downstream inlet pressure and gas phase volume fraction The appropriate model is adaptively selected from the single-phase liquid incompressible orifice plate model, the single-phase gas compressible isentropic critical flow model, the two-phase homogeneous HEM model, and the API-14B orifice plate model to calculate the throttling mass flow rate. For gas or two-phase flow, when the pressure ratio is lower than the critical pressure ratio, the upper limit of the critical flow is used to limit the mass flow rate. In the two-phase homogeneous HEM model, the Wood sound velocity approximates the mixed-phase sound velocity to determine the upper limit of the critical flow.

[0056] Specifically, in the single-phase liquid incompressible model, the mass flow rate expression is shown in equation (4), where A is the effective throttling area. For effective pressure difference.

[0057]

[0058]

[0059] (4) In the formula, The throttling mass flow rate is calculated using a single-phase liquid incompressible model. Indicates the density of the liquid phase; Indicates the effective flow diameter of the nozzle (or orifice), and is determined by... Calculate the effective area of ​​throttling.

[0060] In the single-phase gas isentropic model, the critical pressure ratio is given by equation (5), and the upper limit of the critical mass flow rate is given by equation (6). When the pressure ratio is lower than the critical pressure ratio, the flow enters the critical flow. When the flow is not critical, the solution is obtained according to the isentropic nozzle relationship.

[0061] (5) (6) In the formula, This represents the critical pressure ratio at which gas undergoes critical flow. Indicates the specific heat capacity ratio of a gas (adiabatic index); The throttling mass flow rate calculated using a single-phase gas isentropic model; This indicates the gas phase density under upstream operating conditions of the nozzle.

[0062] In the two-phase HEM model, the mixed-phase sound velocity is calculated by the Wood approximation (see Equation (7)) and the mass flow rate is limited by the critical flow upper limit (see Equation (8)).

[0063] (7) (8) In the formula, This represents the density of a homogeneous gas-liquid mixture; Indicates the speed of sound in a mixture; Indicates the gas phase volume fraction. Indicates the liquid volume fraction; Indicates the speed of sound in the gas phase. Represents the square of the speed of sound in the gas phase; Represents the speed of sound in the liquid phase. This represents the square of the speed of sound in the liquid phase.

[0064] 4.4) To demonstrate the impact of throttling temperature drop on hydrate risk, this invention provides a throttling temperature drop calculation: When energy coupling is enabled, the throttling temperature drop is calculated based on an isenthalpic (rapid isenthalpic) or isentropic model to obtain the temperature after throttling. .

[0065] 4.5) The throttling mass flow rate is used as the pipeline inlet mass flow rate boundary, and the temperature after throttling is used as the pipeline inlet temperature boundary, in order to participate in subsequent heat transfer and hydrate calculations.

[0066] In step 5) above, the multiphase transient solution is executed according to the inlet boundary conditions and the hierarchical solution schedule, including the following steps: 5.1) Within the time window, the corresponding solution kernel is selected according to the hierarchical solution scheduling to perform transient solution.

[0067] In this embodiment, the pipeline transient solver supports two types of kernels: long-term thermal kernel (first solver kernel) and strong transient kernel. Kernel (Second Solver Kernel).

[0068] 5.2) When in the gradual change phase, the first solution kernel is used for solving. The first solution kernel uses quasi-steady-state hydraulics as the framework and couples a radial RC thermal resistance-heat capacity network to solve the transient energy equation to simulate the thermal inertia of the pipe wall, insulation layer and external medium. The radial RC thermal resistance-heat capacity network abstracts the pipe wall, insulation layer and external medium into series and parallel thermal resistance and distributed heat capacity, and updates the heat transfer flux between the pipe wall and the fluid through time step.

[0069] Specifically, using quasi-steady-state hydraulics as the framework, the transient energy equation is solved, and a radial RC thermal resistance-thermal capacity network is coupled to simulate the thermal inertia of the pipe wall / insulation / soil, which is used for shutdown cooling and long-term stable production evolution.

[0070] 5.3) When in the valve closing ramp, restart ramp or risk triggering stage, the second solution kernel is used for solving. The second solution kernel uses fully implicit two-fluid conservation equation discretization and solves the mass, momentum and energy conservation equations through backward Euler discretization and Newton iteration to capture compressibility and pressure wave propagation.

