Cost Effective and Reliable Energy System for Kathmandu University Complex

Thèse de Master , 2018 , 99 Pages , Note: 4.0

Ingénierie - Technique énergétique

This thesis presents the techno-economic feasibility study and reliability analysis of a hybrid energy system for Kathmandu University (KU), Dhulikhel, Nepal by using the resources available at premises. For the analysis, both on-grid and off-grid cases are considered. Various energy mix models are identified and optimized for the comparison purposes using HOMER (Hybrid Optimization Model for Multiple Energy Resources) software.

The results are also analyzed and compared based on their reliability using Monte Carlo Simulation in RAPTOR software. Based on various combinations and simulations, best cost-effective and reliable model has been proposed.

From this study, it is found that being a cost-effective model doesn’t mean it’s reliable too. In order to get the reliable supply, one may have to bear the extra cost to provide standby or emergency supply system. So, the cost plays an important role to determine what level of reliability is required.

Table of contents

ACKNOWLEDGMENTS

ABSTRACT

GLOSSARY OF ABBREVIATIONS

LIST OF SYMBOLS

LIST OF FIGURES

Chapter 1
INTRODUCTION
1.1 Background
1.2 Case Study Area
1.3 Statement of Problem
1.4 Research Objectives
1.5 Research Approaches
1.6 Research Nobility
1.7 Scope and Limitations of the Thesis
1.8 Thesis Organization

Chapter 2
LITERATURE REVIEW
2.1 Hybrid Energy Model
2.2 Cost Analysis
2.3 Reliability Analysis
2.4 HOMER software for optimization
2.5 RAPTOR software for Reliability and Availability analysis
2.6 Comparison between On-grid and Off-grid Hybrid System
2.7 Challenges and Issues of Interconnecting Grid and Embedded Generation

Chapter 3
RESEARCH METHODOLOGY
3.1. Problem formulation
3.2. Literature Review
3.3. Input data gathering
3.4. Model Building
3.5. Result Analysis

Chapter 4
RESULTS AND DISCUSSION
4.1 Cost Analysis of Energy Mix Models
4.2 Reliability and Availability analysis of different Energy Mix Models

Chapter 5
CONCLUSION AND RECOMMENDATION
5.1 Conclusion
5.2 Recommendations & Future Work

REFRENCES

ANNEX

Annex 1: TOD meter data of KU for year 2073/74.

Annex 2: NPC of each component

Annex 3: Cash flow of different models

Annex 4: Raptor Monte Carlo Model

Annex 5: Results from Raptor Simulations

GLOSSARY OF ABBREVIATIONS

Abbildung in dieser Leseprobe nicht enthalten

LIST OF SYMBOLS

Abbildung in dieser Leseprobe nicht enthalten

LIST OF FIGURES

Figure 1. 1 Map of Kathmandu University [6]

Figure 1. 2 Satellite view of Kathmandu University [7]

Figure 2. 1 Hybrid Energy System [15]

Figure 2. 2 Two state model

Figure 2. 3 Components in series

Figure 2. 4 Components in parallel

Figure 2. 5 Steps for Monte Carlo Simulation

Figure 3. 1 Block diagram of summary of activities carried out

Figure 3. 2 Load pattern of Kathmandu University

Figure 3. 3 Global horizontal radiation

Figure 3. 4 Extraterrestrial horizontal radiation daily profile

Figure 4. 1 Energy mix model case I

Figure 4. 2 Cash flow summary for model 1, Case I

Figure 4. 3 Nominal Cash flows in each year, model 1, Case I

Figure 4. 4 Monthly average energy contribution by components, model 1, Case I

Figure 4. 5 Monthly average electricity production for model 1, Case I

Figure 4. 6 Battery bank state of charge, model 1, Case I

Figure 4. 7 Inverter output for model 1, Case I

Figure 4. 8 Rectifier output for model 1, Case I

Figure 4. 9 Energy mix model, Case II

Figure 4. 10: Cash flow summary, model 1, case II

Figure 4. 11 Nominal Cash flow in each year, model 1, case II

Figure 4. 12 Monthly average energy contribution by components, model 1, Case II

Figure 4. 13 Solar output for model 1, Case II

Figure 4. 14 Diesel generator 1 output for model 1, Case II

Figure 4. 15 Diesel generator 2 output for model 1, Case II

Figure 4. 16 Battery Bank state of charge for model 1, Case II

Figure 4. 17 Battery Bank state of charge frequency histogram, model 1, Case II

Figure 4. 18 Battery Bank SOC percentage monthly statistics, model 1, Case II

Figure 4. 19 Inverter output for model 1, case II

Figure 4. 20 Inverter output for model 1, case II

Figure 4. 21 Cash flow summary of each component, model 1, case III

Figure 4. 22 Nominal Cash flow in each year, model 1, case III

Figure 4. 23 Solar output, model 1, case III

Figure 4. 24 Converter output, model 1, case III

Figure 4. 25 Converter output, model 1, case III

Figure 4. 26 Raptor Monte Carlo Model for Model 8, Case I

Figure 4. 27 Raptor Monte Carlo Model for Model 2, Case II

LIST OF TABLES

Table 3. 1 Input values of Cost Estimation

Table 3. 2 TOD Meter Tariff [26]:

