Online fault diagnosis method and system for instrument air system

Abstract

The invention discloses an online fault diagnosis method and system for an instrument air system, and belongs to the field of oil and gas field development, and the method comprises the steps: arranging the pipe network structure of the instrument air system, and classifying and layering the pipe network structure of the instrument air system according to attributes; data acquisition: acquiring process data in a pipe network structure of the instrument air system; branch pipeline abnormity diagnosis: diagnosing the state of a branch valve according to the process data of the branch upstream and downstream process equipment so as to obtain the branch leakage condition; and air compression station abnormality diagnosis: diagnosing the air compression station abnormality according to the process data of each substructure of the air compression station in the specific process equipment. The on-line monitoring means is used for replacing manual inspection, manual analysis is reduced through model diagnosis, the working efficiency of station workers is effectively improved, pipeline leakage and other abnormalities are rapidly diagnosed, and energy loss caused by idling of the air compressor is effectively reduced.

Description

Technical Field

[0001] This invention belongs to the field of oil and gas field development, and specifically relates to an online fault diagnosis method and system for an instrument ventilation system. Background Technology

[0002] Instrument air is dried and extremely clean compressed air, requiring oil-free, dust-free, and dry air with a dew point temperature below -40℃. As the fourth most widely used energy source in the industrial sector, instrument air systems are central to many industrial product systems. Providing clean, dry, and reliably pressurized instrument air to air-consuming equipment greatly improves equipment performance and efficiency, thereby directly reducing production costs, improving equipment management and automation levels, and is crucial for ensuring the safe operation of petrochemical plants and achieving energy conservation and consumption reduction. A stable instrument air supply is also one of the key factors for realizing automated control.

[0003] Therefore, real-time monitoring of the instrument ventilation system is necessary to promptly detect issues such as downtime, leaks, and substandard quality. Currently, the main methods rely on manual intervention. One approach involves regular inspections and tests using checkpoints and instrument measurements. Another method involves setting thresholds on online instruments for manual analysis. While this method allows for continuous online monitoring, it requires data transmission instruments, and its analytical capabilities depend heavily on the user's skill level, resulting in slow problem detection, low location accuracy, and numerous redundant alarms.

[0004] Current existing technologies, such as CN201110458194 "Intelligent Detection Method and System for Compressed Air System Pipeline Leaks," provide an intelligent detection method and system for compressed air system pipeline leaks. This includes: sensors installed on the terminal delivery pipeline collect the pressure, temperature, and flow rate of compressed air, transmitting the signals to a low-pass filter circuit for coarse filtering, and then transmitting the signals back to the host computer via an A / D conversion circuit. The host computer uses the actual measured start-end point data as a basis to obtain the relationship between gas flow parameters and time and pipeline length, and then compares the theoretical output with the actual output to achieve pipeline leak detection. This achievement is based on a real-time transient model, using pipeline hydraulic and thermal models to compare theoretical and actual data to detect pipeline leaks, and requires the additional installation of two flow, pressure, and temperature sensors on each pipeline.

[0005] CN202311180775 discloses a fault detection algorithm for compressed air systems based on the random projection PLS method, relating to the field of data-driven fault detection. This work uses a data-driven modeling approach, requiring a large amount of sample data for training and separate modeling for specific processes.

[0006] This research project presents a remote fault diagnosis and prediction technology for air compressors. Taking screw compressors as the research object, it uses a backpropagation (BP) neural network as the analysis method to diagnose compressor faults and establish a remote monitoring system. This achievement utilizes the BP neural network method for modeling, requires sample data for training, and its maintenance depends on the development team. Summary of the Invention

[0007] The purpose of this invention is to provide an online fault diagnosis method and system for instrument ventilation systems to improve the efficiency of diagnostic work.

[0008] The beneficial effects of this invention are achieved through the following technical means: an online fault diagnosis method for an instrument ventilation system, comprising the following steps: The instrument ventilation system pipeline structure is organized and classified and layered according to attributes. The pipeline structure is divided into main lines, branch lines, valves, specific process equipment and substructures of specific process equipment, and the upstream and downstream of each level are marked. Data acquisition involves collecting process data from the instrument ventilation system's piping network structure. Branch pipeline anomaly diagnosis involves diagnosing the status of branch valves based on process data from upstream and downstream process equipment, thereby identifying branch pipeline leakage. Air compressor station anomaly diagnosis involves diagnosing anomalies in the air compressor station based on the process data of each substructure within the specific process equipment. In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is an air-lock valve, the diagnosis rule is that the downstream flow rate is >0; the upstream pressure suddenly drops by more than 20%; and the upstream liquid level continues to drop, indicating that the valve status changes from closed to open, that is, there is a leak in the branch pipeline.

