A deep water underwater gas production tree safety monitoring and early warning method and system

Abstract

This invention discloses a method and system for safety monitoring and early warning of deep-water subsea gas production trees. The early warning method quantifies three indicators: peak-to-average ratio, energy operator, and steady-state deviation, and uses Bayesian inference to calculate the probability of risks occurring in gas production tree components and the system. The deep-water subsea gas production tree safety monitoring and early warning system includes a gas production tree data acquisition and transmission unit, an underwater data processing unit, a risk assessment and early warning unit, and an early warning display unit. This invention can fuse multi-source features and perform transient analysis, capturing weak transient signals and multi-parameter coupled changes that characterize early failures, achieving early warning of potential risks. By constructing a dynamic risk matrix and introducing trend prediction and evolution analysis, the diagnostic dimensions are more comprehensive.

Description

Technical Field

[0001] This invention belongs to the field of marine engineering, specifically relating to a method and system for safety monitoring and early warning of deep-sea subsea gas production trees. Background Technology

[0002] The deepwater subsea gas production tree is one of the most crucial pieces of equipment in a subsea production system, playing a vital role in well control, pressure regulation, and production distribution. Its safe operation directly determines the output of deep-sea gas fields and the safety of personnel and equipment. With technological advancements, subsea gas production tree control methods have evolved from traditional hydraulic control to more advanced electro-hydraulic hybrid control. Compared to traditional subsea hydraulically controlled gas production trees, subsea electro-hydraulic hybrid control gas production trees offer the following advantages: compact structure and light weight; combining the high output capacity of hydraulic actuators with the precise control and two-way communication capabilities of an electro-hydraulic system; enabling real-time and ultra-long-distance control; better deep-water adaptability; and easier maintenance and upgrades of the control system.

[0003] The submersible electro-hydraulic combined gas production tree (DBM) control system ensures the normal operation of the DBM. However, current safety monitoring and early warning methods for DBMs primarily rely on passive alarm mechanisms that set fixed thresholds for single physical quantities. These mechanisms struggle to capture transient characteristics and multi-parameter coupled changes that characterize early-stage faults, lacking the ability to predict fault evolution trends and resulting in delayed warnings. Furthermore, safety conditions such as valve leakage, control signal interruption, or seal failure can lead to the loss of function or structural damage to critical equipment in the DBM, potentially threatening the lives of surface workers through a chain reaction. Therefore, a deep-water submersible DBM safety monitoring and early warning system is essential. Summary of the Invention

[0004] This invention is proposed to solve the problems existing in the prior art, and its purpose is to provide a method and system for safety monitoring and early warning of deep-water subsea gas production trees.

[0005] This invention is achieved through the following technical solution: A method for safety monitoring and early warning of deep-water subsea gas production trees includes the following steps: S1. Collect raw sensor data from the deep-water subsea gas production tree; S2. Preprocess the raw sensor data of the deep-water subsea gas production tree collected in step S1 to obtain standardized data; S3. A parallel computing architecture is adopted to confirm the integrity of the data, obtain continuous data from the sensor, and extract key features from the standardized data. S4. Based on the key sensor features extracted in step S3, calculate the transient change difference and steady-state deviation of each sensor, and normalize the transient change difference and steady-state deviation to obtain the comprehensive transient difference. S5. Calculate the probability of a risk occurring to the deep-water subsea gas production tree component; S6. Based on the posterior probability of safety risk of the deep-water subsea gas production tree component calculated in step S5, compare the posterior probability of safety risk of the deep-water subsea gas production tree component with the preset risk threshold to determine the degree of risk of the deep-water subsea gas production tree component. S7. Based on the degree of risk of the deep-water subsea gas production tree component obtained in step S6, construct a dynamic risk map, and combine it with the safety operation specifications to comprehensively determine the overall risk level, set three-level early warning trigger rules, and carry out differentiated early warning output.

[0006] In the above technical solution, the preprocessing in step S2 includes amplifying and filtering the raw sensor data, calibrating the switching quantity, and aligning the data from different sensors.

[0007] In the above technical solution, step S3 specifically includes the following steps: S31, Set the sliding time window [T] k ,T k+1 The standardized data obtained in step S2 is divided into multiple independent data segments according to a time window, and each data segment is distributed in parallel to multiple computing nodes of the distributed architecture corresponding to the parallel computing architecture. S32. Within each computing node, independently verify the continuity of timestamps and the validity of values ​​for the allocated data segments. S33. If the computing node detects missing values ​​or abnormal null values ​​exceeding physical limits in the data segment, it performs anomaly type determination; and repairs occasional anomalies using linear interpolation. The repair formula for the linear interpolation method is as follows: In the formula: i is the number of the i-th sensor in the deep-water subsea gas production tree monitoring system; t m The target sampling time where sporadic anomalies exist; t m-1 For the target sampling time t m Previously, distance t m The most recent valid sampling time; t m+1 For the target sampling time t m After that, distance t m The most recent valid sampling time; x i (t m (t) represents time t m Valid data after linear interpolation repair; x i (t m-1 (t) represents time t m-1 Valid sampled data; xi (t m+1 (t) represents time t m+1 Valid sampling data; S34. Each computing node performs multi-threaded concurrent processing on the verified complete data segment to calculate the current time window [T]. k ,T k+1 Key features of the i-th type of sensor; The key features include local average values ​​and local absolute peak values; The specific method for calculating the local average value is as follows: Let the sliding time window [T] be... k ,T k+1 Within the range, if the i-th type of sensor contains N discrete sampling points, then the sliding time window [T]... k ,T k+1 The formula for calculating the local average value within [ ] is: In the formula: For sliding time window [T k ,T k+1 Local average value within ]; N represents the valid sensor data within the sliding time window; N is the sliding time window [T] k ,T k+1 Within the sliding time window, j represents the total number of discrete valid sampling points of the i-th sensor; j is the index of the discrete sampling point within the sliding time window. Sliding time window [T k ,T k+1 The formula for calculating the local absolute peak value within the area is: In the formula: P i (T k ) represents the sliding time window [T] k ,T k+1 Local absolute peak value within ]; x i (t j (t) represents time t j Sensor data; [T k ,T k+1 [This represents a sliding time window;] S35. Collect the local average value and local absolute peak value output by each computing node through the aggregation node, as well as the sensor data segments verified in step S32. Splice and reassemble them according to the original time series, and output a high-dimensional time series containing complete signal features as continuous sensor data. The continuous sensor data is used as the input benchmark for the comprehensive transient calculation in step S4.

[0008] In the above technical solution, the criterion for determining the anomaly type in step S32 is as follows: If the number of consecutive outliers is ≤3 and the adjacent sampling points are valid data, they are marked as occasional outliers and repaired using linear interpolation. If the number of consecutive outliers exceeds 3, it is marked as a fault, interpolation repair is not performed, and the abnormal data segment is removed.

[0009] In the above technical solution, in step S4, the transient change difference uses the peak-to-average ratio to quantify the amplitude of sensor data change and the energy operator to quantify the degree of transient energy concentration. The formula for calculating the peak-to-average ratio is: In the formula: R i (t) represents the peak-to-average ratio of the i-th type of sensor within the time window at time t; x i (t) represents the real-time sampled value of the i-th type of sensor; [T k ,T k+1 [This represents a sliding time window;] i (T k ) represents the i-th type of sensor in window [T] k ,T k+1 The average value within ]; It is a very small positive number. ; The calculation formula for the energy operator is: In the formula: x is the energy operator value of the i-th type of sensor at time t; i (t) represents the real-time sampled value of the i-th type of sensor; Let be the sampling period of the i-th type of sensor; The formula for calculating the steady-state deviation is: In the formula, D i (t) represents the average deviation of points within the selected time period; S i (t k B represents the valid sampled value corresponding to the k-th sampling time selected from the continuous data of the sensor; i (t k ) represents the i-th sensor at t k The continuous baseline value corresponding to the time.

[0010] In the above technical solution, the formula for calculating the comprehensive instantaneous difference is: In the formula, TDI i (t) represents the composite instantaneous difference value. To measure the weight of the peak-to-average ratio, To measure the weight of the energy operator, The weight used to measure steady-state deviation; R i (t) represents the peak-to-average ratio of the time window in which the i-th type of sensor is located at time t; D represents the energy operator value of the i-th type of sensor at time t; i (t) represents the average deviation of points within the selected time period.

[0011] In the above technical solution, the formula for calculating the posterior probability of a safety risk occurring in the deep-water subsea gas production tree component is as follows: In the formula, Given the combined transient difference (TDI) of all n sensors, this represents the posterior probability of a safety risk occurring in the target monitoring component of the deep-water subsea gas production tree. The TDI represents the combined transient error of the i-th sensor at time t under the condition that a safety risk occurs in the target monitoring component of the deep-water subsea gas production tree. i The conditional probability; This indicates that, under the condition that the target monitoring component of the deep-water subsea gas production tree is in a risk-free and safe state, the comprehensive instantaneous difference value of the i-th sensor at time t is TDI. i The conditional probability; G represents the prior probability of a safety risk occurring in the target monitoring component of the deep-water subsea gas production tree; G represents the Boolean value of the component's risk occurrence. G=1 indicates a safety risk to the deepwater subsea gas production tree target monitoring component, while G=0 indicates that the deepwater subsea gas production tree target monitoring component is in a safe, risk-free state; TDI i This represents the combined instantaneous difference value of the i-th sensor at time t.

[0012] In the above technical solution, the threshold is obtained through expert decision-making; When the posterior probability of a safety risk is greater than the threshold, the probability of the component being at risk is considered high, and it is considered to be in a risky state. When the posterior probability of a security risk occurs When the threshold is reached, the probability of the component being at risk is determined to be low, and it is considered to be in a safe or low-risk state.

