Long-distance pipeline construction equipment operation and maintenance method and system

Abstract

The present application relates to the technical field of long-distance pipeline operation and maintenance, and provides a long-distance pipeline construction equipment operation and maintenance method and system, comprising the following steps: step one, collecting equipment operation state data in real time through a multi-source sensor disposed on the construction equipment, wherein the operation state data at least includes vibration signals, temperature, hydraulic pressure, oil particle pollution degree, and real-time geographical position of the equipment; step two, using an edge computing node arranged at the site of the construction unit to pre-process and extract degradation features of the collected operation state data, and generating an equipment degradation feature vector with a time label; step three, when the edge computing node is connected with a cloud digital twin server. Through multi-source sensing, edge computing, digital twin prediction and spatio-temporal coupling decision, the equipment state is sensed in real time, fault is accurately predicted, and maintenance scheme is scientifically optimized, thereby reducing unplanned downtime and maintenance cost, avoiding construction delay, and significantly improving operation and maintenance efficiency and construction continuity.

Description

Technical Field

[0001] This invention relates to the field of operation and maintenance technology for long-distance pipeline construction equipment, specifically to operation and maintenance methods and systems for long-distance pipeline construction equipment. Background Technology

[0002] Long-distance pipelines, as key infrastructure for energy transportation, are widely used for the long-distance transport of media such as oil and natural gas. The construction routes span large distances, have complex geographical environments, and long construction periods. They involve a large number of heavy construction equipment such as fully automatic welding machines, beveling machines, hydraulic pipe bending machines, and pipe lifting machines. The continuous and stable operation of the equipment directly determines the efficiency, quality, and schedule of pipeline construction. Therefore, the efficient operation and maintenance of construction equipment is of utmost importance.

[0003] Currently, most long-distance pipeline construction equipment adopts traditional periodic maintenance or post-construction repair modes, relying on manual inspections and experience to judge equipment status. This lacks real-time perception and analysis of multi-dimensional operational data such as equipment vibration, hydraulic pressure, and oil levels. Furthermore, existing maintenance technologies do not incorporate equipment degradation patterns and stochastic process theory for accurate lifespan prediction, making it difficult to quantify equipment failure risks. Moreover, they fail to fully integrate construction schedules and spatial constraints into maintenance plans, easily leading to sudden equipment failures, inappropriate maintenance timing, and wasted maintenance resources. This results in construction interruptions, project delays, and increased costs, failing to meet the maintenance requirements of large-scale, high-efficiency, and high-reliability long-distance pipeline construction. To address the problems of incomplete equipment status perception, low fault prediction accuracy, disconnect between maintenance decisions and construction schedules, and unreasonable resource scheduling in existing technologies, this invention proposes a method and system for the maintenance of long-distance pipeline construction equipment. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for the operation and maintenance of construction equipment for long-distance pipelines, which solves the problems of incomplete equipment status perception, low accuracy of fault prediction, disconnect between operation and maintenance decisions and construction progress, and unreasonable resource scheduling in existing technologies.

[0005] To achieve the above objectives, the present invention provides a method for the operation and maintenance of long-distance pipeline construction equipment, comprising the following steps: Step 1: Collect real-time equipment operating status data using multi-source sensors deployed on the construction equipment. The operating status data includes at least vibration signals, temperature, hydraulic pressure, oil particle contamination level, and the real-time geographical location of the equipment. Step 2: Using the edge computing nodes set up at the construction unit's site, preprocess the collected operating status data and extract degradation features to generate equipment degradation feature vectors with time tags; Step 3: When establishing a connection between the edge computing node and the cloud digital twin server, the degradation feature vector is synchronized to the digital twin server through a breakpoint resume method to update the degradation state parameters in the device digital twin model. Step 4: The digital twin model is based on a stochastic degradation process. It performs Bayesian state estimation on each device and predicts the probability density function of the device's remaining useful life and the cumulative probability of failure in the future. Step 5: Obtain the pipeline construction schedule and spatial constraints of each process. With the goal of minimizing the expected total delay of the entire welding operation caused by equipment maintenance downtime, use the accumulated failure probability obtained in step S4 to construct a time-space coupled maintenance decision model and solve for the optimal maintenance start time window and the corresponding maintenance resource scheduling scheme. Step 6: Send the optimal maintenance start time window and maintenance resource scheduling plan to the maintenance personnel's terminal and execute the maintenance operation.

