CN-121981489-A - Internet of things perception driven offshore facility group toughness maintenance scheduling and ship path collaborative optimization method

CN121981489ACN 121981489 ACN121981489 ACN 121981489ACN-121981489-A

Abstract

The invention discloses a method for performing toughness maintenance scheduling and ship path collaborative optimization on an offshore facility group driven by the perception of the Internet of things, and relates to the technical field of ocean engineering operation and maintenance and intelligent optimization. The method comprises the steps of constructing a maintenance decision hierarchy of the sensing support of the Internet of things, carrying out structural modeling on an offshore facility group operation and maintenance network, carrying out uncertainty parameter calibration on navigation time, operation duration and an reachable time window based on sensing data of the Internet of things, constructing a time integral optimization target based on toughness guidance, establishing a constraint system of accumulated propagation of time uncertainty along a path and feasibility guarantee, and adopting a layered iteration solving and local enhancement strategy output scheme. The method can realize closed-loop fusion of the perception of the Internet of things and decision optimization, effectively improve the execution robustness of a scheduling scheme and the system toughness recovery capability, and provide decision support for the operation and maintenance of facilities such as offshore wind farms, oil-gas platform groups and the like.

Inventors

CUI XINHAO
XIAO YIYONG
XIAO FANGCHENG
LI BO
JI ZIGUANG
ZHANG YUE
REN YI

Assignees

北京航空航天大学

Dates

Publication Date: 20260505
Application Date: 20260209

Claims (10)