[0013] In the above technical solution, the three-level early warning triggering rule is as follows: When there are only 1 to 2 risk components in the risk map, it is identified as a potential risk and a level 1 warning is triggered. When 3 to 4 risk components appear in the risk map, it is judged as an abnormal risk, triggering a level 2 warning, activating a local alarm, and simultaneously recording the risk development curve and related signal change data. When five or more risky components appear in the risk map, it is judged as a failure risk, triggering a level three warning. While the local audible and visual alarms continue, the three-dimensional risk map, real-time monitoring data and fault prediction information are uploaded to the remote operation and maintenance platform and pushed to the operation and maintenance personnel's terminal.

[0014] A deep-water subsea gas production tree safety monitoring and early warning system based on the aforementioned method includes a gas production tree data acquisition and transmission unit, an underwater data processing unit, a risk assessment and early warning unit, and an early warning display unit. The gas-producing tree data acquisition and transmission unit is connected to the underwater control unit via a signal cable; the underwater data processing unit is connected to the gas-producing tree data acquisition and transmission unit via a communication cable; the risk assessment and early warning unit is connected to the underwater data processing unit via a communication cable; and the early warning display unit is connected to the risk assessment and early warning unit via a signal cable. The gas-gathering tree data acquisition and transmission unit includes a data acquisition and statistics module, a data statistics preprocessing module, a data storage module, and a data transmission module; the data acquisition and statistics module is connected to the control signal processing module via a signal cable; the data statistics preprocessing module is connected to the data acquisition and statistics module via a signal cable; the data storage module is connected to the data statistics preprocessing module via a signal cable; and the data transmission module is connected to the data statistics preprocessing module via a signal cable. The underwater data processing unit includes a signal receiving module, an acquisition timing coordination module, a signal conditioning and shaping module, a continuous and transient dual-channel generation module, a continuous baseline extraction module, a transient detection and feature extraction module, a feature alignment and standardization module, a transient analysis module, a historical data recording module, and a data output module. The signal receiving module is connected to the data transmission module via a signal cable. The acquisition timing coordination module is connected to the signal receiving module via a signal cable. The signal conditioning and shaping module is connected to the acquisition timing coordination module via a signal cable. The continuous and transient dual-channel generation module is connected to the signal conditioning and shaping module via a signal cable. The continuous baseline extraction module is connected to the continuous and transient dual-channel generation module via a signal cable. The transient detection and feature extraction module is connected to the continuous and transient dual-channel generation module via a signal cable. The feature alignment and standardization module is connected to both the continuous baseline extraction module and the transient detection and feature extraction module via signal cables. The transient analysis module is connected to the feature alignment and standardization module via a signal cable. The historical data recording module is connected to the transient analysis module via a signal cable. The data output module is connected to the transient analysis module via a signal cable. The risk assessment and early warning unit includes a multi-source feature fusion module, a trend prediction and evolution analysis module, a dynamic risk matrix construction module, a hierarchical early warning triggering module, a potential risk alert module, an abnormal trend warning module, a failure risk alarm module, an early warning feedback and strategy optimization module, and a risk scoring module. The multi-source feature fusion module is connected to the data output module via signal cables; the trend prediction and evolution analysis module is connected to the multi-source feature fusion module via signal cables; the dynamic risk matrix construction module is connected to the multi-source feature fusion module via signal cables; the hierarchical early warning triggering module is connected to both the trend prediction and evolution analysis module and the dynamic risk matrix construction module via signal cables; the potential risk alert module is connected to the hierarchical early warning triggering module via signal cables; the abnormal trend warning module is connected to the hierarchical early warning triggering module via signal cables; the failure risk alarm module is connected to the hierarchical early warning triggering module via signal cables; the early warning feedback and strategy optimization module is connected to the potential risk alert module, the abnormal trend warning module, and the failure risk alarm module via signal cables; and the risk scoring module is connected to the early warning feedback and strategy optimization module via signal cables. The early warning display unit includes a core early warning information display module, a data visualization module, a response measure guidance module, and a historical early warning retrospective module. The core early warning information display module is connected to the risk scoring module via a signal cable; the data visualization module is connected to the risk scoring module via a signal cable; the response measure guidance module is connected to the risk scoring module via a signal cable; and the historical early warning retrospective module is connected to the risk scoring module via a signal cable.

[0015] The beneficial effects of this invention are: This invention provides a method and system for safety monitoring and early warning of deep-water subsea gas production trees. It can fuse multi-source features and perform transient analysis, capture weak transient signals and multi-parameter coupled changes that characterize early failures, and achieve early warning of potential risks. By constructing a dynamic risk matrix and introducing trend prediction and evolution analysis, the diagnostic dimensions are more comprehensive. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method for safety monitoring and early warning of deep-water subsea gas production tree according to the present invention; Figure 2 This is a system composition diagram of the deep-water subsea gas production tree to which this invention is applied; Figure 3 This is a system composition diagram of the deep-water subsea gas production tree safety monitoring and early warning system of the present invention.

[0017] in: 101. Surface Control Unit; 102. Uninterruptible Power Supply Module; 103. Power Supply Module; 104. Logic Control Module; 105. Communication and Network Module; 106. Data Processing Module; 107. Electro-hydraulic Integration Module; 108. Human-Machine Interface Module; 109. Surface Hydraulic Control Module; 110. Deepwater Subsea Gas Production Tree; 111. Subsea Control Unit; 112. Subsea Hydraulic Control Module; 113. Natural Gas Hydrate Inhibitor Valve; 114. Chemical Injection Metering Valve; 115. Natural Gas Hydrate Inhibitor Isolation Valve; 116. Methanol Injection Valve; 117. Methanol Injection Isolation Valve; 118. Scale Inhibitor Valve; 119. Scale Inhibitor Isolation Valve; 120. Deepwater Subsea Gas Production Tree Body; 121. Deepwater Subsea Gas Production Tree 122. Packer test valve, 123. Annular isolation valve, 124. Annular main valve, 125. Annular vent valve, 126. Annular wing valve, 127. Surface control downhole safety valve, 128. Downhole chemical injection valve, 129. Wellhead connection module, 130. Wellhead module, 131. Production wing valve, 132. Production isolation valve, 133. Wet gas flow meter, 134. Production outlet module, 135. Production main valve, 136. Production throttle valve, 137. Workover valve, 138. Vent monitoring valve, 139. Switching valve, 140. Natural gas hydrate inhibitor valve control module, 141. Methanol injection valve control module, 142. Scale inhibitor valve control module, 143. Annular wing valve control module, 144. 145. Annular Isolation Valve Control Module; 146. Annular Main Valve Control Module; 147. Workover Valve Control Module; 148. Production Wing Valve Control Module; 149. Switching Valve Control Module; 150. Production Main Valve Control Module; 151. Control Signal Processing Module; 152. Communication and Data Acquisition Module; 153. Electro-hydraulic Distribution Module; 154. Power Distribution Module; 201. Signal Receiving Module; 202. Acquisition Timing Coordination Module; 203. Signal Conditioning and Shaping Module; 204. Continuous and Transient Dual-Channel Generation Module; 205. Subsea Data Processing Unit; 206. Continuous Baseline Extraction Module; 207. Transient Detection and Feature Extraction Module; 208. Feature Alignment and Normalization Module; 209. Transient Analysis Module; 210. The system includes the following modules: 1. Historical data recording module; 2. Data output module; 2. Multi-source feature fusion module; 2. Trend prediction and evolution analysis module; 2. Dynamic risk matrix construction module; 2. Hierarchical early warning triggering module; 2. Potential risk alert module; 2. Abnormal trend warning module; 2. Failure risk alarm module; 2. Early warning feedback and strategy optimization module; 2. Risk assessment and early warning unit; 2. Risk scoring module; 2. Core early warning information display module; 2. Data visualization module; 2. Response measure guidance module; 2. Historical early warning retrospective module; 2. Early warning display unit; 2. Data acquisition and statistics module; 2. Data statistics preprocessing module.229. Data storage module; 230. Data transmission module; 231. Data acquisition and transmission unit.

[0018] For those skilled in the art, other related figures can be obtained from the above figures without any creative effort. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0020] The invention is applied to deep-water subsea gas production trees, such as... Figure 2 As shown, the system includes a wellhead module 130 located below the seabed, a deepwater subsea gas production tree 110 installed above the wellhead module 130, a surface control downhole safety valve 127 installed in the downhole pipeline, a downhole chemical injection valve 128 installed in the downhole pipeline, and a surface control unit 101 installed on the surface platform. The deepwater subsea gas production tree 110 is located above the seabed. The wellhead module 130 is connected to the deepwater subsea gas production tree 110 via a tubing hanger, a sealing system, and a connector to guide the production gas below the seabed. The deepwater subsea gas production tree 110 is connected to the surface control unit 101 via an umbilical cable for deepwater subsea gas production. The valves of the subsea gas production tree 110 are controlled to open and close, completing the subsea gas production process, operation, and maintenance. The wellhead module 130 guides the production gas to the deepwater subsea gas production tree 110. Under the control of the surface control unit 101, the deepwater subsea gas production tree 110 delivers the production gas to the platform above the sea surface. The surface-controlled downhole safety valve 127 is connected to the wellhead module 130 through a pipeline and is used to close the wellbore in an emergency to ensure safe production. The downhole chemical injection valve 128 is connected to the surface-controlled downhole safety valve 127 through a pipeline and is used to precisely control the chemical injection process to solve the production risks caused by changes in the physicochemical properties of the wellbore fluid during the extraction and transportation process.

[0021] The deepwater subsea gas production tree 110 includes a wellhead connection module 129 installed at the wellhead, a deepwater subsea gas production tree body 120 connected to the wellhead connection module 129, a subsea control unit 111 connected to various valves, a moisture flow meter 133 installed in the deepwater subsea gas production tree production loop, a production throttle valve 136 installed in the deepwater subsea gas production tree production loop, a production isolation valve 132 installed in the deepwater subsea gas production tree production loop, a production outlet module 134 installed at the terminal of the deepwater subsea gas production loop, and a module installed in the annular space return loop. The circuit includes a packer test valve 122, an annular vent valve 125 installed in the annular space circuit, a natural gas hydrate inhibitor valve 113 installed in the chemical injection circuit, a chemical injection metering valve 114 installed in the chemical injection circuit, a natural gas hydrate inhibitor isolation valve 115 installed in the chemical injection circuit, a methanol injection valve 116 installed in the chemical injection circuit, a methanol injection isolation valve 117 installed in the chemical injection circuit, a scale inhibitor valve 118 installed in the chemical injection circuit, and a scale inhibitor isolation valve 119 installed in the chemical injection circuit.