[0006] Preferably, the degradation feature vector in step two includes vibration time-domain kurtosis, spectral envelope energy, hydraulic pressure drop rate, and a wear index based on oil particle contamination, wherein the wear index... The calculation formula is: ; in, This refers to the number of particles with a diameter between 15 μm and 25 μm per milliliter of oil. This refers to the number of particles with a diameter greater than 25 μm. Total number of particles, weighting coefficient , And satisfy .

[0007] Preferably, in step four, the digital twin model uses a Wiener process with drift to describe the amount of equipment degradation. : ; in, The drift coefficient, Where is the diffusion coefficient. This is standard Brownian motion; When the current time is obtained Degradation characteristics observation sequence Then, the drift coefficients are updated online using Kalman filtering or particle filtering. The posterior distribution is obtained, and its posterior mean is derived. With variance Let the failure threshold be... Then the remaining service life The probability density function is: ; Future period Cumulative probability of internal equipment failure for: .

[0008] Preferably, the objective function of the time-space coupled maintenance decision model in step five is: ; in, The number of welding work surfaces affected by maintenance. For the first Delay weighting for each work area The maintenance start time is the time that caused the welding delay at this work site. Time of arrival of maintenance resources and spare parts delivery routes The function; The constraints include: from the current decision time Up to the planned maintenance start time The cumulative probability of failure.

[0009] The operation and maintenance system for long-distance pipeline construction equipment includes: The data acquisition module is used to acquire multi-source operating status data and location information of construction equipment in real time; The edge computing module is deployed at the site of each construction unit and connected to the data acquisition module. It is used to preprocess the acquired data and calculate degradation features to generate degradation feature vectors, and has a local real-time early warning function. The digital twin module, deployed on a cloud server, communicates with the edge computing module via an intermittent network. It is used to update the device degradation digital twin model using the degradation feature vector and predict the remaining service life and failure probability of the device based on stochastic process theory. The operation and maintenance decision module, connected to the digital twin module, is used to integrate pipeline construction progress data and geospatial information to generate maintenance start time windows and resource scheduling schemes with the goal of minimizing disturbance to continuous welding operations. The communication module, embedded between the edge computing module and the digital twin module, provides an asynchronous data synchronization mechanism based on breakpoint resumption.

[0010] Preferably, the data acquisition module specifically includes: The vibration signal acquisition unit is installed on the engine main bearing housing and the hydraulic pump housing to acquire triaxial vibration acceleration. Temperature acquisition unit is used to monitor engine coolant temperature, hydraulic oil temperature and ambient temperature; The pressure acquisition unit is used to measure the pressure of each circuit in the hydraulic system in real time. The oil quality acquisition unit has a built-in particle counter and moisture sensor to output the oil particle contamination level and water content. The GPS / BeiDou positioning unit is used to obtain the device's real-time coordinates and movement trajectory.

[0011] Preferably, the edge computing module specifically includes: The data preprocessing unit is used to denoise, remove invalid values, and align the time of the raw sensor signals. The time-frequency feature extraction unit is used to calculate the envelope spectrum, wavelet energy, and slope change of the hydraulic pressure of the vibration signal. A degradation feature synthesis unit is used to fuse multiple features into the degradation feature vector and calculate the real-time wear index according to the wear index formula of claim 2. The local threshold comparison unit is used to compare the degradation feature vector with the preset multi-level alarm threshold. When the threshold is exceeded, the unit's audible and visual alarm is triggered directly.

[0012] Preferably, the digital twin module specifically includes: The equipment model building unit creates a digital image of each construction piece of equipment, including physical properties, degradation equations, and observation equations. The Bayesian state estimation unit executes Kalman filtering or particle filtering algorithms to update the posterior distribution of drift coefficients online. The remaining lifetime prediction unit uses the remaining lifetime probability density function formula to calculate and output the remaining lifetime distribution and the cumulative failure probability in real time. The model synchronization unit is responsible for importing the accumulated degradation feature vectors from the edge side in batches each time a connection is successfully established, and for updating the state of the digital twin model.

[0013] Preferably, the operation and maintenance decision module specifically includes: The construction progress interface unit is used to acquire and parse pipeline construction schedule data; The maintenance-construction coupled optimization unit solves the objective function and outputs the maintenance time window that minimizes the expected total welding delay, the required number of spare parts types, and the optimal route for maintenance personnel. The spare parts and personnel scheduling unit automatically generates spare parts outbound instructions and maintenance personnel work orders based on the optimization results. The visualization unit displays the thermal distribution of equipment health status, predicted fault risk levels, and recommended maintenance windows on a GIS map in real time.

[0014] Preferably, the construction progress interface unit identifies the start and end mileage, time interval, and logical relationship between key processes for each welding operation surface.