Table 3. 3 MTBF & MTTR of different Components

Table 4. 1 Various energy mix model that can fulfill the demand, Case I

Table 4. 2 Costs of the different models, Case I

Table 4. 3 Cost Summary for model 1, Case I

Table 4. 4 Average electricity Contribution by components in model 1, Case I

Table 4. 5 Solar PV output data, Model 1, Case I

Table 4. 6 Battery details for model 1, Case I

Table 4. 7 Converter data for model 1, Case I

Table 4. 8 Grid purchases for model 1, Case I

Table 4. 9 Emissions of different gases for model 1, Case I

Table 4. 10 Various energy mix model that can fulfill the demand, Case II

Table 4. 11 Costs of the different models, Case II

Table 4. 12 Total NPC Cost of different Components for model 1, case II

Table 4. 13 Average electricity Contribution by components, model 1, case II

Table 4. 14 Solar output data for model 1, case II

Table 4. 15 Diesel generator 1 output data, model 1, case II

Table 4. 16 Diesel generator 2 output data, model 1, case II

Table 4. 17 Battery data, model 1, case II

Table 4. 18 Converter data for model 1, case II

Table 4. 19 Emission data for model 1, case II

Table 4. 20 Energy mix model, Case III

Table 4. 21 Cost of the different models, Case III

Table 4. 22 Total NPC Cost of different Components for model 1, case III

Table 4. 23 Energy produced by each component, model 1, case III

Table 4. 24 Solar output data, model 1, case III

Table 4. 25 Converter data, model 1, case III

Table 4. 26 Grid data, model 1, case III

Table 4. 27 Emission data, model 1, case III

Table 4. 28 Best models comparison in each case

Table 4. 29 Availability & MTBF of Various energy mix model, Case I

Table 4. 30 MTTR & System failures of various energy mix model, Case I

Table 4. 31 Results from Raptor Simulations, Model 8, Case I

Table 4. 32 Availability & MTBF of Various energy mix model, Case II

Table 4. 33 MTTR & System Failures of various energy mix model, Case II

Table 4. 34 Results from Raptor Simulations, Model 2, Case II

Table 4. 35 Availability & MTBF of Various energy mix model, Case III

Table 4. 36 MTTR & System failures of various energy mix model, Case III

Table 4. 37 Best reliable models comparison in each case

DEDICATION

With a Prayer I dedicate this dissertation work to my father Rajendra Kumar Sah, my mother Rana Kumari Sah and my brother Suraj Sah.

ACKNOWLEDGMENTS

First and foremost, I would like to thank God, the Almighty, for nothing is possible without His will.

I would like to express my profound gratitude and sincere appreciation to my advisor Ashish Shrestha, Lecturer, Department of Electrical and Electronics Engineering, Kathmandu University for his encouragement, thought proving inspiration and continuing guidance for accomplishing this valuable task. I would also like to thank Associate Professor Brijesh Adhikary, head of Department of Electrical and Electronics Engineering, Kathmandu University and whole department for helping me through the whole process of the thesis.

I would also like to express my deepest gratitude to Anil Adhikari, Electrical Engineer at Nepal Electricity Authority, Bhaktapur DCS, Naresh Manandhar, Electrical Engineer at Nepal Electricity Authority, Kavre DCS, Ganesh Kumar Sah, Electrical Engineer, Project Management Directorate, Nepal Electricity Authority and Sagar Mahat, generator operator at Kathmandu University for providing me necessary data required for the thesis. I truly want to thank Ramesh Paudel, Asst. Manager of Power Trade Department, Nepal Electricity Authority for the idea regarding the thesis topic. I would also like to thank Mohamed EL-Shimy, professor of electrical energy systems in department of Electrical Power and Machines, Ain Shams University, Cairo, Egypt for guidance throughout the completion of the thesis work. I would also like to thank all of my friends especially Sabin Oli, Ajay Singh and Pankaj Kumar Rauniyar for their help in various occasions.

Finally, I must express my very profound gratitude to my Parents for providing me with unfailing support and continuous encouragement throughout my years of study and through the process of researching and writing this thesis. This accomplishment would not have been possible without them. Thank you.

ABSTRACT

Cost effective and Reliable energy system is the need of any mankind. In this study, cost effective hybrid energy mix model from different energy resources available at Kathmandu University (KU) has been identified which can fulfill the load demand of KU throughout the year. The resources available at KU are grid, Diesel Generator 1 (DG1) of 32 kW, DG2 of 100 kW, DG3 of 120 kW, and Solar PV of 65 kWp with battery. Energy mix models for three cases i.e. on-grid with capital investments, off-grid with capital investments and on-grid without capital investments has been identified for comparison purposes. On-grid Case I and III contained the eight different energy mix models whereas off-grid case II contained only six possible energy mix models which can fulfill the load demand of KU. Load profile of KU was calculated from the data collected from Nepal Electricity Authority (NEA) for a whole year. The cost analysis was performed using Hybrid Optimization Model for Electric Renewable (HOMER) software, a freeware tool for designing and analyzing hybrid power systems based on the load profiles. HOMER was used to optimize the system based upon the Total Net Present Cost (TNPC). Cash flow summary for all the energy mix models for each case was obtained for the optimal cost allocation of each individual component present in the system. After the cost analysis, reliability and availability analysis of each model for each cases was done using Monte Carlo Simulation in RAPTOR software, a free software for the analysis of Reliability, Availability and Maintability.

In case I, the model containing grid, solar PV and six batteries at 24 V DC has been found as the cheapest energy mix model having least NPC as $ 463,665 and COE as 0.129 $/kWh whereas the model containing grid, DG1, DG2, DG3, and solar PV with six batteries at 24 V DC has been found as the most reliable model having MTBR 997.42 hours among 1000 hours, availability of 0.9999 and system failure of 0.06 times but this model was found as the most expensive model for case I. When capital cost was excluded as of case III, then the same reliable model of case I was found as the cheapest model with total NPC as $ 385,293 and COE as $ 0.105 $/kWh and the most reliable model as well. When grid was not available as of case II, the model containing DG1, DG2 and solar with 192 batteries in the architecture of 32 parallel string each containing six batteries has been found as the best cost effective model with TNPC as $ 4,242,027 and COE as 1.198 $/kWh. But the reliable model in this case was found as the model containing DG1, DG2, DG3 and solar with 192 batteries in the architecture of 32 parallel string each containing six batteries with MTBF of 967.32 hours and availability of 0.9983. So, at the present scenario for the KU when the initial capital cost has already been invested, the model containing grid, solar PV with battery and DG1, DG2, DG3 as backup units is the best cost effective as well as reliable model.