[0071] Specifically, a fully implicit two-fluid mass / momentum / energy equation (backward Euler BE discretization + Newton iteration) is used for the compressibility and pressure wave / boundary strong coupling window during the restart phase.

[0072] 5.4) When switching between different solution kernels and time windows, inherit the pressure, temperature, liquid holdup and hydrate deposition state along the process to ensure numerical continuity.

[0073] Specifically, such as Figure 11As shown, in radial RC heat transfer, the system abstracts the pipe wall, insulation layer and external medium (seawater or soil) into a series and parallel thermal resistance and distributed heat capacity network. The heat transfer flux between the pipe wall and the fluid is updated by time step, thereby obtaining a shutdown cooling time constant that is more consistent with the field.

[0074] In a radial RC thermal resistance-heat capacity network, the typical thermal resistance expression is shown in equation (9); the internal convection heat transfer correlation is shown in equation (10), which is used to calculate the internal wall convection heat transfer coefficient.

[0075]

[0076]

[0077] (9) (10) In the formula, This represents the convective thermal resistance of the fluid to the inner wall of the pipe; Indicates the convective heat transfer coefficient inside the pipe; This indicates the effective inner diameter after deposition; Indicates the radial thermal resistance of the pipe wall; Indicates the outer diameter of the pipe wall; Indicates the inner diameter of the pipe wall; Indicates the thermal conductivity of the pipe wall; This indicates the total thermal resistance from the outside of the insulation layer to the external environment; Indicates the thermal resistance of the insulation layer; Indicates the thermal resistance of the heat transfer on the external medium side of the insulation layer; represents the Nusselt number; Pr represents the Prandtl number; Re represents the Reynolds number.

[0078] When using backward Euler (BE) discretization, the thermal balance expressions for solid nodes are shown in equations (11) and (12), and the heat loss term is substituted back into the energy equation to achieve implicit coupling.

[0079] (11) (12) In the formula, This represents the equivalent heat capacity of the tube wall nodes within the current axial element; Indicates the temperature of the pipe wall nodes; This indicates the discrete-time layer number, where n+1 is the next time layer; Indicates fluid temperature; This represents the convective thermal resistance between the fluid and the pipe wall; Indicates the temperature of the insulation layer nodes; This represents the equivalent heat capacity of the insulation layer node; This indicates the external ambient temperature.

[0080] The kernel adopts a fully implicit discretization of the two-fluid conservation equations, and the conservation forms of mass, momentum and energy are shown in equations (13) to (15), respectively; where the source term includes the interphase exchange term, wall friction and heat transfer term.

[0081] (13) (14) (15) In the formula, This represents the volume fraction of the k-th phase; This represents the density of the k-th phase; This represents the axial velocity of the k-th phase; This represents the mass source term of the k-th phase (including interphase mass transfer). This indicates shared pressure between the two phases; This represents the source term of the momentum equation for the k-th phase; This represents the source term of the energy equation for the k-th phase; This indicates the enthalpy of the k-th relative value.

[0082] In step 6) above, the calculation of clogging probability and clogging segment events by coupling hydrate equilibrium temperature, kinetics, and deposition feedback includes the following steps: 6.1) Calculate the hydrate equilibrium temperature based on the current pressure and gas composition, combined with salinity correction and inhibitor concentration correction.

[0083] In this embodiment, the hydrate equilibrium temperature Used to determine the driving force of hydrate formation. Hydrate equilibrium temperature. Determined by pressure P and gas composition, and further considering salinity and inhibitor correction: (1) Based on pure methane hydrate equilibrium data points (1) Interpolation / extrapolation; (2) Linear correction for components such as C2 / C3 / C4, CO2, and N2; (3) Empirical cooling correction for salinity; (4) Hammerschmidt-type empirical correction for thermodynamic inhibitors such as MEG / methanol, with inhibitor concentration expressed as mass fraction. enter.

[0084] 6.2) Determine the degree of supercooling or superheating based on the difference between the fluid temperature and the hydrate equilibrium temperature. The mass transfer coefficient at the gas-liquid interface is the main control factor. The empirical mapping of supercooling is used to represent the hydrate formation term. The heat transfer limitation at the solid-liquid interface is the main control factor. The empirical mapping of superheating is used to represent the hydrate decomposition term. Output the rate of change of hydrate volume fraction.

