Thoughts on Power Utilities and Their Planning Process


I’ve not come to bury Ceaser, but to praise him. Somebody famous wrote that a while ago. Was it Ben Franklin? Kevin Bacon, some guy who lived in Stratford upon Avon?

Dominion Resources made the news this week by ranking in the bottom 3 in several measures of renewable energy support including amount sold, amount generated, and support for various things like clean power certificates, etc. Meanwhile, other utilities, for example National Grid, UK owned New England utility operator, ranked near the top in these same measures. Utility performance correlated directly with the utility service territory’s energy policies. Utilities in service territories having strong renewable energy policies ranked well while utilities in the Southeast, service territories with slack regulatory goals, fared poorly on these measures.

What I’d like to do in this paper is to explain the utility business a bit and some of the realities that constrain it. This is in no way meant to give Dominion a pass but to help you, dear readers, understand why nothing happens quickly in the business of electric power.

Renewable energy is fast becoming low cost of capital generation, is zero fuel cost generation, and has a definite role in a utility’s fleet subject to the limitations of its variable and relatively uncontrolled nature. Dominion is short sighted in its aversion to renewable sources. These sources have the advantage of being added in small increments of capacity but are subject to interruption by weather and their limited ability to be dispatched.


  1. retrieved 6/29.

Power Utility Business Model

Electric power utilities are in the business of turing capital and income into power and more income. Utilities retain some of the profits for reinvestment in the capital plant and distribute some to widows, orphans, and pensioners as dividends.

The most valuable thing the utility has is a well oiled process for making this happen. The second most important thing it has is all that stuff that turns money into power at your home or business.

That stuff includes engineers who plan for the future, design, purchase, and operate the plant, and manage to bill you and collect your payments and pay the suppliers and workers.

The Utility Asset Life Cycle

Utilities have a long planning cycle. In this section, I’ll describe the cycle and why it is the way it is.

  • Dominion is alway planning 15 years ahead
  • It takes Dominion 10 years to purchase and commission new assets
  • Dominion expects assets to last 40 years over which it recovers the cost of the asset and earns money from its operation
  • Dominion borrows money to purchase new assets
  • Dominion cannot include the capital costs of the new asset in its rate calculations until the unit enters commercial operation.

These times deeply permeate Dominion’s management DNA. Dominion is very risk adverse and “do what we know” oriented. Dominion wants to own and control its generation and transmission and doesn’t want to be bothered by the pesky clouds, calm days, and too windy days that interfere with the dispatch of renewable generation.

In fact, engineering models of power system behavior treat renewable generation sited in the retail distribution as negative loads that magically produce power rather than as managed generation that participates in the utility’s control strategy. This is largely a result of the lack of energy storage in renewable devices.

  • Even partial clouding of an array affects output of all of the series connected panels in the array.
  • Wind guesting causes continually changing power output that power electronics can only partially correct.
  • When there is too much wind, wind turbines are feathered to protect the machine from excessive torques.

Addition of batteries to the turbine power electronics can help here. GE has a pedestal storage package that stores sufficient energy to allow the turbines using it to participate in frequency control. Solar installations can benefit from similar addition of a small amount of storage that maintains output as shading fluctuates.

It is only recently that control protocols and design standards have been developed that allow deployment of more than token amounts of properly designed renewable generation in the retail distribution.

Stranded Assets

Utilities have a bias toward controlling generation and distribution assets because they are expensive, have lengthy acquisition cycles, lengthy lifetimes, and gradual but uncertain cost recovery. Any change in the environment that interrupts this orderly process of acquisition, operation, cost recovery, and return on investment is bad for business. When power from a source becomes more expensive than the going rate, the utility looses money on any power that asset produces. When this condition persists, the asset is uneconomical to operate. It has become a stranded asset.

Many older nuclear generating stations require expensive modernizations to replace control systems, worn plant components, etc. With the decline in power prices made possible by combined cycle natural gas plants and plentiful natural gas, the cost of wholesale power has dropped below the level that permits the nuclear station operator to recover costs and make a profit. This phenomena appears to be long term and is forcing retirement of nuclear generating stations commissioned in the ’60s and ’70s.