[0015] This invention provides a method and system for the operation and maintenance of construction equipment for long-distance pipelines. It has the following beneficial effects: This invention uses multi-source sensors to collect real-time operational data such as equipment vibration, temperature, hydraulic pressure, oil particle contamination level, and geographical location. It relies on edge computing to complete data preprocessing and degradation feature extraction, effectively solving the problems of unstable network, high data transmission latency, and untimely local response at long-distance pipeline construction sites. This significantly improves the real-time performance, accuracy, and reliability of equipment status perception, providing high-quality data support for subsequent equipment health assessment and life prediction.

[0016] This invention introduces cloud-based digital twin technology, based on drift Wiener process and Bayesian state estimation, to achieve accurate modeling of equipment degradation process and probabilistic prediction of remaining service life. It breaks through the limitations of traditional periodic maintenance, which is either "over-maintenance" or "under-maintenance," and realizes the transformation from passive maintenance and periodic maintenance to predictive maintenance. This significantly reduces the risk of unplanned equipment downtime, reduces maintenance costs and spare parts inventory pressure, and ensures construction continuity.

[0017] This invention constructs a time-space coupled maintenance decision model that integrates construction progress, process space constraints, and equipment failure probability. With the goal of minimizing the expected total welding delay across the entire line, it optimizes maintenance timing and resource scheduling, solving problems such as conflicts between equipment maintenance and construction progress, unreasonable allocation of maintenance resources, and insufficient consideration of spatial constraints. This significantly improves the scientific nature of operation and maintenance decisions and the efficiency of construction organization, ensuring the efficient and orderly progress of long-distance pipeline projects. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: This invention provides a method for the operation and maintenance of construction equipment for long-distance pipelines, including the following steps: Step 1: Collect real-time equipment operating status data using multi-source sensors deployed on the construction equipment. The operating status data should include at least vibration signals, temperature, hydraulic pressure, oil particle contamination level, and the real-time geographical location of the equipment. Specifically, at the construction site of the long-distance pipeline, sensors are installed and calibrated on each key construction equipment (including but not limited to: fully automatic pipeline welding machine, beveling machine, hydraulic pipe bending machine, pipe hoisting machine, crawler excavator, air compressor, and hydraulic pump station). The vibration sensor adopts a triaxial MEMS accelerometer with a range of ±50g and a frequency response of 0.5~10kHz. It is installed in three locations: the engine main bearing housing, the hydraulic pump housing, and the gearbox housing. It is bolted and tightly attached to the metal substrate to avoid suspended installation. The sampling frequency is set to 10kHz. The temperature sensor is a PT100 platinum resistance thermometer with a measurement range of -40~125℃. It is installed at the engine coolant outlet, the hydraulic oil tank return port, and the shaded area. The sampling period is 1s. The hydraulic pressure sensor is a piezoelectric pressure transmitter with a range of 0~40MPa and an accuracy of 0.25%FS. It is installed at the hydraulic pump outlet, the main control valve inlet, and the rodless chamber of the actuator. The sampling frequency is 100Hz. Oil quality sensor: Online oil particle counter (ISO4406 standard), with a built-in moisture sensor, installed before the return oil filter in the hydraulic system, providing real-time output of particle count and water content for particles ≥4μm, ≥6μm, ≥14μm, ≥21μm, and ≥25μm; sampling cycle 5s. Positioning module: Beidou + GPS dual-mode positioning, positioning accuracy ≤1m, 1Hz output of device latitude, longitude, elevation, movement speed, and heading angle. All sensors are connected to an edge computing gateway via CAN bus or industrial Ethernet, with a local cache capacity ≥32GB and data retention ≥30 days after power failure.