[0085] In this embodiment, the hydrate kinetics adopt a mass / heat transfer-constrained engineering model: the generation term is based on the gas-liquid interface mass transfer coefficient. The main controller is the supercooler. The empirical mapping indicates that the decomposition terms are mainly controlled by the limited heat transfer at the solid-liquid interface, driven by superheat. Indicates. Dynamic output. It is compatible with numerical solvers and can apply a scaling boost to the generated terms during the pause phase. It represents the gas-liquid interface area (specific interface area) per unit volume of a mixture. Indicates the mass transfer coefficient on the liquid side. The hydrate equilibrium temperature is represented by T under the current pressure and composition conditions, where T represents the fluid temperature. Indicates the volume fraction of hydrate. Represented in hours, representing the rate of change. The unit of time.

[0086] 6.3) The hydrate deposition state is updated using a two-pool exchange model of slurry pool and sediment pool: the growth of sediment pool is controlled by the sedimentation coefficient and the velocity inhibition scale, and the resuspension is controlled by the resuspension coefficient and the velocity enhancement scale; the effective inner diameter is calculated based on the sediment volume fraction in the sediment pool, and the effective inner diameter is fed back to the pressure drop calculation and heat transfer calculation.

[0087] In this embodiment, as Figure 6 As shown, the sedimentation feedback adopts a two-pool model: slurry pool. With sediment reservoir Sedimentary reservoir growth is determined by the sedimentation coefficient. With flow rate suppression scaling control, resuspension is determined by the resuspension coefficient. With velocity-enhanced scaling control; and effective inner diameter calculated based on the sediment pool. Typical form is = ,in The proportion of deposits that cause diameter reduction. This feedback is further reflected in friction, pressure drop, and heat transfer area, and can be superimposed on roughness growth feedback.

[0088] 6.4) Based on the uncertainty of the sediment volume fraction, the basic blockage probability is calculated using either transcendental probability calculation or Monte Carlo uncertainty propagation. The blockage probability is obtained by adaptively correcting the working conditions based on the flow rate, liquid holdup, tilt angle and effective inner diameter.

[0089] In this embodiment, as Figure 7 As shown, the probability of congestion It's not the solver state variables, but rather the post-processing risk mapping of the state variables. (Using...) Two types of uncertainty quantification with Monte Carlo methods: (1) Assuming sediment volume fraction There is log-normal uncertainty, given the mean With coefficient of variation Calculate the value exceeding the critical value Exceeding probability (2) Monte Carlo: On The input is sampled and propagated, and the conditional probability is calculated for each sample and the mean is taken.

[0090] exist In the equation, the parameters σ and μ are defined as shown in equation (16), and the expression for the exceedance probability is shown in equation (17). The basic probability is then multiplied by the operating condition correction term. The correction term for the flow stage considers at least the flow velocity V, liquid holdup HL, tilt angle θ, and effective inner diameter D. Different correction logics are used for the shutdown stage.

[0091] (16) (17) In a specific implementation, such as Figure 9 As shown, the operating condition correction terms adopt a piecewise analytical formula to balance interpretability and deployability: the correction factor in the flow stage is composed of the product of the velocity factor, liquid holdup factor, tilt angle factor, and pipe diameter factor; different segment weights for liquid holdup and tilt angle are used in the shutdown stage. The coefficients of each segment are obtained from the calibration of historical operating conditions in the field.

[0092] (18) In Monte Carlo simulations, the system sets a random seed to ensure reproducibility and maps the conditional probability of each sample to a piecewise S-shaped curve: when When Pb is in the low value range, it rises slowly. After exceeding the threshold, Pb rapidly approaches 1; subsequently, V / HL / θ / D corrections are applied to obtain the sample. ,final .

[0093] (19) 6.5) Using a deposition volume fraction exceeding a first threshold and its growth rate exceeding a second threshold as the criterion for clogging segment formation, adjacent trigger grids are merged, and the location, length, formation time, and severity of the clogging segment are output. Figure 8 As shown.