The Utility Day

A typical utility day begins with low over night power demand. As families and businesses wake up, demand for power increases and peaks toward the end of the work day while businesses are still active but family members are returning home and upping power consumption for heating, cooling, and cooking. As businesses close and homes settle for the night, demand returns to the base level. The over night load is called “base load” because it represents the amount of power that the utility can expect to supply day in, day out, come blizzards, ice storms, hurricanes and thunder.

But electricity has to be produced at the rate it is consumed. There is currently no economical way to make inventory for sale at busy times as is possible in other industries. The utility must produce or import the amount needed right now. No more or no less.

Automatic controls on the generators maintain generation in balance with load. When load increases, the controls of a load-dispatched generator increase the generator’s output. The remaining generators are said to be base loaded and respond only to upsets.

So as the day goes on, the utility is increasing output, bringing generators on line, reducing output, and idling generators to follow the load. The utility operates its lowest cost generators continuously and brings on its most expensive units last and secures them first. This is called economic dispatch. When power is needed, the more costly assets operate. When power is not needed, costly assets snooze.


In the beginnings of Thomas Edison’s DC power, the vision was that each industrial plant would make electricity locally. DC electricity could be transmitted over short distances of a mile or so because it had to be made at the voltage needed by the loads served. And each load type, lighting, small motors, big motors, needed its own voltage to keep current to reasonable values and reduce transmission losses to resistance in the wires. It quickly became apparent that this design would not permit the economies of scale needed.

Nicola Tesla and Westinghouse had a different idea. Make alternating current and use transformers to change the voltage around from a high voltage that allowed low current transmission of bulk power to the voltages that various kinds of customer equipment needed. AC distribution, multiple generators, and organization of the electric system into generation, transmission, and retail distribution tiers quickly developed and became the technical norm.

Stability and Security

Connecting a lot of generators in parallel requires that the load change in an orderly manner and that the generator voltage control and speed control behave in compatible orderly ways. Otherwise, voltage and frequency variations would make utility equipment and customer equipment unhappy. In severe cases, current fluctuations or power fluctuations could cause a circuit breaker to trip resulting in loss of generation or load.

When production and consumption become badly mismatched by failure of a generator, loss of a transmission line or customer transformer, etc minor governor adjustments can’t correct the imbalance and more drastic action occurs. Unit loss of load protection trips the unit to protect it. Reverse power flow from the network into a generator causes it to trip. Transmission lines trip when current flow through them exceeds design limits.

If the amount of load or generation lost is large enough, the sudden imbalance could provoke further instabilities, the disturbance would grow, and the system would collapse. This first happened in a dramatic way in November 1965 with the great east coast blackout.

Utility systems from Washington DC north into New York and New England collapsed. The disturbance began on the US-Canada border. Engineers had changed trip relay set points for four circuits carrying power from Canada into the US. Three of the set points were raised correctly. One was left at its former value. As load grew, system operators allowed the load flow across the boarder to increase toward the new value. The relay that had not been re-calibrated tripped its circuit. The remaining 3 circuits tried to carry the load flow meant for 4 and tripped.

When this happened, Canada had too little load, units oversped and tripped. South of the boarder, there was too little supply so other circuits tried to pick up the slack, overloading them. The protection sensed the  overload and tripped the circuit as part of  a cascading failure sequence lasting 2o minutes or so. The NBC evening news originated from Washington DC with no Huntley, no Brinkley , and the big story was the system collapse of the east coast power grid.

So one of the responsibilities of the utility planning process is to design the system to be stable. In a stable design and operating mode, a single failure leads to a new steady state in which customers affected by the failed component loose power but the system ad a whole returns to a new equilibrium that serves the remaining customers. Any transients are short lived and die out. Such a system satisfies stability and security criteria. If the failure cascades like falling dominoes, the system is unstable and complete collapse can occur.

Planning Process Goals

Utilities, like Dominion, operate sophisticated planning processes that guide the utility’s equipment purchases, station dispatch order, fuel orders, maintenance outage plans, and rate filings. Each year the utility planning process does the following technical activities

  • Estimate the amount and location of load growth
  • Estimate future energy costs
  • Estimate future regulatory requirements
  • Evaluate system structure to ensure that demand can be served
  • Evaluate system stability and security
  • Evaluate the environmental burden of operation, usually as a monetary cost

The utility uses model based design processes to do all of theses things using computer models. The utility models economic growth, population growth, load growth and distribution, future energy prices, and equipment operating history to determine if it has adequate generation, that the generation is economical to operate, and the needed power can be produced within emissions caps, and that power can be delivered to all of the customers in its service territory.

