S DRAFT Primer for Gas-Electric Modeling in MISO`s Phase III Clean

S
DRAFT Primer for Gas-Electric
Modeling in MISO’s Phase III
Clean Power Plan Study
June 2015
DRAFT - 1 | P a g e
Contents
1
Introduction.................................................................................................................................... 5
2
Representation of Natural Gas Infrastructure ............................................................................... 7
2.1
Gas Supply (Producing Fields) ....................................................................................................... 8
2.2
Gas Pipelines ............................................................................................................................... 10
2.3
Gas Nodes ................................................................................................................................... 13
2.4
Gas Demand (Non-Power) .......................................................................................................... 13
2.5
Gas Demand (Electric Power) ..................................................................................................... 14
2.6
Gas Storage ................................................................................................................................. 14
2.7
Future Developments ................................................................................................................. 15
3
Co-optimization ........................................................................................................................... 18
4
Outputs ........................................................................................................................................ 19
5
Summary & Next Steps ............................................................................................................... 20
Appendix A: Common PLEXOS Electric Generator Properties (Units) and Descriptions .................. 21
Appendix B: PLEXOS Natural Gas Pipelines ..................................................................................... 22
Appendix C: Pipeline Capacity and Expansion Properties ................................................................. 24
Appendix D: PLEXOS Natural Gas Storage Fields ............................................................................ 25
Appendix E: PLEXOS Database Structure ......................................................................................... 28
DRAFT - 2 | P a g e
List of Figures
Figure 1. PLEXOS Gas Model Development Phases and Approximate Timeline ........................................ 7
Figure 2. Gas Modeling Objects: Symbols, Classes, and Descriptions ........................................................ 7
Figure 3. Generic Production Cost Bands in the Phase II PLEXOS Gas-Electric Model ............................. 9
Figure 4. Energy Transfer gas network including Panhandle Eastern Pipeline (PEPL) System (left) and
natural gas producing basins and plays (right) ............................................................................................. 9
Figure 5. Receipt Points for Panhandle Eastern Pipeline Market Zone ..................................................... 10
Figure 6. US Natural Gas Pipeline Network…………………………………….……………………………….11
Figure 7. Partial representation of Panhandle Eastern Pipeline system definition in PLEXOS .................. 11
Figure 8. Panhandle Eastern Pipeline System (Source: Energy Transfer, 2015) ...................................... 12
DRAFT - 3 | P a g e
Note: This document is not intended to be a user manual for the PLEXOS Integrated Energy Model (“gaselectric model”); it is rather intended as a primer for MISO stakeholders who are interested in learning
about the Model’s application within the context of MISO’s study of the Clean Power Plan. It is in draft
form and will be updated as the model evolves and pending review by Energy Exemplar, the vendor of
PLEXOS.
DRAFT - 4 | P a g e
1 Introduction
In 2014, MISO carried out preliminary investigations into the potential reliability and economic impacts
of the draft Clean Power Plan (CPP) on the generation fleet. These studies (Phase I and II1) provided
insight into generator dispatch and capacity expansion trends for a range of strategies to comply with
the CPP, including the EPA’s Building Blocks2. Phase I and II analyses were performed with a resource
forecasting software called Electric Generation Expansion Analysis System (EGEAS)3. The use of this tool,
which does not model electric transmission, allowed for the simulation of more than 1,000 sensitivities
around policy, economic and technological conditions, over a 20-year planning horizon, in the context of
CPP compliance.
The Phase III study builds upon its predecessors, with a scope informed by results from the Phase I and II
analyses. It also incorporates stakeholder input, namely by addressing requests for consideration of
electric transmission and natural gas infrastructure, as well as state-level CO2 constraint modeling. Phase
III study needs drove the selection of a different analytical tool from Phase I and II—specifically, one that
allows for 1) explicit representation of rate-based CO2 targets as a constraint on a user-defined
collection of generators, e.g. applicable generators in a given state and 2) simultaneous modeling of the
gas and electric systems.
This tool, the PLEXOS Integrated Energy Model, is an Energy Exemplar optimization platform for energy
market simulation and analysis4. MISO has used the production cost functionality of PLEXOS for several
major studies, including the Manitoba Hydro-Wind Synergy Study5 and the Minnesota Renewable
Energy Integration Study (MRITS)6. The PLEXOS gas model is a relatively new addition to the Integrated
Energy Model, with increasingly granular representations of the gas system under development by the
vendor. Gas and electric infrastructure are interconnected in the Integrated Model via gas-fired electric
generation and co-optimized dispatch or capacity expansion can be simulated. This functionality allows
for examination of the interdependencies between the gas and electric infrastructure.
Large-scale co-optimized simulation of the gas and electric systems is relatively new not only to MISO,
but also to the power industry overall. MISO fully acknowledges the learning curve associated with this
endeavor and plans to collaborate with and leverage the expertise of its stakeholders and the broader
industry throughout the process. MISO’s CPP gas-electric modeling aims to approximate real world
operation of the gas and electric systems. The end goal of this exercise is to inform stakeholders on both
electric and gas system impacts of the CPP, including potential infrastructure expansion needs under
different approaches to compliance, and the associated cost and timelines for implementing this
infrastructure.
This document provides an overview of the gas portion of the PLEXOS Integrated Energy Model, in the
context of MISO’s CPP study.
1
See https://www.misoenergy.org/WhatWeDo/EPARegulations/Pages/EPAStudies.aspx
See http://www2.epa.gov/carbon-pollution-standards/fact-sheet-clean-power-plan-framework.
3
See http://www.epri.com/abstracts/Pages/ProductAbstract.aspx?ProductId=000000000001016192.
4
See http://energyexemplar.com/software/plexos-desktop-edition/.