From this study, it is concluded that being a cost effective model does not mean it is reliable too. In order to get the continuity of supply, one may have to bear the extra cost to use standby and emergency power systems. So the cost plays an important role to determine what level of reliability and availability of the system is required.

Chapter 1 INTRODUCTION

1.1 Background

Electric Energy is one of the most important aspect for progress and development of any nation. Whole country rely on electricity 24/7. Reliable power supply at acceptable cost is vital for the economic growth and global development of any country [1]. The national grid of Nepal is unable to meet the load demand and there is accidental power cut time to time. Unreliable power supply is one of the major constraint for the economic growth of the nation. According to [2], the economic loss in Nepal’s industrial sector is US$ 24.7 million a year due to planned and unplanned outages which is 4.43% of the industrial sector GDP or 0.47% of the national GDP in 2001. Also 76% of the industries have standby generations which are able to supply 74% of the full load electricity requirements only. Interruption of power supply results in the loss of time and money.

Different facilities have different requirements for reliability of electric power supply. Reliable power system is very essential for the sensitive organization such as industries, hospitals, college, universities, offices etc. For the country like Nepal, the interruption of power supply is very often due to which the day to day work and daily life of the people is highly affected. There is huge loss of time and money due to this. Standby and emergency power systems are installed at such premises to serve the load whenever there is interruption of power supply [3]. In order to get the continuity of supply, one may have to bear the extra cost to use standby and emergency power systems. The choice of various alternatives depends upon the reliability and the cost. These alternatives may be the various configurations of utility grid, diesel generator, solar PV, wind turbine generator etc. Cost and Reliability analysis of various configurations is important to choose the best cost effective and reliable configurations. Cost plays an important role to choose the different level of reliability as required.

1.2 Case Study Area

Kathmandu University (KU) central campus is chosen as the site for the case study in this thesis. KU was established in 1991 AD with the motto "Quality Education for Leadership". KU is an autonomous, non-government, not for profit institution which maintains high standards of academic excellence. It develops leaders in professional areas through quality education. KU provides so many undergraduate and postgraduate programs in the fields of engineering, medical sciences, science, management, education, arts, music and law. It is about 30 kilometers east of Kathmandu having round-the-year pleasant climate and panoramic Himalayan Views [4].

It is located in a mountainous landscape in Dhulikhel Municipality, Kavrepalanchowk district, a part of province No.3. KU lies at 27°37'09.5"N and 85°32'19.5"E with an altitude of 1511 meter above sea level [5]. Map and satellite view of KU is shown below in the Figure 1.1 and 1.2.

Image deleted due to copyright issues

Figure 1. 1 Map of Kathmandu University [6]

Image deleted due to copyright issues

Figure 1. 2 Satellite view of Kathmandu University [7]

1.3 Statement of Problem

The national grid of Nepal cannot meet the load demand of the country due to which there is frequent unplanned accidental power cut time to time. In KU, there are certain emergency loads which are very sensitive to interruptions such as loads during lab, presentation, seminars, classes and other official works. So the only choice left during the interruption is to use the emergency and standby power systems. Operation cost and reliability is different for different configurations. Cost and Reliability play an important role to choose the various alternatives and their different configurations. The possible choices are various configurations of utility grid, solar PV with battery and different sizes of diesel generator. So it is essential to find the best cost effective and reliable configuration of different choices which can fulfill the load demand of Kathmandu University.

1.4 Research Objectives

The main objective of this Research is to identify the best cost effective and reliable energy mix model using resources available at Kathmandu University which can fulfill the load demand of the Kathmandu University.

The specific objectives of the study are:

(i) To identify and compare the various energy mix models based on its levelized cost.

(ii) To identify the best energy mix model based on its reliability.

1.5 Research Approaches

Various Energy mix model was identified using HOMER (Hybrid Optimization Model for Multiple Energy Resources) software. HOMER is a simulation software which simulates and check all the possible combinations of different sources according to the input in a single run, and then arrange the different models on the basis of TNPC of each model which can fulfill the load demand. Homer simulates hundreds or even thousands of systems depending on how we set up our problem and finally it gives the best optimum model based on the TNPC of each model [8].

After the simulation, the results were also analyzed and compared based on their reliability and availability. Reliability and Availability of each model was calculated using Monte Carlo simulation in RAPTOR software. RAPTOR software is a modelling software used for the analysis of Reliability, Availability and Maintability of a system containing so many blocks in different series parallel combinations. It works on the basis of Monte Carlo Simulation which randomly checks the operation of the equipment before it fails and it also checks the random time that will be required to repair the failed equipment on each run [9].

1.6 Research Nobility

This thesis introduces new approaches for the modelling of energy for cost effective and reliable electricity supply for KU. Different Energy mix models were identified on the basis of TNPC and cost of energy. Reliability and Availability of each model was calculated using Monte Carlo simulation. This thesis has a novel idea to develop a cost effective and reliable power system for KU.

1.7 Scope and Limitations of the Thesis

There are some limitations and assumptions considered during conducting this thesis.

i) The load data of the KU was taken of the year 2016/2017 AD and assumed the same load pattern during the whole project life cycle.

ii) The possible accurate cost of the components and its Operation & Maintenance (O & M) cost were considered using various techniques and websites. However actual current price might be different. Also, Integration cost has been excluded.

iii) Only unplanned failure (non-momentary) of the feeder were included in this study for reliability analysis. It was assumed that all the failures could be repaired.

1.8 Thesis Organization

The dissertation consists of five chapters.

Chapter 1 presents a brief introduction about the importance of reliable power system for the development of any country and at reasonable cost. It explains about the demerits due to planned and unplanned outages. It also explains about the case study area, problem statement, Research objectives, Research approaches, Research nobility, Scope and limitations of the thesis.

Chapter 2 presents literature review on cost effective and reliable power system. It also explains about the hybrid energy model and its advantages. It explains about the cost and reliability analysis approaches with the software used for the simulations.

Chapter 3 proposes the methodology carried out during completion of the thesis. It explains step by step activities that is carried out during this research.