Planning System Products

When the system fails to satisfy operational, economic, or regulatory constrains, the engineers try alternate system structures to find one that can satisfy the utility’s obligations and goals. They propose a new asset, include it in the model, and evaluate the modified model until they find a workable objective system. The utility translates the model study results into a number of products.

  • Rate filings
  • New generating station plans
  • Existing generating station retirements
  • New  transmission lines plans

Proposing New Assets

The utility planning process proposes new asset development  to provide  continuing service to the electric power market believed to exist at the target date. As with any exercise in futurism, these studies are only as good as the assumptions made. Also, specific proposals make assumptions about

  • where facilities may be sited and the right of ways that may be available for transmission lines.
  • A generating station site must be geologically suitable, must provide an adequate ultimate heat sink, must be accessible from transportation networks to permit installation of equipment and delivery of fuel.
  • A viable right of way from the generating station to a transmission substation must exist to allow the power produced to be brought to market.
  • The site must have adequate retail power distribution available to supply the power necessary to start the station.






Individual Income in the US

This article came about as a result of watching the Democratic Presidential Debate last Thursday. Senator Sanders and others rail about income inequality in the United States. Their chief complaint is that the rich are sucking up all the money in the country while middle class incomes are stagnant. Senator Sanders offers this problem, probably without any understanding of how income is actually distributed or what it costs to live in the US. In this post, I’ll look briefly at these two topics.

To this point, The Moocher is unaware of any specific proposal to curtail wealth accumulation by the upper extreme of the income distribution.

Revision History

I revised this article to describe the relationship between income density and income distribution and to attribute and describe the social classes used later in the article. I also added the statement that Senator Saunders had not proposed a way to address income inequality.

The Key Points to Take Away

The basic thesis of this article is that the real problem is not that there are statistical tranches of income or that increasing income and inheritance taxes on those earning more than $1,000,000/year will not, by itself, change the fundamental structure of the economy. The problem is that those at the bottom of the economy, about the bottom half of individual incomes are just getting by or struggling. In this structure, the cost of a basic standard of living is out of reach of over 25% of our population, our underclass and our working poor.

If you use the numbers at the end of this post to estimate an income for a 2 adult, 2 child, stay at home mom household, you get a value of $70,000 for a secure income with recommended savings rates. Reference [1] shows that the 2012 median household income was $62,000 so less than half of our households meet this standard.

Our economy is an emergent phenomenon resulting from literally trillions of decisions occurring each day. Our affordability trap results from the way these choices fit together, it is an emergent property that results from all of the trade flows in our economy, particularly those influencing the cost of housing, health care, transportation, food, and energy.

I think we will be hard-pressed to find one simple lever to move to fix the personal income/cost of living mismatch. These flows are continually changing in response to new means of production, new desirable goods and commodity costs.

Structural changes (increasing manufacturing automation  and clerical task automation) are reducing the number of jobs that yield a working class and middle class incomes . This is largely responsible for the erosion in middle class jobs.

Big C Communism’s attempts to have a centrally planned economy, i.e. to control these emergent phenomena, were a miserable failure. Communism created artificial shortages and excesses as a result of the inability to predict the proper flows of goods and services while attempting to control them. I don’t expect that the US Congress will do any better. The problem is just too complex.


This article wouldn’t be pseudo-scholarly without references so here are a couple.


A bit about statistics

Statistics offer one way to make sense of your income. Where are you ranked in the total US population of incomes? There’s you, those who make less, and those who make more. Reference 1 shows the cumulative income distributions and has a calculator that will figure out your place in this spread.

Their figure 1 shows the total US income summed up from zero to $1,000,000. Your doctors and dentists are in the $200,000 to $400,000 range so this figure includes them but it excludes entertainers, paid athletes, and corporate and hedge fund executives. But it does catch about 96% of the country’s population. Those lofty folk not included account for 4% of the population.

Where the data comes from

The US Census has several sources of data. First it conducts annual surveys. Second, the Internal Revenue Service makes anonymous income data sets available to other agencies, researchers, actuaries, and such. The data used in Reference 1 is very reliable.