5
See:
https://www.misoenergy.org/Library/Repository/Meeting%20Material/Stakeholder/Planning%20Materials/
Manitoba%20Hydro%20Wind%20Synergy%20TRG/Manitoba%20Hydro%20Wind%20Synergy%20Study
%20Final%20Report.pdf.
6
See https://mn.gov/commerce/energy/images/final-mrits-report-2014.pdf.
2
DRAFT - 5 | P a g e
-
Section 2: Representation of Natural Gas Infrastructure overviews the representation of natural
gas production, demand and storage in the Integrated Energy Model, along with data sources
and underlying modeling assumptions for the CPP Phase III study
-
Section 3: Co-optimization provides a high-level description of co-optimization in the Integrated
Energy Model, including the objective function and its constraints
-
Section 4: Outputs briefly discusses key metrics produced by gas-electric co-optimization
-
Section 5: Summary & Next Steps outlines future modeling efforts and opportunities for
stakeholder engagement in model developments and results review for the CPP Phase III study
DRAFT - 6 | P a g e
2 Representation of Natural Gas Infrastructure
The development of the gas portion of the PLEXOS Integrated Energy Model is on-going. The initial
release of the gas model included state-level representation of gas supply, demand and transportation.
In other words, pipelines were represented with state-level segments per pipeline, all connected to a
single state gas node. Additionally, the aggregate gas demand for residential, commercial and industrial
sectors per state was tied to the same single gas node.
The second iteration disaggregated these elements into separate components, interconnected via
hundreds of gas nodes. Future versions of the model will incorporate additional granularity, for
example, representation of gas contracts. Figure 1 provides an approximate timeline for historical and
future gas model development.
Preliminary
(State) Model
Aggregated (state-level)
representation of gas
[add ininfrastructure
graphic here]
Current (Nodal) Model
Disaggregated gas infrastructure;
gas-fired electric generators
mapped to individual pipelines
Future (Enhanced)
Model
Pipeline tariffs incorporated;
additional model granularity
2014 ----------------------------- 2015 ----------------------------- 2016 -------------------------------------------------------Figure 1. PLEXOS Gas Model Development Phases and Approximate Timeline
-----The following elements of natural gas infrastructure are currently represented in the PLEXOS gas model:
-
Gas producing fields
Gas pipelines
Gas demand
Gas nodes
Gas storage
These elements are briefly described in Figure 2 and further defined in subsequent sections.
Symbol
Class
Gas Field
Description
Field from which gas is extracted
Gas Storage
Storage where gas can be injected and extracted
Gas Pipeline
Pipeline for transporting gas
Gas Node
Connection point to gas network
Gas Demand
Demand for gas covering one or more nodes
Gas Zone
A collection of Gas Nodes
Figure 2. Gas Modeling Objects: Symbols, Classes, and Descriptions
7
7
Image sourced from Energy Exemplar, 2015
DRAFT - 7 | P a g e
Much of the data for the gas model’s underlying database is sourced from the Eastern Interconnection
Planning Collaborative (EIPC) Gas-Electric System Interface Study8. Some of this data is publically
available and published in the EIPC Study report. A portion of the data gathered for the EIPC study is
from Federal Energy Regulatory Commission (FERC) Critical Energy Infrastructure CEII) forms, and can be
obtained by qualifying entities via request to FERC. The remainder of the data used in the EIPC Study is
proprietary.
The PLEXOS gas network is represented in extensive detail, with thousands of gas infrastructure
elements. Proprietary forecasts are used to develop gas production volumes and costs, as well as
profiles for non-power gas demand. The following sections provide details around each major gas
system component in the model, both as modeled by the vendor and as modified by MISO, if applicable,
for purposes of the CPP study.
2.1 Gas Supply (Producing Fields)
Natural gas supply in the current version of the model is represented with ~1,000 individual gas injection
points across the pipeline network. These injection points are a proxy for physical production fields. For
each “gas field” (i.e. gas supply injection point), the user can add definition around production,
including:
-
Initial volume (MMcf): the volume of gas in the field at the start of the simulation9
Production cost ($/MMBtu): the cost to produce gas at the field
o Can be defined as a curve or in bands, e.g. increasing costs as field is depleted
Production volume (MMcf): the volume of production at the field
o Can be defined as production bands, e.g. decreasing production over time
Min/max production (MMcf): the min or max level of production per field, defined for userselected time interval, e.g. max daily or hourly production
Initial volume for gas production at a given gas field is set by the vendor10 to a generic value (1x10^16
MMcf). The underlying assumption is that there is sufficient gas supply to serve gas demand. The
implication of this assumption is that gas supply will not be a limiting factor on dispatch of gas-fired
generators in the model. For Clean Power Plan (CPP) modeling, initial volume will remain at the generic
value, with an underlying assumption that the limiting factor on gas supply from a given field or injection
point is the maximum daily rate of production (MMcf/d).
The min production property is set to zero by the vendor and this value will be retained for MISO’s use
of the model in the CPP study. MISO will model max (daily) production capacity for each field based on
proprietary IHS CERA forecasts for productive capacity11 (forecasts are daily values, thus metric in model
is daily). Production volume and production cost are modeled by the vendor with ten generic bands ().
The production bands will be replaced by proprietary forecasts from IHS CERA12 for the CPP study.
8
See http://www.eipconline.com/Gas-Electric_Documents.html.
Property definitions are taken from the PLEXOS user manual.
10
Energy Exemplar is the vendor for PLEXOS.
11
Productive capacity is defined by IHS CERA as an estimate of the pipeline-grade dry natural gas that a
given play, basin or region can produce in a given year and that can be carried to market on the
infrastructure assumed to exist at that time.
