Abstract: This paper details the design of two modules in the intelligent controller of a subsea oil pipeline inspection crawler, and presents the design tasks and schemes for these two modules. Keywords: pipeline inspection; intelligent crawler; intelligent controller; defect location Introduction Oil pipelines are essential transportation equipment for the energy sector. After long-term use, the pipe walls are thinned due to erosion and corrosion, often leading to leaks, causing huge economic losses, environmental pollution, and ecological damage. Therefore, effective in-service inspection of oil pipelines to prevent such accidents has become a top priority for the energy sector. Because traditional inspection methods are very difficult for subsea oil pipelines, extensive research and development efforts have been undertaken both domestically and internationally to overcome this challenge [1-7]. Currently, developed countries mainly use pipeline inspection crawlers for in-service inspection, employing methods such as magnetic flux leakage detection and ultrasonic testing. In 2002, the Shengli Oilfield Administration, Shanghai Jiaotong University, Harbin Institute of Technology, and Shanghai University of Electric Power jointly applied for and received funding for the National 863 Program project "Inspection and Maintenance Technology of Subsea Pipelines". The key technologies of this project include defect detection, detector diameter adjustment, precise positioning, signal acquisition and processing, power supply and drive, intelligent crawler, intelligent control system, pipeline defect evaluation, and deployment and retrieval. Shanghai University of Electric Power is mainly responsible for research on intelligent control of electric crawlers, strapdown inertial navigation systems and precise geodetic positioning, and detection power supplies. This paper only introduces the design problems of two modules in the intelligent controller: defect location process control and abnormal situation analysis and rescue. The task of these two modules is to perform system self-checks. If a problem is found, the abnormal situation is analyzed and the rescue module handles the abnormality; otherwise, the defect location process control module operates normally. The design work of these two modules is described in detail below. 1. Design of Defect Location Process Control Module 1.1 Design Tasks The main tasks of the defect location process control module are: 1. Automatically start the crawler at the detection starting point and automatically stop at the end point; 2. Automatically switch between crawling speed and detection speed; 3. Provide a positioning signal and relative distance during precise positioning and determine the positioning status; 4. After accurately locating the defect, the detection device automatically stops, and the isotope emission source is aligned with the defect to emit isotopes; 5. Automatically start crawling after precise positioning. 1.2 Design Scheme 1.2.1 Speed ​​Switching and Positioning Process Design of the Detection Device The primary task of the detection device is to locate the defect, precisely position it, align the isotope emission source with the defect, and then automatically stop the detection device. The motion process involved in this task is shown in Figure 1. It can be divided into several parts: First, driven by the crawler, the detection device starts from the initial position, reaches point 1 (selected based on experience), and then travels at its maximum speed. The position information fed back from the odometer wheel causes it to slow down to point 3 at the predetermined deceleration position (detection range), simultaneously beginning to detect defect information. After detecting a defect at point 4, the position control system uses fuzzy control or PID control to stop the detection device at point 5 (within the accurate positioning zone), within the allowable detection position range. Then, the control system precisely measures the relative distance between the radiation device and the defect center, sends a precise positioning signal, and adjusts at low speed until the radiation device stops at the defect center. In Figure 1, V represents the crawler speed. During the startup to maximum speed process, to ensure a smooth transition between states, a ramp-up control law can be used (the initial allowable displacement for the ramp-up control at the start of the state transition is set at 30±5m); similarly, during the transition from crawling to detection state, ramp-down control is also used (the initial allowable displacement for the deceleration length is set at 15±5m); constant speed control (high speed, low speed) is used in both crawling and detection states; after the defect signal is detected, to accurately locate it, it should first be rapidly decelerated to zero. To ensure smooth operation, fuzzy control or PID control methods can be used to keep the detection device within the allowable positioning area (the allowable displacement is initially set at 5±1m). Since the detection device cannot generate vibration inside the pipe, the speed control loop cannot have overshoot. 1.2.2 Defect Identification During Detection The defect identification process is shown in Figure 2. In the first online detection process, the approximate location and characteristics of the defect can be determined. In the second precise detection and positioning, it is necessary to complete the comparison and confirmation of the measured defect signal category. The detection and confirmation of the signal is achieved by identifying some characteristic indicators of the defect signal (such as peak value, average value, etc.). In the identification method, an artificial neural network is chosen to identify and confirm defects. Common defects are classified and analyzed to find their characteristic parameters. Then, a suitable algorithm (such as the BP algorithm) is selected to construct a three-layer network. This network is trained to achieve offline identification, which is then applied to online identification. Simultaneously, when a defect is confirmed, the location of the defect center can be accurately pinpointed so that the subsequent radiation device can be aligned. 1.2.3 Input and Output Information 1. Input information includes defect judgment signals from the detection section, odometer wheel displacement information, defect center location information, and predicted defect location. 2. Output information includes crawler crawling speed commands, crawler crawling distance and distance from the endpoint signals, and X-ray gate opening or closing commands. 