[0020] Step 2: Using edge computing nodes set up at the construction unit's location, preprocess and extract degradation features from the collected operational data to generate a time-stamped equipment degradation feature vector. This degradation feature vector includes vibration time-domain kurtosis, spectral envelope energy, hydraulic pressure drop rate, and a wear index based on oil particle contamination. The wear index... The calculation formula is: ; in, This refers to the number of particles with a diameter between 15 μm and 25 μm per milliliter of oil. This refers to the number of particles with a diameter greater than 25 μm. Total number of particles, weighting coefficient , And satisfy ; Specifically, edge computing nodes are deployed in the container room of the construction unit's residence, using industrial-grade edge servers (8-core CPU, 16GB RAM, 512GB SSD), running a Linux real-time operating system. Data preprocessing includes: vibration signals: using a 5th-order Butterworth bandpass filter (10Hz~5kHz) to remove DC components and power frequency interference; using the 3σ criterion to remove abnormal spikes; framing in 1-second increments with 50% inter-frame overlap; temperature or pressure: using a sliding window mean filter (10-point window) to remove jump values ​​caused by sensor disconnection or poor contact; time alignment unified to a 1Hz timescale; oil data... : Remove instantaneous fluctuation values ​​and retain the 5-second steady-state average to ensure alignment with the vibration or pressure characteristic timestamp; Degradation feature extraction: Vibration temporal kurtosis: Calculate the kurtosis for each frame of vibration acceleration signal, approximately 3.0–3.5 under normal conditions, rising to 5–15 when bearings or gears deteriorate; Spectral envelope energy: Perform Hilbert envelope + FFT on the vibration signal to extract the energy in the 1–5 kHz frequency band as an indicator of wear impact intensity; Hydraulic pressure drop rate: Linear fitting slope of the steady-state value of the hydraulic pump outlet pressure over 30 consecutive minutes, normally ≤0.2 MPa / h, ≥0.5 MPa / h when there is internal leakage or seal deterioration; Wear index : ; Weighting This refers to the number of particles with a diameter between 15 μm and 25 μm per milliliter of oil. This refers to the number of particles with a diameter greater than 25 μm. Total number of particles, weighting coefficient , And satisfy Construction of degenerate eigenvectors It outputs one degradation feature vector with timestamp t every 1 minute, which is cached locally and will not be lost when the network is disconnected.

[0021] Step 3: When establishing a connection between the edge computing node and the cloud digital twin server, the degradation feature vector is synchronized to the digital twin server through a breakpoint resume method to update the degradation state parameters in the device's digital twin model. Specifically, the edge and cloud utilize 4G / 5G + satellite dual-mode communication. When the network at the construction site is unstable, an asynchronous synchronization mechanism with breakpoint resumption is activated. Edge-side caching: Degraded feature vectors are packaged hourly, with each package ≤10MB, and the data for the most recent 7 days is stored locally. Connection establishment strategy: A cloud connection is attempted every 15 minutes; after network recovery, unsynchronized data packets are uploaded first, using breakpoint resumption to avoid duplicate uploads. Model update triggering: After the cloud digital twin server receives the complete time-series degraded feature vector, it automatically appends it to the device's observation sequence, updating the model's degradation state parameters, with an update cycle ≤5 minutes. Extreme network outages: Local early warning is run independently on the edge side, without relying on the cloud; after network recovery, all feature data from the outage period is re-uploaded in batches.

[0022] Step 4: The digital twin model, based on a stochastic degradation process, performs Bayesian state estimation on each device and predicts the probability density function of the device's remaining useful life and the cumulative probability of failure in future periods. In Step 4, the digital twin model uses a Wiener process with drift to describe the amount of device degradation. : ; in, The drift coefficient, Where is the diffusion coefficient. This is standard Brownian motion; When the current time is obtained Degradation characteristics observation sequence Then, the drift coefficients are updated online using Kalman filtering or particle filtering. The posterior distribution is obtained, and its posterior mean is derived. With variance Let the failure threshold be... Then the remaining service life The probability density function is: ; Future period Cumulative probability of internal equipment failure for: ; Specifically, the digital twin model is deployed on a cloud server (dual-socket Xeon processor, 64GB RAM), with an independent degradation model established for each device. Parameter initialization is based on historical data from the same model of device. Degradation process modeling: employing the Wiener process with drift. ; Drift coefficient: initial mean =0.02 / day, variance 0.001 2 , Diffusion coefficient: fixed at 0.01. For standard Brownian motion: standard Brownian motion, The overall degradation amount is obtained by weighted fusion of degradation feature vectors; Online Bayesian update of observation sequence: one observation x1, x2, ..., x_t every 1 minute; Algorithm: Use Kalman filtering when the noise is low, and switch to particle filtering (500 particles) when there is strong nonlinearity or non-Gaussianness. Output: Real-time updates Posterior mean and variance; Remaining lifetime (RUL) prediction.

