Research On Control Of Temperature Of Oil Pipeline Based On Skin Effect Current Tracing System

This paper sets pipeline tracing at low temperatures as research background.The crude oil usually contains a lot of wax,whose coagulation point is relatively high.When crude oil is transported in low-temperature environment,it is very likely to congeal and may clog up the pipeline.So the flow rate of oil is related to oil temperature.To solve this problem,this paper chooses skin effect heat tracing system to warm up the pipeline and oil in it.The work in this papaer includes:This paper establishes the pipeline electromagnetic field model with Maxwell's equations.Then,the relationship between skin depth and frequency,conductivity and permittivity is analyzed based on it.After that,with skin depth,it is natural to deduce power equivalent resistance for the pipeline and relationship between heating power and frequency.Furthermore,we can establish the model between current and oil temperature.All above research can be used to design control method for skin effect current tracing system.To control skin effect current tracing system,there are some challenges that need paying attention.Firstly,the quantity of crude oil transported by pipeline is so large that the control system of oil temperature is a system with great inertia.Secondly,the conflict between temperature overshoot and heating rate need different control parameters in different heating stages.Thirdly,the ambient temperature and initial temperature of oil are not constant,which makes conventional control method with constant control parameters unsuitable.To address these problems,in this paper,a new control method,which is based on genetic algorithm and deep neural network with Bayesian regularization,is proposed.To achieve training data for neural network,this paper use genetic algorithm to optimize control parameters in different conditions.Then,after trained with these data,deep neural network with Bayesian regularization will generate different control parameters for different conditions to make sure that oil in pipeline can be heated to target temperature in different conditions.This control method will be validated with simulation.The parameter design of skin effect current tracing system plays an important role for its performance.Firstly,given that the system of pipeline and cable is inductive and the current needs filtering,in this paper,a LCL filter is designed based on inductance of pipeline and cable.Secondly,to meet the requirement of heating power and avoid the damage of IGBTs,selecting a proper frequency is very important.Lastly,this paper design PI control parameters for inner loop to ensure a fast and stable response.To validate the control method and parameters design,an experiment platform with the core of DSP2812 is built up.The performance of skin effect heat tracing system designed in this paper will be tested in different working conditions.The results show that the temperature error is below 0.5? and barely has no overshoot.The current error is below 0.1A and its harmonic content is less than 15%.The skin effect current tracing system designed in this paper works stably and can adjust control parameters according to the change of environmental parameters.It can meet the requirement of heating crude oil at low temperatures.