Sri Lanka Renewable Power Project

1
Sri Lanka Renewable Power Project

• Phase I
• Justin Coe
• Erik Hofland
• Zachary Minteer
• Marco Olson
• Delaney Packard

2
Phase I Description

• Phase I of this Capstone project is a general
  method of renewable power system design and
  optimization that can be applied to any specific
  set of resource and load distributions.
• This method consists of two parts
• Expected Monetary Value (EMV) optimization
• System Resource Distribution (SRD) analysis

3
EMV - Theory

• Description
• Determines optimal sizes of power sources based
  on load and resource profile averages
• Program minimizes the EMV by choosing the
  configuration which uses the least amount of
  diesel

4
EMV Decision Tree
5
EMV Configuration Output

• For each generation configuration (Q), the
  average output is calculated by the sum of the
  products of the probability arrays (Pi) and the
  resource availability arrays (Xi).
• Value ?(PiXiQ), for all i.
• Value is calculated for wind, solar and
  batteries.
• EMV (Load - ?Values)Cost of Diesel

6
EMV Resource

• Example resource profile array
• Pwind .3, .2, .2, .1, .2
• Xwind 1, .75, .5, .25, 0
• This profile indicates for 30 of the time, the
  turbine operates at full power, 20 of the time
  at 75 power, 20 of the time at 50 power, 10
  of the time at 25 of the power, and 20 of the
  time at 0 power.
• The wind and solar resource profile arrays are
  compiled from specific site data.

7
EMV - Battery Probability

• The probability of the batteries being charged
  (Pb) is approximated by considering the size of
  the wind and solar installations, the size of the
  load, and the size of the batteries as shown
  below
• The probability is between 0 and 1.

8
EMV - Calculation

• The EMV of each configuration calculates the cost
  of diesel required to power the amount of load
  not supplied by available renewable sources and
  batteries.
• The EMV is calculated for every possible source
  configuration the optimal configuration has an
  EMV which is closest to 0.

9
EMV - Interpretation

• The smallest EMV represents the least expensive
  operating solution. The long term cost of the
  power system can be determined by multiplying the
  EMV (given in cost per hour of operation) by the
  number of hours in a given period. Total system
  cost is this operating cost plus the installation
  costs. The best power system configuration is the
  one which minimizes the total system cost.

10
EMV Approximations and Assumptions

• Maintenance costs are ignored.
• Load is an average for a given hour.
• Resource profiles are average values.
• Battery probability is approximated.

11
EMV - Example

• Load72.5 kW
• Pwind0.25,0.25,0.5
• Xwind1,.50, 0
• Qwind 20 30 40 50 60 70 80 90 100
• Psolar0.4,0.1,0.5
• Xsolar1,0.5,0
• Qsolar 20 30 40 50 60 70 80 90 100
• Qbattery 3 6 9 12 15 18 21 24 27

12
EMV - Example
13
SRD

• Hourly analysis of system performance
• Load and resource profiles can be arbitrarily
  defined to represent different scenarios
• Useful to account for location specific anomalies
• Verification of EMV solution
• Based on Linear Programming Upper Bound (LPUB)
  optimization method

14
SRD - Optimization

• Minimize
• Ctotal CwXw CsXs Cb Xb CdXd
• Subject to
• Xw Xs Xb Xd gt Demand
• Xi lt Xi,max

15
SRD - Example
16
SRD - Example
17
SRD No Wind Example
18
SRD No Wind Example, Day 2
19
SRD No Wind Example

• Day 1 Diesel Usage 759 kWh each day
• Day 2 Diesel Usage 1150 kWh
• Total Diesel Used 1909 kWh
• About 150 gallons of Diesel
• Amount of diesel storage is site specific

20
Phase I Conclusion

• EMV analysis provides a starting point for
  determining optimal power system sizing
• SRD analysis can provide insight into how well
  the system performs in specific situations
• An iterative approach to the SRD analysis should
  be used for final sizing decisions