[0023] Step 5: Obtain the pipeline construction schedule and spatial constraints of each process. With the objective of minimizing the expected total delay in welding operations due to equipment maintenance downtime, utilize the accumulated fault probability obtained in Step S4 to construct a time-space coupled maintenance decision model. Solve for the optimal maintenance initiation time window and the corresponding maintenance resource scheduling scheme. The objective function of the time-space coupled maintenance decision model in Step 5 is: ; in, The number of welding work surfaces affected by maintenance. For the first Delay weighting for each work area The maintenance start time is the time that caused the welding delay at this work site. Time of arrival of maintenance resources and spare parts delivery routes The function; The constraints include: from the current decision time Up to the planned maintenance start time The cumulative probability of failure; Specifically, the operation and maintenance decision module interfaces with the construction project management system to automatically obtain the overall welding construction schedule, the start and end mileage of each welding work surface, critical path information, and logical constraints between processes. The entire pipeline is divided into independent welding work surfaces every 10 kilometers and numbered sequentially. The delay weight for work surfaces on the critical path is set to 1.0, and the delay weight for work surfaces on non-critical paths is set to 0.3–0.7. The standard maintenance time for a single piece of equipment is set to 4–8 hours, with 3 professional maintenance teams and 2 central spare parts warehouses configured. The transportation speed of maintenance personnel and spare parts is calculated at 30 km / h. With the goal of minimizing the expected total delay of welding operations across the entire line caused by equipment maintenance downtime, a time-space coupled maintenance decision model is constructed by combining the cumulative failure probability output in step four. The model constraint is set so that the cumulative failure probability of equipment from the current decision time to the planned maintenance start time does not exceed 0.1. A genetic algorithm with a population size of 100 and 150 iterations is used to solve the model, and the optimal maintenance start time window, the required spare parts type and quantity, the optimal arrival path of maintenance personnel, and the spare parts delivery plan are output to ensure that the maintenance schedule is highly matched with the construction progress and spatial location.

[0024] Step 6: Send the optimal maintenance start time window and maintenance resource scheduling plan to the maintenance personnel's terminals and execute the maintenance work; Specifically, the operation and maintenance decision-making module distributes the optimal maintenance start time window, maintenance resource scheduling plan, and spare parts list obtained from the solution to the handheld tablet terminal of maintenance personnel via 4G / 5G network, and simultaneously sends a secondary notification via SMS to ensure information delivery. The terminal interface clearly displays the equipment number, on-site geographical location, equipment health level (green / yellow / orange / red), recommended maintenance time window and allowable adjustment range, spare parts material code and requisition location, standardized maintenance process card, safety operation precautions, and equipment historical fault and maintenance records. After arriving on-site, maintenance personnel start work by scanning the code on the terminal, and finish work by scanning the code again, simultaneously uploading the information of replaced parts and automatically updating the equipment electronic ledger. After maintenance is completed, the system automatically compares the sensor data before and after maintenance to verify the effect of equipment health status recovery, and sends the maintenance results back to the cloud server to update the baseline parameters of the digital twin model, forming a data closed loop and continuously optimizing the model accuracy.

[0025] Example 2: The operation and maintenance system for long-distance pipeline construction equipment includes: The data acquisition module is used to acquire multi-source operating status data and location information of construction equipment in real time. The data acquisition module specifically includes: The vibration signal acquisition unit is installed on the engine main bearing housing and the hydraulic pump housing to acquire triaxial vibration acceleration. Temperature acquisition unit is used to monitor engine coolant temperature, hydraulic oil temperature and ambient temperature; The pressure acquisition unit is used to measure the pressure of each circuit in the hydraulic system in real time. The oil quality acquisition unit has a built-in particle counter and moisture sensor to output the oil particle contamination level and water content. GPS / BeiDou positioning unit is used to obtain the device's real-time coordinates and movement trajectory; Specifically, the data acquisition module consists of a vibration signal acquisition unit, a temperature acquisition unit, a pressure acquisition unit, an oil quality acquisition unit, and a GPS / BeiDou positioning unit. The overall protection level reaches IP67, adapting to humid, dusty, and vibrating conditions in the field. The vibration signal acquisition unit uses a triaxial MEMS accelerometer, securely bolted to the engine main bearing housing and hydraulic pump housing to ensure rigid connection with the base and reduce transmission errors. The temperature acquisition unit uses a PT100 platinum resistance thermometer with shielded compensation wires and waterproof sealing at the connectors, monitoring engine coolant, hydraulic oil, and ambient temperature. The pressure acquisition unit uses a stainless steel pressure-resistant interface, directly connected in series in the hydraulic main circuit, monitoring pump outlet, control valve, and actuator pressure in real time. The oil quality acquisition unit incorporates a high-precision particle counter and a moisture sensor, connected in series in the hydraulic return line, with a pressure resistance of no less than 32MPa and an operating temperature range of -20℃ to 80℃. The GPS / BeiDou positioning unit has an external high-gain antenna, fixed in an unobstructed position on the top of the cab, outputting real-time equipment coordinates, trajectory, and attitude, providing basic data for subsequent spatial scheduling.

