Efficiency.org - Bringing Energy Efficiency to Market
 
With many initiatives in full swing to put ratings on homes and buildings across the country, there may be a lesson to be learned from recent cases where consumers have prevailed in suing car companies for promising an MPG that is not based on real-world driving.

Before you read any farther, I want to make clear that these models are not the problem, it is actually how they are being applied.  Applying any predictive model to an individual building is the problem.  It is up to us to use these tools as a way to manage this risk in pools, and avoid driving that risk down to homeowners.  Great examples of this can be seen in the success of solar PPA and Leasing models.  

Most labels, including the National RESNET HERS and California HERSII labeling system and the DOE Home Energy Score, are in fact asset scores that, similar to an MPG, score a house based on a set of average users.  These scores have potentially wide variance for any particular building, and have a tendency in many climate zones to over-predict savings.

A recently released study called Modeled vs. Actual Savings for Energy Upgrade California Retrofits, analyzes predicted savings based on the CEC’s required energy modeling software, against actual results from customer bills.  The study shows that California HERSII is over-predicting by and average of 50%, with a huge amount of variance between winner and losers.  Resulting in more than 78% of homeowners not achieving the savings being predicted.
A recent LBNL report on the Home Energy Score called "Accuracy of the Home Energy Saver Energy Calculation Methodology" was announced with an email headline declaring that HES is now within 1% accuracy on average, which is a great result.  Of course on closer examination it is clear that there is wide variance for any individual project - perhaps even wider than other tools in the study.  Which of course goes back to the original issue that homeowners are truly not interested in the average, especially if they get the bad end of the stick.  
This issue is not exclusive to the United States.  The Green Deal in the UK has also suffered from estimated savings that are outpacing reality.  In an April 11, 2012 article in The Telegraph, called “Green measures for homes 'save less on fuel bills than forecast

"An official pilot study of 67 homes in Sutton, South London, found all of those who took the 25-year payback package rather than ten years - a third of the owners - faced repayments higher than the savings.

Another study of 139 council houses in Sunderland found savings on energy bills were just 12 per cent rather than the expected 19 per cent.

Luciana Berger, Labour's climate change spokesman, said the proposals were "complex, confusing and leave customers exposed to mis-selling".

"Before anyone takes out the Green Deal they have a right to know how much it will cost them and how much they will save. Relying on guesswork just isn't good enough. If people are promised savings which never arrive, they will think the Green Deal is a con."


Both the CA and UK rating systems are based on a faulty notion that relative scores are more important than accuracy.  In striving to achieve a relative indicator of performance, and removing behavior (how people actually use their homes) we are left with a system that is really more about policy and theory then what matters to real people - which generally boils down to, how much does it cost, and what will I save.

The idea that we are going to put MPG stickers on every home in CA, or the US, at great expense (It will cost $5B to label home in CA alone) is a mistake.  At least with a real MPG on a car it is based on actually testing the vehicle, and you are not having to test every single vehicle on the road, vs. labeling buildings where we are attempting to derive energy use from physics calculations and every house needs an expensive custom test.  

Here is an article on the topic of MPG from the Huffington Post, Hyundai Fuel Economy Lawsuit Filed Alleging Misleading Ads, that I think should have us all worried:

"For carmakers, the new trendy thing is to have a vehicle in the lineup that gets 40 mpg. One huge problem is that fuel efficiency figures are not based on real-world driving. And automakers opt to advertise with the fuel economy figures that are most impressive -- for highway driving -- rather than lower city or average mileage calculations, which would make their cars look less efficient.

But the Hyundai lawsuit is the second one in recent months to challenge automakers over lofty fuel economy claims. In February, California attorney Heather Peters sued Honda in small claims court over the fuel economy claims for her 2006 Honda Civic hybrid. She said she never got anything close to the 50 mpg she was promised. A judge awarded her $9,867 in the case, which Honda is appealing."


In the end our goal is to save energy and drive consumer adoption.  There is a distinct risk that all of our efforts may backfire when consumers come to understand how imperfect our ratings are, and when the financial community tries to underwrite investments based on energy savings that don’t really exist.

It is time that we start focusing on real data and actual savings, rather than more complicated regulatory schemes.  All this is not to say that energy efficiency does not work, instead it should tell us that we need to move from energy efficiency expressed through code and complicated ratings or scores, and instead focus on turning savings into a resource that can be valued and traded.

We are in a unique moment in time where we can move past regulatory frameworks, to engage markets that can finance our long-term goals, which are simply too expensive to achieve driven primarily with public dollars.  

This article is not to say that modeling is worthless or wrong.  In point of fact, it is far better at predicting savings than just simple one size fits all deemed savings, and can be very good predictors of a large pool of buildings.  However, we need to rapidly move from a model where we have policy anointed solutions based on long regulatory process and instead start measuring and valuing savings predictions based on actual results.  Once we have a system that can measure real savings versus predictions, markets can step in, invest, and manage savings risk so that building owners and households don't have to.

If you look at the Solar industry as a guide, where in CA residential Energy Service Contracts (Leases and PPAs) are currently 75% of the market, building owners on completely insulated from risk through performance guarantees, private capital is flowing, and quality has become a function of industry.  