Chapter 4 studies the best cost effective and reliable energy mix model for the case study area for different cases.

Chapter 5 provides the conclusion and recommendations for future work.

Chapter 2 LITERATURE REVIEW

2.1 Hybrid Energy Model

Hybrid Energy Model is the combination of two or more than two energy resources which can fulfill the load demand. The sources may include solar PV, wind energy, diesel generators, micro-hydro etc. Because it is the combination of two or more than two energy sources, non-availability of one source will be fulfilled by the availability of another sources. For example, for any hybrid system containing solar and wind energy, non-availability of solar energy during night time will be fulfilled by the availability of wind energy.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2. 1 Hybrid Energy System [15]

Advantages of Hybrid Energy System are [16]:

i) Hybrid Systems achieve better utilization of renewable resources, thus minimizing the cost of energy.

ii) High system reliability and availability.

iii) Better load matching.

iv) Less maintenance and uses less fuel.

Whenever load increases, its capacity can be increased by adding renewable energy or diesel generator or both. Some renewable resources such as micro hydro and diesel generator produce AC power whereas solar PV and wind turbine produce DC power which can be managed and supplied using converting devices like converter. Excess energy can be stored in energy storage device which can be used during peak hours or during less generation.

In this research, both off grid and on grid hybrid system was modelled using solar PV with battery and three different sizes of diesel generator. Converter is used to convert the DC power into AC and vice versa. Different configurations of energy mix model was identified using HOMER software and compared on the basis of their operation costs. Also the reliability and availability of each model was calculated using Monte Carlo Simulation in RAPTOR software and the best model was identified.

2.2 Cost Analysis

It is very important to determine the cost of the projects for the utilization and allocation of the proper resources. For the best project planning and management, accurate cost estimates of the projects are very important. In [10], the authors discussed on optimization results and effective cost analysis of Hybrid Renewable Energy System (HRES) using HOMER. They showed the combination of PV, Wind Generator (WG), and battery as the optimal choice off grid system on the basis of costs for their locations. In [11], the authors optimized both off-grid and on-grid wind-solar hybrid power system using HOMER software based on load profile, solar irradiance and wind speed. They concluded that off grid power system is more expensive than grid connected power system to serve the same load. In [12], the optimum combination of off grid HRES system were found by the authors on the basis of cost analysis for their location using HOMER software. In [13], the authors had made comparison between hybrids PV-wind on grid system with off grid system. The authors also performed Sensitive analysis using HOMER software to check the variation in grid energy costs on its total costs. Finally they concluded that the grid connected hybrid system of PV-wind energy is most economical for their region and total net present cost was also less than the off grid system. In [14], HOMER software was used to model a cost effective off grid hybrid system containing solar PV, wind energy and diesel generator for a remote island.

2.3 Reliability Analysis

Reliability studies are important for the system planning, and day to day operating decisions. Reliability analysis of various options is important for the proper selection of standby power systems [3]. Reliability tells us how much failure to expect [9]. Reliability is a measure of the probability that an item will perform its intended or required function or mission for a specified interval under stated conditions [16]. There are two factors used for the measuring the reliability, availability and maintainability of any system. The first factor that demonstrates the reliability is either the Mean Time Between Failure (MTBF) for repairable systems, or Mean Time To Fail (MTTF) for non-repairable systems and the second factor is the failure rate (λ) which is the reciprocal of MTTF or MTBF. So MTBF or MTTF is a direct measure of reliability. More the MTBF or MTTF, more is the reliability. The repair rate (µ) is the reciprocal of Mean Time To Repair (MTTR).

Abbildung in dieser Leseprobe nicht enthalten

Figure 2. 2 Two state model

Figure 2.2 shows the two state model for a system. Where repair rate and failure rate is given by:

λ= ..(i) where, MTBF= Total uptime/ No. of failures and

µ= .(ii) where, MTTR= Total downtime/No. of failures

The hybrid resources is connected together in different series-parallel configurations to improve the availability of the power supply. Each component has its own failure and repair rates.

The probability of availability and unavailability of a component having failure and repair rates λ & µ is given as:

Abbildung in dieser Leseprobe nicht enthalten

Abbildung in dieser Leseprobe nicht enthalten

a) When two components are in series:

Each component has its own failure rate and repair rate. The equivalent failure and repair rates are given in equation (v) and (vi) respectively.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2. 3 Components in series

Abbildung in dieser Leseprobe nicht enthalten

b) When two components are in Parallel:

Similarly, when two components are connected in parallel as shown in Figure 2.4, the equivalent failure and repair rates are given in equation (vi) and (vii).

Abbildung in dieser Leseprobe nicht enthalten

Figure 2. 4 Components in parallel

Abbildung in dieser Leseprobe nicht enthalten

A system or device is said to be working good if it doesn’t fail during its time of operation. But the system or devices are expected to fail during its operation and after successful repair it is returned to the service during the entire useful life of the device or systems. In this case the most appropriate measure of reliability is the availability of the device which tells us how much time to expect [9].

2.3.1 Reliability evaluation approaches

There are two reliability evaluation approaches i.e. deterministic and Probabilistic (or stochastic) approaches. In deterministic model, actual result is found from the conditions under which the experiments are carried out. We use physical considerations to predict an almost accurate outcome. It is meant to yield a single solution. The major drawbacks of this approach is that it doesn’t account for the stochastic nature of system behavior. Whereas Probabilistic model predicts more of a probable outcome i.e. it gives distribution of possible outcomes and gives some measure of how likely each is to occur. It will determine a probabilistic behaviour of the observable outcome. There are two main approaches for probabilistic approaches which is analytical approaches and Monte Carlo Simulation (MCS). Analytical techniques uses direct analytical solutions to calculate the reliability indices by representing the system in mathematical models whereas Monte Carlo Simulation estimates reliability indices by simulating the actual random behaviour of the system [17]. Analytical approaches becomes unrealistic due to its rough representation that doesn’t fit to the real system and takes too much time for the development of the analytical model and is impracticable. So simulation is required for analyzing complex systems which is close to reality. The failures in power systems are random in nature. MCS can be used to simulate these failures. MCS is a probabilistic method that can be used to predict the behaviour of the system.