[0094] Specifically, congestion segment detection is used to... This translates sedimentary evolution into events that are readable in engineering terms. (Using "sedimentary volume fraction") Exceeding the first threshold and the growth rate "Exceeding the second threshold" is used as the formation criterion, and the trigger points of adjacent grids are merged into a blockage segment. The center position, length, formation time, and severity of the blockage segment are output. The severity can be based on... Threshold group and pressure drop gradient along the path Threshold groups are classified.

[0095] Engineered difference in time series Based on the spatial grid x, the deposition growth rate and pressure drop gradient are calculated according to equations (20) and (21), respectively; the first frame is set as To avoid false triggering, adjacent trigger grids are then grouped consecutively to obtain the length and center position of the blocked segment.

[0096] (20) (twenty one) The risk index and the expected congestion time are calculated according to equations (22) and (23), respectively.

[0097] (twenty two) (twenty three) In the formula, This represents the i-th discrete time point; Indicates the position of the j-th axis mesh node; Indicates the spacing between adjacent axial mesh nodes; Indicates time ,Location Pressure at the location; This represents the normalized congestion risk index, with values ​​limited to [0,1]. This indicates the estimated remaining time for blockage based on the current rate of deposition growth.

[0098] The above embodiments also include step 8: restarting feasibility assessment based on production baseline pressure drop. For reference, the pressure drop required for restart is estimated based on the cross-sectional area ratio. And predict pressure drop With available pressure differential In comparison, if the predicted pressure drop Exceeding available pressure differential If so, a restart risk warning will be output.

[0099] In this embodiment, the restart feasibility assessment is used to avoid relying solely on subjective judgment based on curves. (Based on production baseline pressure drop) For reference, the minimum cross-sectional area ratio and the predicted restart pressure drop are calculated according to formula (24); if the predicted pressure drop exceeds the available pressure difference, a restart risk warning is output.

[0100] (twenty four) In the formula, This represents the minimum effective inner diameter after considering hydrate deposition; This represents the ratio of the minimum effective flow cross-sectional area to the clean pipe flow cross-sectional area, i.e. ; This indicates the nominal inner diameter of the cleaning pipe.

[0101] In one embodiment of the present invention, a digital twin early warning system for the risk of hydrate blockage in a subsea pipeline coupled with an oil nozzle is provided, comprising: The data preprocessing module acquires online monitoring data and basic engineering data, and preprocesses the acquired data to form an observation sequence. The continuous state identification module performs multi-condition continuous state identification based on the observation sequence to determine the current and predicted working condition stages and output the state probability or confidence level and the time of change. The time window planning and hierarchical solution scheduling module establishes or updates the digital twin case structure based on the results of continuous state identification, and performs adaptive time window planning and hierarchical solution scheduling. The nozzle throttling coupling module updates the dynamic parameters of the nozzle, heat transfer, friction and hydrate online when a throttling element is present at the inlet. It also establishes the coupling boundary between the throttling element and the subsea pipeline and calculates the throttling mass flow rate and throttling temperature drop. The calculation results are used as the pipeline inlet boundary conditions. The multiphase transient solution module executes multiphase transient solutions within a time window based on the entry boundary conditions and hierarchical solution scheduling, and inherits state variables between different solution kernels; The blockage probability assessment and event-based module couples hydrate equilibrium temperature, kinetics, and deposition feedback to calculate blockage probability and blockage segment events. The early warning output module compares the congestion probability, severity, or expected congestion time with a threshold and outputs early warning information. It also adaptively adjusts subsequent time window planning based on the early warning information to form a closed-loop scheduling.

[0102] In the above embodiments, multi-condition continuous state identification includes: Construct an observation feature vector within a sliding window. The observation feature vector includes inlet flow rate or production, nozzle or valve opening, the difference between upstream inlet pressure and outlet back pressure or equivalent pressure difference, and the difference terms of each parameter. Change point detection is performed based on feature vectors to identify change points in the observation sequence; Based on the state transition matrix, the state probability is recursively calculated to determine the posterior probability of the current state belonging to each working condition, and the state probability or confidence level is output. The state output is constrained by a combination of dwell time and hysteresis threshold: entering the target state requires the state probability to exceed the entry threshold and to be maintained for a first preset number of sampling periods; exiting the current state requires the state probability to be lower than the exit threshold and to be maintained for a second preset number of sampling periods, so as to suppress state jitter.