[0022] The wellhead connection module 129 includes a guide base fixedly connected to the upper connecting flange of the wellhead module 130, a tree body connecting flange disposed on the guide base, a locking mechanism, a sealing assembly, and a fluid channel assembly; the guide base provides a guiding reference for the lowering and docking of the gas production tree; the tree body connecting flange is connected to the bottom flange of the gas production tree body 120 by bolts; the locking mechanism is a hydraulically driven locking ring structure used to circumferentially clamp and axially pre-tighten the wellhead after the gas production tree is in place; the sealing assembly is a metal sealing ring disposed at the connection interface, used to ensure deep sealing. The system maintains a seal between the wellhead and the gas production tree under high water pressure. The fluid channel assembly includes a riser interface and an inner cavity channel communicating with the production casing, annular channel, and chemical injection channel, enabling fluid communication between the fluid inside the wellbore and the production circuit of the deepwater subsea gas production tree. The structure of the wellhead connection module 129 is a conventional gas production tree wellhead connection device structure in this field. The connection interface between the wellhead connection module 129 and the wellhead module 130 is the upper connecting flange of the upper shell of the wellhead module 130, i.e., the load-bearing and sealing connection part located at the top of the wellhead module 130. The lower end of the guide base of the wellhead connection module 129 engages with the positioning step on the outer surface of the upper shell of the wellhead module 130 via a positioning shoulder. The lower end of its tree body connecting flange is locked to the outer cone of the upper connecting flange of the wellhead module 130 via a locking mechanism, and a metal sealing ring is provided between the two to form a structural load-bearing and sealing interface between the wellhead module 130 and the deepwater subsea gas production tree body 120. The deep-water subsea gas production tree body 120 is connected to the wellhead connection module 129 via a production pipeline for guiding and controlling the flow of production gas into the production loop. The subsea control unit 111 is connected to the deep-water subsea gas production tree body 120, the natural gas hydrate inhibitor valve 113, the methanol injection valve 116, and the scale inhibitor valve 118 via cables and hydraulic lines, respectively, for controlling the opening and closing of each valve. The moisture flow meter 133 is connected to the deep-water subsea gas production tree body 120 via a production pipeline for measuring the real-time flow rate of the water-containing production gas flowing out of the gas production tree. The production throttle valve 136 is connected to the moisture flow meter 133 via a production pipeline. The production gas control module 133 is connected to the production isolation valve 136 via the production pipeline, and is used to safely disconnect the production gas and ensure the safety of other modules during special operations. The production outlet module 134 is connected to the production isolation valve 132 via the production pipeline, and is used to transport the production gas in the production loop to the subsea manifold. The packer test valve 122 is connected to the annular space in the wellhead connection module 129 via a pipeline, and is used for pressure and sealing tests at the bottom of the gas production tree. The annular vent valve 125 is connected to the packer test valve 122 and the vent monitoring valve 138 via a pipeline. The following valves are connected: a natural gas hydrate inhibitor valve 113, which is connected to the production throttle valve 136 via a pipeline, and a chemical injection metering valve 114, which is connected to the natural gas hydrate inhibitor valve 113 via a pipeline, and a chemical injection metering valve 115, which is connected to the chemical injection metering valve 114 via a pipeline, and a methanol injection valve 116, which is connected to the production throttle valve 136 ..., and a methanol injection valve 116, which is connected to the production throttle valve 136 via a pipeline, and a methanol injection valve 116, which is connected to the production throttle valve 136, and a methanol injection valve 116, which is connected to the production throttle valve 136, and a methanol injection valve 116, which is connected to the production throttle valve 136, and a methanol injection valve 116, which is connected to the production throttle valve 136, and a methanol injection valve 116, and a methanol injection valve 115, which is connected to the production throttle valve 136 via a pipeline, and a methanol injection valve 116, which is connected to the production throttle valve 136, and a methanol injection valve 116, and a methanol injection valve 115, which is connected to the production throttle valve 136, and a methanol injection valve 116, and a methanol injection valve 116, and a methanol injection valve 116, and a methanol injection valve The process involves the following steps: Chemical injection metering valve 114 is connected to methanol injection valve 116 via a pipeline to control the methanol injection flow rate; methanol injection isolation valve 117 is connected to chemical injection metering valve 114 via a pipeline to control and isolate the methanol injection pipeline; scale inhibitor valve 118 is connected to the production pipeline via a pipeline to regulate and control scale inhibitor injection, prevent scaling, and ensure unobstructed production gas flow; chemical injection metering valve 114 is connected to scale inhibitor valve 118 via a pipeline to control the scale inhibitor injection flow rate; and scale inhibitor isolation valve 119 is connected to chemical injection metering valve 114 via a pipeline to control and isolate the scale inhibitor injection pipeline. The deepwater subsea gas production tree body 120 includes a production main valve 135 installed at the inlet of the deepwater subsea gas production tree production loop, a production wing valve 131 installed at the outlet of the deepwater subsea gas production tree production loop, a switching valve 139 installed in the deepwater subsea gas production tree switching loop, an annular main valve 124 installed at the inlet of the deepwater subsea gas production tree annular space loop, an annular wing valve 126 installed at the outlet of the deepwater subsea gas production tree annular space loop, a workover valve 137 installed in the annular space loop, an annular isolation valve 123 installed in the annular space loop, a vent monitoring valve 138 installed in the annular space loop, and a sealing valve installed at the upper outlet of the production pipeline. A deep-water subsea gas production tree cap 121 is provided at the upper outlet of the production pipeline. The deep-water subsea gas production tree cap 121 includes a cap body, a sealing plug, a locking mechanism, a sealing ring, and a connecting interface. The cap body is connected to the upper outlet of the gas production pipeline through the locking mechanism. The sealing plug is inserted into the production channel and forms a pressure seal with the inner wall of the channel through the sealing ring, which is used to close the production channel when the gas production tree is shut down, workover, or under testing conditions. The connecting interface is used to dock with offshore operating equipment during installation or disassembly. The structure of the deep-water subsea gas production tree cap 121 is a conventional subsea gas production tree cap structure in the art and is not limited in this invention. The production outlet module 134 includes an outlet flange, an outlet flow channel, a one-way valve assembly, and a detection interface. The outlet flange is used to achieve mechanical connection and sealing with the inlet end of the underwater manifold. The outlet flow channel is used to guide the production gas output from the production isolation valve 132 into the underwater manifold. The one-way valve assembly is located in the outlet flow channel to prevent fluid backflow into the gas-producing tree when pressure fluctuations or back pressure conditions occur on the underwater manifold side. The detection interface is used to install pressure, temperature, or flow monitoring sensors at the outlet end to monitor the outlet conditions. The specific structural form of the production outlet module 134 belongs to the conventional gas-producing tree outlet connection structure in this field. The main production valve 135 is connected to the production pipeline in the wellhead connection module 129 via a production pipeline, and is used to guide and control the flow of production gas into the production loop. The production wing valve 131 is connected to the main production valve 135 via a production pipeline, and is used to guide and control the flow of production gas into the subsea manifold. The switching valve 139 is connected to the pipeline between the main production valve 135 and the production wing valve 131 via a pipeline, and is also connected to the pipeline between the annular main valve 124 and the annular wing valve 126 via a pipeline, and is used to control the flow of gas in the annular space loop to the production loop. The annular main valve 124 is connected to the wellhead connection module 129 via a pipeline, and is used to control the outflow of gas from the annular space and regulate the pressure within the annular space. The annular wing valve 126 is connected to the annular main valve 124 via a pipeline, and is used to control the connection and isolation between the annular space loop and other modules. The workover valve 137 is connected to the pipeline in the wellhead connection module 129 via a pipeline, and is used for workover and inspection. Safety isolation during special operations such as repair; the annular isolation valve 123 is connected to the workover valve 137 via a pipeline for complete gas isolation between the annular space and the production pipeline; the vent monitoring valve 138 is connected to the wellhead connection unit 129 via a pipeline for monitoring, flow control and safety isolation of production gas during venting; the pipeline in the wellhead connection module 129 is not a new independent component, but an internal cavity channel in the fluid channel assembly within the wellhead connection module 129 connected to different circuits; the production main valve 135 and the production wing valve 131 are connected to the production channel in the wellhead connection module 129 connected to the production casing via the production pipeline; the annular main valve 124, the annular wing valve 126, the workover valve 137 and the annular isolation valve 123 are connected to the annular channel and branch channel in the wellhead connection module 129 connected to the annular pipeline; and the vent monitoring valve 138 is connected to the internal cavity channel in the wellhead connection module 129 connected to the annular vent branch. The underwater control unit 111 includes an underwater hydraulic control module 112, a control signal processing module 150, a communication and data acquisition module 151, an electro-hydraulic distribution module 152, a power distribution module 153, a natural gas hydrate inhibitor valve control module 140, a methanol injection valve control module 141, a scale inhibitor valve control module 142, an annular wing valve control module 143, an annular isolation valve control module 144, an annular main valve control module 145, a workover valve control module 146, a production wing valve control module 147, a switching valve control module 148, and a production main valve control module 149. The underwater hydraulic control module 112 includes a hydraulic pump, hydraulic valves, pressure sensors, flow regulators, hydraulic pipelines, and seals. The hydraulic pump provides the required hydraulic energy; the hydraulic valves regulate the flow rate and direction of the fluid; the pressure sensors monitor the working pressure of the hydraulic system in real time; the flow regulators adjust the flow rate according to control signals; the hydraulic pipelines connect various hydraulic actuators; and the seals ensure the mechanical sealing of the hydraulic system. The control signal processing module 150 includes a signal processor, signal amplifier, filter, analog-to-digital converter, digital-to-analog converter, and communication interface. The signal processor receives and processes signals transmitted from the electro-hydraulic integrated module 107 and sensors; the amplifier and filter are used for... The signal is filtered and amplified; the analog-to-digital converter converts the analog signal into a digital signal; the digital-to-analog converter converts the digital signal into an analog signal; the communication interface is used to transmit the processed signal to the communication and network module 105; the communication and data acquisition module 151 includes a data acquisition unit, a data storage device, and a communication interface; the data acquisition unit is used to acquire real-time data from the sensor; the data storage device is used to store the acquired data; the communication interface transmits the data to the electro-hydraulic distribution module 152 via a wired connection; the electro-hydraulic distribution module 152 includes a power supply, a distribution valve, a regulating valve, and hydraulic pipelines; the power supply module provides the required electrical energy for the entire electro-hydraulic system; the distribution valve and regulating valve are used to precisely control the direction and flow rate of the fluid; the hydraulic pipelines transmit the fluid... The power is transmitted to the execution unit; the power distribution module 153 includes a power converter, a regulated power supply, a distribution circuit, and a protection device; the power converter converts the external power supply into the required voltage; the regulated power supply ensures voltage stability; the distribution circuit distributes power to each module; the protection device ensures safe system operation and avoids overload and short circuit; the natural gas hydrate inhibitor valve control module 140 includes an electric actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor; the electric actuator opens and closes the natural gas hydrate inhibitor valve 113 according to the control signal; the control signal input terminal receives instructions from the upper control system; the valve body is used to control fluid flow; the pressure sensor and flow sensor are used to monitor the pressure and flow of the fluid to ensure the natural gas hydrate... The methanol injection valve 116 is designed to ensure the accuracy and safety of its operation. The methanol injection valve control module 141 includes an electric actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor. The electric actuator opens and closes the methanol injection valve 116 according to the control signal. The control signal input terminal receives instructions from the upper control system. The valve body controls fluid flow. The pressure and flow sensors monitor the fluid pressure and flow rate to ensure the accuracy and safety of the methanol injection valve 116's operation. The scale inhibitor valve control module 142 includes an electric actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor. The electric actuator opens and closes the scale inhibitor valve 118 according to the control signal. The control signal input terminal receives instructions from the upper control system.The valve body is used to control fluid flow; pressure and flow sensors are used to monitor the pressure and flow of the fluid to ensure the accuracy and safety of the operation of the anti-scaling agent valve 118; the annular wing valve control module 143 includes an electric actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor; the electric actuator opens and closes the annular wing valve 126 according to the control signal; the control signal input terminal receives instructions from the upper control system; the valve body is used to control fluid flow; pressure and flow sensors are used to monitor the pressure and flow of the fluid to ensure the accuracy and safety of the operation of the annular wing valve 126; the annular isolation valve control module 14 ... The annular isolation valve 123 is opened and closed by a control signal; the control signal input terminal receives instructions from the upper control system; the valve body is used to control fluid flow; pressure and flow sensors are used to monitor the pressure and flow of the fluid to ensure the accuracy and safety of the operation of the annular isolation valve 123; the annular main valve control module 145 includes an electric actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor; the electric actuator opens and closes the annular main valve 124 according to the control signal; the control signal input terminal receives instructions from the upper control system; the valve body is used to control fluid flow; pressure and flow sensors are used to monitor the pressure and flow of the fluid to ensure the accuracy and safety of the operation of the annular main valve 124; the workover valve control module 146 includes an electric actuator... The system includes an actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor; the electric actuator opens and closes the workover valve 137 according to the control signal; the control signal input terminal receives instructions from the upper control system; the valve body is used to control fluid flow; the pressure sensor and flow sensor are used to monitor the pressure and flow of the fluid to ensure the accuracy and safety of the workover valve 137 operation; the production wing valve control module 147 includes an electric actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor; the electric actuator opens and closes the production wing valve 131 according to the control signal; the control signal input terminal receives instructions from the upper control system; the valve body is used to control fluid flow; the pressure sensor and flow sensor are used to monitor the pressure and flow of the fluid to ensure the accuracy and safety of the production wing valve operation. The accuracy and safety of operation of the switching valve 139 are ensured. The switching valve control module 148 includes an electric actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor. The electric actuator opens and closes the switching valve 139 according to the control signal. The control signal input terminal receives instructions from the upper control system. The valve body controls fluid flow. The pressure and flow sensors monitor the pressure and flow of the fluid to ensure the accuracy and safety of the switching valve 139's operation. The main production valve control module 149 includes an electric actuator, a control signal input terminal, a valve body, a pressure sensor, and a flow sensor. The electric actuator opens and closes the main production valve 135 according to the control signal. The control signal input terminal receives instructions from the upper control system. The valve body controls fluid flow.Pressure and flow sensors are used to monitor the pressure and flow of fluids to ensure the accuracy and safety of the operation of the main valve 135. The electro-hydraulic distribution module 152 is connected to the surface control unit 101 via an umbilical cable, and is used to receive and distribute electrical control signals and hydraulic power. The power distribution module 153 is connected to the electro-hydraulic distribution module 152 via a power cable, and is used to provide the power required by the underwater hydraulic control module 112 and the communication and data acquisition module 151. The underwater hydraulic control module 112 is connected to the electro-hydraulic distribution module 152 via a hydraulic pipeline, and is used to provide the hydraulic power required by the control module. The control signal processing module 150 is connected to the electro-hydraulic distribution module 152 via a signal cable, and is used to convert and send surface control commands. The communication and data acquisition module 151 is connected to the control unit 101 via a signal cable. A signal processing module 150 is connected to transmit control commands and acquire control module data; a natural gas hydrate inhibitor valve control module 140 is connected to a communication and data acquisition module 151 and an underwater hydraulic control module 112 via signal cables and hydraulic lines, and is used to control the natural gas hydrate inhibitor valve 113; a methanol injection valve control module 141 is connected to a communication and data acquisition module 151 and an underwater hydraulic control module 112 via signal cables and hydraulic lines, and is used to control the methanol injection valve 116; a scale inhibitor valve control module 142 is connected to a communication and data acquisition module 151 and an underwater hydraulic control module 112 via signal cables and hydraulic lines, and is used to control... The annular wing valve 118 is connected to the communication and data acquisition module 151 and the underwater hydraulic control module 112 via signal cables and hydraulic lines to control the annular wing valve 126; the annular isolation valve control module 144 is connected to the communication and data acquisition module 151 and the underwater hydraulic control module 112 via signal cables and hydraulic lines to control the annular isolation valve 123; the annular main valve control module 145 is connected to the communication and data acquisition module 151 and the underwater hydraulic control module 112 via signal cables and hydraulic lines to control the annular main valve 124; and the workover valve control module 146 is connected to the communication and data acquisition module 151 and the underwater hydraulic control module 112 via signal cables and hydraulic lines to control the annular main valve 124; The communication and data acquisition module 151 and the underwater hydraulic control module 112 are connected to control the workover valve 137; the production wing valve control module 147 is connected to the communication and data acquisition module 151 and the underwater hydraulic control module 112 via signal cables and hydraulic lines to control the production wing valve 131; the switching valve control module 148 is connected to the communication and data acquisition module 151 and the underwater hydraulic control module 112 via signal cables and hydraulic lines to control the switching valve 139; and the production main valve control module 149 is connected to the communication and data acquisition module 151 and the underwater hydraulic control module 112 via signal cables and hydraulic lines to control the production main valve 135.

