NATURAL GAS ADAPTIVE SCHEDULING

1
NATURAL GAS ADAPTIVESCHEDULING
Peak Energy Solutions, Inc.
for APA Group Perth, WA
Peak Energy Solutions, Inc. Phone (719)
265-6621 Email stevehoyle_at_peakenergysolutions.co
m Support Email support_at_PeakEnergySolutions.com
2
Overview

• Introduction to Peak Energy Solutions, Inc.
• What is Natural Gas Adaptive Scheduling (NGAS)?
• NGAS Implementation for Gas Clients

3

• Highlights

Provides fast-running, customized, gas
scheduling software Years of experience in
natural gas scheduling and process
optimization Familiar with NAESB and a wide
variety of pipeline business rules
4

• Current NGAS Customers

El Paso (and Mojave) since 1994 Northwest
Pipelines since 1996 Kern River Gas Transmission
since 2000 Texas Gas (Boardwalk Pipelines) since
2004
5

• Principal Associates

Steve Hoyle, Ph.D. Stanford University, Mathematic
al Programming First use of AMPL/CPLEX in
natural gas scheduling in 1994 Alan Magner,
Bachelor of Business Studies in Economics
Accounting, University of Limerick 20 years in
Lean Six Sigma process improvement Deb Parker,
M.S. Stanford University, Operations Research
Experience in natural gas scheduling since
1997 Steve Dziuban, Ph.D. Rensselaer Polytechnic
Institute, Operations Research and Statistics
Experience in natural gas scheduling since 1997
Linda Tully, M.S. University of
Colorado, Computer Science Experience in
natural gas scheduling since 2004
6
Natural Gas Distribution Environments

• Physical View
• Capacity limits at receipt, throughput, and
  delivery locations are enforced.

• Ownership View
• Gas ownership transitions from upstream supplier,
  through pipeline (including pools), and to the
  downstream market.
• Total gas flow is determined by supply/ demand in
  accordance with tariffs, laws, regulations, and
  agreements.

Rec
Suppliers
Throughput
Pool
Pool
Del
Del
Markets
7
What is NGAS?

• Natural Gas Adaptive Scheduling
• A scheduling engine that interfaces with your
  system to provide optimal scheduling solutions
• Uses state-of-the-art optimization methods
• Implemented with powerful AMPL/CPLEX linear
  programming software -- essential for large,
  complicated pipelines
• Runs on PC or UNIX platforms
• Most effective when used to perform all
  scheduling adjustments (cuts)

8
Advantages of Using NGAS

• Outperforms industry running times typical run
  times of 2 to 5 minutes
• Cost effective -- full-time AMPL/CPLEX developers
  not needed
• Higher pipeline throughput via unique
  reinstatement optimization step
• Displacement handled automatically
• Most business rules are data-driven (such as
  Priority of Service for capacity cuts and number
  and length of cycles)
• Gas clients own customized code

9

• Traditional Scheduling
• Double cuts decrease throughput
• Always occur when sequential cuts are made
• Result from physical and ownership cuts
• Ineffective use of displacement gas decreases
  throughput

NGAS Increases Throughput
Automatically uses displacement gas Detects
slack in schedule created by earlier cuts Uses
this slack to reinstate gas on double-cut
noms/contracts without violating any constraints
10
NGAS Avoids Double Cuts
A (rank 2)
C (IT)
nom 150
nom 120
RULES
Pool
- Pool Balance Flow in equals flow out -
Operator Physical or contractual - Capacity
Physical limitation
nom 90
nom 80
B (rank 1)
D (firm)
SEQUENTIAL APPROACH A B C D 1. Pool
Balance Cut A to 110 110 90 120 80 2. Operato
r D is above limit by 10 so cut 100 90 120 70
it to 70 then also cut A by 10 more 3. Capacity
B is over by 20 so cut to 70
and, 100 70 100 70 since D is firm, cut C by
20 ADDITIONAL NGAS STEP 4. By considering all
constraints simultaneously, 120
70 120 70 NGAS recognizes slack created by
double cuts and reinstates 20 to noms A and C
Traditional Results
170 Throughput
NGAS Results
190 Throughput
11
APA WA and NGAS working together
Does APA in WA have optimized shipping in its
network?
How would improved transmission throughput
benefit APA in WA?
As client volumes and infrastructure are
changing, is the APA transmission optimized?
Can gas shipments be scheduled between
disconnected pipelines?
12
APA Group and NGAS working together
Are the requirements for APA on the east coast
different from WA?
Fast data I/O
13
Gas Client System and NGAS work together
Database/GUI Piece
NGAS Piece
Fast data I/O
Fast data I/O
Fast running algorithms
Well suited to accounting
Well suited to accounting
Optimal, accurate scheduling
Friendly customer interface
Friendly customer interface
Maximize throughput
14

• Implementation Highlights

Several initial pool balancing and rebalancing
methods available Modularity facilitates
maintenance, such as changing the cut
order Capacity techniques make verification and
explanation of cuts much easier Previously
scheduled amounts protected at contract level or
at nomination level, or can be cut Implements
complex, customized segmentation rules
15
NGAS and Gas ClientSystem Integration

• Gas Client System
• Gathers and stores information (contract data,
  nominations, etc.)
• Provides input data to NGAS using flat files or
  ODBC
• NGAS sequentially determines and labels cuts with
  accurate reason codes
• Cuts are passed to database via NGAS flat files
  or ODBC
• Reinstatement cuts are labeled as such
• NGAS provides outputs (by nomination, contract,
  point, etc.)
• Gas Client System uploads output files
• Updates database tables
• Prepares outputs and reports for display to
  end-users

16
Notional Implementation Sequence for Gas Client

• Develop software requirements
• Describe database interface
• Build NGAS model
• Run test sets for verification of results
• Train Gas Client personnel as needed
• 24/7 support agreement available

17
Requirements Considerations
Contract or market carriage (combination?) Priorit
y of Service (firm forward haulage vs.
Interruptible haulage pro-rata rules) System Use
Gas by contract or nomination Pool structure and
balancing rules Tiebreaking Rules (Ranks, Dates,
etc.) Force Majeure and Assignment rules Park
Services (physical, contractual,
seasonal) Physical pipeline model (grid vs. soda
straw)
18
Requirements Considerations
Capacity cuts (Physical limits at receipt,
delivery, throughput, net, bi-directional hourly
vs. daily) Non-Capacity Cuts (Operator balancing,
contract limits, Park/Lend) Imbalances Data
Input/Output Reports (cuts, contract amounts,
point flows) Run times needed Anticipated
development timeline (SRS, data, test, implement,
support)