[0103] In the above embodiments, establishing or updating the digital twin instance structure and performing adaptive time window planning and hierarchical solution scheduling includes: A unified computational example structure is constructed, which includes pipeline geometry and mesh, environmental boundaries, fluid properties, stable production, shutdown and restart scheduling, throttling elements and model parameters. The operating condition stage information in the computational example structure is updated based on the continuous state identification results. Adaptive time window planning includes: using the stage boundary, the end time of the ramp segment, the change point time, and the risk trigger time as candidate window boundaries; when the candidate boundary falls into the ramp segment, it will be aligned to the end time of the ramp; and the window length will be adjusted segment by segment according to the maximum congestion probability within the window, where the higher the maximum congestion probability, the shorter the window length. The hierarchical solution scheduling includes: using the first solution kernel and the first time step for the gradual change phase; using the second solution kernel and the second time step for the valve closing ramp, restart ramp or risk trigger phase, the second time step is smaller than the first time step, and inheriting the pressure, temperature, liquid holdup and hydrate deposition state along the process when switching solution kernels.

[0104] In the above embodiments, the coupling boundary between the throttling element and the subsea pipeline is established, and the throttling mass flow rate and throttling temperature drop are calculated. The calculation results are used as the pipeline inlet boundary conditions, including: Obtain the equivalent orifice diameter, flow coefficient, and opening degree of the throttling element; Obtain upstream pressure, upstream temperature, and downstream inlet pressure, and calculate the gas phase volume fraction based on fluid properties; Based on the upstream pressure, downstream inlet pressure, and gas phase volume fraction, the appropriate model is adaptively selected from the single-phase liquid incompressible orifice plate model, the single-phase gas compressible isentropic critical flow model, the two-phase homogeneous HEM model, and the API-14B orifice plate model to calculate the throttling mass flow rate. Among them, for gas or two-phase flow, when the pressure ratio is lower than the critical pressure ratio, the critical flow upper limit is used to limit the mass flow rate. In the two-phase homogeneous HEM model, the Wood sound velocity approximates the mixed phase sound velocity to determine the critical flow upper limit. The throttling temperature drop is calculated based on the isenthalpic or isentropic model to obtain the temperature after throttling; The throttled mass flow rate is used as the boundary of the pipeline inlet mass flow rate, and the temperature after throttling is used as the boundary of the pipeline inlet temperature.

[0105] In the above embodiments, the multiphase transient solution is executed according to the inlet boundary conditions and the hierarchical solution scheduling, including: Within the time window, the corresponding solution kernel is selected based on the hierarchical solution scheduling to execute the transient solution: When in the gradual change phase, the first solution kernel is used for solving. The first solution kernel uses quasi-steady-state hydraulics as the framework and couples a radial RC thermal resistance-heat capacity network to solve the transient energy equation to simulate the thermal inertia of the pipe wall, insulation layer and external medium. The radial RC thermal resistance-heat capacity network abstracts the pipe wall, insulation layer and external medium into series and parallel thermal resistance and distributed heat capacity, and updates the heat transfer flux between the pipe wall and the fluid through time stepping. When in the valve-closing ramp, restart ramp, or risk-triggered phase, a second solution kernel is used for solving. The second solution kernel uses a fully implicit two-fluid conservation equation discretization and solves the mass, momentum, and energy conservation equations through backward Euler discretization and Newton iteration to capture compressibility and pressure wave propagation. When switching between different solution kernels and time windows, the pressure, temperature, liquid holdup, and hydrate deposition state parameters along the process are inherited to ensure numerical continuity.

[0106] In the above embodiments, the coupling of hydrate equilibrium temperature, kinetics, and deposition feedback to calculate the clogging probability and clogging segment events includes: Based on the current pressure and gas composition, and combined with salinity correction and inhibitor concentration correction, the hydrate equilibrium temperature is calculated. The degree of supercooling or superheating is determined based on the difference between the fluid temperature and the hydrate equilibrium temperature. The mass transfer coefficient at the gas-liquid interface is the main control factor. The hydrate formation term is represented by the empirical mapping of the degree of supercooling. The heat transfer limitation at the solid-liquid interface is the main control factor. The hydrate decomposition term is represented by the empirical mapping of the degree of superheating. The rate of change of hydrate volume fraction is output. A two-pool exchange model of slurry pool and sediment pool is used to update the hydrate deposition state: sediment pool growth is controlled by the sedimentation coefficient and velocity inhibition scale, and resuspension is controlled by the resuspension coefficient and velocity enhancement scale; the effective inner diameter is calculated based on the sediment volume fraction in the sediment pool, and the effective inner diameter is fed back into the pressure drop calculation and heat transfer calculation; Based on the uncertainty of the sediment volume fraction, the basic clogging probability is calculated using either the transcendental probability calculation or the Monte Carlo uncertainty propagation. The clogging probability is then obtained by adaptive correction based on the operating conditions, such as flow velocity, liquid holdup, tilt angle, and effective inner diameter. Using the deposition volume fraction exceeding the first threshold and its growth rate exceeding the second threshold as the criteria for blockage formation, adjacent trigger grids are merged, and the location, length, formation time, and severity of the blockage are output.