[0023] The waterborne control unit 101 includes an uninterruptible power supply (UPS) module 102, a power supply module 103, a logic control module 104, a communication and network module 105, a data processing module 106, an electro-hydraulic integration module 107, a human-machine interface module 108, and a waterborne hydraulic control module 109. The UPS module 102 is connected to an external power cable via a power cable to provide power in the event of a power outage. The power supply module 103 is connected to the UPS module 102 via a power cable to supply power to the control system. The logic control module 104 is connected to the power supply module 103 via a power cable to execute preset logic actions. The communication and network module 105 is connected to the logic control module 104 via a signal cable to execute data communication... The system includes a high-reliability network for data transmission; a data processing module 106 connected to a logic control module 104 via a signal cable for processing, analyzing, and storing data from collected sensors and equipment; a surface hydraulic control module 109 connected to a logic control module 104 via a signal cable for generating and regulating high-pressure hydraulic power; a human-machine interface module 108 connected to a logic control module 104 via a signal cable for displaying the operation interface and sending commands; and an electro-hydraulic integration module connected to a communication and network module 105 and a data processing module 106 via a signal cable, connected to a surface hydraulic control module 109 via a hydraulic pipeline, and connected to an underwater control unit 111 via an umbilical cable for integrating electrical control and hydraulic power and transmitting them to the seabed. The uninterruptible power supply module 102 includes a battery pack, a battery charger, a voltage regulator, and a power switching device. The battery pack provides backup power to the water control unit 101 when the external power supply is interrupted or fluctuates. The battery charger charges the battery pack and maintains its charge within a set range. The voltage regulator stabilizes the output voltage. The power switching device automatically switches between the external power supply and the battery pack based on the external power status and battery charge to ensure continuous and reliable power supply to the system. The power module 103 includes a power converter, a voltage regulator, and a protection device. The power converter converts external AC power into DC operating voltage suitable for each functional module. The voltage regulator stabilizes the DC voltage. To reduce the impact of power grid fluctuations on the control system; protection devices are used to protect the power supply circuit from overcurrent, overvoltage, or short circuit; the logic control module 104 includes a microprocessor, memory, input / output interface, sensor interface, and actuator interface; the microprocessor is used to process the acquired signals and generate control commands; the memory is used to store the control program and operating parameters; the input / output interface is used to realize signal interaction with other control modules; the sensor interface is used to receive sensor signals; the actuator interface is used to output control commands to the valve drive device or other actuators to realize valve opening and closing or related action control; the communication and network module 105 includes a communication interface, circuit board, network protocol module, and data encryption device; communication The interface is used to realize wired data exchange with the logic control module 104, the control signal processing module 150, and other external modules; the circuit board is used to carry the communication circuit and complete signal processing and forwarding; the network protocol module is used to support the Modbus industry standard protocol to realize data encoding, parsing, and transmission; the data encryption device is used to encrypt and verify the integrity of the transmitted data to ensure communication security and anti-interference performance; the data processing module 106 includes a processor, a memory, and a data interface; the processor is used to perform calculations, processing, and analysis on the collected operating data such as pressure, temperature, and flow; the memory is used to store the control program, operating parameters, and historical operating data; the data interface is used to realize wired data exchange with the logic control module 104, the control signal processing module 150, and other external modules. 4. Data exchange between communication and data acquisition modules 151; the electro-hydraulic integration module 107 includes hydraulic valves, an electronically controlled drive device, and a feedback sensor; the hydraulic valves are used to switch and regulate the pressure, flow rate, and direction of hydraulic oil; the electronically controlled drive device is used to drive the corresponding hydraulic valves to operate according to the control signals output by the logic control module 104; the feedback sensor is used to collect the pressure, flow rate, and other status parameters of the hydraulic system and transmit them back to the surface control unit 101 to realize the monitoring and closed-loop regulation of the hydraulic system; the human-machine interface module 108 includes a touch screen, a display unit, control buttons, and an alarm device; the touch screen and display unit are used to display the operating parameters, alarm information, and historical records of the deep-water subsea gas-producing tree and the surface control unit in real time;Control buttons are used for operators to input control commands such as start, stop, valve opening and closing, and mode switching; alarm devices are used to issue audible and visual alerts when abnormal operating conditions or fault warnings occur; the surface hydraulic control module 109 includes a hydraulic pump, an electro-hydraulic drive device, a valve controller, and a pressure sensor; the hydraulic pump is used to provide the required hydraulic power to the surface and underwater actuators; the electro-hydraulic drive device is used to adjust the flow direction and flow rate of hydraulic oil according to the control signal output by the logic control module 104; the valve controller is used to drive the opening and closing of relevant valves; the pressure sensor is used to collect the pressure status of the hydraulic circuit and transmit it back to the surface control unit 101 for hydraulic system monitoring and safety protection; The above modules are all general functional units known in the art. The specific chip models and schematic diagrams are not the substantive innovations of this invention. Those skilled in the art can select existing mature devices to complete specific circuit designs based on the composition and function of each of the above modules and the main components and their functions within the modules.

