By Robert Bradley Jr. — January 25, 2022
Editor’s note: Part I yesterday described Krebs’s work to level the playing field against the Department of Energy’s Office of Energy Efficiency and Renewable Energy (EERE). Today’s post continues by examining how EERE modeling skews the results towards “deep decarbonization” (electricity over gas, via appliance regulation).
Q. What are some “tricks” by DOE’s Office of Energy Efficiency and Renewable Energy for finding “net benefits” to imposing stringent efficiency standards on gas appliances?
A. EERE’s “determinations” have come from highly subjective life-cycle costing (LCC) analyses. Remember: these modelers are in the business of finding different answers from what self-interested consumers are determining. They make their models “smarter” than the market for mutual benefit.
Tricks include inflated energy-price forecasts (for the disfavored energy, gas), lowballed maintenance costs for favored appliances (electric), and so on. Modeler bias has traditionally been aided and abetted by using exclusive, expensive and/or proprietary software that can be easily gamed by modelling staff to “determine” that forecasted LCC benefits exceeded costs.
A whole new cost/benefit modeling approach is required. James Broughel (Mercatus Center) commented:
The problem with this approach is that from an economic standpoint, it makes much more sense to look for the option with the most benefits in excess of costs, thereby achieving the most bang for consumer and business bucks. Instead, DOE interprets its legal obligations as meaning it should identify the most stringent option with any benefits over costs—no matter how small the difference might be.
Q. The modeling should not be done at all, probably. But how to correct such an exercise?
A. More realistic, transparent and robust variables, which if honestly utilized by EERE, would cancel benefits and produce negatives. Analytical variables and assumptions thereof are supposed to be transparent and routinely tested for sensitivity per the Code of Federal Regulations (CFR) under 10 CFR Appendix A to Subpart C of Part 430. But they are not, making analysis very difficult for anyone wanting to challenge EERE’s regulatory hegemony.
Q. Has gas been victimized?
A. Regarding my involvement, the American Gas Association in 1997 asked the President of my company (Laclede Gas, at the time) if I would represent the gas utility industry on EERE’s newly formed “Advisory Committee on Appliance Energy Efficiency Standards” (ACAES, since disbanded). I agreed and was appointed by DOE Secretary Penà for my first term and reappointed by Secretary Richardson for a second term.
During that time, I endeavored to get EERE to start analyzing consumer energy savings potential as they occur based upon reduced consumption as reflected by utility billing structures. We called it Consumer Marginal Energy Rates (CMER). This approach provided an accurate cross-check to EERE’s overly complicated and opaque life-cycle costing (LCC) methodologies.
We also advised EERE to start using Monte Carlo analyses to implement more robust and transparent “what if” sensitivity analyses. While EERE published a draft report about CMER in 1999, they never really employed it for gas utility bills. Why? Because CMER methodologies to calculate consumer savings were consistently and significantly lower than savings “determined” by EERE’s gamed LCC methodologies which, in turn, enabled EERE to establish more stringent appliance energy efficiency standards.
Q. Are factors not in favor of electric-over-gas just being ignored also?
A. Yes, EERE doesn’t include what it should re cost effectiveness. Such is the case of reality checking LCC “determinations” against simple paybacks. Spires’s review for NAS ver FEB102020 provides a real-world methodology that anyone can follow for calculating their own utility-based energy savings potential from appliance efficiency upgrades, no software or computer scientists necessary.
One thing EERE did implement was the ACAES’s recommendation to use Monte Carlo analyses. Well, kind of anyway. They did so in a manner that weaponized it to defy transparency and game the system to further artificially bolster economic benefits. For example, in the case of commercial boilers, EERE assumed consumers are unconcerned with maximizing investment paybacks.
——————–
Mark Krebs, a mechanical engineer and energy policy analyst, has been involved with energy efficiency design and program evaluation for more than thirty years. He has served as an expert witness in dozens of State energy efficiency proceedings, has been an advisor to DOE, and has submitted scores of Federal energy-efficiency filings. Mark’s first article was published in the Public Utilities Fortnightly and was entitled “It’s a War Out There: A Gas Man Questions Electric Efficiency” (December 1996).
For more about Mark, see his MasterResource archive.
This series began yesterday and continues with Part III tomorrow and Part IV on Thursday.
Thé progressive insanity continues.
Different country, different systems, but from here in Australia your President appears to have far too much power.
Michael. Vk 5 ELL.
As do your state premiers.
There is a credible argument under a strict interpretation of our Constitution (w/r/t Enumerated Powers and the concept Federalism) whether he actually does have the power that he has been exercising.
But as Obama infamously said “I have a phone and a pen.”
The president is the head of the executive branch, and hence has the power to tell the executive branch how to it’s job.
The problem is that the executive branch was never intended to have it’s fingers into nearly every aspect of everyday life.
