Monte-carlo based robust analytical method for optimal sizing and reliability of hybrid renewable energy system

Abstract

The need for a more reliable power from the utility grid and ever-increasing concerns on Greenhouse Gas (GHG) emission effect has globally promoting Renewable Energy Sources (RES). RES is increasingly being adopted in complementing traditional fossil fuels in the energy power supplies. Hybrid Renewable Energy (HRE) systems incorporating wind and solar sources offers lower costs, higher reliability, reduced investment risks, fuel diversification etc. However, wind speed and solar radiation are characterized by their limitations of inherent intermittency and variability. These limitations have led to the concept of optimal sizing and reliability assessments to maintain a balance between generated power and the system loads. Nonetheless, RES reliability assessment studies are site-specific, but existing studies are inexhaustive given the capacity availability and reliability requirements of various sites as well as their performance evaluations. This thesis presents the optimal sizing and reliability assessment of a hybrid solar and wind energy systems for a selected location. Weibull statistical method and air temperature amplitude based statistical models are adopted for wind and solar energy potential assessments of the selected site. The Weibull parameters were estimated using standard deviation method for wind energy potential assessment. Moreover, the air temperature based models of Hargreaves and Samani; Allen; Samani; and Bristow-Campbell models were used for solar energy potential assessment. Simulation of the uncertainty in the wind speed and its probability distribution is performed by using Auto-Regressive Moving Average (ARMA) model to improve wind speed normal distribution. In this approach, the best normal distribution for the simulated wind speed for the reliability analysis is chosen. To improve the performance of the Photovoltaic (PV) module, a single diode six parameter model is developed. First, the P-V and I-V curves were used to generate the required constraints. These constraints were then used to obtain the solution vector of the six parameters using MATLAB and System Advisor Model (SAM). Also, the system’s capacity availability and reliability was assessed using Monte Carlo (MC) simulation. Finally, the result of the MC reliability assessment is later served as Loss of Power Supply Probability (LPSP) constraints to Artificial Bee Colony (ABC) algorithm for the system’s optimal sizing and enhanced reliability assessment. Results from the study show that both wind and solar energy potential of the selected site is high and can generate power at utility level. The ARMA simulated wind speed shows an improvement of 21.8% in standard deviation over the measured wind speed. The adoption of the negative components in the ARMA model transformation resulted in least error of 23.34% in the final wind simulation. Results obtained based on the six parameter solution vector gives improved performance of the PV module. Using the developed MC technique, capacity availability of 100% and LPSP of zero is achieved. The developed ABC algorithm resulted in system reliability improvement of 98.92% when the MC results are constraint into the ABC for the optimal sizing. Various results were validated at appropriate sections and finally, the optimal sizing results of PV/battery RES power system is found to give the best reliability. Such a system has great reliability and can be implemented in facilities requiring constant power supplies such as critical infrastructure.