Resource Optimization

Case Study Analysis

EGMP49_RO_G19

Duke Energy Coal

Allocation

Case Study Submitted by:

GROUP NO: RO_G19
EGMP49063 – Taru Raaj Agarwal
EGMP49036 – Prasad Kamath Panambur
EGMP49064 – Tarun Singh Nahar
Table of Contents
Case Problem: ............................................................................................................................................................2
Duke Energy Coal Allocation: ................................................................................................................................2
Model Development: .................................................................................................................................................5
Define the Decision Variables: ..............................................................................................................................5
Define the Objective Function:..............................................................................................................................5
Define the Constraints: ..........................................................................................................................................6
Supply side Constraints: ....................................................................................................................................6
Demand Side constraints: ..................................................................................................................................6
Optimal Solution with Managerial report details: (All 6 questions):...................................................................7
The Sensitivity Analysis & Implications: ............................................................................................................ 10
Executive Summary: ........................................................................................................................................... 10

Case Problem:
Duke Energy Coal Allocation:
Duke Energy manufactures and distributes electricity to customers in the United States and Latin America. Duke
recently purchased Cinergy Corporation, which has generating facilities and energy customers in Indiana,
Kentucky, and Ohio. For these customers Cinergy has been spending $725 to $750 million each year for the fuel
needed to operate its coal-fired and gas-fired power plants; 92% to 95% of the fuel used is coal. In this region,
Duke Energy uses 10 coal-burning generating plants: five located inland and five located on the Ohio River. Some
plants have more than one generating unit. Duke Energy uses 28–29 million tons of coal per year at a cost of
approximately $2 million every day in this region.

The company purchases coal using fixed-tonnage or variable-tonnage contracts from mines in Indiana (49%),
West Virginia (20%), Ohio (12%), Kentucky (11%), Illinois(5%), and Pennsylvania (3%). The company must
purchase all of the coal contracted for on fixed-tonnage contracts, but on variable-tonnage contracts it can
purchase varying amounts up to the limit specified in the contract. The coal is shipped from the mines to Duke
Energy’s generating facilities in Ohio, Kentucky, and Indiana. The cost of coal varies from $19 to $35 per ton and
transportation/delivery charges range from $1.50 to $5.00 per ton.

A model is used to determine the megawatt-hours (mWh) of electricity that each generating unit is expected to
produce and to provide a measure of each generating unit’s efficiency, referred to as the heat rate. The heat
rate is the total BTUs required to produce 1 kilowatt-hour (kWh) of electrical power.

Coal Allocation Model

Duke Energy uses a linear programming model, called the coal allocation model, to allocate coal to its generating
facilities. The objective of the coal allocation model is to determine the lowest-cost method for purchasing and
distributing coal to the generating units. The supply/availability of the coal is determined by the contracts with
the various mines, and the demand for coal at the generating units is determined indirectly by the
megawatthours of electricity each unit must produce.
The cost to process coal, called the add-on cost, depends upon the characteristics of the coal (moisture content,
ash content, BTU content, sulfur content, and grindability) and the efficiency of the generating unit. The add-on
cost plus the transportation cost are added to the purchase cost of the coal to determine the total cost to
purchase and use the coal.

Current Problem

Duke Energy signed three fixed-tonnage contracts and four variable-tonnage contracts. The company would like
to determine the least-cost way to allocate the coal available through these contracts to five generating units.
The relevant data for the three fixed-tonnage contracts are as follows:

For example, the contract signed with RAG requires Duke Energy to purchase 350,000 tons of coal at a price of
$22 per ton; each pound of this particular coal provides 13,000 BTUs.

The data for the four variable-tonnage contracts follow:

For example, the contract with Consol, Inc., enables Duke Energy to purchase up to 200,000 tons of coal at a
cost of $32 per ton; each pound of this coal provides 12,250 BTUs.

The number of megawatt-hours of electricity that each generating unit must produce and the heat rate provided
are as follows:

For example, Miami Fort Unit 5 must produce 550,000 megawatt-hours of electricity, and 10,500 BTUs are
needed to produce each kilowatt-hour.The transportation cost and the add-on cost in dollars per ton are shown
here:
MANAGERIAL REPORT:

Prepare a report that summarizes your recommendations regarding Duke Energy’s coal allocation problem. Be
sure to include information and analysis for the following issues:

1. Determine how much coal to purchase from each of the mining companies and how it should be allocated to
the generating units. What is the cost to purchase, deliver, and process the coal?

2. Compute the average cost of coal in cents per million BTUs for each generating unit (a measure of the cost of
fuel for the generating units).

3. Compute the average number of BTUs per pound of coal received at each generating unit (a measure of the
energy efficiency of the coal received at each unit).