12
Breakeven prices (USD per Mcf) are forecasted by IHS CERA for 2015 for eight US regions, as well as
portions of Canada. Per IHS CERA’s definition, full-cycle unit break-even prices are not normalized to
Henry Hub but reflect economics for a play at the point of entry into the pipeline grid. Break-even prices
(without and with natural gas liquids [NGLs] credits) are calculated at the play level for the “typical” well
and include leasehold, finding and development (F&D), operating expenses (opex), royalty, taxes, and
9
DRAFT - 8 | P a g e
Figure 3. Generic Production Cost Bands in the Phase II PLEXOS Gas-Electric Model
For the Phase III study, gas field objects will serve as proxies for real world pipeline pooling points, or
aggregations of multiple gas injection points for a given pipeline or pipeline segment. These pooling
points serve as liquid trading hubs for gas, and may be fed by gas injection from a variety of
interconnected gas fields or pipelines. For example, gas supply on the Panhandle Eastern (PEPL) system
(Figure 4) is represented in the model with gas field objects in the states of Kansas, Missouri, Illinois
Michigan and Ohio.
Figure 4. Energy Transfer gas network including Panhandle Eastern Pipeline (PEPL) System (left)
15
natural gas producing basins and plays (right)
14
and
return. Capital costs are success-weighted and based on equipment needed for the “typical” well.
Weighted-average cost of capital (WACC) is assumed to be 10%. Taxes are based on tax benefits
available to all producers. Well useful life is assumed to be 20 years.
14
Image sourced from Energy Transfer, 2015; see
http://tgcmessenger.energytransfer.com/InfoPost/resources/documents/2014PEMarketerMeeting.pdf.
15
“Lower 48 states shale plays” image sourced from the Energy Information Association (EIA), 2015; see
http://www.eia.gov/oil_gas/rpd/shale_gas.jpg.
DRAFT - 9 | P a g e
The right image in Figure 4 shows the natural gas production fields from which PEPL shippers16 could
theoretically access gas supply directly. Figure 5 shows actual receipt points for the PEPL Market Zone
(yellow triangles). The complexity of real world gas supply is challenging to capture in a model,
especially one that spans the Eastern Interconnection. While the proxy pooling points in the model may
not line up one-for-one with physical hubs they offer a reasonable approximation of aggregated gas
supply.
Figure 5. Receipt Points for Panhandle Eastern Pipeline Market Zone
17
2.2 Gas Pipelines
The North American natural gas pipeline network is represented in the model by hundreds of separate
gas network elements, generally based on gas pipeline tariff definitions (e.g. a given pipeline is divided
into Producing Zones A and B, and Market Zones C, D and E)18. Pipeline segments are interconnected
with one another and with corresponding gas producing fields/gas supply points via gas nodes. Likewise,
segments of two different pipelines, physically interconnected in the real world, are joined via nodes in
the model.
For reference, Figure 6 shows the natural gas pipeline network (major interstate and intrastate
pipelines) in the US.
16
Gas pipeline tariff language generally refers to customers of pipeline capacity as “shippers”.
Image sourced from Energy Transfer, 2015; see
http://peplmessenger.energytransfer.com/ipost/PEPL/maps/market-zone.
18
See Appendix B for a full listing of pipelines defined by Energy Exemplar in the PLEXOS gas model.
17
DRAFT - 10 | P a g e
Figure 6. US Natural Gas Pipeline Network
An example of individual pipeline representation in PLEXOS is illustrated in Figure 7, which shows a
partial representation of the Panhandle Eastern Pipeline (PEPL).
Interconnect w/
MichCon (MI)
Interconnect
w/NGPL (IL)
Interconnect
w/ANR (OH)
PEPL Market 2
Receipt
PEPL
Field
PEPL
Market 1
PEPL
Market 2
PEPL
Market 3
PEPL Market 3
Receipt
2nd Interconnect
w/ANR (OH)
Figure 7. Partial representation of Panhandle Eastern Pipeline system definition in PLEXOS
For reference, the following pipeline interconnections and segments are modeled in PLEXOS for the
PEPL system:
DRAFT - 11 | P a g e
-
PEPL individual interconnections (and associated states) with ANR (OH), Col Gas (OH), E Ohio
Gas (OH), Enogex (OK), KMI EN (KS), MichCon (MI), NGPL (IL), Nisource (IN), NNG (KS), ONG (OK),
Sstar (KS), TW (TX), Union Gas (MI) and Trunkline (IL)
-
PEPL individual pipeline segments include:
o Field to Market 1
o Market 1 to Market 2
o Market 2 to Market 3
o Market 2 Receipt to Market 2
o Market 3 Receipt to Market 3
o Market 3 to Export
o Import to Market 3
PEPL’s actual system map is provided in Figure 6.
Figure 8. Panhandle Eastern Pipeline System (Source: Energy Transfer, 2015)
Pipeline production properties defined in PLEXOS include:
-
Flow charge ($/MMBtu): the incremental cost of extracting gas from the pipeline, charged
against the flow in each user-defined interval (e.g. hour or day)
o The total cost of this charge is reported as production cost for a given pipeline.
DRAFT - 12 | P a g e
o
-
-
Flow charge is not defined by the vendor in the current iteration of the model. For
further discussion of the impacts of this modeling assumption, as well as future plans for
consideration of the flow charge variable, see Section 5.7.
Initial volume (MMcf): the volume of gas in the pipeline at the start of the simulation
Min/max volume (MMcf): lower and upper limits on the volume of gas that can be stored in the
pipeline
o This metric can be used to represent pipeline linepack.
Max flow (MMcf): the maximum quantity of gas that can be extracted from the pipeline (as a
rate)
o This includes gas flowing through the pipeline as well as gas taken from storage in the
pipeline, i.e. it puts an upper limit on the flow of gas.
o The max flow property could be used to model de-rates for maintenance or outages, for
example.
o As like other time interval properties, the user can customize the interval, e.g. hour, day,
week.