[0026] The edge computing module, deployed at each construction unit's location, connects to the data acquisition module. It preprocesses the acquired data and calculates degradation features, generating degradation feature vectors and providing local real-time early warning capabilities. The edge computing module specifically includes: The data preprocessing unit is used to denoise, remove invalid values, and align the time of the raw sensor signals. The time-frequency feature extraction unit is used to calculate the envelope spectrum, wavelet energy, and slope change of the hydraulic pressure of the vibration signal. The degradation feature synthesis unit is used to fuse multiple features into a degradation feature vector and calculate the real-time wear index based on the wear index formula. The local threshold comparison unit is used to compare the degradation feature vector with the preset multi-level alarm thresholds. When the threshold is exceeded, the unit's audible and visual alarm is triggered directly. Specifically, the edge computing module is deployed in the container room at the construction site, using industrial-grade ruggedized servers and running a Linux real-time operating system. It includes a data preprocessing unit, a time-frequency feature extraction unit, a degradation feature synthesis unit, and a local threshold comparison unit. The data preprocessing unit receives multi-source sensor data in real time, performs bandpass filtering, noise reduction, outlier removal, and time-domain framing on vibration signals, and performs sliding mean filtering, jump removal, and time alignment on temperature and pressure signals, uniformly generating 1Hz standard time-series data. The time-frequency feature extraction unit uses a hardware acceleration module to perform Hilbert envelope and fast Fourier transform on the vibration signals, outputting envelope spectrum energy, wavelet energy, and pressure change slope in real time. The degradation feature synthesis unit calculates vibration kurtosis, spectral envelope energy, hydraulic pressure drop rate, and wear index in real time according to preset formulas, fusing them to generate a standardized degradation feature vector and adding a timestamp. The local threshold comparison unit presets multiple health thresholds, compares the degradation feature vector in real time, and immediately triggers on-site audible and visual alarms and sends SMS alerts to the unit manager if the threshold is exceeded, achieving rapid edge-side response and ensuring equipment operation safety without relying on the cloud.

[0027] The digital twin module, deployed on a cloud server, communicates with the edge computing module via an intermittent network. It is used to update the device's degradation digital twin model using degradation feature vectors and predict the device's remaining lifespan and failure probability based on stochastic process theory. The digital twin module specifically includes: The equipment model building unit creates a digital image of each construction piece of equipment, including physical properties, degradation equations, and observation equations. The Bayesian state estimation unit executes Kalman filtering or particle filtering algorithms to update the posterior distribution of drift coefficients online. The remaining lifetime prediction unit uses the remaining lifetime probability density function formula to calculate and output the remaining lifetime distribution and the cumulative failure probability in real time. The model synchronization unit is responsible for importing the accumulated degradation feature vectors from the edge side in batches each time a connection is successfully established, and for updating the state of the digital twin model. Specifically, the digital twin module is deployed on a high-performance cloud server and consists of an equipment model building unit, a Bayesian state estimation unit, a remaining lifetime prediction unit, and a model synchronization unit. The equipment model building unit creates a lightweight 3D digital image for each construction piece of equipment, binding the equipment's physical parameters, structural composition, component life curves, and historical maintenance records to form a complete digital archive. The Bayesian state estimation unit receives degradation feature sequences uploaded from the edge side and automatically switches between Kalman filtering and particle filtering algorithms based on the data noise characteristics, updating the posterior distribution of the equipment degradation drift coefficients online and continuously correcting the model parameters. The remaining lifetime prediction unit calculates the probability density function of the remaining lifetime and the cumulative probability of failure at different future time periods in real time based on the updated degradation model parameters and preset failure thresholds, outputting the median remaining lifetime, confidence interval, and risk level. The model synchronization unit adopts a breakpoint resume mechanism, batch synchronizing cached data from the edge side when the network recovers, automatically verifying data integrity and completing model state updates, ensuring that the cloud-based digital twin model is highly consistent with the actual degradation state of the equipment on site, providing reliable model support for accurate prediction and decision-making.