4. Suppose that Duke Energy can purchase an additional 80,000 tons of coal from American Coal Sales as an “all
or nothing deal” for $30 per ton. Should Duke Energy purchase the additional 80,000 tons of coal?

5. Suppose that Duke Energy learns that the energy content of the coal from Cyprus Amax is actually 13,000
BTUs per pound. Should Duke Energy revise its procurement plan?

6. Duke Energy has learned from its trading group that Duke Energy can sell 50,000 megawatt-hours of
electricity over the grid (to other electricity suppliers) at a price of $30 per megawatt-hour. Should Duke Energy
sell the electricity? If so, which generating units should produce the additional electricity?
Model Development:
Define the Decision Variables:
As defined in the Case Problem the primary objective is to determine the Lowest cost(minimize) to allocate the
coal available through these seven mining companies contracts to the five generating units.

A Liners Programming analysis can be prepared and decision variables would be defined with tons of Coal
purchased from all the suppliers (x) and used by the generating units (y).

There are total 7 suppliers and 5 generating units from the case Problem and therefore we will have total of 35
decision Variables:

X= RAG(1); Peabody Coal(2); American Coal(3); Consol Inc(4); Cyprus Amax (5); Addington Mining(6);
Waterloo(7);

Y = Miami Fort Unit 5 (1); Miami Fort Unit 7(2); Beckjord unit 1(3); East bend Unit 2(4); Zimmer unit 1(5);

Decision variables (Amount of coal shipped from X to Y)

Miami Fort Unit 5 Miami Fort Unit Beckjord East Bend Zimmer Unit 1
(Mi5) 7 (Mi7) Unit1 (Be1) Unit 2(Ea2) (Zi1)

RAG(Ra) Ra-Mi5 Ra-Mi7 Ra-Be1 Ra-Ea2 Ra-Zi1

Peabody(Pe) Pe-Mi5 Pe-Mi7 Pe-Be1 Pe-ea2 Pe-Zi1
American(Am) Am-Mi5 Am-Mi7 Am-Be1 Am-Ea2 Am-Zi1
Consol(Co) Co-Mi5 Co-Mi7 Co-Be1 Co-Ea2 Co-Zi1
Cyprus(Cy) Cy-Mi5 Cy-Mi7 Cy-Be1 Cy-Ea2 Cy-Zi1
Addington (Ad) Ad-Mi5 Ad-Mi7 Ad-Be1 Ad-Ea2 Ad-Zi1
Waterloo(Wa) Wa-Mi5 Wa-Mi7 Wa-Be1 Wa-Ea2 Wa-Zi1

Define the Objective Function:

Objective function as stated is to minimize total cost of purchasing, processing and distributing the coal.

To define coefficients of Objective function would be the total cost($) of Coal per ton and derived by adding all
three inputs = Cost of purchasing coal from supplier + Cost of shipping units from supplier to generating unit +
cost of burning the coal at generating unit.

Therefore objective function table would be as following:

Co-efficients in $ Miami Fort Miami Fort Beckjord Unit1 East Bend Unit Zimmer Unit 1
Unit 5 (Mi5) Unit 7 (Mi7) (Be1) 2(Ea2) (Zi1)

RAG(Ra) 37.00 37.00 36.75 32.00 32.75

Peabody(Pe) 39.75 39.75 40.50 35.75 36.50
 American(Am) 38.00 38.00 39.75 34.00 33.75
Consol(Co) 45.25 45.25 45.85 42.25 41.85
Cyprus(Cy) 50.00 50.00 49.75 45.00 45.75
Addington (Ad) 38.25 38.25 39.00 37.25 37.00
Waterloo(Wa) 46.00 46.00 45.60 42.00 43.60

The main objective function to minimize the cost of buying and distribution:

Minimize Z = 37XRa-Mi5 + 39.75XPe-Mi5 + 38XAm-Mi5 + 45.25XCo-Mi5 + 50XCy-Mi5 + 38.25XAd-Mi5 + 46XWa-Mi5

+ 37XRa-Mi7 + 39.75XPe-Mi7 + 38 XAm-Mi7 + 45.25XCo-Mi7 + 50XCy-Mi7+ 38.25XAd-Mi7+ 46XWa-Mi7

+ 36.75XRa-Be1 + 40.50XPe-Be1+39.75XAm-Be1 + 45.85X Co-Be1 + 49.75XCy-Be1 + 39XAd-Be1 + 45.6XWa-Be1

+ 32X Ra-Ea2+ 35.75XPe-Ea2 + 34XAm-Ea2 + 42.25XCo-Ea2 + 45XCy-Ea2 + 37.25XAd-Ea2 + 42XWa-Ea2

+ 32.75XRa-Zi1 + 36.5XPe-Zi1 + 33.75XAm-Zi1 + 41.85XCo-Zi1 + 45.75XCy-Zi1 + 37XAd-Zi1 + 43.6XWa-Zi1