2.3.2 Probability distributions in reliability evaluation

For reliability evaluation, the parameters associated are described by probability distribution. The system or its components will not all fail at a fixed time but will fail at different times in the future depending on their type, manufacture and operating condition. So, the time to failure will follow a probability distribution which may or may not be known and describes the probability of certain component failing or not within a time specified. This probability value is a function of the time that is specified or considered. Similarly, a system that is failed and is being repaired is unlikely to have a constant repair time and times-to-repair are distributed according to a probability distribution which again may be known or not. In all practical cases, the appropriate probability distribution must be deduced from sample testing or from a data collection scheme associated with operation of the components, devices or systems.

There are mainly two types of distributions i.e. discrete and continuous. Discrete distribution represent random variables that can assume only certain discrete values whereas continuous distribution represent random variables that can assume any value within a finite range. The continuous distributions include Normal or Gaussian, Exponential, Weibull, Gamma, Rayleigh, Lognormal whereas Binomial and Poisson distributions fall in the discrete distribution.

Common probability distributions are [18]:

i) Normal Distribution: In this type of distribution, the mean or expected value and a standard deviation to describe the variation about the mean is required. It is symmetric. Values in the middle near the mean are most likely to occur.
ii) Lognormal Distribution: In this type of distribution, Mean value and standard deviation of the values are entered. Values are positively skewed but not symmetric like a normal distribution.
iii) Uniform Distribution: In this type of distribution, all values have an equal chance of occurring, and minimum and maximum values are entered.
iv) Triangular Distribution: In this type of distribution, minimum, most likely, and maximum values are needed. Values around the most likely are more likely to occur.
v) Exponential Distribution: In this type of distribution, the time between independent events which happens at a constant average rate is modelled. Exponential mean value is entered.
vi) Discrete Distribution: In this type of distribution, specific values that may occur and the likelihood of each is required.

2.3.3 Monte Carlo Simulation

Monte Carlo Simulation (MCS) is named after the city of Monte Carlo in Monaco, which is famous for gambling. It simulates the actual system process and random behavior of the system. This method allows input parameters to occur in completely random manner over a certain range of values [19]. Thus for any n random numbers, each appears with equal probability i.e. independent and identically distributed. MCS uses a range of values which is a probability distribution for the analysis. It calculates results again and again up to thousands or tens of thousands times each time using different set of random values from the probability functions to complete the simulation. Variables can have different probabilities on the basis of different outcomes occurring.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2. 5 Steps for Monte Carlo Simulation

During Monte Carlo simulation, values are sampled at random from the input probability distributions. In each experiment of Monte Carlo Simulation possible values of input random variables are generated according to their distributions. Then the output value is calculated according to the samples of input random variables. After a number of experiments, a set of samples of output variable is obtained for the statistical analysis. There are three steps required for the Monte Carlo Simulation:

Step I: Sampling on random input variables X

Where, X = (X1, X2, X3,….,Xn)

Step II: Evaluating model output Y

Where, Y=g(X)

Step III: Statistical analysis on model output

Three steps of Monte Carlo simulation is shown in the Figure 2.5.

In this research Monte Carlo simulation was done to evaluate the reliability and availability of the different energy mix model. In [17], the authors analyzed their system by calculating the reliability indices of standalone hybrid micro grid of different configurations of renewable energy resources such as Photovoltaic (PV), Wind Turbine Generator (WTG), Micro Gas Turbine (MGT) and Diesel Generator (DG) using Monte Carlo Simulations. In [18], Reliability assessment method of wind integrated system connected to local loads had been introduced. Random outages were imposed to load chronological data to simulate grid failures using Monte Carlo Simulation methods and a comparison between different energy mix models was done on the basis of availability, average interruption duration, and average interruption frequency. In [20], the authors also calculated reliability indices of their systems containing different configurations of WTG, PV, and MTG to find the best reliable configuration. In [21], the reliability of supply for a hybrid PV-Wind system was calculated using Monte Carlo simulation. MCS approach had been used to include a number of random variables and their interactions. In [22], Monte Carlo method along with enumeration technique was used as an effective and powerful approach for reliability evaluation of large scale system. In [23], the authors calculated availability of each component as well as availability of overall system of various configurations containing transformers and transmission lines of various sizes using Monte Carlo simulations.

2.4 HOMER software for optimization

HOMER (Hybrid Optimization of Multiple Energy Resources) is developed by National Renewable Energy Laboratory (NREL) in United States [8]. It is generally used to optimize the hybrid system on the basis of techno as well as the economic basis. Its input consists of resources data, technical data, cost data and energy demand data and its output will be the optimized results on the basis of the cost. The technology information is also included in this software. It is based on bottom up approaches. Sensitivity analysis can also be done using this software. It also provides the data such as emission of different gases in the environment. Simulation, optimization and sensitive analysis are its three main tasks.

2.4.1 Cost analysis procedure by HOMER [13]

i) Net Present Cost (NPC): It is the installation cost and the operating cost of the system throughout its life time. It is given as:

Abbildung in dieser Leseprobe nicht enthalten

TAC= Total annualized cost ($)

CRF= Capital recovery factor

i= interest rate in percentage

Rprj=Project life time in year

ii) Total annualized cost: It is the sum of the annualized costs of every component including capital, O & M cost, replacement and fuel cost.

iii) Capital recovery factor: It is the ratio which is used to calculate the present value of a series of equal annual cash flows. It is given as:

CRF={i*(1+i)n} / {(1+i)n-1 } Where,

n= Number of years

i= annual real interest rate

iv) Annual real interest rate: It is the function of nominal interest rate. It is given as:

i=(i’ –F) / (1+F) Where,

i= real interest rate

i’= nominal interest rate

F= annual inflation rate

v) Cost of energy (COE): It is the average cost/kWh of useful energy produced by the system. It is given as:

Abbildung in dieser Leseprobe nicht enthalten

2.5 RAPTOR software for Reliability and Availability analysis

RAPTOR software is a modeling software program for Reliability, Availability and Maintability. RAPTOR works using a Monte Carlo model which allows randomness in equipment’s operation before it fails and also allows randomness in how long it takes to repair the failed equipment so that we can visualize what’s happening to our overall system and components in the system [9]. A graphical representation of system model can be designed using Reliability Block Diagram (RBD) in RAPTOR software. RBD consists of six different types of elements i.e. blocks, events, nodes, links, hierarchies and markers. A block usually represents the physical piece of equipment. Events represent actions that either do or do not occur. Nodes and links tie the blocks and events together such that the failure logic of a system is defined i.e. nodes and links define the series-parallel sub structures. Hierarchical elements represent major subsystems of the system being modelled. Markers define the beginning and end of an RBD structure and its substructures. RAPTOR software consists of four modes i.e. workspace view, simulation view, weak link analysis view and failure effects view. After RBD is modelled, MTBF and MTTR along with its life distribution are entered in each block. If we have no clue about the life distribution, exponential should be chosen as the default failure distribution because it is the most common distribution used in reliability community and the location should be left as zero. Similarly, for repair distribution lognormal should be chosen as the default repair distribution [9]. By the way, all units for MTBF and MTTR must be same, for e.g. let’s say hours. The required number of runs and time is entered for the simulation. After successful completion of the number of runs, the results are displayed as output. RAPTOR is useful for the engineers for converting technical details into money for life cycle cost considerations.

2.6 Comparison between On-grid and Off-grid Hybrid System

On-grid hybrid system is grid connected power system integrated with other embedded generation such as Solar PV, Diesel Generator, and Wind Turbine Generator etc. When energy resources such as Solar PV is integrated along with the grid, grid doesn’t have to serve the entire load because part of the load will be served by the solar PV too. If there is excess energy than required, it can be sent to electricity grid. We can have a net metering system set with our electricity supplier to receive the compensation for the excess electric energy supplied back to the grid. Efficiency is higher and there is almost unlimited storage in the grid. Expensive battery back-up is not required for the grid-tied system to store any excess energy. If it is required to store the energy to use during power outage (if any), battery backup may be installed along with the battery charger and solar charge controller. More maintenance may be required in this case. However, the on-grid system is the cheapest option of all because the COE is minimum in this case. The amount is paid only for the energy consumed. Hence, if grid is available, it is more economic in sense.

In the other hand, if access to grid is not available, no option is left than to go off grid. Off grid hybrid system is independent of utility grid. It is the best options in areas where grid is not available. In this case, there may be different configurations of Solar PV with battery, WTG etc. Diesel Generator as backup units are kept in case if the current hybrid system doesn’t meet the current load requirements when the production is minimal or absent. Standalone systems are more complex than on-grid system and is less flexible. Also, there is power loss in the form of charging and discharging of battery system. COE is also higher in this case than that of on grid system if diesel generator is operated because it consumes certain amount of fuel every hour which is costly. So, grid connected hybrid system will be the optimal choice from a monetary point of view rather than off-grid hybrid system.

2.7 Challenges and Issues of Interconnecting Grid and Embedded Generation

There are different challenges and issues of interconnecting grid and embedded generation. Some important challenges and issues are [24]:

(i) Synchronization of Diesel Generator with Grid
(ii) Operation and Control
(iii) Change of Short Circuit Capacity
(iv) Stability
(v) Harmonics
(vi) Power Quality
(vii) Unbalancing
(viii) Protection System Requirements
(ix) Frequency Regulation
(x) Uncertainty in Power Production

(i) Synchronization of Diesel Generator with Grid

In order to synchronize Diesel Generator and grid following four conditions should be met:

(a) Phase sequence
(b) Voltage Magnitude
(c) Frequency
(d) Phase Angle

(a) Phase Sequence

The phase sequence of three phases of the grid must be same with the phase sequence or phase rotation of the three phases of the generator.

(b) Voltage Magnitude

Magnitude of the sinusoidal voltage of the grid must be equal to the sinusoidal voltage produced by the generator. If the two voltages are not equal, the generator will either put out MVAR if the generator voltage is higher than the grid voltage (over excited) or absorb MVAR if the generator voltage is less than the grid voltage (under-excited).

(c) Frequency

The frequency of the sinusoidal voltage produced by the grid must be equal to the frequency of the sinusoidal voltage produced by the generator.

(d) Phase Angle

The phase angle between the voltage produced by the grid and the generator must be zero.

Need of Synchronization

(a) To make perfect connection of two AC power sources without any disturbance.

(b) Damage to the prime mover due to heavy acceleration can be avoided.

(c) Damage to the generator and its windings can be avoided.

(d) Proper synchronization can reduces the power and voltage variations.

(ii) Operation and Control

There is the risk of operation and control due to the penetration of embedded generation. As there is variations in solar power and wind energy, there may be adverse effect on voltage control functionality of the system due to increase in variations between the maximum and minimum voltage level, compared to the situation when embedded generation is not available.

(iii) Change of Short Circuit Capacity

The level of short circuit capacity (SCA) also increases with the increase in penetration of embedded generation. Although, it is desirable to have a high SCC capacity sometimes, increase in SCC may causes problems.

(iv) Stability

Stability issues increases with the increase in penetration level. Transient stability, dynamic stability, and voltage collapse needs to be considered.

(v) Harmonics

Level of Harmonic Voltage distortion is influenced due to the presence of electronic converter and rotating machine. Although the Power electronics interfaces offer advanced system possibilities, it injects harmonics currents into the system. Harmonics can also be injected by rotating generators due to the design of the winding and non-linearity of the core.

(vi) Power Quality

Different embedded generations have different characteristics and thus create different power quality issues. Power quality may be improved due to the increase in network fault level by different penetrations. But a single large diesel generator on a weak network may lead to the power quality issues particularly during starting and stopping. Also excessive use of power electronics devices and modern controls leads to the power quality problems.