1. The method is characterized in that a hierarchical processing flow integrating sensing fusion, toughness quantification and collaborative optimization is adopted, uncertainty parameter calibration driven by data of the Internet of things, accumulated propagation constraint modeling of time uncertainty along a path, optimization target design of toughness guidance and joint decision of maintenance scheduling and a ship path are comprehensively arranged in the same frame, and the method comprises the following steps: step one, constructing a maintenance decision-making hierarchical structure of the perception support of the Internet of things, wherein the maintenance decision-making hierarchical structure comprises a perception layer, a data layer, a model layer and an application layer; step two, carrying out structural modeling on the offshore facility group operation and maintenance network, and defining facility node sets, base node sets, ship sets, route arc segment sets and related parameters and decision variables; Step three, calibrating uncertainty parameters of navigation time, operation duration and an reachable time window based on historical operation and maintenance data and real-time information flow acquired by the sensing network of the Internet of things; Step four, constructing an optimized objective function based on toughness guidance, and taking the time integral of capacity recovery as a core measure to maximize expected toughness contribution; Establishing a constraint system of uncertainty accumulated propagation and feasibility guarantee, and explicitly tracking the accumulated propagation effect of time uncertainty along a ship access path; step six, adopting a hierarchical iteration solving and local enhancement strategy, and outputting maintenance task assignment, a ship access sequence and a time sequence scheduling scheme.
2. The method of claim 1, wherein the sensing layer in the first step deploys sensors according to facility types and collects equipment health data and sea state observation data, the sensors at least comprise one or more of vibration sensors, temperature probes, pressure transmitters, corrosion monitoring devices and structural health monitoring systems, the sea state observation equipment at least comprises one or more of weather buoys, anemometers, wave heights and ocean current meters, the collected data are preprocessed by edge computing nodes and then transmitted to a shore-based data center, the data layer cleans, fuses and extracts features of multi-source heterogeneous data, and utilizes statistics to estimate degradation urgency indexes of each facility, function recovery contributions are defined according to the facility types, the model layer builds a mixed integer linear programming model and solves the model, and the application layer outputs an executable scheme and supports rolling optimization.
3. The method according to claim 1, wherein the second step comprises determining a set of offshore facility nodes N to be maintained, i and j being used as indexes, determining a set of base nodes D to maintain berthing and departure of a ship, D being used as indexes, determining a set of complete vertexes V=NU D, determining a set of schedulable heterogeneous maintenance ships K being used as indexes, determining a set of feasible route arc segments A for connecting vertexes, wherein facility characteristic parameters comprise a function recovery contribution c i , a urgency index u i , an accessibility time window [ e i , l i ] and a ship compatibility parameter w ik , and time random parameters comprise a desired value mu ij k and a standard deviation sigma ij k of navigation time and a desired value mu i p and a standard deviation sigma i p of operation duration.
4. The method of claim 1, wherein in the third step, parameter estimation is performed on navigation time uncertainty calibration, marine weather observation history, marine automatic identification system track data and marine log statistical samples are utilized, the randomness of the navigation time is derived from ship additional resistance caused by a wave field, forward and backward flow effect generated by the marine current field and wind pressure difference effect caused by the wind field, the navigation time approximately obeys normal distribution N (mu ij k , (σ ij k ) 2 ), parameter estimation is performed on operation duration uncertainty calibration, a fault mode identification result output by a fault diagnosis model of equipment is combined with historical similar maintenance work order records, the operation duration approximately obeys normal distribution N (mu i p , (σ i p ) 2 ), and an reachable time window parameter [ e i , l i ] is dynamically corrected according to rolling update of a weather forecasting system and state change of sensing equipment of the internet of things.
5. The method according to claim 1, wherein the quantification of the toughness contribution in the fourth step is based on the principle that the earlier the maintenance operation is completed, the longer the maintenance facility resumes normal operation, the greater its cumulative contribution to the total capacity of the system, and the toughness contribution is defined as the time integral of the recovery capacity over the planning period: , The urgency weight u i of the internet of things drive is introduced and combined, and the optimization goal is set to maximize the desired toughness contribution: , Where R i is the toughness contribution variable of facility i, the target incentive preferentially completes the tasks of high urgency, high capacity contribution and as early as possible to extend the duration of capacity recovery.
6. The method according to claim 1, wherein the fifth step comprises establishing a desired completion time recurrence constraint for a first facility i accessed by the vessel k after starting from the base d, the desired completion time being a base-to-facility desired travel time plus a facility desired job duration: , Achieving time sequence dependency relationship among tasks through recursion accumulation for subsequent facilities j ; Establishing a variance accumulation propagation constraint: , based on the assumption that the navigation time of each navigation section and the operation time of each facility are mutually independent, the accumulated variance is in a superposition recurrence relation along the path, and the accumulated variance of the completion time of the task which is more backward on the path is larger: ; Establishing a mapping constraint of the accumulated standard deviation s i and the accumulated variance v i : 。
7. The method according to claim 1, wherein the fifth step further comprises establishing a time window probability satisfying constraint by introducing a risk control mechanism based on normal distribution alpha quantiles, and the alpha quantiles at the completion time of the facility i are: , Wherein z α is a standard normal quantile corresponding to the confidence level alpha, the random time window constraint is converted into a deterministic equivalent form, and the probability that the maintenance completion moment falls into the permission time window under the given confidence level alpha is ensured to meet the corresponding requirement: , , The higher the confidence level α, the greater the time margin requirement and the more conservative and robust the scheme.
8. The method of claim 1, wherein the step five is characterized by linearizing the non-linear square root relationship in the cumulative variance-to-standard deviation mapping constraint using a piecewise cut approximation method, comprising determining the scaling ζ based on a maximum allowable relative approximation error ε: , The minimum number of segments η required to cover the variance interval [ V min , V max ] is calculated: , Determining a start breakpoint V p , a slope coefficient k p and an intercept coefficient b p of each segment of the cut line: , the interval to which the segment corresponds is indicated by the binary variable λ ip , into which the cumulative variance of the facility i falls: , , , , Ensuring that the facility assigned the task activates exactly one segment, calculating the cumulative standard deviation from the activated segment, the linearization method ensures that the approximation error does not exceed a preset tolerance epsilon.
9. The method of claim 1, wherein step five further comprises establishing task assignment and path connectivity constraints. Wherein the task-unique constraint ensures that each facility is served at most by one vessel: ; the task compatibility constraint ensures that the task assignment meets the ship capability requirement: ; The path traffic conservation constraint requires that each allocated facility be either a path start point or exactly one in-arc and each allocated facility be either a path end point or exactly one out-arc: , , To ensure connectivity of the ship path.
10. The method of claim 1, wherein the hierarchical iterative solution and local enhancement strategy in the sixth step includes generating an initial feasible solution by adopting a construction heuristic, calculating a composite priority index u i ·c i /(l i −e i of each facility to be maintained, performing greedy allocation by traversing the facilities in descending order of priority, selecting facilities to form a subset to be optimized by adopting an urgency frequency priority search operator according to the descending order of the ratio u i /f i of urgency to frequency, releasing associated assigned variables and arc variables to form a sub-problem with controlled scale, adopting a space neighborhood search operator by adopting an inner layer, identifying sensitive nodes with accumulated variance exceeding a threshold, defining facility synchronous release decision variables in the neighborhood by taking the sensitive nodes as the center, accurately solving the sub-problem formed after release of the variables according to geographical distance, accepting a new solution if the objective function value is improved, moderately expanding the scale neighborhood otherwise, shrinking the neighborhood or switching the search direction, and continuing iteration until the maximum iteration number is reached or continuous several times are terminated.