Define the Constraints:

In this problem there are two main types of constraints 1) Supply side constraints 2) Demand side constraints

Supply side Constraints:

In the supply side constraints it limits the amount of coal can be brought by two different types of contracts one
is fixed tonnage contract and other is Variable tonnage. There are totally 7 suppliers and therefore we conclude
that there are 7 supply side constraints:

First of the 3 constraints are for the supplier with fixed tonnage contracts are:
Ra-Mi5 + Ra-Mi7 + Ra-Be1 + Ra-Ea2 + Ra-Zi1 = 350,000 (RAG)
Pe-Mi5 + Pe-Mi7 + Pe-Be1 + Pe-ea2 + Pe-Zi1 = 300,000 (Peabody)
Am-Mi5 + Am-Mi7 + Am-Be1 + Am-Ea2 + Am-Zi1 = 275,000 (American)

Next 4 constraints for the suppliers sides with variable tonnage contract, which implies maximum amount
purchased can be equal or less than amount specified in the contract are:
Co-Mi5 + Co-Mi7 + Co-Be1 + Co-Ea2 + Co-Zi1 <= 200,000 (Conosol)
Cy-Mi5 + Cy-Mi7 + Cy-Be1 + Cy-Ea2 + Cy-Zi10 <= 175,000 (Cyrus)
Ad-Mi5 + Ad-Mi7 + Ad-Be1 + Ad-Ea2 + Ad-Zi1 <= 200,000 (Addington)
Wa-Mi5 + Wa-Mi7 + Wa-Be1 + Wa-Ea2 + Wa-Zi1 <= 180,000 Waterloo

Demand Side constraints:

The demand side constraints specify the number of mWh of electricity that must be generated by each
generating unit.

Let Exy = mWh hours of electricity generated by a ton of coal purchased from supplier x and used by generating
unit y.

The quantity for purchasing coals is not given directly and needs to be calculated as follows:
All the units used for weights are pounds and thus we convert to BTU/ton by multiplying by 2000 to convert to
BTU/ton (1 US Ton = 2000 Pounds) to standardize the calculations.

Therefore for Supplier RAG, BTU/lb * 2000 = 13,000*2000 = 26,000,000 BTUs/Ton, similarly all Suppliers
BTUs/ton is calculated in Demand-Supply-Cost Tab, Supplier table.

Similarly Megawatt hours of electricity that each generating unit must produce and heat rate provided (in kWh)
needs to be standardized to mWh and required BTUs needs to be calculated.

Eg: For Miami fort unit 5, Hear rate 10,500 kWh is multiplied by 1000 to standardize to mWh and further
Required BTU is calculated by multiplying the Heat Rate in BTU/mWh * Electricity produced.

With these we get the desired Demand Side constraints prepared for Optimization.

Optimal Solution with Managerial report details: (All 6 questions):

1. Determine how much coal to purchase from each of the mining companies and how it should be
allocated to the generating units. What is the cost to purchase, deliver, and process the coal?

After the analysis and coal allocation Model Development for Duke Energy in the RO- Analysis tab in Excel,
Following table shows the quantities of coal to be purchased from each of the mining companies and
allocation of these quantities to various generating units:

Eg: Below table it indicates that from RAG, Bekjord Unit 1 should purchase and allocate 61538.46 coal to
process. Similarly below all the tables can be inferred in this way.

Ra-Mi5 Ra-Mi7 Ra-Be1 Ra-Ea2 Ra-Zi1

0 0 61538.46154 288461.5385 0

Pe-Mi5 Pe-Mi7 Pe-Be1 Pe-ea2 Pe-Zi1

217105.2632 11278.19549 71616.54135 0 0

Am-Mi5 Am-Mi7 Am-Be1 Am-Ea2 Am-Zi1

0 0 0 0 275000

Co-Mi5 Co-Mi7 Co-Be1 Co-Ea2 Co-Zi1

0 0 33877.55102 0 166122.449

Cy-Mi5 Cy-Mi7 Cy-Be1 Cy-Ea2 Cy-Zi1

0 0 0 0 0

Ad-Mi5 Ad-Mi7 Ad-Be1 Ad-Ea2 Ad-Zi1

 0 200000 0 0 0

Wa-Mi5 Wa-Mi7 Wa-Be1 Wa-Ea2 Wa-Zi1

0 0 98672.56637 0 0

The total cost to purchase, deliver, and process the coal is $ 53,407,249.33

2. Compute the average cost of coal in cents per million BTUs for each generating unit (a measure of the
cost of fuel for the generating units).

To Compute the Average cost of coal in cents per million, we first need to Calculate the total cost of purchase by
the generating units, this is calculated in the Case Analysis Tab. We calculate this by multiplying the amount of
coal shipped purchased and Total Cost of coal per Ton.