The following production properties, though not defined in the current version of the model, can also be
specified by the user:
-
-
-
Flow charge back ($/MMBtu): the incremental cost of extracting gas at the pipeline sending
node, charged against the flow backward in each user-defined interval
o The total cost of this charge is reported as production cost for a given pipeline.
o This charge can only be applied to bi-directional pipelines.
Volume imbalance (MMcf): the absolute value of the difference between delivery volume into
the pipeline and the redelivered volume off the pipeline (i.e. Volume Imbalance = end volume –
initial volume)
Imbalance charge ($/MMBtu): the charge applied to the volume imbalance
In addition to pipeline production properties, capacity and expansion properties19 can be defined by the
user. The vendor’s definition of the natural gas pipeline system will be used for MISO’s CPP study. As
mentioned previously in this document, the vendor definition of the natural gas pipeline system is based
on that used for the EIPC gas-electric study. For specifics on the definition of pipeline characteristics in
the EIPC study, see Exhibit 7 of Target 2 of the study report20.
2.3 Gas Nodes
Gas nodes serve as the interconnectors of gas infrastructure elements in the model, as well as the tie
point between gas-fired electric generators and gas pipelines. Nodes can be defined by assigning them
properties, such as a constraint on the max flow through the node.
Currently, gas node property definitions from the vendor do not include max flow day; MISO will use this
definition, allowing interconnects to set flow constraints rather than nodes in the model.
2.4 Gas Demand (Non-Power)
Residential, commercial, industrial and transportation gas demand is currently modeled in PLEXOS with
individual state-level demand profiles (RCIT aggregate demand profile for each state). Each profile
19
For a list of gas pipeline capacity properties, as defined by Energy Exemplar, see Appendix D.
Expansion properties will not be addressed in this document, given its focus on the production cost
functionality of PLEXOS.
20
See http://www.eipconline.com/uploads/Exhibit_7_GPCM_13Q4base_Database.pdf
DRAFT - 13 | P a g e
contains a value for each hour of the year, based on historical demand. State-level residential,
commercial and industrial (RCI) gas demand will be distributed amongst gas nodes within the state, per
LDC gas demand forecasts, as gathered for the EIPC gas-electric study (see Exhibit 15 of the Target 2
study report21 for LDC forecast source documentation). The LDC forecasts will determine the nodal
distribution factors in the CPP study model; the shape of the hourly profile of the demand from the
vendor will be normalized and retained; the shape will then be applied to proprietary demand forecasts
from IHS CERA22 to produce an 8760-hourly profile.
Canadian and Mexican gas imports/exports will not be modeled in the CPP study, with the exception of
liquefied natural gas (LNG) exports. LNG exports will be represented with state-level profiles, again
based on IHS CERA proprietary forecasts.
The demand shortage price is assumed by the vendor to be $100/MMBtu; this value will be retained for
the CPP study.
2.5 Gas Demand (Electric Power)
Gas-fired electric generators are interconnected with gas infrastructure in the model via gas nodes; they
are likewise interconnected with electric infrastructure via an electric bus. In the current iteration of the
model, generators with multiple real-world interconnections to natural gas infrastructure are
represented with a single pipeline interconnect (arbitrarily selected amongst the interconnections) in
the model. The mapping of gas-fired generators to gas pipelines in the model is based on data collected
via the EIPC study24.
Electric power demand for natural gas is dynamic in the PLEXOS model, determined via co-optimization
(see Section 3 for additional details). Common generator properties are listed in Appendix A.
2.6 Gas Storage
Gas storage fields are also represented in the PLEXOS gas model, tied to gas pipelines via gas nodes, and
defined by the following key properties:
-
-
-
Initial volume (MMcf): the volume of gas in the storage field at the start of the simulation
Min/max volume (MMcf): the min/max volume of gas allowed in storage
o “Working gas” is the difference between the max volume and the min volume
Max withdrawal/injection (MMcf): the maximum volume that can be withdrawn from/injected
into storage in a given interval, e.g. daily or hourly
Withdrawal/injection charge ($/MMBtu): the incremental cost to withdraw gas from/inject gas
into storage
o Can be defined as a curve or in bands, e.g. increasing costs as field is depleted
Target (MMcf): sets the storage volume for the end of the interval
o An associated target penalty ($/MMBtu) can also be set for violating the target, to drive
storage utilization trends in the model.
Max ramp (MMcf): the maximum allowed rate of change in storage end volume between one
dispatch interval and the next
21
See http://www.eipconline.com/uploads/Exhibit_15_LDCs_Gas_Demand_Forecasts.pdf
The CERA forecast is a 20-year projection (single daily demand value per month) per state.
24
See Appendices 1 through 6 at http://www.eipconline.com/Gas-Electric_Documents.html for generatorto-pipeline mapping.
22
DRAFT - 14 | P a g e
In the current version of the model, there are hundreds of storage field objects26. For these fields, the
vendor has included definition of initial storage levels (for January and May), max volume and max
withdrawals. These values are based on the EIPC study dataset (see Table E7-5 in Exhibit 7. GPCM
13Q4base Database27) and they will be retained for the CPP study.
The generic withdrawal charge as set by the vendor for state-level gas storage is equal to 10% of the
cost to produce gas in a given state. Injection charges are not currently modeled in PLEXOS. For the
Phase III study, MISO will replace the state-level values for the cost to produce gas with forecasted
values for individual gas basins (which will be applied to nearby pooling points per pipeline, as described
in Section 5.1). MISO proposes to retain the 10% storage charge-to-production cost assumption, using
the forecasted production cost values for the Phase III study but welcomes feedback on more
appropriate methods of modeling gas storage charges in future applications of the PLEXOS model.
Storage injections and withdrawals are not modeled as profiles; rather, the model will consider storage
withdrawal/injection as it makes dispatch decisions.