[0028] The operation and maintenance decision module, connected to the digital twin module, integrates pipeline construction progress data and geospatial information to generate maintenance initiation time windows and resource scheduling plans with the goal of minimizing disruption to continuous welding operations. The operation and maintenance decision module specifically includes: The construction progress interface unit is used to acquire and parse pipeline construction scheduling data, and identify the start and end mileage, time interval, and logical relationship between key processes for each welding operation. The maintenance-construction coupled optimization unit solves the objective function and outputs the maintenance time window that minimizes the expected total welding delay, the required number of spare parts types, and the optimal route for maintenance personnel. The spare parts and personnel scheduling unit automatically generates spare parts outbound instructions and maintenance personnel work orders based on the optimization results. The visualization unit displays the thermal distribution of equipment health status, predicted fault risk levels, and recommended maintenance windows on a GIS map in real time. Specifically, the operation and maintenance decision-making module is deployed in the cloud and includes a construction progress interface unit, a maintenance-construction coupling optimization unit, a spare parts and personnel scheduling unit, and a visualization unit. The construction progress interface unit connects to the project management system through a standard data interface, automatically reads and parses the construction schedule data of the entire line, identifies the start and end mileage, time interval, critical path, and logical relationship between processes for each welding operation, and establishes a space-time construction network model. The maintenance-construction coupling optimization unit aims to minimize the expected total welding delay caused by maintenance downtime. It integrates equipment failure probability, construction space constraints, and resource allocation information to construct a time-space coupled maintenance decision model, uses a genetic algorithm to solve it efficiently, and outputs the optimal maintenance time window, spare parts list, personnel path, and delivery plan. The spare parts and personnel scheduling unit connects to the enterprise ERP inventory system, automatically generates spare parts outbound instructions, transport orders, and maintenance personnel work orders, and pushes them synchronously to relevant execution terminals. The visualization unit is based on a GIS geographic information platform and displays equipment distribution, health status heatmap, failure risk level, and recommended maintenance window on an electronic map in real time. It supports zooming, positioning, querying, and path planning, providing managers with an intuitive and comprehensive operation and maintenance command interface.

[0029] The communication module, embedded between the edge computing module and the digital twin module, provides an asynchronous data synchronization mechanism based on breakpoint resumption. Specifically, the communication module is embedded between the edge computing module and the digital twin module, employing a dual-link redundancy design with 4G / 5G as the primary and satellite communication as a secondary, ensuring stable data transmission in complex outdoor environments. The module has a built-in breakpoint resume protocol, which packages degraded feature data into fixed-size fragments, each with a checksum and breakpoint marker. After a network interruption, the transmission position is automatically recorded, and the upload resumes from the breakpoint when the connection is restored, avoiding duplicate transmissions and data loss. The communication process uses the AES-256 encryption algorithm to encrypt the data, while enabling two-way authentication and data integrity verification mechanisms to prevent data tampering, leakage, or unauthorized access. Edge-side cached data is cyclically overwritten in chronological order, storing the most recent 7 days of original data and feature data. After the network is restored, data from the network outage period is automatically retransmitted in batches, ensuring data consistency between the edge and the cloud, supporting real-time updates of the digital twin model and accurate and reliable operation and maintenance decisions.

[0030] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for operating and maintaining long-distance pipeline construction equipment, characterized in that, Includes the following steps: Step 1: Collect real-time equipment operating status data using multi-source sensors deployed on the construction equipment. The operating status data includes at least vibration signals, temperature, hydraulic pressure, oil particle contamination level, and the real-time geographical location of the equipment. Step 2: Using the edge computing nodes set up at the construction unit's site, preprocess the collected operating status data and extract degradation features to generate equipment degradation feature vectors with time tags; Step 3: When establishing a connection between the edge computing node and the cloud digital twin server, the degradation feature vector is synchronized to the digital twin server through a breakpoint resume method to update the degradation state parameters in the device digital twin model. Step 4: The digital twin model is based on a stochastic degradation process. It performs Bayesian state estimation on each device and predicts the probability density function of the device's remaining useful life and the cumulative probability of failure in the future. Step 5: Obtain the pipeline construction schedule and spatial constraints of each process. With the goal of minimizing the expected total delay of the entire welding operation caused by equipment maintenance downtime, use the accumulated failure probability obtained in step S4 to construct a time-space coupled maintenance decision model and solve for the optimal maintenance start time window and the corresponding maintenance resource scheduling scheme. Step 6: Send the optimal maintenance start time window and maintenance resource scheduling plan to the maintenance personnel's terminal and execute the maintenance operation.

2. The operation and maintenance method for long-distance pipeline construction equipment according to claim 1, characterized in that, The degradation feature vectors in the second step include vibration time-domain kurtosis, spectrum envelope energy, hydraulic pressure drop rate, and wear index based on oil particle contamination, wherein the wear index is calculated by the following formula: ； wherein is the number of particles having a size between 15 and 25 pm per ml of oil, is the number of particles having a size greater than 25 pm, is the total number of particles, the weighting factor , and satisfies .