For eg: Total Cost of Coal purchased for Miami Fort Unit 5 is 8,629,934, which is only from Peabody mines as its
clearly observed from question1. Then we derive the cost per million BTU by dividing total cost by one million.

Total cost of coal Cost per million

Generating Unit Alias purchased BTU
Miami Fort Unit 5 Mi5 8,629,934 8.63
Miami Fort Unit 7 Mi7 8,098,308 8.10
Beckjord unit 1 Be1 11,214,763 11.21
East bend Unit 2 Ea2 9,230,769 9.23
Zimmer unit 1 Zi1 16233474.49 16.23

3. Compute the average number of BTUs per pound of coal received at each generating unit (a measure
of the energy efficiency of the coal received at each unit).

To Calculate the BTUs/pound of Coal received at generating unit, we need to get the total amount of coal
purchased at each generating unit(Decision variables). We already have the total BTUs calculated in required
BTS for each generating unit, which is now divided by the Total cost of coal purchased in Pound(by multiplying
by 2000).

Below are the Average number of BTUs per pound of coal at each generating unit.

Total Coal BTU/ Pound of

Genearting Unit Alias Purchased(Ton) Total BTUs coal

Miami Fort Unit 5 Mi5 217,105 5,775,000,000,000 13300.00

Miami Fort Unit 7 Mi7 211,278 5,100,000,000,000 12069.40

Beckjord unit 1 Be1 265,705 6,565,000,000,000 12353.92

East bend Unit 2 Ea2 288,462 7,500,000,000,000 13000.00

 Zimmer unit 1 Zi1 441,122 11,000,000,000,000 12468.19

4. Suppose that Duke Energy can purchase an additional 80,000 tons of coal from American Coal Sales as
an “all or nothing deal” for $30 per ton. Should Duke Energy purchase the additional 80,000 tons of
coal?

Considering that Duke Energy can again purchase additional coal from New American coal sales, we include that
as a new constraint 80000 <= 80000 in our solver solution in Q4 tab. After running the solver we can observe
that Object function of minimum cost has reduced after purchase new coal from American Coal sales.

53008710.77 < 53407249.33

Thus we recommend that, Duke Energy should purchase 80,000Tons from American Coal Sales to minimize
overall cost.

5. Suppose that Duke Energy learns that the energy content of the coal from Cyprus Amax is actually
13,000 BTUs per pound. Should Duke Energy revise its procurement plan?

The Case Analysis Linear programming Model is revised and re-solved after revising the model considering
Cyprus Amax is 13,000 BTUs per pound in the excel Tab Q5.

We observe that overall Cost is minimized with above condition, 53174799.5 than the original case cost which
was 53407249.33.

Which based on the current solution we can Recommend that Duke energy should revise the procurement plan.

The Solver also automatically, identified the revised plan of the procurement.

Instead of procuring from (Waterloo) Wa-Be1, quantity of 98672.56637

The procurement of coal will be done from Cyprus Amax Cy-Be , quantity of 85769.23077

6. Duke Energy has learned from its trading group that Duke Energy can sell 50,000 megawatt-hours of
electricity over the grid (to other electricity suppliers) at a price of $30 per megawatt-hour. Should Duke
Energy sell the electricity? If so, which generating units should produce the additional electricity?

TBD
The Sensitivity Analysis & Implications:

1. RAG contract can be increased to 4,35,769 tons of coal to decrease the total cost by $15.7 for every unit
increase.
2. Peabody contract can be increased to 3,83,835 tons of coal to decrease the total cost by $13.1 for every
unit increase.
3. American contract can be increased to 3,63,492 tons of coal to decrease the total cost by $12.98 for
every unit increase.
4. Consol Inc contract can be increased to 2,91,020 tons of coal to decrease the total cost by $3.58 for
every unit increase.
5. Addington Mining contract can be increased to 2,12,500 tons of coal to decrease the total cost by $9.49
for every unit increase.
6. Hence its best to increase the RAG contract and reduce procurement from Variable contact providers to
reduce the cost further.
7. Also, all the units have the capacity to produce more electricity.

Executive Summary:
Duke Energy purchases coal approximately 60% coal from the suppliers on Fixed contracts and 40% from the
suppliers on Variable contracts. There is different quality of coal provided by each supplier in terms of BTU
content, Sulphur content, etc. hence, the cost of producing electricity varies. The cost of transportation is varies
between suppliers and power plants. Considering all these constraints we have arrived on the optimal solution
of purchase of coal by each Power plant from different suppliers in order to minimize the total cost of
procurement of coal.

The solution and sensitivity analysis concludes that Duke Energy should increase their Fixed contracts with RAG,
Peabody and American if the requirement is stable so that cost can be further decreased and profits can be
increased because the variable contracts suppliers are charging very high prices.