2.7 Future Developments
The PLEXOS gas model is continually evolving.
Planned and potential improvements (and approximate timelines) include representation of the
following:
-
Multiple interconnections between gas-fired generators and gas pipelines (near-term)
o In the current version of the gas model, gas-fired electric generators that are
interconnected with multiple pipelines are represented with a single interconnection,
arbitrarily selected amongst the multiple interconnects.
o Future versions of the model will attempt to reflect multiple interconnects.
-
Gas storage tariffs (mid-term)
o Currently, modeled storage withdrawal/injection rates do not reflect cost differentials
across gas storage fields; future versions of PLEXOS may include customized
representation of storage costs, such as deliverability, capacity and withdrawal/injection
charges.
-
Gas pipeline tariffs (mid-term)
o In the current version of the model, fuel transportation contracts are not captured, i.e.
regardless of the contract arrangements made in the real world, each gas-fired
generator in the model has the same access to the pipeline capacity with which it is
interconnected. Essentially all gas-fired generators have interruptible transportation
service in this version of the model.
o More specifically, there is no gas transport cost (captured via the flow charge property
on individual pipelines or pipeline segments) reflected in the model. Thus, the order of
dispatch of gas-fired electric generators in the model is driven by 1) the cost to supply
gas at a given pooling point, 2) the economic and physical characteristics of generators
and 3) the physical constraints on the pipeline network and the electric transmission
system.
26
See Appendix E for a list of the individual gas storage fields defined in the PLEXOS model.
See http://www.eipconline.com/uploads/Exhibit_7_GPCM_13Q4base_Database.pdf for definition of
individual storage properties (capacity, MDth; max injection rate, MDth/d; max withdrawal rate, MDth/d),
27
DRAFT - 15 | P a g e
o
o
-
MISO notes the significant role the difference in the cost of transport from one region to
another (i.e. basis) plays in the real world dispatch of gas-fired generators. Future
iterations of the model will attempt28 to reflect transport cost variations from one
pipeline to the next, by assigning pipeline tariff rates as transport costs via the Flow
Charge property to each pipeline segment modeled.
Likewise, LDC contracts to serve RCI load are not captured in the current version of the
model; however, because RCI demand is modeled with a demand profile, it is essentially
equivalent to firm service and reduces the amount of pipeline capacity available for gasfired electric generators.
Local Distribution Companies (LDCs) (mid-term)
o RCI demand behind the LDC city gate is currently represented as part of a state-wide
demand profile that is then distributed across gas nodes in the state per LDC demand
forecasts. This demand will require access to pipeline capacity on an hourly basis at a
level corresponding to the level of hourly demand. The remainder of the pipeline
capacity is available to serve electric demand for gas. The underlying assumption of this
representation of RCI demand is that LDCs will release contractual capacity not needed
to serve load and that electric generators will have access to this capacity, on an hourly
basis. While this is a simplified representation of real world allocation of unused
capacity, it is a reasonable assumption for this stage of the model.
o Electric power demand from gas-fired generators behind LDC city gates is currently
represented in the same manner as electric power demand with direct interconnects.
This is to say that regardless of whether the generator behind the city gate has a firm
contract with an LDC, it is still modeled on par with generators that are directly
interconnected with the interstate pipeline system.
o Future versions of the model may more closely represent LDC load, including electric
generation. This aspect of the model is still under development.
Other potential areas for future model development include:
-
Customized gas demand shortage penalty price
o There is currently a placeholder penalty for unserved gas demand (represented as “gas
demand shortage” in the objective function) in the model of $100/MMBtu; absent an
industry standard or explicit market indicator for the value of lost gas load, this figure
will serve as a generic approximation for the whole of the Eastern Interconnect.
-
Priority order for pipeline access/generator dispatch
o Each gas-fired generator in the current version of the model is mapped to a single
pipeline (future versions will account for multiple interconnects).
o Each individual pipeline is represented as interconnected pipeline segments, based on
tariff-defined major production and market zones, tied together via gas nodes.
o At this level of granularity, there may be multiple gas-fired generators connected via the
same node to a given segment of pipe. For existing gas-fired generators, individual
28
Each pipeline company offers their own suite of services, at rates customized to the cost recovery
needs of that pipeline. While pipelines are required to list min and max rates, shippers may also receive
discounted (negotiated rates). MISO will work with stakeholders, the gas industry and the vendor of
PLEXOS to determine an appropriate approach to modeling transport costs in future iterations of the
model.
DRAFT - 16 | P a g e
o
o
generator characteristics (i.e. heat rate) will dictate which of the similarly mapped gens
would be dispatched first.
For new, generic (all with the same characteristics and costs) combined cycle and
combustion turbines in the model mapped to the same node, the first listed will be
dispatched.
In future versions of the model, the type of contract a given generator holds may be the
deciding factor for which unit/s gets dispatched amongst several generators mapped to
the same gas node.
MISO welcomes stakeholder input on the current representation of gas infrastructure in the PLEXOS
model, as well as on planned and potential model developments.
DRAFT - 17 | P a g e
3 Co-optimization
The production cost module of PLEXOS performs chronological, security-constrained unit commitment
and economic dispatch. The integration of the PLEXOS gas model, described in the preceding sections,
with the PLEXOS electric model allows for co-optimized gas-electric dispatch. This means that the model
considers both gas and electric infrastructure characteristics and constraints when formulating the
dispatch stack.