[0009] In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is an air-to-open valve, the diagnosis rule is that the downstream flow rate decreases by more than 50%; the upstream pressure continues to rise; the upstream liquid level continues to rise; and the downstream liquid level stops fluctuating or continues to drop. This indicates that the valve status changes from open to closed, meaning that there is a leak in the branch pipeline.

[0010] In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is a regulating valve, the diagnosis rule is that the interlock control condition is triggered. After the control system sends a command to the valve, if the process parameters of the downstream process equipment do not change, it indicates that there is a problem with the valve, that is, there is a leak in the branch pipeline.

[0011] When the regulating valve detects the valve opening, if the valve opening remains unchanged or the response is slow after the control system sends a command to the valve, it indicates that there is a problem with the valve, that is, there is a leak in the branch line.

[0012] The upstream is an air compressor station, and the downstream is specific process equipment.

[0013] When the number of leaks in the downstream branch lines of the main line exceeds the set value, it indicates that there is a leak in the main line, and a main line leak alarm is triggered.

[0014] The specific diagnosis of air compressor station anomalies is as follows: Within time T, the pressure in the instrument air tank drops rapidly, and the pressure value is ≤ limit A; the air compressor current is normal, the air compressor runs continuously for more than a certain period of time, and the compressor temperature rises. If at least one of the above situations occurs, it indicates that the air compressor pipeline is leaking. Within time T, the pressure in the instrument air storage tank drops rapidly, and the pressure value is less than or equal to the limit B. The current of multiple air compressors drops to zero simultaneously, indicating that the air compressors have stopped. If, under single-tower regeneration conditions, the outlet temperature is ≤ limit C, or the regeneration time is ≤ limit D, it indicates that the drying process is unqualified.

[0015] An online fault diagnosis system for an instrument ventilation system, including The data acquisition module is electrically connected to the PLC and instruments on site to perform real-time data acquisition; The model editing module edits and modifies model rules based on an online fault diagnosis method for an instrument ventilation system. The parameter selection module, based on the rules of the model editing module, matches the objects and parameters that the rules require to be retrieved; The anomaly diagnosis module calculates the collected parameters based on the model rules defined by the model editing module to obtain the anomaly diagnosis results.

[0016] The beneficial effects of this invention are as follows: Modeling using examples and mechanistic models is closer to field applications; the dual-driven modeling strategy of data and model, along with the monitoring system that mutually verifies and collaboratively analyzes instrument data and production process parameters, achieves precise and efficient monitoring and positioning. It not only has low requirements for instruments but can also automatically adjust according to production conditions, offering advantages such as low investment, rapid implementation, and simple maintenance. Furthermore, it establishes for the first time a mechanism for constructing and applying anomaly diagnosis models for instrument ventilation systems in oil and gas processing plants. Through the organic combination of preset rules and custom programming, the adaptability of the results is significantly improved, enabling it to flexibly adapt to the actual conditions of different sites. Attached Figure Description

[0017] Figure 1 Flowchart of online fault diagnosis method for instrument ventilation system; Figure 2 This is a schematic diagram of the online fault diagnosis system for the instrument ventilation system. The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Detailed Implementation

[0018]

Example 1

[0019] As shown in the table, the entire instrument ventilation system piping structure is outlined. The air compressor station serves as the power source, and its internal substructures constitute the upstream process equipment. Then downstream, the main pipelines are connected to trunk line 1, trunk line 2, trunk line 3, etc.; all the trunk lines are considered as one layer.

[0020] Pipelines are connected to different process equipment via branch lines. The downstream of the main line is the branch line, the valves on the branch line are considered as one layer, and the specific process equipment downstream of the branch line is considered as another layer.

[0021] As shown in the table, the upstream and downstream structure of each process equipment can be obtained, such as the complete path of air compressor station - main line 1 - branch line 1 - valve 1 - separator, which helps to determine which parameters of upstream and downstream equipment need to be tested for subsequent diagnosis. In practical applications, modifications can be made according to the actual structure of the factory area; for example, there may be a main line 2 and subsequent branch lines. Alternatively, branch line 1 may branch into branch line 1-1 and branch line 1-2, which are connected to different process equipment. When diagnosing anomalies, the diagnosis will be performed on branch line 1-1 and branch line 1-2, rather than branch line 1. This will allow the anomaly to be located on a specific branch line and valve, resulting in more accurate results.