Cut back the government to what is allowed in the constitution and the problem with Executive Orders goes away.
It doesn’t matter if the Constitution limits his power as long as everyone LETS him exercise that power.
He does, but little of it was actually granted by our constitution.
Michael, you sound like my Social Studies teacher (far too many decades ago), claiming the Prime Minister has only one vote in Parliament and therefore has limited power. But in Practice, the PM is able to control his caucus with the lure of one day being appointed a Minister with its pay increases and perks OR keeping his Ministers in line with the threat of firing them. As long as the ruling party is near the top in the polls, they fall in line with the PM. Anyway, that is how it works in Canada. ie the PM has more power than the President (in Practice)
And worse yet, here in Canada, with a number of political parties, the PM can be a dictator even though only 31% of population voted for his party.
Governments should be required to meet all of the sustainability goals that they promulgate and this should extend to government employees even in their off time. So, for example, if a government employee built up frequent flier miles traveling on government “business” any travel resulting from those “earned” miles would count against the government.
Governments should go to net zero first and foremost.
Not just continues, it’s accelerating.
Engage, ludicrous speed.
It appears that even lunatics can detect lunacy
“Airlines flying near-empty ‘ghost flights’ to retain EU airport slots
Analysis from Greenpeace finds deserted flights are generating millions of tons of harmful emissions.”
https://www.theguardian.com/environment/2022/jan/26/airlines-flying-near-empty-ghost-flights-to-retain-eu-airport-slots
Modelling modelling modelling..
Program into a computer what you think/believe/want to hear and then, with all the unassailable authority that computers have, it recites it all right back at you and the world
Just like the school-playground wimp coming to school every day with Tyson Fury at his side.
You gonna pick a fight? huh. are ya are ya are ya?
Nah. Thought not.
It’s been a long time since I wrote about this but it seems appropriate to do so again:
The first program I wrote was at Hamden High in Hamden Conn in 1968. Our school had invested in a computer for student learning. It was the size of a refrigerator and used the pre-Fortran language Focal. The program was produced on a tape with punch holes in it which was then feed into the computer. Our assignment was to write a program that worked. I decided, the menace that I was at the time, to use it to play a cruel joke on my friend. So I wrote a simple if/then program I called “The Personality Analyzer”. The program called for all kinds of input about a person but the only thing that mattered was the height number. Knowing who I was going to use as an example and knowing their heights, I programed the machine to spit out a glowing description of two of my friends and a horrible description of the one I was playing a joke on. When you input 69 inches for the height you would get the same horrible description regardless of any other input you used. Frank, the object of the joke, went off on the machine declaring “How does it know”. The teacher gave me an A, after all, I did write a program that did exactly what I programed it to do. I have had a very hard time accepting computer output since that time, especially if I suspect less than honorable motives.
Frank would probably get a glowing description today.
Frank has been in therapy since.
Re early Fortran versions:
We had one at Uni called Forgo, which was replaced by Watfor (U of Waterloo Fortran). These worked well, but Watfor was better.
Who got to name these early Fortrans? I proposed BeforGo and OnceUponaTran.
Re the pre-Fortran language Focal – we didn’t like it, and called it SweetFocal. 🙂
This is why I ended up with a 2015 Volvo which shuts the engine off automatically a few seconds after a stop. The owner cannot permanently disable this dangerous “feature” as it takes a few ms to start the engine and bring it up to speed. Sometimes those ms can make the difference between a near miss and a fatal accident. All in the name of CAFE standards.
Also found out that it is not possible to purchase a Volvo in the US with a manual transmission, can’t even buy it in Europe and have it shipped over…
Manual transmissions are too difficult/complicated to operate for most new US drivers anymore.
As a user of computer models for more than a half century, I’ve learned a lot from them (my first models were on vacuum tube analog computers). The first lesson was that they are almost always very wrong. This usually starts with improper programming, but then ultimately progresses to inadequate input parameters. “Inadequate” includes not only not enough of them, but also improper values used for them. No matter how simple the system, the model is never complete. Even the simplest of systems, which can be solved with a closed form equation, are never exact. The foremost rule of modelling is “Predictions are always wrong”, or as Casey Stengel so eloquently put it: “Never make predictions, especially about the future”.
That said, a lot can be learned from modelling. Foremost is a better understanding of the system that is modelled. What is now called ‘tuning’ the model gives insight into which variables are critical, and which are irrelevant, and which combinations of inputs are important. In the business world where I come from, this insight is what usually leads to ultimate product decisions that prompted the need for the modelling in the first place. Rarely is a model an end product, precisely because of the inherent limitations of models.