(vii) Unbalancing

There may be unbalancing of loads in phases which should not go beyond the permissible limit.

(viii) Protection System Requirements

Depending upon the characteristics such as rated power, mode of operation, and technology used, the impact of hybrid resources on over current protection may vary. The protection system must be properly designed and must be properly coordinated.

(ix) Frequency Regulation

The system frequency must be maintained within strict limits to ensure smooth operation of the system.

(x) Uncertainty in Power Production

There is uncertainty in power production due to seasonal variation. There is variation in wind speed and solar radiation due to which power production is affected.

Chapter 3 RESEARCH METHODOLOGY

On the comprehensive level, this research was carried out in the following five methodological framework:

1. Problem formulation
2. Literature review
3. Input data gathering
4. Model building
5. Result analysis

The summary of activities carried out during this research is shown in block diagram in Figure 3.1.

Abbildung in dieser Leseprobe nicht enthalten

Figure 3. 1 Block diagram of summary of activities carried out

3.1. Problem formulation

The problem of Nepalese grid is that it is weak, unreliable and of poor quality which cannot meet the load demand of the country due to which there is frequent unplanned accidental power cut as well as planned load shedding time to time. In KU, there are some emergency loads which are very sensitive to interruptions such as loads during lab, presentation, seminars, classes and other official works. The only choice left during the interruption is to use the emergency and standby power systems. Reliability and Cost plays an important role to choose the various alternatives of standby and emergency power systems and their different configurations. The possible choices for KU is the various configurations of utility grid, solar PV with battery and different sizes of diesel generator. So it is essential to find the best cost effective and reliable configuration.

3.2. Literature Review

Detailed literature review was conducted with the help of research papers, national and international reports, books, libraries, journals and through internet. The literature related to HOMER and RAPTOR software was studied and working of these software’s were also studied.

3.3. Input data gathering

3.3.1 Electric Load data

The hourly load data of Kathmandu University was collected from Nepal Electricity Authority (NEA) for the year 2016/2017 AD. Then the load pattern and various resources available at Kathmandu University were analyzed. Majority of the load is served by the grid. In case of grid failure or during outage, there are three diesel generators of size 40 kVA, 125 kVA, and 200 kVA which serve the load. Similarly Solar PV of 65 kWp is also installed at the university premises. So the resources available at KU was taken in this research for the various configurations of energy mix model.

Various Cost Effective Energy mix model was identified using the HOMER (Hybrid Optimization Model for Multiple Energy Resources) software. HOMER examines all possible combinations of system types in a single run, and then sorts the systems according to the optimization variable of choice. The average daily load pattern of Kathmandu University for each month is shown in Figure 3.2.

Abbildung in dieser Leseprobe nicht enthalten

Figure 3. 2 Load pattern of Kathmandu University

3.3.2 Solar data

Nepal is located at favorable location for solar resource. The average global solar radiation in Nepal varies from 3.6-6.2 kWh/m2 day. The sun shines for about 300 days a year which is almost 2100 hours per year and average insolation intensity is about 4.7 kWhm-2 day-1 which is 16.92 MJ/m2 day [25]. The solar radiation of Kathmandu University was imported from the HOMER software which extracts data from the online data of NASA Surface Meteorology and Solar Energy website. From data, the annual average for daily solar radiation of the Kathmandu University was found to be 4.76 kWh/m2/day with average clearness index of 0.531. Global horizontal radiation and Extraterrestrial horizontal radiation daily profile is shown in Figure 3.3 and Figure 3.4 respectively.

Abbildung in dieser Leseprobe nicht enthalten

Figure 3. 3 Global horizontal radiation

Abbildung in dieser Leseprobe nicht enthalten

Figure 3. 4 Extraterrestrial horizontal radiation daily profile

3.3.3 Diesel generator set and others

Diesel generator set has become a popular solution as a backup/standby power sources. The generators can be used for small loads such as in homes as well as for larger loads like industries, plants, commercial buildings, hospitals etc. They are available in various specifications and sizes. In this research, three different sizes of diesel generator i.e. 40 kVA, 125 kVA and 200 kVA were considered for the analysis. The cost of diesel per liter was taken as 0.77 $/L [26].

A hybrid charge controller was used between AC and DC bus bars to exchange the power. The hybrid energy model had an option of multiple string having 6 batteries (Copy of Surrette 4KS25P) of nominal voltage 4V and nominal capacity of 1900 Ah, 7.6 kWh. Annual interest rate of project was assumed to be 6% on the basis of Nepal inflation rate in 2018 [27] and project period was assumed to be 25 years.

3.3.4 Cost input for the HOMER software

All the data for Grid, TOD meter Tariff Rate, Load data, Diesel Generators data, Solar PV data, Converter data, Battery data were entered into the HOMER software. The generator cost was taken from the website of NepKart, an online trading portal and marketplace for Nepal. O & M cost of each generator was calculated from the data which was obtained from the generator log book of generator house at Kathmandu University. Similarly the remaining component cost was retrieved from the website of Alibaba Global Trader and from the reference. There might be some cost variations (maximum 10%) due to transportation cost and discount in bulk orders. Table 3.1 shows the input Parameters into the HOMER software. Similarly Table 3.2 shows the TOD meter tariff rate.

Table 3. 1 Input values of Cost Estimation [28, 29, 30]

Abbildung in dieser Leseprobe nicht enthalten

Table 3. 2 TOD Meter Tariff [31]:

Abbildung in dieser Leseprobe nicht enthalten

3.3.5 MTBF and MTTR data

MTBF & MTTR for all components is listed in Table 3.3. MTBF and MTTR was calculated from the data collected from various sources. Grid data was collected from Nepal Electricity Authority (NEA). Diesel generator data was collected from the generator house of Kathmandu University. Similarly, for solar PV, battery and converter, the data was collected from the literatures. The life distribution was chosen as the exponential for MTBF and Lognormal for MTTR with default Standard Deviation (SD) of the RAPTOR software [9].

