Prototype control and monitoring system safety device from leakage ammonia at marine loading arm with comparison of Neural Network (NN) and Extreme Learning Machine (ELM) method

This paper presents design and research studies in marine loading arm plant system. Artificial Neural Network (NN) and ELM (Extreme Learning Machine) methods are used and compared in this valve control system by implement it in a prototype using microcontroller. This prototype use value of temperature sensor and value of ammonia gas sensor in the furnace as parameter of heat to control the flow of air and valve of safety device. The temperature sensor used in this research is the type of DHT11. The ammonia gas sensor is MQ sensor. This prototype also uses fan and servo as the actuator. Fans are used to supply the oxygen and servo is used to control the valve of ammonia. From the experimental result, the data shows that the optimization of safety device system using ELM method works better compared with NN. The control system has a very good response and it can work well (percentage of error is less than 0.4%). Hence, if the system is applied in the marine loading arm plant, it could improve the performance of safety device control systems and save the leakage of ammonia gas.

[1]  C. L. Liu,et al.  Robust microgrid power flow using particle swarm optimization , 2013, International School on Nonsinusoidal Currents and Compensation 2013 (ISNCC 2013).

[2]  Kuo Lung Lian,et al.  Microgrid power flow using Homotopic and Runge-Kutta Method , 2015, 2015 IEEE 2nd International Future Energy Electronics Conference (IFEEC).

[3]  Tsai-Hsiang Chen,et al.  A distribution power flow using particle swarm optimization , 2012, PES 2012.

[4]  Adi Soeprijanto,et al.  Incremental Particle Swarm Optimizer with local search for Optimal Power Flow Subjected to Digital GCC based on Neural Network , 2012 .

[5]  Adi Soeprijanto,et al.  Digital Generator Capability Curve for Improving Optimal Power Flow based on IPSO , 2013 .

[6]  Adi Soeprijanto,et al.  Comparison methods of Fuzzy Logic Control and Feed Forward Neural Network in automatic operating temperature and humidity control system (Oyster Mushroom Farm House) using microcontroller , 2016, 2016 International Symposium on Electronics and Smart Devices (ISESD).

[7]  Adi Soeprijanto,et al.  Neural network implementation for invers kinematic model of arm drawing robot , 2016, 2016 International Symposium on Electronics and Smart Devices (ISESD).

[8]  Adi Soeprijanto.  Neural Network Optimal Power Flow(NN-OPF) based on IPSO with Developed Load Cluster Method , 2010 .

[9]  Adi Soeprijanto,et al.  Optimal Placement and Sizing of Distributed Generation for Minimize Losses in Unbalance Radial Distribution Systems Using Quantum Genetic Algorithm , 2014 .

[10]  Adi Soeprijanto,et al.  Smart-Meter based on current transient signal signature and constructive backpropagation method , 2014, 2014 The 1st International Conference on Information Technology, Computer, and Electrical Engineering.

[11]  Adi Soeprijanto,et al.  Rotor bars fault detection by DFT spectral analysis and Extreme Learning Machine , 2016, 2016 International Symposium on Electronics and Smart Devices (ISESD).

[12]  Kuo Lung Lian,et al.  Improved Robustness of Sequential Three Phase Power Flow Using Homotopic Method , 2013 .

[13]  Adi Soeprijanto,et al.  Real-time unbalanced load flow development using direct-ZBR method and modified lambda iteration for on-line monitoring and control , 2013 .