In the world of climate models, the ‘science’ is clearly unsettled. The proof of that is in the outputs – they’re all different. The second is that there is little system knowledge apparent in the participants. The assumption seems to be that all variables are irrelevant, except for CO2. Yet, very simple equilibrium climate sensitivity models which vary only CO2 are discarded as too simple. It seems to me that many participants have failed to learn from their own models. This leads me to place them in the same category as the Cargo Cults of New Guinea. They’ll never fly.
To confirm some of Mr. Krebs’ insights:
When objectively structured, Monte Carlo modeling is great. Indeed, it yields insights far beyond those achievable with conventional scenario analysis.
But, one can game many ways the inputs’ (aka assumptions) characteristics to achieve a preconceived conclusion. Key characteristics include the “most likely” value for an input (point of central tendency is statistics jargon), range of possible outcomes (confidence interval) and the range’s shape (distribution).
This comment is about electric light bulbs, but it illustrates the same non-sense. When we began replacing, or were forced to replace, incandescent bulbs with compact fluorescents we noted the following: they were expensive, they couldn’t produce quite the same number of lumens for a bulb meant to fit into a recess or lamp, they would degrade more quickly than expected and become dark, they had some difficulty igniting in the cold. In short, they violated every promise made and were a source of mercury contamination to boot.
As we began to transition to LED bulbs I got into the habit of writing the installation date on the bulb neck as I put them in and then writing the date of failure as I removed them. The bulbs are advertised as lasting some amount of time (a MTTF measure), for example 9.1 years if used 3 hours perday — or a bit short of 10,000 hours.
I have no bulbs so far that are coming close to this figure. The results depend on manufacturer, but the WalMart brand bulbs, for which I have the most data, are running at about 1500 hours (assuming around 3 hours per day). There are an inordinate number failing on the early part of the Weibull distribution. Once again, the green product does not meet its promise(s).
In the late 80s I converted regular incandescent bulbs to ones rated at 130V. I often got over ten years of usage. Never understood why they didn’t become the standard.
Thank the Phoebus Cartel. It is a fascinating story. There are a few YouTube videos hanging about regarding this.
Phoebus cartel – Wikipedia
Such as: This Is Why We Can’t Have Nice Things https://youtu.be/j5v8D-alAKE?t=1
I used to buy the little diode-disks that you’d insert on the end of an incandescent bulb that converted it to DC (and halfed the wattage). With those diodes, the bulbs would last for many years of constant use. I doubt you can buy/find them anymore.
There is also a fallacy about incandescent bulb efficiency. When you need more light as in winter, heat is a real but small benefit. Outdoor water pump structures for example often use this backup and lots of 100+ bulbs were hoarded when they were to be illegal. Similar problem in origin with toilets.
Of course that’s watts.
I’ve been buying name brand LED bulbs with a lot of success. I’ve got them in every fixture except the ones that only get turned on infrequently, and in the last 10 years, I’ve only lost one or two.
What brand pray tell? I am also using GE bulbs which are not much better than the WalMart Great Value brand. The GE bulbs fail in an interesting manner. They start blinking. They are made in China and I joked to my wife that these may be coded pleas for help from Uyghur workers.
how about an electric light bulb with over 1 million hours?
They would quickly be out of business due to lack of repeat sales.
To be fair, there’s only one. Over 120 years so far.
I can’t recall where this bulb is except that it is in central California, but they are nursing it along at present and have been for years.
Livermore, in a fire station: http://centennialbulb.org/
The real problem is there should not be a DOE or EERE at all. Most of the USA government should not exist all, and the few things it is supposed to do, border control, law enforcement, it either refuses to do or deliberately performs incompetently. It is silly to complain DOE/EERE is doing the job wrong; it was created to do the job wrong.
i am curious if there are any robust studies on heat pumps/mini splits with tankless water heat vs traditional forced air home heating systems and tanked water heaters? My own experience building a new modern home with the former including LED lighting has my power bill at 1/10th of previous homes I have owned in WA State. If consumption is able to decrease by mandated efficient systems, then emissions reductions follow. Seems like a more logical and cost-effective first step in energy conservation.
Or, companies could be allowed to out compete each other to produce better, less expensive products. Then, consumers could decide what works best for them.
And yet would the system that works for Washington work in Wisconsin or Arizona or Florida? We have huge variation in climate and geography and personal preferences in this country. The government should not be setting rules at all and certainly not one size fits all rules.
The “one size fits all rules” are much easier to understand and apply for the people doing the planning and development.
If they knew what they were doing they’d be running multibillion dollar companies instead.
Your energy use reduction has more to do with the insulation and air-tightness standards for new construction than anything else. The most efficient mix is gas for heating, cooking, and clothes drying and electric for the balance. This is because any operation using heat directly doesn’t waste 60% of the fuel heating value converting to electricity. Heat pumps don’t work well once you hit about 32F. They rely on backup resistive heating elements to keep you warm at that point. In fact, once they hit a COP of less than about 3 they are far less efficient than burning the gas at the house.