Study On Smart Wax Cleaning Pig

The crude oil produced in China has the characteristics of lower temperature,higher freezing point and higher wax content.Pipeline transportation is one of the most common way in the of transporting crude oil.China has large quantities of pipelines which would generate more and more wax depositions as the service time increases.That's would increase the resistance and the energy consumption of systems.In some areas where the situation of wax deposition is severe,there even has some possibilities of occurring wax blockage accident.Also,traditional wax cleaning pigs have many disadvantages such as wax plug formation,low efficiency of wax cleaning and repeated operation.In order to solve those problems,it is of great significance to designing a brand-new type of wax cleaning pig for the maintenance of pipeline networks.Aim at those pipelines with a diameter of 377 mm and have been used for a long time,a new type of intelligent wax cleaning pig has been researched and designed,it contains three parts including velocity controlling unit,wax cleaning unit and sealed joint unit.The velocity controlling unit is responsible for controlling the pig to run in the oil pipeline with an ideal velocity through the algorithm of Kalman filter and Incremental PID control.The wax cleaning unit is responsible for cleaning wax deposition with a higher efficiency and preventing secondary deposition of wax particles.In order to reach this expectation,a method of combining wax cutting with jet action is proposed.The sealed joint unit is responsible for connecting the two units mentioned above by Hooke hinge and sealing connecting sleeve and increasing the trafficability of the whole equipment.In this thesis,the key part of the velocity controlling unit and the wax cleaning unit are designed in detail,the strength of key components of these two units were checked by Abaqus2010.Combined with the actual condition of crude oil pipeline,the key screw pair which is used for controlling the bypass institution is calculated in detail.Furthermore,the Kalman filter algorithm and increment PID algorithm,that are used for processing data which is collected by the sensor and controlling the whole equipment with an ideal velocity,are detailedly described.At the same time,The Explicit module of Abaqus is utilized for simulating the process of cleaning wax deposition.A number of comparative experiments are carried out,it is found that the process of wax cleaning is relatively stable when the whole equipment run at the speed about 1m/s.In order to test the performance of the wax cleaning motor,the module of Simulink in Matlab2014 b is utilized for simulating the whole system,through some procedure of optimization,the system not only has favorable dynamic characteristics but some degree of robustness.Also,in this thesis,the main control circuit,temperature pressure and speed detection circuit,power module,communication module and other key circuits in the control system of intelligent wax cleaning pig are designed in detail.In order to test these circuits' performance and weather it could arrive the application requirements,the Multisim is utilized to simulating the actual condition to do these functional verifications.Last but not the least,in this thesis,a brand-new wax cleaning algorithm is raised which is based on Open CV and deep convolutional neural networks Firstly,deformation parameters are collected by measuring arm and transformed into contact position coordinates.Next,the Matplotlib of Python site packages is used for doing some data augmentation,this could reduce the possibility of overfitting in the training process of the whole model,enhance the generalization and robustness.Finally,building deep convolutional neural networks through Tensorflow and regulating the key parameters to get the optimal accuracy and the minimum variance.For verifying the advantages of deep learning,this method is compared with traditional machine learning algorithm(SVM)and deep feedforward neural network.The result demonstrated that this method has a better performance.The whole process lays a foundation for the later testing in actual working conditions.