Description

Internet of things perception driven offshore facility group toughness maintenance scheduling and ship path collaborative optimization method Technical Field The invention belongs to the technical field of ocean engineering operation and maintenance and intelligent optimization, and particularly relates to a method for scheduling toughness maintenance and ship path collaborative optimization of an offshore facility group driven by perception of the Internet of things. Aiming at maintenance and guarantee requirements of various offshore facility groups distributed in a wide sea area, the method comprehensively considers multi-source uncertain factors such as dynamic change of marine meteorological environment, uncertainty of a maintenance task reachable time window, random fluctuation of operation execution duration and the like, monitors and degenerates prediction of facility states driven by sensing data of the Internet of things, dynamically quantifies maintenance priority of toughness guidance, and jointly optimizes decision of ship paths and task time sequences, improves resolvability and engineering applicability of large-scale examples by means of a layered iteration solving mechanism, and realizes overall optimal configuration and high-efficiency toughness operation of an offshore facility group maintenance system. Background With the rapid development of marine economy, the construction scale of various offshore facilities continues to expand. These groups of offshore facilities share the common feature of being geographically dispersed, remote from shore-based security, and exposed to harsh marine environments for long periods of time. The operation and maintenance of offshore facilities also presents special challenges. On the one hand, facilities are distributed in wide sea areas with large mutual distance, maintenance ships need to navigate among a plurality of facilities, and travel time accounts for a significant proportion of total operation time. On the other hand, the randomness of the marine meteorological conditions causes the limitation of maintenance operation window, and personnel can safely climb a hill operate only when the indexes such as wave height, wind speed, ocean current and the like are all within the safety threshold. These windows tend to be limited in duration and difficult to predict accurately, placing higher demands on the timeliness and robustness of the maintenance schedule. In addition, different types of facilities have differentiated maintenance requirements and priorities, and coordinating the maintenance scheduling of heterogeneous facility clusters becomes a complex combinatorial optimization problem. The progress of the Internet of things brings new opportunities for the operation and maintenance of offshore facilities. Modern offshore facilities are commonly provided with vibration sensors, temperature probes, structural health monitoring systems and other sensors, and key components and structures can be continuously monitored. By combining with external data sources such as weather buoys, marine observation satellites, ship automatic identification systems and the like, operators can acquire comprehensive information covering the internal health state and external environmental conditions of equipment. After the data are cleaned, fused and extracted in characteristics, a maintenance strategy based on a state can be supported, and the transition from regular overhaul to on-demand intervention is realized. However, how to effectively convert mass data collected by the internet of things into an optimal maintenance scheduling and ship path scheme is still a difficult problem in engineering practice. The prior art mainly has the defects of decision separation, uncertainty neglect, lack of toughness view angle and the like. Therefore, there is a need for a method for collaborative optimization of marine facility group maintenance scheduling and ship path that can comprehensively utilize the perceived data of the internet of things, reasonably quantify the toughness contribution, explicitly model the cumulative propagation of time uncertainty, support collaborative scheduling of heterogeneous facilities, and have good computational scalability. Disclosure of Invention (1) The purpose of the invention is that: The invention aims to provide a method for performing toughness maintenance scheduling and ship path collaborative optimization on an offshore facility group driven by the perception of the Internet of things. Aiming at the problems that maintenance tasks in the existing offshore operation and maintenance technology are determined to be mutually fractured with ship route arrangement, dynamic change of a sea condition accessibility time window and randomness of operation duration are difficult to systematically describe, solving efficiency and scheme quality are difficult to consider under the condition of facility type isomerism and scale expansion, and the l