Customer Outage Costs

Customer outage effects are a critical part of the consequence of failure for most assets. When, for example, a cable faults, the utility is faced with the cost to repair the fault, but more importantly customers are without power. In almost all cases, the customer effects carry more significance.

Conceptual approach to customer outage costs

Customer costs do not involve real dollars changing hands between Toronto Hydro and its customers. Rather, these costs are placeholders in the feeder model to ensure that the customers’ perspective is explicitly included in economic analyses and spending decisions. A good analogy is a car manufacturer deciding whether to include automatic windows in a particular model. If he knows the automatic windows will increase cost by $1000 per car, then he must ask whether he thinks his customers value the feature at more than $1000. If they do he should install it; if they don’t he should not. Putting the customer at the center of spending decisions is a fundamental component of Asset Management decisions.

Ideally, outage cost should capture the amount the customer would have been willing to spend to avoid the outage (or the amount he would want to be compensated to accept it). Ideally, we would like to simply ask customers how much they would pay to avoid an outage of, say, two hours at a random time in the next year. This concept is represented by the figure below.

In this scenario, if there were an outage, the total cost to the customers would be $25 for direct repair, which would be captured in their rates, and a total of $110 in subjective cost due to lost service, which is not collected by the utility or anyone else. This is utilitarian approach to quantifying outage cost. We say that any investment whose total benefit outweighs its total cost is justified, regardless of who receives the benefits or bears the costs. (This approach is commonly used by government agencies in their cost benefit analyses.)

Unfortunately, it is not usually possible to simply ask electric energy customers their perceived cost of an outage. In the case of residential customers, they probably do not know and would have difficulty estimating, and both residential and commercial/industrial customers are unlikely to answer such a question without bringing their own strategic considerations to bear. For example, they may believe you will raise their rates based on their answer. Numerous surveys have been performed to estimate customer outage cost (also called “damage function”). The results have been mixed and mostly unusable for our purposes. Because of this difficulty, Toronto Hydro uses assumed values of customer cost for assessing the consequences of outages.

Practical approach

In Toronto Hydro’s view, customer outages have three parameters that affect their importance: the outage event (“SAIFI” effect), the outage duration (“SAIDI” effect), and the type of customers affected. Currently, it is difficult to correlate numbers of customers or types of customers with particular assets. Because of this, Toronto Hydro’s approach to quantifying and dollarizing customer outage costs is based on the peak load affected by any failure.

For example, in the conceptual diagram shown above if the breaker fails the entire feeder will be out of service for some period of time. Therefore, the customer outage cost for the breaker failure would be based on the peak load for the feeder, which information is readily available. Where peak load is not known, say in the case of pole-top transformers or line reclosers, connected kVA is used instead. Connected kVA can be defined as the sum of the nameplate ratings of the connected transformers that will be impacted by an outage. In this case, connected kVA is used as a proxy to represent both the quantity of customers, as well as the importance of those customers that will be impacted by an outage.

Most utilities think of customer interruptions costs as having two distinct components: the “event” cost, which corresponds to SAIFI effect, and the “duration” cost, which corresponds to SAIDI effect. This is a natural framework, since SAIDI and SAIFI are used universally among electric utilities. The “event” and “duration” costs are often specified for a given customer type (e.g., residential, commercial), and for non-residential customers they may be functions of the size of the service. For simplicity, the “duration” cost is usually assumed to be constant, so that the fourth hour of an outage is just as costly as the third. However, because the first hour includes the “event” cost as well as the first hour of “duration,” the first hour is always the most costly. These approximations may not be valid for very long outages, where customers take steps to mitigate their cost, such as leaving town or installing a backup generator.

Toronto Hydro has decided on a cost of $30 for each kVA of peak load interrupted. This is the outage “event” or “SAIFI” cost related to an asset failure. For the “duration” or “SAIDI” cost, Toronto Hydro has decided on a cost of $15 per kVA per hour of interruption. So, a failure caused a three-hour outage for customers representing 100kVA of peak load would represent a cost of 100kVA x $30 for the “event” and a cost of 100kVA x 3 hours x $15 for the “duration” effect, for a total outage cost of $3000 + $4500 = $7500. This would be in addition to any direct costs to repair or replace failed equipment. In cases where peak load is not available, connected kVA of the downstream transformers or other nameplate ratings are used to approximate peak load.

These costs will be reviewed periodically and updated as Toronto Hydro’s senior management sees fit or as new information becomes available. Bear in mind that this approach is intended to approximate the customers’ perceived cost, so there is nothing hard-and-fast about the methodology used. It is possible in the future that a better approach, perhaps based on new survey data, will become available. In that case, the feeder model will be updated to reflect this new approach.

Continue to destructive vs non-destructive.


Fatal error: Allowed memory size of 201326592 bytes exhausted (tried to allocate 35 bytes) in C:\Inetpub\vhosts\\httpdocs\wikis\bis\inc\auth\plain.class.php on line 281