[0024] The wellhead module 130 includes a wellhead base, casing head, tubing head, hanger, internal and external sealing devices, wellhead connecting flange, wellhead pressure monitoring port, and a fluid channel communicating with the lower wellbore. The wellhead base provides base support. The casing head and tubing head are used to fix the hangers for the casing and tubing, respectively. The internal and external sealing devices are used to achieve wellbore sealing in deep water and high-pressure environments. The wellhead connecting flange is used to achieve mechanical connection with the tree flange at the bottom of the gas production tree. The wellhead pressure monitoring port is used to install wellhead pressure and temperature sensors to collect wellhead operating conditions. The structural design of the above wellhead module belongs to the conventional structural form of deep water wellhead equipment. Example 1

[0025] like Figure 1 As shown, a method for safety monitoring and early warning of deep-water subsea gas production trees includes the following specific steps: S1. Collect raw sensor data from the deep-water subsea gas production tree to obtain its working status; S2. Process the raw sensor data of the deep-water subsea gas production tree collected in step S1. Amplify and filter the raw sensor data, calibrate the switching quantity, and align the data of different sensors to ensure that the data meets the input of the subsequent processing module. Obtain standardized data as the input of step S3. Step S2 specifically includes the following steps: S21. For different types of weak raw data, a three-level architecture of pre-amplifier fixed gain, programmable automatic gain (PGA), and interface matching is adopted. First, an industrial-grade instrumentation amplifier is used to build a pre-amplifier differential amplifier circuit. A matching fixed gain is set for different sensors to suppress common-mode interference in long-distance transmission. The preamplified data is fed into a programmable gain amplifier, which automatically matches the gain level according to the sensor range. Then, it automatically switches the range according to the real-time amplitude and performs 24-hour gain self-calibration to eliminate the influence of temperature drift. A rail-to-rail operational amplifier is used to convert the amplified data into single-ended data, and the amplitude is precisely matched to the input range of the subsequent ADC to ensure that the amplified data fully covers the effective input range of the ADC and maximizes the resolution of analog-to-digital conversion. S22. It adopts a two-stage architecture of analog hardware pre-filtering and digital software filtering to selectively filter out power frequency interference, fluid pulsation noise, mechanical vibration noise, and baseline drift in deep water conditions. The hardware uses a bandpass circuit to filter out baseline drift and power frequency interference, while the software performs classification processing according to the characteristics of the variables. For slow variables such as pressure / temperature, amplitude limiting and moving average filtering are used to eliminate random jumps, and for fast variables such as vibration, Kalman filtering is used to filter out Gaussian noise while retaining dynamic fault characteristics. S23. For switch quantities such as valve position and alarm, potential normalization and electromagnetic crosstalk are achieved through opto-isolation. Industrial-grade optocouplers are used to complete the electrical isolation between the field side and the system side, eliminate strong electromagnetic crosstalk underwater, and at the same time, the switch quantity levels of different amplitudes in the field are uniformly converted into levels that the MCU can recognize, so as to avoid high voltage interference that could damage the processing module. By introducing dual-threshold hysteresis comparison logic and dynamic reference calibration technology, the level offset caused by line aging or environmental changes is automatically corrected. Combined with hardware RC debouncing and software continuous sampling confirmation mechanisms, the correct identification of switch status is ensured under complex operating conditions. S24. To address the timing deviations caused by different sampling rates and transmission links, a high-precision temperature-controlled crystal oscillator is set as the global master clock in the surface control unit. Synchronous clock signals are sent to all underwater sensor acquisition nodes to provide a unique global time reference for all sampling and data stamping actions, eliminating clock drift between different acquisition nodes. Analog sensors of the same type adopt the same ADC multi-channel synchronous sampling mode, and the sampling action is triggered at the same time by the global clock. The inherent delay of each sensor channel is calibrated in advance in the laboratory, and the calibration value is written into the acquisition module register to provide a reference for subsequent software compensation. For each raw sampled data from each sensor, an absolute timestamp based on the global master clock is bound, and the sensor ID, sampling rate, and channel calibration delay value are also marked to form a standardized timestamped dataset. The baseline sampling frequency is set according to the early warning requirements, and an equally spaced target anchor point sequence is generated. Different types of data are classified and resampled. The original timestamp is offset in the reverse direction according to the inherent delay of the calibrated channel to eliminate the time deviation caused by transmission lag. The continuity of the timestamp is checked, and abnormal frames with excessive deviation are removed. S3. A parallel computing architecture is adopted to confirm the integrity of the data, obtain continuous data from the sensor, and extract key features from the standardized data. The specific method for step S3 is as follows: S31, Set the sliding time window [T] k ,T k+1 The standardized data obtained in step S2 is divided into multiple independent data segments according to a time window, and each data segment is distributed in parallel to multiple computing nodes of the distributed architecture corresponding to the parallel computing architecture. The number of computing nodes is adapted to the total number of sensor channels of the deep-water subsea gas production tree, the monitoring data sampling rate, and the real-time requirements of early warning. The number of sensor channels allocated to a single computing node shall not exceed 8, and the minimum configuration is 1 main computing node and 8 data processing nodes. S32. Within each computing node, the continuity of timestamps and the validity of values ​​of the allocated data segments are independently verified to obtain continuous data from the sensor, and the continuous data from the sensor is used as the input for step S4. If the computing node detects missing values ​​or abnormal null values ​​exceeding physical limits in a data segment, it first performs an anomaly type determination. If the number of consecutive anomaly points is ≤3 and the adjacent sampling points are valid data, it is marked as an occasional anomaly and repaired using linear interpolation. If the number of consecutive anomaly points is >3, it is marked as a fault anomaly, and no interpolation repair is performed. The abnormal data segment is removed and a self-test process is triggered to ensure the continuity of data collected by the real device. The specific operation method of linear interpolation is as follows: at time t m There are missing values ​​or outlier null values ​​that exceed the physical limits. The nearest valid sampling points before and after them are (t) m-1 ,x i (t m-1 )) and (t m+1 ,x i (t m+1 Then at time t m Missing values ​​or outlier null values ​​x that exceed physical limits i (t m The interpolation repair formula for ) is: In the formula: i is the number of the i-th sensor channel in the deep-water subsea gas production tree monitoring system; t m The target sampling time is defined as the time when missing values ​​or occasional outlier null values ​​exist; t m-1 For the target time t m Previously, distance t m The most recent valid sampling time; t m+1 For the target time t m After that, distance t mThe most recent valid sampling time; x i (t m (t) represents time t m Valid data after linear interpolation repair; x i (t m-1 (t) represents time t m-1 Valid sampled data; x i (t m+1 (t) represents time t m+1 Valid sampling data; S33. Each computing node performs multi-threaded concurrent processing on the verified complete data segment to calculate the current time window [T]. k ,T k+1 Key features of the i-th type of sensor; The key features include local average values ​​and local absolute peak values ​​used for subsequent steps; The specific method for calculating the local average value is as follows: Let the sliding time window [T] be... k ,T k+1 Within a window, if the i-th type of sensor contains N discrete sampling points, then the local average value within that window is... The calculation formula is: In the formula: N represents the valid sensor data within the sliding time window; N is the sliding time window [T] k ,T k+1 Within the sliding time window, j represents the total number of discrete valid sampling points of the i-th sensor; j is the index of the discrete sampling point within the sliding time window. The local absolute peak P within this window i (T k ), used for calculating the peak-to-average ratio in step S4, local absolute peak value P i (T k The formula for calculating ) is: In the formula: P i (T k ) represents the local absolute peak value, x i (t j (t) represents time t j Sensor data; [T k ,T k+1 [This represents a sliding time window;] S34. Synchronously collect the local average values ​​of the outputs from each computing node through the aggregation node. and local absolute peak P i (T kThe sensor data segments, along with the verified data segments, are spliced ​​and reassembled according to the original time series to output a high-dimensional time series containing complete signal features as continuous sensor data. The continuous sensor data serves as the input reference for the comprehensive transient calculation in step S4. S4. Obtain the continuous data from the sensor according to step S3, and calculate the transient change difference and steady-state deviation of each sensor based on the key features extracted in step S3. Then, normalize the transient change difference and steady-state deviation to obtain the comprehensive transient difference. The transient abrupt difference uses the peak-to-average ratio to quantify the amplitude of sensor signal abrupt changes and an energy operator to quantify the degree of transient energy concentration. The formula for calculating the peak-to-average ratio is as follows: In the formula, R i (t) represents the peak-to-average ratio of the i-th type of sensor within the time window at time t; x i (t) represents the real-time sampled value of the i-th type of sensor; [T k ,T k+1 [This represents a sliding time window;] i (T k ) represents the i-th type of sensor in window [T] k ,T k+1 The average value within ]; Minimal positive number ; The calculation formula for the energy operator is: In the formula, x is the energy operator value of the i-th type of sensor at time t; i (t) represents the real-time sampled value of the i-th type of sensor; Let be the sampling period of the i-th type of sensor; The steady-state deviation is used to quantify the cumulative degree of long-term deviation of the signal from the baseline; The formula for calculating the steady-state deviation is: In the formula, D i (t) represents the average deviation of points within the selected time period; S i (t k B represents the valid sampled value corresponding to the k-th sampling time selected from the continuous data of the sensor; i (t k ) represents the i-th sensor at t k The continuous baseline value corresponding to a given time is the reference data for that time under normal steady-state operating conditions of the gas production tree. The initial calibration baseline is calibrated by performing a rated operating condition steady-state calibration test before the deep-water subsea gas production tree is officially put into operation or after annual shutdown maintenance. The gas production tree is controlled to maintain stable operation at its designed rated production pressure, temperature, valve opening, flow rate, and other parameters, without valve action, operating condition adjustments, or malfunctions. Steady-state operating data from the corresponding sensors is continuously collected. After amplification, filtering, and standardization in step S2, a continuous baseline sequence B is formed that matches the sampling rate and time length of subsequent online monitoring data. i (t k ); The formula for calculating the comprehensive instantaneous difference is: In the formula, TDI i (t) represents the composite instantaneous difference value. To measure the weight of the peak-to-mean ratio, To measure the weight of the energy operator, The weights used to measure steady-state deviation; S5. Calculate the probability of risk occurring to the deep-water subsea gas production tree component. The specific calculation method is as follows: In the formula, Given the combined transient difference (TDI) of all n sensors, this represents the posterior probability of a safety risk occurring in the target monitoring component of the deep-water subsea gas production tree. The TDI represents the combined transient error of the i-th sensor at time t under the condition that a safety risk occurs in the target monitoring component of the deep-water subsea gas production tree. i The conditional probability; This indicates that, under the condition that the target monitoring component of the deep-water subsea gas production tree is in a risk-free and safe state, the comprehensive instantaneous difference value of the i-th sensor at time t is TDI. i The conditional probability; G represents the prior probability of a safety risk occurring in the target monitoring component of the deep-water subsea gas production tree; G represents the Boolean value of the component's risk occurrence. G=1 indicates a safety risk to the deepwater subsea gas production tree target monitoring component, while G=0 indicates that the deepwater subsea gas production tree target monitoring component is in a safe, risk-free state; TDI i This represents the combined instantaneous error value of the i-th sensor at time t; S6. Based on the probability of risk P(G|TDI) calculated in step S5, compare the probability of risk occurrence with the preset risk threshold t to determine the degree of risk of the gas production tree component. The threshold t is determined through expert decision-making; When P(G|TDI)>t, the probability of the component being at risk is considered high, and it is considered to be in a risky state. when When the probability of a risk occurring to the component is determined to be low, it is considered to be in a safe or low-risk state. S7. Based on the risk assessment results of each component of the gas production tree obtained in step S6, construct a dynamic risk map and, in conjunction with the data safety operation specifications, comprehensively assess the overall risk level, set three-level early warning trigger rules, and achieve differentiated early warning output: when only 1-2 risk components exist in the risk map, it is judged as "potential risk" and triggers a level one early warning; when 3-4 risk components appear in the risk map, it is judged as "abnormal risk" and triggers a level two early warning, activates a local alarm, and synchronously records the risk development curve and related signal change data; when 5 or more risk components appear in the risk map, it is judged as "failure risk" and triggers a level three early warning. While continuously triggering local audible and visual alarms, the three-dimensional risk map, real-time monitoring data, and fault prediction information are uploaded to the remote operation and maintenance platform and pushed to the operation and maintenance personnel's terminals, ensuring that the operation and maintenance party can quickly obtain risk details and formulate emergency response plans, realizing closed-loop management from risk identification to early warning implementation, and completing the classification of early warning information. Example 2

