| Pig robots driven by gas, liquid and other media, are usually have been used to clean up the pipes. Pig robot is often trapped or lost in the pipeline that will cause safety accident. So it is very necessary to track and locate the pig robot real time to improve the performance and ensure the safety of pipeline. To improve positioning accuracy, the thesis presented a locating system for detecting the pig robot through emitting extreme low-frequency(ELF) magnetic electromagnetic waves which is received by the magnetic sensor outside the pipeline.An ELF electromagnetic intensity distribution model is proposed for the electromagnetic coil of finite length solenoid based on the magnetic vector potential theory. Electromagnetic transmitting coil is equivalent to a current loop for the current magnetic dipole model neglecting the coil size, shape and current distribution a magnetic field distribution model is built on the magnetic vector potential, and the magnetic field is related to the size of the excitation coil which will improve the calculation accuracy of the magnetic field strength. Further, the winding mode of magnetic emission coil is also optimized.According to the spatial relationship of the single-axis magnetic sensor, the pipeline and the distribution model of the spatial magnetic field strength, the relationship between sensing speed and the output voltage waveform of the sensor is considered in time domain. Relationship equation is established for the sensor output voltage, sensor speed and tilt angle. Then, an automatic and fast locating method of one-dimensional coordinates of pig robot is proposed.The correlation analysis between the sensor output voltage and the theoretical output voltage was analyzed for positioning of pipeline robot. Numerical experiments were carried out and the influence of correlation average thresholds on the positioning accuracy is analyzed.A self-adapting position model was proposed to realize the 3D location of pig robot by sensor array. The position of the pig robot and its attitude are figured outthrough the sensor array positioning model, the measurement errors caused by the material, thickness and external environment of the pipeline were eliminated by equating the physical quantity of the pipeline and the environment to the change of the sensor output voltage magnification. An adaptive positioning model of pig robot is established and the localization equation is deduced based on the structure of the symmetric magnetic sensor array.A particle swarm quasi-Newton hybrid algorithm is adopted to solve the adaptive nonlinear equation for the locating trapped robot. Based on the analysis of the characteristics of quasi-Newton algorithm(BFGS) and Particle Swarm Optimization(PSO), a hybrid Particle swarm PSO+BFGS algorithm was designed to solve nonlinear equations, making full use of the search function of PSO global spaces and BFGS local, which improves positioning precision. Then the experimental results show that the average positioning errors are reduced by 4.2cm, 3.8cm, 4.6cm and 3.4cm, 2.6cm, 3.8cm in x, y and z directions, respectively, compared to that of PSO and BFGS.Due to the disadvantages of the alkaline battery and the lithium battery, liquid formic fuel cells with high energy density is proposed as power supplier for pig robot. the hardware of the positioning system is fabricated, and the transmitting antenna, transmitting control circuit, receiving antenna, receiving signal conditioning circuit, control circuit and so on are made respectively.Furthermore, the experimental results show that the positioning system can meet with the requirements of non-contact, online, real time localization. |
PDF Full Text Request