IEEE Power & Energy Magazine - May/June 2016 - 77
and dynamically, is going to be essential in designing the
DSO markets and its integration with wholesale markets.
DSO Infrastructure Alternatives
and Cost Benefit Aspects
Earlier, we mentioned that the first wholesale market systems had development and implementation costs in the tens
of millions of dollars. Today, increased market scale and
complexity might result in implementation costs for a new
market system that approach US$100 million including the
systems used to register participants, publish information,
accept bids, clear the market (including the sophisticated
optimization algorithms), and perform settlements. (Settlements is the process of calculating payments from the market
to suppliers and to the market from load serving entities.) In
the wholesale markets, it is typical today for the settlements
process to include the order of 100 different "charge types"
or settlement calculations based on the number of products
and the types of participants in the market. Larger market
participants have duplicate systems of their own to interface
to the wholesale markets and emulate the clearing and settlements calculation. So the total investment in market systems
by the market operator and the participants is at least twice
the US$100 million that the market operator has to spend.
The high cost of these market systems is justified due to
the large volumes of energy that are transacted in the market. on the order of US$10 billion a year in most ISOs today.
Even a comparatively small improvement in market efficiency will pay back the investment in a year or lower.
Will a DSO cost as much as an ISO market system? And
will it pay for itself in a reasonable period of time? These are
key questions. The design of the DSO markets and operations has to take the costs into account at some point when
making decisions about complexity and scale. We're not
going to project DSO financials in this article, but we will
talk about some challenges to realizing a cost effective DSO.
✔ First, the DSO will have to come into operation with
a fairly complete and attractive set of market products
and functions if it is to achieve the goal of incentivizing DER deployment and market participation. This
means that the bulk of the software development costs
will have to occur up front and that the benefits won't
start accruing until customers and third parties adopt
DER and DSO market participation.
✔ Second, the DSO scale in terms of number of participants will (hopefully) be larger than that of an ISO.
On the one hand, a DSO might only encompass the
territory of one utility-probably 10-25% of the total load served of an ISO. But, there will be many
more retail-level resources to manage than there are
wholesale generators. So where an ISO might have on
the order of 1,000-5,000 market participants, a DSO
might have about 50,000 participants unless aggregators can reduce that. But for reliability purposes the
DSO will need to track that many resources.
✔ Third, the ISO today has to manage reliability and con-
gestion on a transmission network of tens of thousands
of nodes; the largest network model in use at an ISO is
100,000 nodes. Here, the DSO benefits from the largely
radial nature of distribution systems, what happens on
one feeder electrically can largely be separated from adjacent feeders. But the number of feeders in a utility is
large (on the order of thousands for a large utility) and
the number of electrical nodes on each feeder to be represented is potentially very large (also on the order of thousands, particularly if lateral circuits or secondary systems serving residential prosumers are to be considered).
The DSO will likely have at least as many market
products as the ISO. To gain the advantages of wholesale market participation, it needs to emulate most of the
wholesale market, and then it will have unique distribution
Therefore, the scale of the DSO market operations would
be expected to be roughly the same or greater than that of the
ISO market operations. The need to solve a complex optimization problem across the entire network (as opposed to
subnetworks) in a narrow time window may not be an issue,
but all the other aspects will be.
Will development of the DSO benefit from lessons
learned from the wholesale market? Undoubtedly, it will;
plus the state of the art of e-commerce is far more advanced
from when the first ISO systems were implemented, and
development costs are arguably lower. On the other hand,
new development will be required for novel DSO products
Will DSO systems benefit from a standardization effect
wherein multiple utilities will be able to use the same software
and business processes? Will DSO systems be cloud based on
a per-transaction service fee basis that reduces utility investment and costs? These are two arguments for reduced DSO
costs as compared to wholesale markets. Against this must
be placed the tendency for different regulatory agencies to
impose adjustments on tariffs and market design in each state
where a utility has a DSO. And at the retail level, we cannot
forget that other jurisdictions impose other taxes and fees to
factor into retail tariffs and utility bills; each county and city
or township can impose different real estate taxes on utility
assets, for instance. It remains to be seen how this will play
out as DSOs develop.
Communications and control between market resources
(and nonmarket DERs) and the DSO are another cost element. When ISOs were formed, the essential transmission
system monitoring and control technologies were in place
at the utilities. Over time, the communications to generators
were transferred (in most ISOs) from the utilities to the new
generation owners. Direct communications to DERs from
the ISO are generally not in place.
The DSO implementation will face several challenges.
One is whether the communications to distribution feeder
devices and necessary control systems are in place to support
ieee power & energy magazine