[0026] like Figure 3 As shown, a deep-sea subsea gas production tree safety monitoring and early warning system for implementing the method of Embodiment 1 includes a gas production tree data acquisition and transmission unit 231, an underwater data processing unit 205, a risk assessment and early warning unit 220, and an early warning display unit 226. The communication and data acquisition module 151 of the gas production tree data acquisition and transmission unit 231 is connected to the data transmission module 230 of the underwater control unit 111 via a signal cable, for real-time acquisition of multi-source sensor data from the underwater gas production tree, and transmission of the pre-processed data. The underwater data processing unit 205 is connected to the data acquisition and transmission unit 226 via a communication cable. The data acquisition and transmission unit 231 is connected to the underwater data processing unit 205 via a communication cable. It receives underwater data collected by the data acquisition and transmission unit 231 and performs feature extraction and transient analysis on the data to provide a basis for further risk assessment and early warning. The risk assessment and early warning unit 220 is connected to the underwater data processing unit 205 via a communication cable. It performs multi-feature fusion on the data and constructs a dynamic risk matrix. It generates graded early warning signals according to preset early warning thresholds and performs risk assessment. The early warning display unit 226 is connected to the risk assessment and early warning unit 220 via a signal cable. It is used to display early warning information in real time and provide different forms of response measures according to the risk level.

[0027] The data acquisition and transmission unit mainly completes the data acquisition step; the underwater data processing unit mainly completes the data processing step, parallel computing step, and transient calculation step; the risk assessment and early warning unit mainly completes the risk calculation step, risk assessment step, and graded early warning step.

[0028] The gas sampling tree data acquisition and transmission unit 231 includes a data acquisition and statistics module 227, a data statistics preprocessing module 228, a data storage module 229, and a data transmission module 230. The data acquisition and statistics module 227, developed based on a digital signal processor (DSP), is connected to the control signal processing module 150 via a signal cable. It extracts data frame headers, parses data acquisition instructions, polls the buffer registers of each sensor channel, reads the sensor ID, distributes data to the corresponding classification array in the DSP memory, sets a fixed time window, and performs statistical analysis on the data in each classification array, calculating the maximum, minimum, average, and number of valid data frames within that period. It is used to receive data acquisition instructions from the surface control unit 101 and classify and statistically analyze the data from each acquisition module. The data statistics preprocessing module 228, also developed based on a DSP, is connected to the data acquisition and statistics module 227 via a signal cable. It uses a median filtering algorithm to identify and remove sharp, abrupt, or dead values ​​in the classification array that significantly deviate from the reasonable physical range. It also calls the DSP's internal FIR low-pass filtering algorithm to filter out high-frequency electrical noise interference and outputs clean data. The sensor digital signal is used to preprocess the statistical data to obtain clean sensor data. The data storage module 229 is developed based on a solid-state drive (SSD) controller and NAND flash memory chip. It is connected to the data statistical preprocessing module 228 via a signal cable, receives clean data from the data statistical preprocessing module 228 and writes it into the internal high-speed RAM cache. When the cache reaches the page write threshold, it calls the wear leveling algorithm of the flash memory controller to write the data along with ECC to the NAND flash physical page for storing the collected sensor data. The data transmission module 230 is developed based on a field-programmable gate array (FPGA). It is connected to the data statistical preprocessing module 228 via a signal cable, uses a FIFO queue to receive data from the data statistical preprocessing module 228, adds frame headers, frame tails and CRC check codes to form standard transmission frames, and uses 8B / 10B encoding technology to perform line encoding on the framed data to ensure DC balance of the signal. Through the high-speed SERDES interface inside the FPGA, it converts the parallel data into a high-speed serial bit stream for transmitting the collected sensor data to the signal receiving module 201.

