We've been meaning to write this post for a while about a very exciting new company out there called GridReason, founded by a former colleague of this author. To date, most Critical Peak (CP) prediction services have been driven by models that rely, at least in part, on daily human judgement to determine the risk of an occurrence of a peak load event.. This author has designed and/or managed such programs for ISO-NE, ERCOT, and Ontario.
As a quick refresher, a CP is an interval when a system reaches peak load for a given period. You may hear terms such as 1 CP, 5 CP, or 12 CP bantered about. When you hear the term "1 CP", it refers to a single critical peak which is the peak load interval for a system during a given period. Oftentimes this period is one year. By the same logic, 5 CP would be the five highest load intervals in a given period. In New England, you may hear 12 CP since ISO-NE Transmission charges are calculated based on monthly system peaks and there are 12 months in a year (but Trans cost allocation is way too complex for this blog post). A Coincident Peak refers to the load of a customer during the system CP interval(s). Capacity Tags are assigned to end use customers based on their Coincident Peak(s) and determine the amount that they will have to pay for capacity on the supply portion of their bill.
Anyone who works in deregulated energy markets should understand how capacity charges are allocated based on CPs. The table below provides a brief overview of the markets where costs are allocated to end users based on usage during CPs.
Once an end user understands how their capacity tags are set, they can decide if they'd like to actively manage them. The decision to manage capacity tags is largely driven by opportunity costs and operational flexibility. A company making a low margin commodity product may have a low opportunity cost of interruption and would be happy to drop load during potential CP hours in exchange for the ability to reduce their electricity spend. To the contrary, a pharmaceutical plant or a Just-in-Time manufacturer has a very high opportunity cost of interruption (batches get junked, schedules get blown) and would typically be unsuitable for a capacity tag management program.
For those who can manage their capacity tags, GridReason becomes an awesome resource. They've developed powerful statistical algorithms to predict the likelihood of a CP event and are available at a fraction of the cost of existing predictive services. If you are curious to learn more about managing capacity tags and their offering, call them up and tell them ETE sent you.
If you'd like to learn more about capacity tags, how they work, and what capacity costs in various deregulated markets, please call us. We'd be happy to give you a proposal for whatever information you are looking for.