Magnetic levitation is an example of a nonlinear system that is naturally unstable. In the control system field, it can be also used to check the effectiveness of control methods. Unlike other researches overcoming by analytical method, this research investigates the solution of the nonlinearity of the magnetic field by placing two effect hall sensors on the top and bottom of magnetic coils. After decreasing the effect of nonlinearity, two PID control tuning methods, Ziegler Nichols (ZN) and Cohen Coon (CC), are compared to obtain an appropriate control structure and its initial parameters. The experimental result shows that the best control structure of ZN method is PID whereas its initial parameters, Kp, Ti, and Td, are 15.33, 0.036 and 0.009, respectively. The rise-time and settling time of the response are 0.54s and 0.59s. The appropriate control structure is PI control whereas Kp and Ti are 30.05 and 0.012 from CC method. The rise-time and settling time resulted are 0.8s and 0.67s. The ZN method with PID control structure has a better response with smaller rise-time and settling time than PI control from CC Method. This research is benefit for solving a problem dealing with magnetic levitation control.

In a process control system influenced by several parameters, which are usually located and separated in a large area, the one-central controller is not effective anymore. It is computationally hard for the controller hardware to do many tasks at the same time, therefore, they are considered to distribute in Local Control Units (LCUs). This research investigates the distributed PID controllers in the water-mixing process. The main controlled variables (SVG) are the level and temperature influenced by three parameters (flow-in of cold water, flow-in of hot water, and flow-out of the main tank), which are then distributed to LCU-1 to LCU-3. To maintain the SVG, the master controller is then designed based on the difference between SVG and feedback of the main-tank (PVG) so that it will deliver SV for each LCU. The results show that with control parameters in LCU-1 (Kp, Ti, Td are 2, 10, 0.46, respectively), LCU-2 (Kp, Ti, Td are 8, 6, 0.6), and LCU-3 (Kp, Ti, Td are 3, 70, 20), the response of water temperature (measured based on settling-time [Ts], % overshoot [OS], and rise-time [Tr]) are 42s, 0%, 45.3s. In the water level control, the response can be maintained with the 60s, 0%, 62s, for Ts, %OS, Tr, respectively.