[0029] The underwater data processing unit 205 includes a signal receiving module 201, an acquisition timing coordination module 202, a signal conditioning and shaping module 203, a continuous and transient dual-channel generation module 204, a continuous baseline extraction module 206, a transient detection and feature extraction module 207, a feature alignment and normalization module 208, a transient analysis module 209, a historical data recording module 210, and a data output module 211. The signal receiving module 201 is developed based on a digital signal processor (DSP) and is connected to the data transmission module 230 via a signal cable. It performs 10B / 8B decoding and CRC verification on the serial data on the external interrupt response bus. After the verification is successful, the load data is extracted to the memory cache. The flow zone is used to receive sensor data from the gas source tree data transmission module 230 and collect operational status information of the deep-water subsea gas source tree 110. The acquisition timing coordination module 202 is developed based on a digital signal processor (DSP) and is connected to the signal receiving module 201 via a signal cable. Utilizing the high-precision RTC inside the DSP, it reads the current absolute time at the moment when the module 201 triggers an interrupt after receiving a complete data frame, and uses this as a timestamp to append to the header of the data frame, ensuring strict alignment of multi-source heterogeneous data on the time axis. The signal conditioning and shaping module 203 is also developed based on a DSP and is connected to the acquisition timing coordination module 202 via a signal cable. It employs polynomial fitting calculations... The algorithm calculates the slowly varying trend line of the data sequence and subtracts it from the original signal to eliminate baseline drift. The Min-Max normalization algorithm is used to linearly map sensor values ​​with different physical dimensions to the dimensionless interval [0,1] or [-1,1] to eliminate baseline drift and dimensional differences. The continuous and transient dual-channel generation module 204, developed based on a digital signal processor (DSP), is connected to the signal conditioning and shaping module 203 via a signal cable. Using a discrete wavelet transform algorithm, it decomposes the conditioned signal into a multi-scale signal channel, reconstructing the low-frequency approximate components into a continuous signal channel representing steady-state conditions and the high-frequency detail components into a transient signal channel representing abrupt events. The continuous baseline extraction module... Block 206 is developed based on a digital signal processor (DSP) and is connected to the continuous and transient dual-channel generation module 204 via a signal cable. It extracts baseline feature values ​​representing the normal steady-state operation of the system from the continuous signal channel. The transient detection and feature extraction module 207 is also developed based on a DSP and is connected to the continuous and transient dual-channel generation module 204 via a signal cable. It applies the Teager energy operator to the transient signal channel. When the energy value exceeds a preset dynamic adaptive threshold, it is marked as a sudden event. Then, the peak factor, peak coefficient, and instantaneous energy within the time period of the event are extracted as transient feature information for detecting sudden events and extracting their feature information in the transient signal channel.The feature alignment and normalization module 208, based on a digital signal processor (DSP), is connected to the continuous baseline extraction module 206 and the transient detection and feature extraction module 207 via signal cables. Using timestamps as keys, it matches continuous baseline feature values ​​and transient feature information at the same time point. The Z-score algorithm is then used to further normalize the two sets of features, concatenating them to form a unified feature vector. The transient analysis module 209 is based on an ARM processor. Developed using a Cortex-A processor, the system connects to the feature alignment and standardization module 208 via a signal cable. It calculates the Mahalanobis distance between the transient feature vector and the continuous baseline feature vector within the aligned feature vector. This distance value is used as a quantification score of the deviation, used to calculate the deviation of transient features relative to the continuous baseline based on the aligned feature vector, thus quantifying transient data. The historical data recording module 210, developed based on a mechanical hard disk drive (HDD) array storage and RAID control card, connects to the transient analysis module 209 via a signal cable. It establishes a FIFO circular file buffer. When the storage capacity reaches a set upper limit threshold, the control card automatically overwrites the file block with the oldest timestamp, achieving cyclic recording within a certain time window. This is used to cyclically store the original signal, processed features, and analysis results within a certain time window. The data output module 211, developed based on a digital signal processor (DSP), connects to the transient analysis module 209 via a signal cable. It serializes and packages the processed feature vector and transient quantization score into JSON format or a custom binary structure, and sends it to the host computer via Ethernet or serial port for final risk assessment and early warning.

[0030] The risk assessment and early warning unit 220 includes a multi-source feature fusion module 212, a trend prediction and evolution analysis module 213, a dynamic risk matrix construction module 214, a hierarchical early warning triggering module 215, a potential risk alert module 216, an abnormal trend warning module 217, a failure risk alarm module 218, an early warning feedback and strategy optimization module 219, and a risk scoring module 221. The multi-source feature fusion module 212 is developed based on a digital signal processor (DSP) and is connected to the data output module 211 via a signal cable. It is used to receive and fuse multi-sensor feature vectors from the underwater data processing unit 205 to generate a comprehensive health status indicator. The prediction and evolution analysis module 213, developed based on an ARM Cortex-A processor, is connected to the multi-source feature fusion module 212 via a signal cable. It analyzes the comprehensive risk score and time series characteristics to predict the short-term risk development trend. The dynamic risk matrix construction module 214, developed based on a digital signal processor (DSP), is connected to the multi-source feature fusion module 212 via a signal cable. It fills the calculation results into a higher-dimensional state space in real time to form a dynamic risk matrix. The hierarchical early warning triggering module 215, developed based on an STM32 series MCU, is connected to both the trend prediction and evolution analysis module 213 and the dynamic risk matrix construction module 214 via signal cables. Based on the current risk status and future risk trends, and according to preset strategies, different levels of early warnings are automatically triggered. The potential risk alert module 216, developed based on an STM32 series MCU, is connected to the hierarchical early warning trigger module 215 via a signal cable. It triggers an alert when the risk matrix indicates a potential anomaly. The anomaly trend warning module 217, also developed based on an STM32 series MCU, is connected to the hierarchical early warning trigger module 215 via a signal cable. It triggers a warning when the risk matrix indicates a clear trend of performance degradation. The failure risk alarm module 218, also developed based on an STM32 series MCU, is connected to the hierarchical early warning trigger module 215 via a signal cable. The system is connected to the risk matrix, which triggers an emergency alarm when the device is nearing failure. The early warning feedback and strategy optimization module 219 is developed based on a digital signal processor (DSP) and is connected to the potential risk warning module 216, the abnormal trend warning module 217, and the failure risk alarm module 218 via signal cables. It is used to collect feedback on the accuracy of the early warning results and adaptively optimize the judgment threshold and early warning strategy of the risk matrix. The risk scoring module 221 is developed based on an ARM Cortex A processor and is connected to the early warning feedback and strategy optimization module 219 via signal cables. It calculates a quantitative risk score for the health status of the gas production tree and provides an intuitive risk level assessment.

[0031] The early warning display unit 226 includes a core early warning information display module 222, a data visualization module 223, a response measure guidance module 224, and a historical early warning retrospective module 225. The core early warning information display module 222 is connected to the risk scoring module 221 via a signal cable, and receives and parses data packets containing risk scores, risk levels, and abnormal component IDs in real time, highlighting the highest priority information from the risk scoring module 221, including the current risk level, risk score, and abnormal components. The data visualization module 223 is connected to the risk scoring module 221 via a signal cable, receives multi-source feature data and time series from the risk scoring module 221, dynamically maps physical quantity values ​​to the two-dimensional pixel coordinate system of the screen, and uses Bézier curves for smoothing. The algorithm plots historical trend curves of feature data over time in real time, presenting risk scores, historical trends, and multi-source feature data in a graphical way such as curves, charts, and 3D models. The response measure guidance module 224 is connected to the risk scoring module 221 via a signal cable. It extracts the currently triggered "warning level" and the specific "abnormal component ID" as joint query keywords. In the locally built relational database, it uses the built-in emergency plan table as the matching target and retrieves the standard operating procedure (SOP) or maintenance suggestion text corresponding to the keywords through SQL statements. The retrieved emergency steps are arranged in chronological order of execution to generate a structured task guidance list, which is displayed in a dedicated area of ​​the human-computer interaction interface to guide the operator in risk reduction operations. The historical early warning and retrospective module 225 is connected to the risk scoring module 221 via a signal cable. It serializes each triggered early warning event and persistently stores it in the log database on the local hard disk. The log database includes timestamps, risk scores, abnormal components, and response measures taken. It provides a front-end composite time selector and category drop-down box to convert user filtering operations into back-end query statements. It displays the historical result set returned by the database in a structured format in the form of paginated data tables and provides a processing script to export the underlying data to common table files such as CSV or Excel for subsequent accident analysis. The core early warning information display module 222, the data visualization module 223, the response measure guidance module 224, and the historical early warning retrospective module 225 are connected to the human-machine interface module 108 via signal cables. They are used to provide risk information to operators and assist them in taking corresponding measures.

[0032] The applicant declares that the above description is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Those skilled in the art should understand that any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention fall within the protection and disclosure scope of the present invention.

Claims

1. A method for safety monitoring and early warning of deep-water subsea gas production trees, characterized in that: Includes the following steps: S1. Collect raw sensor data from the deep-water subsea gas production tree; S2. Preprocess the raw sensor data of the deep-water subsea gas production tree collected in step S1 to obtain standardized data; S3. A parallel computing architecture is adopted to confirm the integrity of the data, obtain continuous data from the sensor, and extract key features from the standardized data. S4. Based on the key sensor features extracted in step S3, calculate the transient change difference and steady-state deviation of each sensor, and normalize the transient change difference and steady-state deviation to obtain the comprehensive transient difference. S5. Calculate the probability of a risk occurring to the deep-water subsea gas production tree component; S6. Based on the posterior probability of safety risk of the deep-water subsea gas production tree component calculated in step S5, compare the posterior probability of safety risk of the deep-water subsea gas production tree component with the preset risk threshold to determine the degree of risk of the deep-water subsea gas production tree component. S7. Based on the degree of risk of the deep-water subsea gas production tree component obtained in step S6, construct a dynamic risk map, and combine it with the safety operation specifications to comprehensively determine the overall risk level, set three-level early warning trigger rules, and carry out differentiated early warning output.

2. The method for safety monitoring and early warning of deep-water subsea gas production trees according to claim 1, characterized in that: The preprocessing in step S2 includes amplifying and filtering the raw sensor data, calibrating the switching signals, and aligning the data from different sensors.