The objective function for gas-electric production cost modeling in PLEXOS can be expressed as the
minimization of the cost to serve gas and electric load, subject to a number of constraints, or
Minimize:
Subject to:
Electric transmission constraints
Pipeline constraints (design capacity)
Feasible gas production
Feasible electric generation and feasible ancillary services provision (per generator)
Electric generation = Electric demand + Electric losses – Unserved electric load
Gas production = Gas demand + Gas generation demand – Unserved gas load
Ancillary service provision ≥ Ancillary Services Requirements (for the pool)
Where:
PCe = electric production cost
PCg = gas production cost
ASc = ancillary services cost
DSe = electric demand shortage cost
DSg = gas demand shortage cost
The definition of gas production cost in the equation above is the volume of gas produced at the field
multiplied by the short run marginal cost of production. The gas demand shortage cost is defined as the
cost of unserved demand on the gas system. The gas production cost and gas demand shortage cost are
simultaneously considered with the electric production cost and electric demand shortage cost, as well
as with the cost of providing ancillary services, in the optimization.
DRAFT - 18 | P a g e
4 Outputs
The user can select the desired outputs of the production cost simulation from 1,000’s of output
parameters. Each individual parameter can be reported for a given period (time interval) and/or as a
summary for the entirety of the simulation. The graphic below shows the outputs of gas-electric cooptimization at a high level.
Figure 9. High-level inputs and outputs for co-optimized gas-electric dispatch in PLEXOS
The outputs of the gas side of the model can be grouped into two main buckets:
-
Physical (congestion) metrics: quantification of the utilization of capacity at each gas node/for
each pipeline segment, in every interval of the simulation; in other words, the duration, location
& magnitude of pipeline congestion
o For comparison, the electric side outputs of the model include line flows and binding
hours.
-
Economic (cost/price) metrics: quantification of the cost to produce and transport gas; gas spot
prices are provided at each gas node for every interval of the simulation
o For comparison, the electric side outputs of the model include locational marginal prices
(LMPs).
In combination, the physical and economic metrics should allow for identification of candidate areas for
natural gas infrastructure expansion. While much of the MISO stakeholder body is familiar with the
representation of electric transmission system congestion, the characterization of congestion on the gas
network is new territory for both MISO and the majority of its stakeholders. The determination of
indicative gas infrastructure needs in the context of the CPP will be a collaborative effort, amongst
MISO, its stakeholders, including gas industry representation.
DRAFT - 19 | P a g e
5 Summary & Next Steps
Though co-optimized gas-electric production cost modeling of this scale is new, it has the potential to
offer insight into the complex interactions between the two systems. In the case of MISO’s CPP study, it
will be used to quantify future natural gas system impacts, and in doing so, allow MISO to develop a
more comprehensive picture of the cost to achieve compliance.
MISO will continue to work with PLEXOS’ vendor to develop the functionality of the gas model, including
the granularity of the representation of the physical infrastructure, as well as the representation of
financial/contractual aspects of gas system operation. Furthermore, MISO will continue to engage with
stakeholders on the application of the gas-electric model in the Phase III CPP study and welcomes
feedback on modeling assumptions, including:
•
Gas transport cost
•
Gas transportation contracts
•
Gas field production cost curves
•
Gas storage withdrawal/injection charges
•
Gas demand shortage cost
DRAFT - 20 | P a g e
Appendix A: Common PLEXOS Electric Generator
Properties (Units) and Descriptions
The following is a list of common properties used to define electric generators in the PLEXOS model. Not
all of these are defined in the current version of the model (not all properties need to be defined to run
the model).
-
-
-
Production
o Max capacity
o Min stable level
o Fuel price
o Load point
o Heat rate
o Variable operation and maintenance charge
o Start cost and start cost time
o Start penalty
o Shutdown cost
o Shutdown penalty
o Rating
o Rating factor
o Min up time
o Min down time
o Fixed load
o Commit
o Max ramp up
o Max ramp penalty
o Max ramp down
o Max ramp down penalty
o Pump efficiency
o Pump load
o Pump units
o Min pump load
o Offer quantity
o Offer price
Capacity
o Fixed operation and maintenance charge
o Firm capacity
o Maintenance rate
o Forced outage rate
o Mean time to repair
Constraints
o Max energy day
o Max energy month
o Max energy year
o Max energy
DRAFT - 21 | P a g e
Appendix B: PLEXOS Natural Gas Pipelines
Acadian Pipeline
Agua Dulce – Frontera
Agua Dulce Hub
Alaska Pipeline Project
Alaska Valdez Pipeline
Algonquin
Alliance Pipeline
American Midstream (AlaTenn)
Anaconda
ANG (TCPL BC system)
ANR
Arkoma Connector
Atmos P/L TX (Lone Star)
Big Sandy
Bison Pipeline
Bluewater
Bridgeline
Brooklyn Union Hub
Brunswick
Canyon Chief
Canyon Express
Carolina Gas Transmission
Carthage Hub
CEGT
Centerpoint SESH
Cheyenne Hub
Cheyenne Plains
Chicago Hub
Chihuahua
Clarington Hub
Cleopatra Gathering
Colorado Interstate Gas
Columbia Gas Transmission
Columbia Gulf Transmission
Commonwealth
Constitution
Consumers Energy