The standard of living results appearing in Reference 2 are also taken from Bureau of Labor Statistics and Census Bureau compilations of results. These too are reliable. The thresholds I cite at the end are from memory and, look as I might, I’ve not been able to find a source for them. But they do square with Reference 2.

Average and Median

The figure is plotted on a semi-logarithmic scale. Researchers and engineers use this plot when one of the variables has a large range and the other has a small range. In this case the Y-axis is percent of total income. The X-axis is the log of individual income. So the vertical axis is “natural” while the horizontal axis is “compressed” to fit on the page.

Those looking at the reference will see that US individual median income is $27,000. I’ll come back to this.

If you look at the plot, you see two horizontal lines that mark the median income and  the average income. The median income is that mythical income that divides the population into two equal parts. Half make less while half make more. The average income is the total income divided by the number of recipients. These two are different with the median lower than the average. This happens when the distribution is skewed with significantly more of the samples being less than the average value than greater than the average.

The lazy-S curve is what statisticians call a cumulative distribution function. The number plotted at each income level is the total percentage of income samples up to that point. In this plot the curve shows the percent of total samples, not the total income. That curve would have a similar shape but would be plotted log-log and is difficult to interpret. To give an example of how the curve works lets consider a couple of ideal cases.

  • Everybody makes $10,000. The plot would show zero up to $9999. At $10,000 it would step to 100%.
  • In reality, US income distribution has the shape shown that reflects a most common individual income near $25,000. It rises sharply to the peak and trails off more slowly with a long tail at between your dentist and your favorite corporate CEO. The cumulative distribution continues to rise because the total number of incomes is continually increasing from the start of the dataset to the end of the dataset.
  • The curve is steep where there are lots of samples and relatively flat where there are few samples, that is at the left and the right ends. The steeper the curve, the more samples in that slice. That gives the S-curve shown. The slope of the curve is the percent of incomes at that level (the value of the slope vs income is the income density curve. The income distribution curve is the integral of the density curve.)


Statisticians often divide groups into subgroups. One of the most common techniques for expressing this division is to talk of percentile levels. Median income is the 50th percentile income, half make more while half make less. 90th percentile income is that amount at which 90% make less and 10% make more. 99th percentile income is that point at which 99% make less and 1% make more.

The funny thing about statistics, is, that no matter what you do, 50th, 90th, and 99th percentile incomes will always exists. That’s simply statistics. Being on one side or the other implies nothing about your character, your hard work, or anything else. That’s just where you are in the 300 million of us.

Class Income Levels

Looking at Dennis Gilbert’s class model in Reference [2], about half of our population are in our working class, working poor, and underclass. From the tables, our working class are about 25% of the population. They have safe housing, adequate diet, and reliable transportation to work but are unable to save significantly. The working poor and the underclass together are about 25% of the population. About 1/8 of us are working poor living in compromised housing, nursing a failing vehicle for transportation, and daily life is a struggle. About 1/8 of us are under class unable to work regularly or unable to work.

The researchers cited in Reference [2] use living conditions, not income level to determine the numbers in the working class, working poor, and underclass. In his model, the underclass are not self-supporting, the working poor have high economic insecurity and risk of poverty. The working class are just scraping by with adequate housing, food, and health care but with limited savings It is coincidence that these levels extend up to the $27,000 median individual income level. The real tragedy is that 1/4 are just getting by, 1/8 are struggling financially and 1/8 are living in serious poverty and 1/4 are unable to save for retirement.

Income Thresholds

A few years ago, I heard an NPR story that gave these income thresholds.

  • $30,000 provides basic subsistence
  • $40,000 allows some saving for retirement and misfortune
  • $10,000 additional is required for each additional dependent sharing a household

The $30,000 provided safe housing, diet, and basic health care and transportation to support travel to and from work. There were no luxuries. The additional $10,000 allows saving for retirement and emergencies. The third amount is the marginal amount needed to add a child or a stay at home adult to the household. It covered the marginal costs associated with the added individual.

A family of 4 with a secure standard of living requires $70,000 in income from one or two wage earners.

The striking fact that stands out in my mind is that half of us earn too little to live independently at a basic standard of living. This is the real problem with income inequality: that 50% of us earn too little to live independently.