1.2.4 Normal Control Section Intelligent Control Structure Its control structure is shown in Figure 3. 2 Abnormal Situation Analysis and Rescue Module 2.1 Design Task During the positioning and detection process, the module monitors the working status and alarm information of each important system in real time, analyzes abnormal situations, and proposes rescue measures. 2.2 Design Scheme 2.2.1 Monitoring System Safety Operation Information: The system continuously monitors the status and alarm signals of various components, including the crawler, power supply, leakage magnetic field and ultrasonic detectors, and tracer ray emitter. If all signals are normal, a "normal" signal is sent to the main control center. 2.2.2 Analyzing Status and Alarm Signals and Determining Fault Types: If the signals at the system's fault points are abnormal, it indicates a system fault. The fault type needs to be determined by analyzing the status and alarm signals. Fault types and determination methods are as follows: 1. Electric Crawler Faults: Primarily motor faults, such as open circuit (no current between terminals) or short circuit (no voltage between terminals); motion obstruction faults, such as zero speed or excessive motor current. 2. Intelligent Controller Faults: Primarily control loop open circuit faults, manifested as no motor speed detection or zero actual motor input voltage; control algorithm failure faults, manifested as excessive deviation between the actual motor speed and the given speed. 3. Power Source Failure: The power battery will run out of power, manifested as the power battery voltage falling below the lower limit; insufficient power battery power, manifested as insufficient remaining power to support the crawler to complete the entire distance; power supply failure, manifested as a drop or zero voltage at a certain point in the power supply system. 4. Sensor Component and Data Storage Failure: Mainly includes magnetic leakage and ultrasonic sensor malfunction; mileage wheel malfunction; rotary encoder malfunction; data storage malfunction. 5. Tracer Ray Emitter Failure: Mainly includes the tracer ray emitter door failing to open; the tracer ray emitter door failing to close. 6. Computer Self-Test and Inter-computer Communication Failure: The microcomputer self-test can detect faults in the CPU, memory, I/O interfaces, and inter-computer communication; inter-computer communication failures mainly include the control center being unable to contact the crawler; the control center being unable to contact the detector; the control center being unable to contact the tracer ray emitter. 2.2.3 Decision on the best rescue plan based on the fault type 1. Crawler fault If the motor is faulty, stop moving. If it can move backward, return. Otherwise, stop and send a distress signal (SOS), disengage the clutch, and wait for rescue. If the movement is obstructed, use pattern recognition to determine the type of obstacle (e.g., turning, deformation, uphill, irregular shape, etc.). Different methods can be used for different obstacles. Generally, reverse a distance, rotate a certain angle, fine-tune the angle of the elastic support arm, and then accelerate forward. If it still cannot pass, stop detecting forward movement and move backward. 2. Intelligent controller fault If the control circuit is open, the handling plan is the same as for the motor fault. If the control algorithm fails, change to another control algorithm. If the control system operates normally, continue detecting forward movement. Otherwise, the handling plan is the same as for the motor fault. 3. Power Source Failure: If the power battery is about to run out of power, immediately stop moving forward, send an SOS distress signal, disengage the clutch, and wait for rescue. If the power battery is insufficient, if the remaining energy is sufficient to support the return, stop moving forward and return; otherwise, stop, send an SOS distress signal, disengage the clutch, and wait for rescue. If the detection power supply fails, activate the backup detection power supply; otherwise, stop detection and move back or forward to the end. Whether to move forward or backward depends on the distance. 4. Sensor Component and Data Storage Failure: Reset the leakage magnetic field or ultrasonic detector; reset the odometer wheel; reset the rotary encoder; reset the data storage. 5. Tracer X-ray Emitter Failure: Reset the tracer X-ray emitter and try again. If the fault cannot be eliminated, activate the emergency device to shut down the tracer X-ray emitter. 6. Computer Self-Test and Inter-computer Communication Failure: Restart the computer; rebuild the communication connection with the crawler; rebuild the communication connection with the detector; rebuild the communication connection with the X-ray device. 2.2.4 The operation plan is determined based on the effect of the rescue measures. Generally speaking, if the effect of the previous decision plan does not achieve the expected goal, the system can adopt the second plan; if it fails again, it will stop and send a tracer distress signal SOS, disengage the clutch, and wait for rescue. 3 Conclusion This paper details the design of two key modules of the intelligent controller of the subsea oil pipeline inspection crawler. Simulation shows the correctness and feasibility of the design of these two modules. Now, these two modules are connected and debugged with other modules of the system. References [1] Cordell JL. The latest developments in pipeline pigging world-wide [J]. Pipes & Pipeline International, 1994, 8(7): 9-16. [2] Raad JA. Comparison between ultrasonic and magnetic flux pigs for pipeline [J]. Pipes & Pipelines International, 1987, 32(1): 7-15. [3] Zhou Ming, He Fengqi, Ma Baiyong. Non-destructive testing method for in-service oil pipelines [J]. Nondestructive Testing, 1999, 21(1): 8-13. [4] Long Wei, Zhou Ming, Huang Jie. Current status and development trend of in-service pipeline ultrasonic testing system [J]. China Mechanical Engineering, 1996, 7(2): 52-54. [5] Sha Jie, Liu Zhanchu, Chen Guofang. Design of control system for small and medium diameter pipeline crawler [J]. Measurement and Control Technology, 2000, 19(4): 27-29. [6] Zhang Xiaohua, Yin Dejun, Deng Zongquan. An autonomous positioning control method for pipeline robot based on visual fuzzy reasoning [J]. Microcomputer Information, 2002, 18(2): 10-11. [7] Jiang Shengyuan, Deng Zongquan, Li Bin, et al. Application of programmable logic controller in pipeline robot control system [J]. Nondestructive Testing, 2001, 23(6): 234-237.