Creole Trail
Crossroads
CrossTex North Texas
Cypress Gas PL
Dauphin Island Gathering
Destin Pipeline
Discovery
Dominion (CNG)
East Breaks Gathering Co
Golden Pass Pipe Line
Great Lakes
Green Canyon Pipeline
GTN TransCanada (NEGT)
Guardian
Gulf Crossing
Gulf South
Gulfstream Pipeline
Henry Hub
High Island Offshore
Houston Pipeline Co
Independence Hub/Trail
Iroquois
Kanda Colman
Katy Hub
Kern River
KM Louisiana Pipeline
KM North Texas PL
KM Tejas Pipeline
KM Texas Pipeline
Lebanon
Leidy Hub
Los Ramones
Louisiana Intrastate (LIG)
Louisiana Resources Pipeline
Mackenzie Valley Pipeline
Manta Ray Offshore Gathering
Maritimes & Northeast
Matargordo Offshore Pipe
Michigan Consolidated (MichCon)
Midcontinent Express Pipeline
Midstream Central Station
Midwestern Gas Transmission
Millennium Pipeline
Mississippi Canyon Gathering
Mississippi River Transmission
Mobile Bay Gathering Misc
Mojave
National Fuel Gas Supply
Nautilus Pipeline
NE Exp
Nemo Pipeline
New York Facilities System
Nexus
NGPL
Nicor
Paiute
Panhandle Eastern
Patriot
PEMEX
Pennstar
Perdido Norte
Perryville Hub
PG&E
Phoenix Gathering
Portland Natural Gas Transmission
Postrock KPC
Questar Pipeline
Ram-Powell Pipeline
Regency
Renaissance
REX East to West Hub
Rockies Express
Ruby
Sabal Trail
Sabine Pipeline
Sasabe-Guaymas
Sea Robin
Seahawk Gathering
Sierrita
So Cal Gas
Southern Natural
Southern Star (Williams Central)
Southern Trails
St. Lawrence Gas
Stagecoach Hub
Stingray
Suffield
Tallgrass Interstate Gas Transmission
Tennessee Gas Pipeline
Texas Eastern
Texas Gas Transmission
Tiger
Topolobampo
Trailblazer
TransCanada Pipeline
TransColorado
Transcontinental
Transgas
Transok
Trans-Union Interstate
Transwestern
DRAFT - 22 | P a g e
East Tennessee
Eastern Shore Natural Gas
El Paso Natural Gas
Elba Express
Empire State Pipeline
Enbridge Magnolia Gathering
Enogex OK
Enterprise Intrastate (Channel)
Enterprise Texas (Valero/Teco)
Equitrans
ET Fuel (Tufco)
Falcon Gas Pipeline
Fayetteville Express
Florida Gas
Foothills
Garden Banks
GC Center
Norte Crossing
North Baja
North Penn
Northern Border
Northern Natural
Northwest Pipeline
NorthWestern Energy (Montana Power)
Nova (TCPL Alberta System)
Okeanos Deepwater System
OKTEX Del Norte
ONEOK WesTex Transmission (Westar)
Opal - W Market Center
Opal Pioneer Plt
Overland Trail Transmission
Overthrust
Ozark Gas Transmission
Pacific Connector
Triton Gathering
Trunkline
Tuscarora
Union Gas Ltd
UT Offshore
Vector Pipeline
Venice Gathering System
Vermont Gas System
Viking
Viosca Knoll
Waha Hub--West TX
Wattenberg Gathering
Westcoast
Westlake GTL Header
White River Hub
Williston Basin Interstate
Wyoming Interstate
Xcel (PSC of Colorado)
DRAFT - 23 | P a g e
Appendix C: Pipeline Capacity and Expansion
Properties
Capacity properties:
-
Maintenance rate (%): the probability that the gas pipeline is out-of-service for a planned
outage, either full or partial
Maintenance frequency: the frequency of maintenance outages in an annual time frame (an
alternative to maintenance rate)
Forced outage rate (%): the probability that the gas pipeline will be out-of-service due to an
unexpected outage, either full or partial
Outage max flow (MMcf): the pipeline max flow during the outage
Outage max flow back (MMcf): the pipeline max flow back during the outage
o This is only applied to bi-directional pipelines.
Min time to repair (hrs): the lower bound on the repair time after an outage
Max time to repair (hrs): the upper bound on the repair time after an outage
Repair time shape: sets the shape parameter of the repair time distribution in the Monte Carlo
outage simulation (for Weibull, lognormal)
Repair time scale: sets the scale parameter of the repair time distribution in the Monte Carlo
outage simulation (for exponential, Weibull, lognormal, SEV, LEV)
Though not used in MISO’s Phase III CPP study, expansion properties can be defined for capacity
expansion modeling, including:
-
Min/max units built: min/max number of units that are automatically constructed in aggregate
over the planning horizon
Min/max units built in year: min/max number of units that are automatically constructed in any
single year of the planning horizon
Project start date: start date of gas pipeline project, for expansion planning
Technical life (yrs): the technical lifetime of the gas pipeline (physical operating period)
Build cost ($): the ‘overnight’ cost of building the pipeline
WACC: the weighted average cost of capital
Economic life (yrs): the economic life of the gas pipeline (period over which fixed costs are
recovered)
Min/max units retired: min/maximum number of units automatically retired in aggregate over
the planning horizon
Min/max units retired in year: min/max number of units that are automatically retired in any
single year of the planning horizon
Retirement cost: the cost of retiring the gas pipeline
DRAFT - 24 | P a g e
Appendix D: PLEXOS Natural Gas Storage Fields
In this version of the PLEXOS Integrated Energy Model, gas infrastructure in the Eastern Interconnection
(EI) is modeled, including gas storage fields. The following states in the EI have gas storage fields: Illinois,
Indiana, Iowa, Kentucky, Louisiana, Maryland, Michigan, Minnesota, Missouri, Montana, New York,
Ohio, Pennsylvania, Texas, Virginia and West Virginia. The gas storage fields and their characteristics
(max withdrawal per day and max volume) are based on the EIPC study data (see Table E7-5. Storage
Facilities in Exhibit 7. GPCM 13Q4base Database of Target 2 Report – Final Draft). The “(1)” or “(2)” label
following the storage field name generally indicates a different owner for a field at a similar location as
the original.