[0107] The above embodiments also include a restart feasibility assessment step: using the production baseline pressure drop as a reference, the pressure drop required for restart is estimated and predicted based on the cross-sectional area ratio, and the predicted pressure drop is compared with the available pressure difference. If the predicted pressure drop exceeds the available pressure difference, a restart risk warning is output.

[0108] The system provided in this embodiment is used to execute the above-described method embodiments. For specific processes and details, please refer to the above embodiments, which will not be repeated here.

[0109] Example 1: The key input parameters and scheduling configurations for this example are shown in Tables 1 to 4; the thresholds and event-based parameters are shown in Table 5.

[0110] Table 1 Key Input Parameters for Example 1

[0111] Table 2 Example 1 Water temperature input with water depth

[0112] Table 3 Restart Step-by-Step Opening Strategy in Example 1 ( )

[0113] Table 4. Segmented Time Step Configuration for Example 1

[0114] Table 5 Thresholds and Event-based Parameters in Example 1

[0115] The system first performs continuous state identification under multiple operating conditions and outputs change points and confidence levels by accessing inlet / outlet temperature, pressure, and opening data online. Then, it automatically plans and schedules time windows based on the identification results and risk triggering conditions. Layered solution, with window boundaries aligned to the end of the ramp segment; system output along the path. , and and calculate Spacetime diagrams (such as) Figure 7 (As shown).

[0116] In this embodiment, and Evolution curve over time as follows Figure 12 As shown; when Exceeding the threshold or When the value falls below the threshold, the system automatically shortens the time window and upgrades the solver level, while simultaneously triggering an alert and outputting suggested actions (such as...). Figure 8 (As shown).

[0117] Example 2, for comparison with inhibitor off. In Example 1, inhibitor injection was turned off (concentration set to 0), while all other parameters remained the same.

[0118] Table 6 Key Differences in Example 2

[0119] Risk comparison results as follows Figure 13 As shown, it can be used to evaluate inhibitor effectiveness and determine the minimum design concentration.

[0120] In one embodiment of the present invention, a computing device is provided. The computing device may be a terminal and may include: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, and the one or more programs include instructions for performing any of the methods in the above embodiments.

[0121] Furthermore, when the aforementioned instructions can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0122] In one embodiment of the present invention, a computer program product is provided, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, and when the program instructions are executed by a computer, the computer is able to perform the methods provided in the above-described method embodiments.

[0123] In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided, which stores server instructions that cause a computer to perform the methods provided in the above embodiments.

[0124] The computer-readable storage medium provided in the above embodiments has a similar implementation principle and technical effect to the above method embodiments, and will not be described again here.

[0125] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0126] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0127] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with an oil nozzle, characterized in that, include: Acquire online monitoring data and basic engineering data, and preprocess the acquired data to form an observation sequence; Based on the observation sequence, multi-condition continuous state identification is performed to determine the current and predicted working condition stages and output the state probability or confidence level and the time of change. Based on the results of continuous state identification, establish or update the structure of the digital twin case, and perform adaptive time window planning and hierarchical solution scheduling. When a throttling element is present at the inlet, the dynamic parameters of the nozzle, heat transfer, friction and hydrate are updated online, and the coupling boundary between the throttling element and the subsea pipeline is established. The throttling mass flow rate and throttling temperature drop are calculated, and the calculation results are used as the pipeline inlet boundary conditions. Within the time window, multiphase transient solutions are executed based on the entry boundary conditions and hierarchical solution scheduling, and state variables are inherited between different solution kernels; Coupled calculations of hydrate equilibrium temperature, kinetics, and deposition feedback were performed to determine clogging probability and clogging segment events. The system compares the probability of congestion, severity, or expected congestion time with a threshold and outputs an early warning message. It then adaptively adjusts subsequent time window planning based on the early warning message to form a closed-loop scheduling mechanism.