Table 3. 3 MTBF & MTTR of different Components [3, 32]

Abbildung in dieser Leseprobe nicht enthalten

3.4. Model Building

Three cases are considered for the model building in HOMER software. They are:

Case I: When grid is connected

Case II: When grid is not connected

Case III: Grid connected but capital cost excluded.

There were eight possible energy mix model for case I, six possible model for case II, and eight possible model for case III. All the energy mix model was optimized by the HOMER software and was categorized on the basis of costs. Similarly, the different energy mix model identified from the HOMER software was modelled again in RAPTOR software to check the reliability and availability of each model.

3.5. Result Analysis

After simulation results, the analysis of different energy mix models which was identified from the HOMER software was done on the basis of their TNPC and COE. Similarly, reliability and availability of each energy mix model was calculated using RAPTOR software for each cases. On the basis of the analysis, suitable conclusions was drawn with recommendations.

Chapter 4

RESULTS AND DISCUSSION

4.1 Cost Analysis of Energy Mix Models

Three different cases were considered for the analysis which can utilize the energy resources available at Kathmandu University and fulfill the demand of Kathmandu University. The cases are as follows:

(i) Case I: When grid is connected

When grid was connected and capital cost was included, eight different models were obtained which can utilize the maximum available resources available at Kathmandu University to fulfill the demand of KU. As shown in Figure 4.1, it contained AC bus voltage of 220 V which was fed from the grid and three diesel generators. Similarly, DC bus voltage of 24 V was fed from solar PV and battery bank. A converter was placed between AC and DC bus bar for power exchange. The load was connected to the AC bus. The annual peak demand was found to be 195 kW and energy demand was found to be 1069 kWh/d as calculated by HOMER software. The random variability of the load was set as the default value of the software: day to day 15 % and time step to time step 20%.

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. 1 Energy mix model case I

The eight possible hybrid energy mix models that could fulfill the load demand by utilizing the available resources available at KU are listed in Table 4.1. Similarly, Total capital cost, TNPC, Total Operation & Maintenance (O & M) Cost and Cost of Energy (COE) is shown in Table 4.2 for each model. As shown in Table 4.2, TNPC and COE is minimum for model 1. So the model 1 containing grid, solar with battery was the best optimized model for the case I.

Table 4. 1 Various energy mix model that can fulfill the demand, Case I

Abbildung in dieser Leseprobe nicht enthalten

Table 4. 2 Costs of the different models, Case I

Abbildung in dieser Leseprobe nicht enthalten

(a) Cost summary of model 1

The cost summary of each components for model 1 is shown in Table 4.3. TNPC for model 1 is $ 463,665 which is minimum of all and so is the best optimized model for case I. Cash flow summary is shown in graph in Figure 4.2.

Table 4. 3 Cost Summary for model 1, Case I

Abbildung in dieser Leseprobe nicht enthalten

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. 2 Cash flow summary for model 1, Case I

(b) Cash flow

Nominal Cash flows in each year is shown in Figure 4.3. Total Capital cost, operating cost, and replacement cost is shown in graph for this energy mix model.

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. 3 Nominal Cash flows in each year, model 1, Case I

(c) Electricity Production by Components

Similarly, average electricity contribution by PV array and grid for model 1 is shown in Table 4.4. PV array contributed 25% of the total energy i.e. 105,615 kWh/yr whereas 75% of required energy i.e. 316,059 kWh/yr is contributed by the grid.

Table 4. 4 Average electricity Contribution by components in model 1, Case I

Abbildung in dieser Leseprobe nicht enthalten

And monthly average electricity production by PV array and grid is shown in graph in Figure 4.4.

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. 4 Monthly average energy contribution by components, model 1, Case I

(d) Solar PV

Monthly average electricity production by solar PV is shown in the Figure 4.5. Table 4.5 shows the detailed output data of solar PV for model 1. Total hours of operation of solar was 4,380 hr/yr with levelized cost of 0.0752 $/kWh. Total energy production was 105,615 kWh/yr.

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. 5 Monthly average electricity production for model 1, Case I

Table 4. 5 Solar PV output data, Model 1, Case I

Abbildung in dieser Leseprobe nicht enthalten

(e) Battery data

The battery model was designed such that there were 6 batteries per string and there was only one such string designed for this case which was connected to 24 V dc bus. The nominal capacity of battery for this model was found to be 45.6 kWh. Average energy cost was found to be 0.101 $/kWh. Detailed battery data is shown in Table 4.6.

Table 4. 6 Battery details for model 1, Case I

Abbildung in dieser Leseprobe nicht enthalten

The battery bank state of charge (SOC) was 100% in each month. The battery bank state of charge is shown in Figure 4.6.

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. 6 Battery bank state of charge, model 1, Case I

(f) Converter data

The capacity of Inverter and Rectifier was 32 kW each. Total hours of operation was 4,383 hr/yr for inverter and 5 hr/yr for converter. Detailed converter data is shown in Table 4.7.

Table 4. 7 Converter data for model 1, Case I

Abbildung in dieser Leseprobe nicht enthalten

Inverter output and rectifier output is shown in Figure 4.7 and 4.8 for each month.

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. 7 Inverter output for model 1, Case I

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. 8 Rectifier output for model 1, Case I

(g) Grid purchases

The grid purchases, energy charge and demand charge for each month for the year 2016/2017 is shown in Table 4.8. Energy purchased was minimum in the month of December which was about 4,687 kWh and maximum in the month of August which was about 57,029 kWh.

[...]

• Université: Kathmandu University
• Cours: Master of Engineering in Electrical Power Engineering
• Note: 4.0
• Auteur: Sanjay Sah (Auteur)
• Année de publication: 2018
• Pages: 99
• N° de catalogue: V437914
• ISBN (ebook): 9783668780446
• ISBN (Livre): 9783668780453
• Langue: anglais
• mots-clé: Renewable Energy, HOMER, Monte Carlo Simulation, RAPTOR, Hybrid System