Accident
Adrian
Alford
Amory
Ancona
Arcadia
Artemas A
Artemas B
Ashmore
Augusta
Austin
Baker
Barnsley
Bear Creek
Beech Hill
Belle River
Belmouth
Bennington
Benton
Bistineau
Blackhawk
Blue Lake 18a
Bluewater
Bluewater (1)
Boardwalk
Bon Harbor
Bond
Boone Mountain
Bridgeport
Brinker
Bunola
Cairo
Carter
Cecilia
Center
Guernsey
Hanson
Hardy
Harrison
Hattiesburg
Hawesville NW
Hayes
Heard
Hebron
Heizer A-1
Henderson
Herscher
Herscher Northwest
Hessen
Hester
Hickory
Hillsboro
Hindustan
Holbrook
Holland
Holmes
Honeoye
Hookdale
Howell
Howesville
Hudson
Hughes
Hunt
Hunters Cave
Ira
Jackson
Jefferson Island
Johnston City
Keelor
Kennedy Lost Creek
Pavonia
Pecatonica
Perry
Perrysburg
Petal
Pine Prairie
Pontiac
Pontiac (1)
Port Barre
Portman
Pratt
Puttygut
Queen
Quinlan
Rachet-Newberne
Rager Mountain
Raleigh City
Rapid River 35
Ray
Redfield
Reed City
Rhodes
Ripley
Riverside
Rockport
Royal Center
Ruston
Sabinsville
Salem
Saltville
Sciota
Sellersburg
Seneca Lake
Shanghai
Sharon
DRAFT - 25 | P a g e
Central Charlton
Centralia
Chippewa
Coco A
Coco B
Coco C
Cold Springs 1
Cold Springs 12
Cold Springs 31
Colden
Collins
Collins Field
Columbiana
Columbus
Columbus City
Comet
Cooks Mills
Corry
Cranberry Lake
Crawford
Crofton East
Dayton North
Derby
Dixie
Dixon
Doe Run
Donegal
Dundee
Early Grove
East Branch
East Diamond
East Independence
East Slaughters
East Unionville
Eaton Rapids
Eden
Egan
Ellisburg
Keota
Kinter
Kirkwood
Lacey
Lake Bloomington
Lanham
Laurel
Lawtons
Lee 11
Lee 2
Lee 8
Leesville
Leidy Tamarack
Lenox
Lexington
Limestone
Lincoln
Lincoln-Freeman
Little Capon
Logansport
Loogootee
Lorain
Loreed
Loudon
Love
Lucas
Lyon 29
Majorsville Dp
Majorsville Sh
Manlove Field
Maple Lake
Markle
Marysville Storage
Mcarthur
Medina
Meeker
Midland
Midway
Sheridan
Shirley
Simpson Chapel
Skin Creek
Sorrento
South Bend
South Chester 15
South Huntingburg
Southern Pines
Spindletop
Spindletop (1)
Spindletop (2)
St Charles
St Marys
St. Jacob
Stagecoach Field
Stark-Summit
Stechman Ridge
Summit
Swan Creek
Swarts Complex
Swede Hill
Switz City
Taggart
Tepe
Terra Alta
Terra Alta South
Thomas Corners
Tilden
Tioga
Troy Grove
Truittsburg
Tuscarora
Unionville
Vardy
Victory A
Victory B
Washington (1)
DRAFT - 26 | P a g e
Ellisburg (1)
Eminence
Epps
Eric
Excelsior 6
Finleyville
Florissant
Four Corners
Four Mile Creek
Freeburg
Gabor Wertz
Galbraith
Gamble Hayden
Glady
Glasford
Goodwell
Goodwin
Graham Lake
Grand Bayou
Grandview
Grass Creek
Greenlick
Greenwood
Mills
Mineral City
Mobley
Monroe City
Moss Bluff
Muldon
Muldraugh
Murrysville (Dice)
Muskie
Muttonville
Napoleonville
Nashville
New Home Dome
North Greenwood
North Summit
Northville
Oakford
Oaktown
Oliver
Onondaga
Overisel
Partello
Washington 10
Waterville
Waverly
Wayne
Weaver
Webster
Wellendorf
Wellington
West Columbus
West Greenville
West Independence
West Unionville
Wharton
White River
Wilfred
Winfield
Winterfield
Wolcott
Woodhull
Worthington
Zane
Zoar
DRAFT - 27 | P a g e
Appendix E: PLEXOS Database Structure
Data in PLEXOS is organized into Objects, Properties and Memberships.
Objects are the fundamental components of the PLEXOS model. Power system (e.g. generators,
transmission lines) and gas system (e.g. gas storage fields, pipelines) elements are all represented as
objects in the model. Each object is assigned a class, name and category. For example, an individual
generator may be defined with class “Generator”, name “Rock Creek: CT1” and category “ABC Energy
Company”. Additional definition (e.g. state, “IA”, fuel type, “natural gas” or region, “MISO”) can be
added by the user for a given object.
Objects are further defined by Properties. Properties can be static (the value is constant) or dynamic
(the value may vary over time, or read from an external file, or change given the application of a
scenario). Properties for a given generator define operational, physical and financial characteristics,
such as max capacity, heat rate, mean time to repair or new build cost. A list of common generator
properties can be found in Appendix A. Likewise, properties for other electric elements as well as gas
elements define physical, operational and financial characteristics.
Memberships dictate the relationship between objects in the model. For example, each generator is a
defined as an individual object, as is each fuel. Each generator requiring a fuel must have an associated
fuel source assigned (e.g. “Pipeline X: SW Leg”). This assignment comprises the membership, defined by
collection, parent name, child name, parent category and child category. For example, the collection for
the fictional Rock Creek unit would be “Generator.Fuels”; the parent name would be the name of the
generator (“Rock Creek: CT1”); the child name would be the name of the fuel source (“Pipeline X: SW
Leg”); the parent category would be the generator owner/area (“ABC Energy Company”); and the child
category would be the fuel (transportation) owner/area (“Pipeline X”).
The memberships that tie together the gas and electric infrastructure in the model are the (gas-fired)
generator-to-gas-node and (gas-fired) generator-to-electric-node.
DRAFT - 28 | P a g e