This research aims to study the using oil palm fiber to develop ceiling board product. The mixture ratios of Gypsum plaster: oil palm fiber: tap water: catalyst are 1 : 0.15 : 1: 0.03 1 : 0.20 : 1: 0.03 1 : 0.25 : 1: 0.03 1 : 0.30: 1: 0.03 and 1 : 0.35 : 1: 0.03 by weight. The ceiling board samples are cast in 30 x 30 x 1 centimeter in dimension. The ceiling board sample testing follows the TIS 219-2552. The results show that the longitudinal and lateral breaking load Nail pull resistance density and thermal conductivity of ceiling board with high quantity of oil palm fiber are lower than ceiling board with low quantity of oil palm fiber while that the deflection and water absorption of ceiling board with high quantity of oil palm fiber are higher. However all of oil palm fiber samples with lower than 0.2 of ratio can pass the standard.

The objectives of this research are to propose an approate and efficient model for analyzing rice corn and sugar yields data. A linear mixed model (LMM) with spatial correlation following the conditional autoregressive model (CAR) was adopted. The results show that for rice yields if the amount of rainfall in a harvest season increases 1 mm the yields will decease by 25.74 tons. if the temperature in the harvest season increases 1 Celsius the yields will decease by 212.60 tons and according to the trend the yield will increase by 15.80 tons each year. For sugar cane yields if the amount of rainfall in the harvest season increases 1 mm the yields will decease by 51.35 tons. if the temperature in the harvest season increase 1 Celsius the yields will incease by 121.30 tons and according to the trend the yield will increase by 29.93.80 tons each year. For corn yield if the amount of rainfall in the havest season increases 1 mm the yields will incease by 73.45 tons. if the temperature in he havest season increases 1 celsius the yields will decease by 40.44 tons and according to the trend the yield will increase by 1.74 tons each year.

The objectives of this research are to propose a proper model for risk assessment for drought in Thailand. A hidden Markov model is adopted. The results indicate that the amount of rainfall depends on the hidden states drought or non-drought state. If the state changes from drought to non-drought the amount of rain will increase. The probability of changing from drought state to non-drought state is higher than from drought state to drought state. The probability of drought occurrence in February is highest followed by in November March December January July April June October September August and May respectively. When the temperature increases the probability of drought occurrence will increase.

The objectives of this research are to propose a proper model for detecting defective machines in auto parts factories to apply the proposed model to the defective product data and to investigate the factors influencing on the machine producing defective products. A Bayesian hidden Markov model (HMM) is adopted. The results show that if the number of defective products increases the machine will be in the hidden bad state. The probability of changing from a hidden bad state to a hidden good state is higher than the probability of changing from a hidden good state to a hidden good state. The highest probability that the machine will be in a good state is at the work process 4 followed by 1 3 and 2 respectively. The highest probability that the machine will be in hidden good state is at the Worker no. 10 followed 9 3 12 7 2 1 6 11 4 8 and 5 respectively and The highest probability that the machine will be in the hidden good state is at the product type 1 followed by 2 and 3 respectively.

This thesis presents an assessment of voltage sags when faults are happened in the 115 kV transmission line of the Provincial Electricity Authority (PEA) when distributed generation (DG) is installed in the system. The computer program PSCAD/EMTDC (Power Systems Computer Aided Design / Electromagnetic Transient in DC system) version 4.2 is applied to provide electrical parameters for the modeling and simulation and to analyze the fault occurred which resulted in voltage sag. Test results from PSCAD/EMTDC show the values of voltage sag when 5 types of fault are occurred (Single Line to Ground Fault Double Line to Ground Fault Three Phase to Ground Fault Line to Fault and Three Phase Fault). For all cases the value of root mean square voltage before fault is 100% or 1.0 pu. It can be seen that for both radial line and closed circuit when DG are installed in the system the value of phase to ground RMS residual voltage and phase to phase RMS residual voltage at the bus of Bang-pa-in 2 substation are more than the values in case of the system without DG.