[0022] Data acquisition involves collecting process data from the instrument ventilation system's piping network structure. Through the plant's Internet of Things (IoT), process data is collected from various process equipment, such as air compressors and their sub-components, as well as downstream separators and venting pipelines. Process data refers to the parameters that need to be monitored during the operation of process equipment, such as pressure, flow rate, liquid level, current, valve opening, temperature, regeneration time, etc., which are selected according to different process equipment.

[0023] Due to explosion-proof and other reasons, the valves used in oil and gas plant areas are often pneumatic valves rather than electronic valves. Therefore, it is impossible to directly obtain the status of valves on each pipeline, and thus it is impossible to know the open and closed status of pneumatic valves. For example, a pneumatic shut-off valve should remain closed when there is no leakage. If there is a leakage, the valve will open. Since it is impossible to directly know the open and closed status of the valve, it is impossible to know whether there is a leakage.

[0024] Branch pipeline anomaly diagnosis involves diagnosing the status of branch valves based on process data from upstream and downstream process equipment, thereby identifying branch pipeline leakage. In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is an air-lock valve, the diagnosis rule is that the downstream flow rate is >0; the upstream pressure suddenly drops by more than 20%; and the upstream liquid level continues to drop, indicating that the valve status changes from closed to open, that is, there is a leak in the branch pipeline.

[0025] In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is an air-to-open valve, the diagnosis rule is that the downstream flow rate decreases by more than 50%; the upstream pressure continues to rise; the upstream liquid level continues to rise; and the downstream liquid level stops fluctuating or continues to drop. This indicates that the valve status changes from open to closed, meaning that there is a leak in the branch pipeline.

[0026] In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is a regulating valve, the diagnosis rule is that the interlock control condition is triggered. After the control system sends a command to the valve, if the process parameters of the downstream process equipment do not change, it indicates that there is a problem with the valve, that is, there is a leak in the branch pipeline.

[0027] When the regulating valve detects the valve opening, if the valve opening remains unchanged or the response is slow after the control system sends a command to the valve, it indicates that there is a problem with the valve, that is, there is a leak in the branch line.

[0028]

[0029] As shown in the table, different valves have different diagnostic criteria, i.e., modeling rules. When a process diagnosis occurs, the valve's status can be determined.

[0030] For example, if the gas shut-off valve is diagnosed as not being closed, it means that there is insufficient gas to keep the valve closed, which means that there is a leak in this branch line, causing the gas shut-off valve to change from closed to open.

[0031] Similarly, for air-to-open valves, if a leak occurs and the gas in the branch line causes the air-to-open valve to be unable to remain open, the corresponding situation in the process data diagnosis will occur. By working backward from this, the status of the air-to-open valve can be obtained from the process data diagnosis, thereby determining whether there is a leak in the branch line.

[0032] A control valve is a valve that automatically adjusts its opening when it receives instructions from the plant's control system, such as when a certain interlock control condition is triggered, like a liquid level interlock. For example, if the separator level reaches 50mm, the valve opening increases to 50%; if the level reaches 70mm, the valve opening increases to 80%.

[0033] If the process parameters of the downstream equipment do not change after the operator issues a manual command, it indicates that there is a problem with the control valve and it cannot be adjusted. Since the control valve is also a pneumatic valve, its inability to be adjusted is often caused by leakage in the branch line.

[0034] The upstream is an air compressor station, and the downstream is specific process equipment.

[0035] Therefore, the state of valves on the branch line can be inferred based on the process parameters of the upstream and downstream process equipment, and further, it can be inferred whether a leak has occurred in the branch line.

[0036] The entire data acquisition process can utilize the existing data acquisition equipment and instruments in the plant area, without requiring extensive new hardware. Furthermore, the general model, combined with modeling tailored to the actual conditions of the plant, allows for faster implementation, lower investment, and self-maintenance by process personnel, eliminating the need for specialized maintenance.

[0037] Air compressor station anomaly diagnosis involves diagnosing anomalies in the air compressor station based on the process data of each substructure within the specific process equipment. When the number of leaks in the downstream branch lines of the main line exceeds the set value, it indicates that there is a leak in the main line, and a main line leak alarm is triggered.