2. The digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with nozzles as described in claim 1, characterized in that, Multi-condition continuous state identification includes: Construct an observation feature vector within a sliding window. The observation feature vector includes inlet flow rate or production, nozzle or valve opening, the difference between upstream inlet pressure and outlet back pressure or equivalent pressure difference, and the difference terms of each parameter. Change point detection is performed based on feature vectors to identify change points in the observation sequence; Based on the state transition matrix, the state probability is recursively calculated to determine the posterior probability of the current state belonging to each working condition, and the state probability or confidence level is output. The state output is constrained by a combination of dwell time and hysteresis threshold: entering the target state requires the state probability to exceed the entry threshold and to be maintained for a first preset number of sampling periods; exiting the current state requires the state probability to be lower than the exit threshold and to be maintained for a second preset number of sampling periods, so as to suppress state jitter.

3. The digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with nozzles as described in claim 1, characterized in that, Establish or update the digital twin instance structure, and perform adaptive time window planning and hierarchical solution scheduling, including: A unified computational example structure is constructed, which includes pipeline geometry and mesh, environmental boundaries, fluid properties, stable production, shutdown and restart scheduling, throttling elements and model parameters. The operating condition stage information in the computational example structure is updated based on the continuous state identification results. Adaptive time window planning includes: using the stage boundary, the end time of the ramp segment, the change point time, and the risk trigger time as candidate window boundaries; when the candidate boundary falls into the ramp segment, it will be aligned to the end time of the ramp; and the window length will be adjusted segment by segment according to the maximum congestion probability within the window, where the higher the maximum congestion probability, the shorter the window length. The hierarchical solution scheduling includes: using the first solution kernel and the first time step for the gradual change phase; using the second solution kernel and the second time step for the valve closing ramp, restart ramp or risk trigger phase, the second time step is smaller than the first time step, and inheriting the pressure, temperature, liquid holdup and hydrate deposition state along the process when switching solution kernels.

4. The digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with nozzles as described in claim 1, characterized in that, Establish the coupling boundary between the throttling element and the subsea pipeline, and calculate the throttling mass flow rate and throttling temperature drop. Use the calculation results as the pipeline inlet boundary conditions, including: Obtain the equivalent orifice diameter, flow coefficient, and opening degree of the throttling element; Obtain upstream pressure, upstream temperature, and downstream inlet pressure, and calculate the gas phase volume fraction based on fluid properties; Based on the upstream pressure, downstream inlet pressure, and gas phase volume fraction, the appropriate model is adaptively selected from the single-phase liquid incompressible orifice plate model, the single-phase gas compressible isentropic critical flow model, the two-phase homogeneous HEM model, and the API-14B orifice plate model to calculate the throttling mass flow rate. Among them, for gas or two-phase flow, when the pressure ratio is lower than the critical pressure ratio, the critical flow upper limit is used to limit the mass flow rate. In the two-phase homogeneous HEM model, the Wood sound velocity approximates the mixed phase sound velocity to determine the critical flow upper limit. The throttling temperature drop is calculated based on the isenthalpic or isentropic model to obtain the temperature after throttling; The throttled mass flow rate is used as the boundary of the pipeline inlet mass flow rate, and the temperature after throttling is used as the boundary of the pipeline inlet temperature.

5. The digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with nozzles as described in claim 1, characterized in that, Based on the inlet boundary conditions and hierarchical solution scheduling, multiphase transient solutions are executed, including: Within the time window, the corresponding solution kernel is selected based on the hierarchical solution scheduling to execute the transient solution: When in the gradual change phase, the first solution kernel is used for solving. The first solution kernel uses quasi-steady-state hydraulics as the framework and couples a radial RC thermal resistance-heat capacity network to solve the transient energy equation to simulate the thermal inertia of the pipe wall, insulation layer and external medium. The radial RC thermal resistance-heat capacity network abstracts the pipe wall, insulation layer and external medium into series and parallel thermal resistance and distributed heat capacity, and updates the heat transfer flux between the pipe wall and the fluid through time stepping. When in the valve-closing ramp, restart ramp, or risk-triggered phase, a second solution kernel is used for solving. The second solution kernel uses a fully implicit two-fluid conservation equation discretization and solves the mass, momentum, and energy conservation equations through backward Euler discretization and Newton iteration to capture compressibility and pressure wave propagation. When switching between different solution kernels and time windows, the pressure, temperature, liquid holdup, and hydrate deposition state parameters along the process are inherited to ensure numerical continuity.