3. The method for safety monitoring and early warning of deep-water subsea gas production trees according to claim 1, characterized in that: Step S3 specifically includes the following steps: S31, Set the sliding time window [T] k ,T k+1 The standardized data obtained in step S2 is divided into multiple independent data segments according to a time window, and each data segment is distributed in parallel to multiple computing nodes of the distributed architecture corresponding to the parallel computing architecture. S32. Within each computing node, independently verify the continuity of timestamps and the validity of values ​​for the allocated data segments. S33. If the computing node detects missing values ​​or abnormal null values ​​exceeding physical limits in the data segment, it performs anomaly type determination; and repairs occasional anomalies using linear interpolation. The repair formula for the linear interpolation method is as follows: In the formula: i is the number of the i-th sensor in the deep-water subsea gas production tree monitoring system; t m The target sampling time where sporadic anomalies exist; t m-1 For the target sampling time t m Previously, distance t m The most recent valid sampling time; t m+1 For the target sampling time t m After that, distance t m The most recent valid sampling time; x i (t m (t) represents time t m Valid data after linear interpolation repair; x i (t m-1 (t) represents time t m-1 Valid sampled data; x i (t m+1 (t) represents time t m+1 Valid sampling data; S34. Each computing node performs multi-threaded concurrent processing on the verified complete data segment to calculate the current time window [T]. k ,T k+1 Key features of the i-th type of sensor; The key features include local average values ​​and local absolute peak values; The specific method for calculating the local average value is as follows: Let the sliding time window [T] be... k ,T k+1 Within the range, if the i-th type of sensor contains N discrete sampling points, then the sliding time window [T]... k ,T k+1 The formula for calculating the local average value within [ ] is: In the formula: For sliding time window [T k ,T k+1 Local average value within ]; N represents the valid sensor data within the sliding time window; N is the sliding time window [T] k ,T k+1 Within the sliding time window, j represents the total number of discrete valid sampling points of the i-th sensor; j is the index of the discrete sampling point within the sliding time window. Sliding time window [T k ,T k+1 The formula for calculating the local absolute peak value within the area is: In the formula: P i (T k ) represents the sliding time window [T] k ,T k+1 Local absolute peak value within ]; x i (t j (t) represents time t j Sensor data; [T k ,T k+1 [This represents a sliding time window;] S35. Collect the local average value and local absolute peak value output by each computing node through the aggregation node, as well as the sensor data segments verified in step S32. Splice and reassemble them according to the original time series, and output a high-dimensional time series containing complete signal features as continuous sensor data. The continuous sensor data is used as the input benchmark for the comprehensive transient calculation in step S4.

4. The method for safety monitoring and early warning of deep-water subsea gas production trees according to claim 3, characterized in that: The criteria for determining the anomaly type in step S32 are as follows: If the number of consecutive outliers is ≤3 and the adjacent sampling points are valid data, they are marked as occasional outliers and repaired using linear interpolation. If the number of consecutive outliers exceeds 3, it is marked as a fault, interpolation repair is not performed, and the abnormal data segment is removed.

5. The method for safety monitoring and early warning of deep-water subsea gas production trees according to claim 1, characterized in that: In step S4, the transient abrupt difference uses the peak-to-average ratio to quantify the amplitude of sensor data abrupt changes and the energy operator to quantify the degree of transient energy concentration. The formula for calculating the peak-to-average ratio is: In the formula: R i (t) represents the peak-to-average ratio of the i-th type of sensor within the time window at time t; x i (t) represents the real-time sampled value of the i-th type of sensor; [T k ,T k+1 [This represents a sliding time window;] i (T k ) represents the i-th type of sensor in window [T] k ,T k+1 The average value within ]; It is a very small positive number. ; The calculation formula for the energy operator is: In the formula: x is the energy operator value of the i-th type of sensor at time t; i (t) represents the real-time sampled value of the i-th type of sensor; Let be the sampling period of the i-th type of sensor; The formula for calculating the steady-state deviation is: In the formula, D i (t) represents the average deviation of points within the selected time period; S i (t k B represents the valid sampled value corresponding to the k-th sampling time selected from the continuous data of the sensor; i (t k ) represents the i-th sensor at t k The continuous baseline value corresponding to the time.

6. The method for safety monitoring and early warning of deep-water subsea gas production trees according to claim 1, characterized in that: The formula for calculating the comprehensive instantaneous difference is: In the formula, TDI i (t) represents the composite instantaneous difference value. To measure the weight of the peak-to-average ratio, To measure the weight of the energy operator, The weight used to measure steady-state deviation; R i (t) represents the peak-to-average ratio of the time window in which the i-th type of sensor is located at time t; D represents the energy operator value of the i-th type of sensor at time t; i (t) represents the average deviation of points within the selected time period.

7. The method for safety monitoring and early warning of deep-water subsea gas production trees according to claim 1, characterized in that: The formula for calculating the posterior probability of a safety risk occurring in the deep-water subsea gas production tree component is as follows: In the formula, Given the combined transient difference (TDI) of all n sensors, this represents the posterior probability of a safety risk occurring in the target monitoring component of the deep-water subsea gas production tree. The TDI represents the combined transient error of the i-th sensor at time t under the condition that a safety risk occurs in the target monitoring component of the deep-water subsea gas production tree. i The conditional probability; This indicates that, under the condition that the target monitoring component of the deep-water subsea gas production tree is in a risk-free and safe state, the comprehensive instantaneous difference value of the i-th sensor at time t is TDI. i The conditional probability; G represents the prior probability of a safety risk occurring in the target monitoring component of the deep-water subsea gas production tree; G represents the Boolean value of the component's risk occurrence. G=1 indicates that the deep-water subsea gas production tree target monitoring component has a safety risk, and G=0 indicates that the deep-water subsea gas production tree target monitoring component is in a safe state without risk. TDI i This represents the combined instantaneous difference value of the i-th sensor at time t.

8. The method for safety monitoring and early warning of deep-water subsea gas production tree according to claim 1, characterized in that: The threshold was determined through expert decision-making. When the posterior probability of a safety risk is greater than the threshold, the probability of the component being at risk is considered high, and it is considered to be in a risky state. When the posterior probability of a security risk occurs When the threshold is reached, the probability of the component being at risk is determined to be low, and it is considered to be in a safe or low-risk state.

9. The method for safety monitoring and early warning of deep-water subsea gas production trees according to claim 1, characterized in that: The triggering rules for the Level 3 early warning are as follows: When there are only 1 to 2 risk components in the risk map, it is identified as a potential risk and a level 1 warning is triggered. When 3 to 4 risk components appear in the risk map, it is judged as an abnormal risk, triggering a level 2 warning, activating a local alarm, and simultaneously recording the risk development curve and related signal change data. When five or more risky components appear in the risk map, it is judged as a failure risk, triggering a level three warning. While the local audible and visual alarms continue, the three-dimensional risk map, real-time monitoring data and fault prediction information are uploaded to the remote operation and maintenance platform and pushed to the operation and maintenance personnel's terminal.

10. A deep-water subsea gas production tree safety monitoring and early warning system for the method described in any one of claims 1 to 9, characterized in that: It includes a gas-gathering tree data acquisition and transmission unit, an underwater data processing unit, a risk assessment and early warning unit, and an early warning display unit; The gas-producing tree data acquisition and transmission unit is connected to the underwater control unit via a signal cable; the underwater data processing unit is connected to the gas-producing tree data acquisition and transmission unit via a communication cable; the risk assessment and early warning unit is connected to the underwater data processing unit via a communication cable; and the early warning display unit is connected to the risk assessment and early warning unit via a signal cable. The gas-gathering tree data acquisition and transmission unit includes a data acquisition and statistics module, a data statistics preprocessing module, a data storage module, and a data transmission module; The data acquisition and statistics module is connected to the control signal processing module via a signal cable; the data statistics preprocessing module is connected to the data acquisition and statistics module via a signal cable; the data storage module is connected to the data statistics preprocessing module via a signal cable; and the data transmission module is connected to the data statistics preprocessing module via a signal cable. The underwater data processing unit includes a signal receiving module, an acquisition timing coordination module, a signal conditioning and shaping module, a continuous and transient dual-channel generation module, a continuous baseline extraction module, a transient detection and feature extraction module, a feature alignment and standardization module, a transient analysis module, a historical data recording module, and a data output module. The signal receiving module is connected to the data transmission module via a signal cable. The acquisition timing coordination module is connected to the signal receiving module via a signal cable. The signal conditioning and shaping module is connected to the acquisition timing coordination module via a signal cable. The continuous and transient dual-channel generation module is connected to the signal conditioning and shaping module via a signal cable. The continuous baseline extraction module is connected to the continuous and transient dual-channel generation module via a signal cable. The transient detection and feature extraction module is connected to the continuous and transient dual-channel generation module via a signal cable. The feature alignment and standardization module is connected to both the continuous baseline extraction module and the transient detection and feature extraction module via signal cables. The transient analysis module is connected to the feature alignment and standardization module via a signal cable. The historical data recording module is connected to the transient analysis module via a signal cable. The data output module is connected to the transient analysis module via a signal cable. The risk assessment and early warning unit includes a multi-source feature fusion module, a trend prediction and evolution analysis module, a dynamic risk matrix construction module, a hierarchical early warning triggering module, a potential risk reminder module, an abnormal trend warning module, a failure risk alarm module, an early warning feedback and strategy optimization module, and a risk scoring module. The multi-source feature fusion module is connected to the data output module via a signal cable; the trend prediction and evolution analysis module is connected to the multi-source feature fusion module via a signal cable; the dynamic risk matrix construction module is connected to the multi-source feature fusion module via a signal cable; the hierarchical early warning triggering module is connected to the trend prediction and evolution analysis module and the dynamic risk matrix construction module via signal cables respectively; and the potential risk alert module is connected to the hierarchical early warning triggering module via a signal cable. The abnormal trend warning module is connected to the hierarchical early warning trigger module via a signal cable; the failure risk alarm module is connected to the hierarchical early warning trigger module via a signal cable. The early warning feedback and strategy optimization module is connected to the potential risk alert module, the abnormal trend warning module, and the failure risk alarm module via signal cables; the risk scoring module is connected to the early warning feedback and strategy optimization module via signal cables. The early warning display unit includes a core early warning information display module, a data visualization module, a response measure guidance module, and a historical early warning retrospective module. The core early warning information display module is connected to the risk scoring module via a signal cable; the data visualization module is connected to the risk scoring module via a signal cable; the response measure guidance module is connected to the risk scoring module via a signal cable; and the historical early warning retrospective module is connected to the risk scoring module via a signal cable.

Images