A system built on real proven savings at the meter will drive innovation and investment, which is the path towards a real and sustainable solution that can achieve the promise of energy efficiency in the build environment.

Enough talk and theory... get the actuarial data and the market will follow!  
 


Comments

08/13/2012 09:52

A recent article made the claim that home energy rating systems such as RESNET’s HERS, California HERSII labeling system and the DOE Home Energy Score were guilty of over-predicting savings that homeowners could enjoy by making their homes more energy efficient.

To enhance the discussion of the accuracy of home energy ratings’ energy use projections I would like to point out a study conducted and published by Advanced Energy on a far larger set of homes. The authors of the study were Michael Blasnik of M. Blasnik & Associates and Shaun Hassel and Benjamin Hannas of Advanced Energy.

The objective of the U.S. Environmental Protection Agency support “Houston Energy Efficiency Study” was to assess the actual energy use of groups of homes built to different energy efficiency specifications in Metropolitan Houston – typical non-program (baseline) homes, ENERGY STAR® homes labeled by a Home Energy Rating and guaranteed performance homes. More than 226,000 homes built from 2002 through 2007 by dozens of different production builders were included in this study. The large dataset also provided the opportunity to analyze how certain construction characteristics are related to actual energy usage.

Data collected for this project included billing data for all new homes built in the CenterPoint utility service territory from 2002 through 2007, information from property assessor databases of four counties, detailed building characteristics for tens of thousands of ENERGY STAR homes from CenterPoint’s ENERGY STAR Homes tracking database, and detailed data files from energy raters including the home energy rating software tool, REM/Rate, input files and building shell and duct leakage test data. The study did not involve any direct data collection in the field but instead relied upon existing data sources. This approach allowed the scope of the study to be much larger in terms of the number of homes analyzed but left some gaps in our understanding of some details, especially of baseline homes. The overall dataset includes hundreds of variables for 226,873 homes, including 114,035 potential baseline homes, 106,197 ENERGY STAR homes and 6,641 guaranteed performance homes.

Although consumption differences across groups of homes are smaller than advertised, ENERGY STAR homes perform very close to the predictions of the models on average, while baseline homes perform better than the reference homes defined by the HERS standard. ENERGY STAR uses a base case reference home defined as minimum local code specifications combined with the least efficient cooling, heating and hot water systems available, a leaky building envelope and a poor duct system. Using this yardstick to measure the performance of the ENERGY STAR houses in the study, they did quite well – showing a strong and fairly consistent relationship between actual and projected performance for both heating and cooling. Therefore the apparent lack of savings is attributable not to underperformance by the ENERGY STAR homes but to the fact that the baseline houses in Houston perform considerably better than the ENERGY STAR reference house.

The relationship between REM/Rate cooling load projections and actual electric usage was examined graphically and statistically for 10,258 homes with sufficient data. REM/Rate projected an average cooling load of 5,506 kWh/yr while the billing analysis estimated average cooling loads at 5,677 kWh/yr, about 3 percent higher – excellent overall agreement. Although the analysis found no systematic bias in the REM/rate cooling projections, there was a large amount of variability in the data. Findings revealed that the correlation was higher between house size and cooling load than between REM/Rate projected cooling load and actual usage. However, the study team feels confident in stating that when using current modeling software with energy-efficient new homes, there is a strong and fairly consistent relationship between actual and projected performance using REM/Rate for both heating and cooling. REM/Rate also estimated the average heating usage of program homes fairly well – only 4 percent lower than the measured loads.

Laurel Elam
RESNET

Reply
tedkidd
08/25/2012 22:16

"You are promising me $700 a year, but you have no idea if your previous promises have delivered?"

The problem of variance may be due to complete lack of accountability. Without pressure to deliver quality in all aspects of delivery; design, modeling, AND implementation, how can we expect anything but wide variance in results?

Because variance exists now doesn't mean we should conclude it is unfix-able. I submit that providing results tracking for contractors will not only allow them to improve their work, it will narrow results variance dramatically.

Currently, savings predictions are a blind guess on the part of programs and contractors. If results are never tracked, how can you have confidence in ability to deliver? Why should the customer?

I don't believe this can't be corrected. But you can't fix failure if you don't realize you've failed.

Under the current approach people are throwing darts at a bulls-eye blindfolded, then walking away from the dart board without finding out what they've hit. They simply assume every dart thrown is a bullseye.

Sounds absurd, but after asking hundreds of contractors I've found very few that have EVER tracked, and NONE that have tracked more than 2 or 3. Everybody is Tiger Woods if you play without a golf card.


Widespread results tracking super aligns interests. Realization tracking could provide huge opportunity for adoption of the comprehensive approach. It would eventually be a viral marketing metric that dramatically grows adoption of comprehensive measures. When homeowners embrace comprehensive, instead of huge likelihood of purchasing an empty promise they'd be able to take their chosen contractors improvement recommendations with their delivery record to the bank.

We need to take the blindfolds off. We need to bring golf cards out on the course. Without measuring it doesn't surprise me at all that variance is huge and over-prediction the norm.

Reply



Leave a Reply


Energy Efficiency, Negawatt