6. The digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with nozzles as described in claim 1, characterized in that, Coupled with hydrate equilibrium temperature, kinetics, and deposition feedback, calculations are performed to determine clogging probability and clogging segment events, including: Based on the current pressure and gas composition, and combined with salinity correction and inhibitor concentration correction, the hydrate equilibrium temperature is calculated. The degree of supercooling or superheating is determined based on the difference between the fluid temperature and the hydrate equilibrium temperature. The mass transfer coefficient at the gas-liquid interface is the main control factor. The hydrate formation term is represented by the empirical mapping of the degree of supercooling. The heat transfer limitation at the solid-liquid interface is the main control factor. The hydrate decomposition term is represented by the empirical mapping of the degree of superheating. The rate of change of hydrate volume fraction is output. A two-pool exchange model of slurry pool and sediment pool is used to update the hydrate deposition state: sediment pool growth is controlled by the sedimentation coefficient and velocity inhibition scale, and resuspension is controlled by the resuspension coefficient and velocity enhancement scale; the effective inner diameter is calculated based on the sediment volume fraction in the sediment pool, and the effective inner diameter is fed back into the pressure drop calculation and heat transfer calculation; Based on the uncertainty of the sediment volume fraction, the basic clogging probability is calculated using either the transcendental probability calculation or the Monte Carlo uncertainty propagation. The clogging probability is then obtained by adaptive correction based on the operating conditions, such as flow velocity, liquid holdup, tilt angle, and effective inner diameter. Using the deposition volume fraction exceeding the first threshold and its growth rate exceeding the second threshold as the criteria for blockage formation, adjacent trigger grids are merged, and the location, length, formation time, and severity of the blockage are output.

7. The digital twin early warning method for hydrate blockage risk in subsea pipelines coupled with nozzles as described in claim 1, characterized in that, It also includes the step of restarting a feasibility assessment: based on production baseline pressure drop. For reference, the pressure drop required for restart is estimated based on the cross-sectional area ratio. And predict pressure drop With available pressure differential In comparison, if the predicted pressure drop Exceeding available pressure differential If so, a restart risk warning will be output.

8. A digital twin early warning system for the risk of hydrate blockage in subsea pipelines coupled with an oil nozzle, characterized in that, include: The data preprocessing module acquires online monitoring data and basic engineering data, and preprocesses the acquired data to form an observation sequence. The continuous state identification module performs multi-condition continuous state identification based on the observation sequence to determine the current and predicted working condition stages and output the state probability or confidence level and the time of change. The time window planning and hierarchical solution scheduling module establishes or updates the digital twin case structure based on the results of continuous state identification, and performs adaptive time window planning and hierarchical solution scheduling. The nozzle throttling coupling module updates the dynamic parameters of the nozzle, heat transfer, friction and hydrate online when a throttling element is present at the inlet. It also establishes the coupling boundary between the throttling element and the subsea pipeline and calculates the throttling mass flow rate and throttling temperature drop. The calculation results are used as the pipeline inlet boundary conditions. The multiphase transient solution module executes multiphase transient solutions within a time window based on the entry boundary conditions and hierarchical solution scheduling, and inherits state variables between different solution kernels; The blockage probability assessment and event-based module couples hydrate equilibrium temperature, kinetics, and deposition feedback to calculate blockage probability and blockage segment events. The early warning output module compares the congestion probability, severity, or expected congestion time with a threshold and outputs early warning information. It also adaptively adjusts subsequent time window planning based on the early warning information to form a closed-loop scheduling.

9. A computer-readable storage medium for storing one or more programs, characterized in that, The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any of the methods described in claims 1 to 7.

10. A computing device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the methods described in claims 1 to 7.

Images