| This thesis presents a kind of new style stroke sensing cylinder based on inside magnetic measuring. Its significance is increasing the degree of automation. Based on published documents, the key technique of stroke sensing cylinder-measurement system is researched. Scale code technique of magnetic scale is presented. Some kinds of magnetic scale code methods is compared and analyzed. A frame superimposed scale coding is presented. From the further studies on the frame superimposed scale code methods, an encoding method is found out. Some examples of this encoding method result are listed. The result of the experiments indicates that magnetic scale coding is feasible. A suitable measuring magnetic circuit structure is selected. To obtain the parameters of measuring magnetic circuit structure, theory research is done. Theory research based on overrelaxation: according to electromagnetic theory fundamental equation, boundary condition is analyzed, iterative formula is deduced, overrelaxation factor is selected and calculation program is composed. Through calculation on varied parameter, a group of suitable parameter is found out. The results of experiments indicate that the theoretical analysis to the magnetic field of the magnetic structure is proper. To remove the result error come from the change of gap between piston rod and sensor, a signal processing method accompanied with machine learning process is researched. The process of fitting sensor's output curve with genetic algorithm is studied. The accuracy error of experiments using this signal processing method is less than 0.2mm.In the first chapter: the state-of-the-arts of stroke measuring technique used cylinder summarized. Based on those, the key techniques and develop trends of stroke sensing cylinder are proposed. The significance of study is stated. At last, the main contents of this dissertation are presented.In the second chapter: based on the principle of stroke sensing, some kinds of measuring magnetic circuit structure are compared and analyzed, and asuitable circuit structure is selected. To obtain the parameter of measuringmagnetic circuit structure, theory research is done. Theory research based on overrelaxation: according to electromagnetic theory fundamental equation, boundary condition is analyzed, iterative formula is deduced, overrelaxation factor is selected and calculation program is composed. Through calculation on varied parameter, a group of suitable parameter is found out.In the third chapter: the conclusion of theory research is verified through experiment. In the experiments, two kind of magnetic sensors are used. The results of the experiments indicate that the theoretical analysis to the magnetic field of the magnetic structure is proper.In the fourth chapter: research on the signal processing method accompanied with machine learning process to remove the influence of the gap change between piston rod and sensor. An axial direction line multi-sensors placed method is presented. In the process of machine learning, use Genetic Algorithm to fit the curve of sensor's output. To improve the convergence speed of genetic algorithm, hill climbing method is joined with.In the fifth chapter: some kinds of magnetic scale code methods is compared and analyzed. Imitating serial communication scale code and frame superimposed scale code are presented. Compared two kinds of scale code, the frame superimposed scale code is better. From the further studies, an encoding method of frame superimposed is found out. Some examples of this encoding method result are listed. The results of the experiments indicate that the scale code is feasible.In the sixth chapter: hardware of signal system is designed. To increase the speed of process, the hardware make of three MCU, two 80C196KC and one 89C51. According to the hardware and method of signal process, work out the software.In the seventh chapter: do general experiments of the stroke sensing cylinder. The experiment's items include: stability experiments, st...
|
PDF Full Text Request