3. The operation and maintenance method for long-distance pipeline construction equipment according to claim 1, characterized in that, In step four, the digital twin model uses a Wiener process with drift to describe the equipment degradation. : ； in, The drift coefficient, Where is the diffusion coefficient. This is standard Brownian motion; When the current time is obtained Degradation characteristics observation sequence Then, the drift coefficients are updated online using Kalman filtering or particle filtering. The posterior distribution is obtained, and its posterior mean is derived. With variance Let the failure threshold be... Then the remaining service life The probability density function is: ； Future period Cumulative probability of internal equipment failure for: 。 4. The operation and maintenance method for long-distance pipeline construction equipment according to claim 3, characterized in that, The objective function of the time-space coupled maintenance decision model in step five is: ; in, The number of welding work surfaces affected by maintenance. For the first Delay weighting for each work area The maintenance start time is the time that caused the welding delay at this work site. Time of arrival of maintenance resources and spare parts delivery routes The function; The constraints include: from the current decision time Up to the planned maintenance start time The cumulative probability of failure.

5. A long-distance pipeline construction equipment operation and maintenance system, using the long-distance pipeline construction equipment operation and maintenance method as described in claims 1-4, characterized in that, include: The data acquisition module is used to acquire multi-source operating status data and location information of construction equipment in real time; The edge computing module is deployed at the site of each construction unit and connected to the data acquisition module. It is used to preprocess the acquired data and calculate degradation features to generate degradation feature vectors, and has a local real-time early warning function. The digital twin module, deployed on a cloud server, communicates with the edge computing module via an intermittent network. It is used to update the device degradation digital twin model using the degradation feature vector and predict the remaining service life and failure probability of the device based on stochastic process theory. The operation and maintenance decision module, connected to the digital twin module, is used to integrate pipeline construction progress data and geospatial information to generate maintenance start time windows and resource scheduling schemes with the goal of minimizing disturbance to continuous welding operations. The communication module, embedded between the edge computing module and the digital twin module, provides an asynchronous data synchronization mechanism based on breakpoint resumption.

6. The operation and maintenance system for long-distance pipeline construction equipment according to claim 5, characterized in that, The data acquisition module specifically includes: The vibration signal acquisition unit is installed on the engine main bearing housing and the hydraulic pump housing to acquire triaxial vibration acceleration. Temperature acquisition unit is used to monitor engine coolant temperature, hydraulic oil temperature and ambient temperature; The pressure acquisition unit is used to measure the pressure of each circuit in the hydraulic system in real time. The oil quality acquisition unit has a built-in particle counter and moisture sensor to output the oil particle contamination level and water content. The GPS / BeiDou positioning unit is used to obtain the device's real-time coordinates and movement trajectory.

7. The operation and maintenance system for long-distance pipeline construction equipment according to claim 5, characterized in that, The edge computing module specifically includes: The data preprocessing unit is used to denoise, remove invalid values, and align the time of the raw sensor signals. The time-frequency feature extraction unit is used to calculate the envelope spectrum, wavelet energy, and slope change of the hydraulic pressure of the vibration signal. The degradation feature synthesis unit is used to fuse multiple features into the degradation feature vector and calculate the real-time wear index according to the wear index formula. The local threshold comparison unit is used to compare the degradation feature vector with the preset multi-level alarm threshold. When the threshold is exceeded, the unit's audible and visual alarm is triggered directly.

8. The operation and maintenance system for long-distance pipeline construction equipment according to claim 5, characterized in that, The digital twin module specifically includes: The equipment model building unit creates a digital image of each construction piece of equipment, including physical properties, degradation equations, and observation equations. The Bayesian state estimation unit executes Kalman filtering or particle filtering algorithms to update the posterior distribution of drift coefficients online. The remaining lifetime prediction unit uses the remaining lifetime probability density function formula to calculate and output the remaining lifetime distribution and the cumulative failure probability in real time. The model synchronization unit is responsible for importing the accumulated degradation feature vectors from the edge side in batches each time a connection is successfully established, and for updating the state of the digital twin model.

9. The operation and maintenance system for long-distance pipeline construction equipment according to claim 5, characterized in that, The operation and maintenance decision module specifically includes: The construction progress interface unit is used to acquire and parse pipeline construction schedule data; The maintenance-construction coupled optimization unit solves the objective function and outputs the maintenance time window that minimizes the expected total welding delay, the required number of spare parts types, and the optimal route for maintenance personnel. The spare parts and personnel scheduling unit automatically generates spare parts outbound instructions and maintenance personnel work orders based on the optimization results. The visualization unit displays the thermal distribution of equipment health status, predicted fault risk levels, and recommended maintenance windows on a GIS map in real time.

10. The operation and maintenance system for long-distance pipeline construction equipment according to claim 9, characterized in that, The construction progress interface unit identifies the start and end mileage, time interval, and logical relationship between key processes for each welding operation surface.

Images