[0038] Based on the branch line alarm results, if 60% of the branch lines trigger alarms, it indicates that the leak may not be in a branch line itself, but rather in the main line of several branches, causing a problem with the main line valve. In this case, a main line leak alarm should be triggered. Simultaneously, branch line alarms should be suppressed.

[0039] The specific diagnosis of air compressor station anomalies is as follows: Within a time interval T, if the pressure in the instrument air tank drops rapidly and the pressure value is ≤ limit A; if the air compressor current is normal and the air compressor runs continuously for more than a certain period of time, and the compressor temperature rises, it indicates that there is a leak in the air compressor pipeline. Usually, the leak is in the external pipeline of the air compressor, and the air compressor station rarely has problems.

[0040] Within time T, the pressure in the instrument air storage tank drops rapidly, and the pressure value is less than or equal to the limit B. The current of multiple air compressors drops to zero simultaneously, indicating that the air compressors have stopped. If, under single-tower regeneration conditions, the outlet temperature is ≤ limit C, or the regeneration time is ≤ limit D, it indicates that the drying process is unqualified.

[0041] Time T and limits A, B, C, and D can all be set according to the actual situation of the factory area.

[0042] like Figure 2 As shown, an online fault diagnosis system for an instrument ventilation system includes... The data acquisition module is electrically connected to the PLC and instruments in the field for real-time data acquisition. It also interfaces with the PLC and instruments to acquire real-time data and perform data filtering. A calculation result information reading function is added to the data acquisition module, enabling it to read alarm information from different models and objects in the anomaly diagnosis module, convert the data type to digital, and use it as real-time data input.

[0043] The model editing module edits and modifies model rules based on an online fault diagnosis method for an instrument ventilation system. The model rule editing adopts a combination of preset rules and custom programming, supports expansion, and provides calculation rules for the anomaly diagnosis module.

[0044] The parameter selection module, based on the rules of the model editing module, matches the objects and parameters that the rules require to be retrieved; The anomaly diagnosis module calculates the collected parameters according to the model rules defined by the model editing module to obtain the anomaly diagnosis results. The high-concurrency computing engine undertakes the computational tasks; the calculation results are used for alarm handling and display, and can also be converted into real-time data as input for the next round of computation, realizing step-by-step calculation and alarm suppression.

[0045] Using online monitoring to replace manual inspections and model-based diagnosis to reduce manual analysis effectively improves the work efficiency of station and depot personnel. It is estimated that this can save personnel in this position 2 hours of work per day, quickly diagnose abnormalities such as pipeline leaks, effectively reduce energy loss caused by air compressor idling, and save each station and depot approximately 30,000 yuan in electricity costs per year. Timely detection of equipment abnormalities can effectively avoid the loss of benefits caused by production stoppages, saving the oilfield approximately 200,000 yuan per year.

[0046] The following explanation will be based on a processing station warehouse of a plant in Xinjiang Oilfield.

[0047] The construction of the diagnostic system first involves developing the diagnostic system based on the oilfield Internet of Things (IoT), and mainly includes: Data acquisition module: Real-time data acquisition uses two methods. One is to directly connect to the DCS system using an OPC gateway and complete data acquisition through the OPC protocol. The other is to connect to field instruments through the controller and acquire instrument and PLC data through the moudbs protocol.

[0048] Model editing module: It adopts a visual interface and simplifies Java scripts into four arithmetic operation rules for model editing.

[0049] Parameter selection module: It adopts the configuration method of object + collection point to realize cross-device data operation mode.

[0050] Anomaly diagnosis module, computing engine: a high-concurrency computing engine developed based on the Flink architecture.

[0051] Results management: The calculation results were integrated with the factory's alarm management platform in the form of alarm information and pushed to the monitoring system.

[0052] Meanwhile, a calculation result information reading function was added to the data acquisition module, which can read alarm information from different models and objects and convert it into numbers as real-time data input.

[0053] Construction of the diagnostic model ① Reviewing the structure of the instrument ventilation system duct network The structure of the instrument ventilation system's piping network at this station has been analyzed, categorized and layered according to attributes, and the driving instruments and upstream and downstream processes have been labeled. A common process flow diagram is shown in the table below:

[0054] ② Branch pipeline anomaly diagnosis modeling Complete the model of the plant based on valve status and process flow production status.

