Monday, November 15, 2010

Smart meters and cognitive biases

The Obama administration's $787 bn stimulus plan contained $3.4 bn to modernize America's electricity grid, including the installation of smart meters. It is estimated that by 2020, there could be as many as 65 million smart meters in the US.

Smart meters have the potential to revolutionize electricity usage and grid management by enabling more efficient, demand-based allocation of power to end-users. They are bi-directional meters that transmit real-time consumption data which can be used by utilities to manage demand using time-of-day pricing and consumers to optimize consumption based on the price signals.

However, the roll-out of smart meters in the US has been accompanied with widespread criticism of inaccuracies in billing. Customers have complained of excess billing and class-action suits have been initiated in a few states. This mirrors similar complaints when Indian utilities replace older meters with the latest digital meters, especially those which have infra-red port-based readings.

I am inclined to the view that such complaints are a result of a cognitive bias called availability heuristic. Consumers' minds have been anchored to a range of values for many years and even decades. Suddenly, when their billed units go up, they struggle to mentally accept it. The availability bias continually draws the old billing pattern to their minds.

In any case, the transition period from one metering system to another will naturally generate discrepancies arising from a variety of commonplace administrative and technical problems. However, in case of smart meters, they are likely to be more significant for the following reasons

1. Smart meters have higher billing accuracy than conventional meters. So it is but natural that input fractions and connection losses at the consumer end that generally get added to utility losses now get, right fully, billed on the consumer.

2. A significant number of old meters will certainly be defective. Their defects range from being sluggish and under-recording units to being stuck-up (and therefore being billed based on average consumption of earlier months). In countries like India, the proportion of such meters can be as high as 60-70%, and even more in utilities where old meters have never been replaced. In one stroke, a smart meter brings out all these suppressed readings.

3. If meters are replaced during summer, then the inevitable seasonal variations (consumption rises in summer) get attributed to the changed meters and its deficiencies.

4. Finally, why should consumers be surprised that their billed units have suddenly increased? The very objective of smart meters is to capture the hidden commercial losses. It is natural that all those charged higher units with the old meters would have had their meters tested and replaced. This only leaves the remaining meters, consisting of the normal ones and the under-recording ones. Their replacement will naturally result in a rise in billed units.

All this only highlights the importance of awareness campaigns to sensitize consumers about the reasons for possible higher billing on installation of smart meters.

Update 1 (16/11/2010)

One of the comments pointed to the concerns over radio frequency radiation arising from smart meters. Green blog points to a new study that compares the radiation levels from various devices and finds out that smart meter radiations are tens of thousands of times smaller than even cell-phone radiations.


J. David Foster said...

Part of the opposition to Smart Meters in the U.S. stems from a growing concern regarding potential radiation hazard associated with these meters. Curiously, many of those most vocally concerned about this potential risk ares still avid users of cell phones.

Jayan said...

Great topic!

Smart meters provides a quantum jump in data collection. Just like UID, it could be a starting point. Wider usage will enable us real time data collection (or say every 15 minutes from all meters) and build a super model to predict the energy requirements for future cities and villages.

It may be more useful to install energy meters on each of the transformers and specific distribution points- say beginning of a street. This will help to identify the leak/theft points.

It is also important to start billing according to time and location. The consumption at 6 pm could be charged at premium than one at 10 am. It is also possible to charge x% more in Jubilee hill(Hyderabad) than one in Hafeezpet(Hyderabad). Finally it should be possible get a report against each lineman/engineer on the recovery rate.

gulzar said...

thanks david. missed out on that. have posted an update.

jayan, yes all those things are possible. a few states have some form of Time of Day (ToD) billing for certain categories of consumers. we do not have bi-directional smart meters in India, but uni-directional automatic meter readers (AMR). some utilities have placed these AMR meters for HT and otehr high-value consumers.

Jayan said...

Good to know we are making progress. albeit slowly!

I guess major bottleneck will be people 's mindset than technology.