[0055] ③ Main pipeline anomaly diagnosis modeling Based on the alarm results of the branch lines, a leak alarm is triggered on the main line pipeline when 60% of the branch lines issue an alarm.

[0056] At the same time, the alarm result will be input into the calculation engine as data for the branch alarm, and the next round of branch alarms will be suppressed.

[0057] ④ Air compressor station anomaly diagnosis Based on the actual conditions of the plant, a diagnostic model for the air compressor station was established.

[0058]

Claims

1. An online fault diagnosis method for an instrument ventilation system, characterized in that: Includes the following steps, The instrument ventilation system pipeline structure is organized and classified and layered according to attributes. The pipeline structure is divided into main lines, branch lines, valves, specific process equipment and substructures of specific process equipment, and the upstream and downstream of each level are marked. Data acquisition involves collecting process data from the instrument ventilation system's piping network structure. Branch pipeline anomaly diagnosis involves diagnosing the status of branch valves based on process data from upstream and downstream process equipment, thereby identifying branch pipeline leakage. Air compressor station anomaly diagnosis involves diagnosing anomalies in the air compressor station based on the process data of each substructure within the specific process equipment.

2. The online fault diagnosis method for an instrument ventilation system according to claim 1, characterized in that: In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is an air-lock valve, the diagnosis rule is that the downstream flow rate is >0; the upstream pressure suddenly drops by more than 20%; and the upstream liquid level continues to drop, indicating that the valve status changes from closed to open, that is, there is a leak in the branch pipeline.

3. The online fault diagnosis method for an instrument ventilation system according to claim 1, characterized in that: In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is an air-to-open valve, the diagnosis rule is that the downstream flow rate decreases by more than 50%; the upstream pressure continues to rise; the upstream liquid level continues to rise; and the downstream liquid level stops fluctuating or continues to drop. This indicates that the valve status changes from open to closed, meaning that there is a leak in the branch pipeline.

4. The online fault diagnosis method for an instrument ventilation system according to claim 1, characterized in that: In the diagnosis of abnormalities in the branch pipeline, if the valve on the branch pipeline is a regulating valve, the diagnosis rule is that the interlock control condition is triggered. After the control system sends a command to the valve, if the process parameters of the downstream process equipment do not change, it indicates that there is a problem with the valve, that is, there is a leak in the branch pipeline.

5. The online fault diagnosis method for an instrument ventilation system according to claim 4, characterized in that: When the regulating valve detects the valve opening, if the valve opening remains unchanged or the response is slow after the control system sends a command to the valve, it indicates that there is a problem with the valve, that is, there is a leak in the branch line.

6. The online fault diagnosis method for an instrument ventilation system according to any one of claims 1-4, characterized in that: The upstream is an air compressor station, and the downstream is specific process equipment.

7. The online fault diagnosis method for an instrument ventilation system according to claim 1, characterized in that: When the number of leaks in the downstream branch lines of the main line exceeds the set value, it indicates that there is a leak in the main line, and a main line leak alarm is triggered.

8. The online fault diagnosis method for an instrument ventilation system according to claim 1, characterized in that: The specific diagnosis of air compressor station anomalies is as follows: Within time T, the pressure in the instrument air tank drops rapidly, and the pressure value is ≤ limit A; the air compressor current is normal, the air compressor runs continuously for more than a certain period of time, and the compressor temperature rises. If at least one of the above situations occurs, it indicates that the air compressor pipeline is leaking. Within time T, the pressure in the instrument air storage tank drops rapidly, and the pressure value is less than or equal to the limit B. The current of multiple air compressors drops to zero simultaneously, indicating that the air compressors have stopped. If, under single-tower regeneration conditions, the outlet temperature is ≤ limit C, or the regeneration time is ≤ limit D, it indicates that the drying process is unqualified.

9. An online fault diagnosis system for an instrument ventilation system, characterized in that: include The data acquisition module is electrically connected to the PLC and instruments on site to perform real-time data acquisition; The model editing module is used to edit and modify model rules according to any one of claims 1-8 of the online fault diagnosis method for an instrument ventilation system. The parameter selection module, based on the rules of the model editing module, matches the objects and parameters that the rules require to be retrieved; The anomaly diagnosis module calculates the collected parameters based on the model rules defined by the model editing module to obtain the anomaly diagnosis results.

Images