Nuclear Renaissance? Probably not…

“Renaissance” is a term frequently used to describe the growth in nuclear power plant construction worldwide.  I recently completed an examination of the nuclear power market and determined that this word (meaning “reborn” in French) has been misapplied. 

No nuclear power plants began construction in the US between 1979 and 2011 – now there are four: 1 in 2011 and three in 2012.  What’s unclear is whether there will be any more.  Fifteen additional plants have been announced, but none have firm construction starts and a few of them have been “indefinitely deferred” or cancelled.  During the 90’s and 2000’s a few plants in Europe began operation, mostly in France.  There are only two currently under construction in France and Finland and both are considerably delayed and over budget. So I would not call what’s happening in Western Europe and the US a “renaissance” by any stretch.  In Russia and Asia, however, it’s not a period of resurging nuclear interest.    Nuclear power plant construction certainly slowed for a while, but by no means did it stop completely.  In fact, a huge growth spurt is happening in South Korea, China, and Russia.  In fact, from 2008 through 2012 China began construction of 30 GW of new plants.  Japan had a number of plants in construction as well, but all have been suspended since Fukushima. 

At present there are 372 GW of nuclear power plants.  They contributed about 12% of the world’s electricity production.  Official forecasts predict an additional 117 GW to be installed by 2020, or an increase of about 23% at a cost of about $344 billion.  Globally, the most likely new nuclear capacity installed by 2020 will be between two thirds and three fourths of the official forecasts and the invested capital about 75% to 85% of official.  In fact, on a probabilistic basis, the official forecasts are higher than the highest metric on the distribution curves. 

Declarations of a renaissance typically originate from organizations whose primary purpose is to promote the technology or are based upon data provided by those organizations.  The source forecasts from the World Nuclear Association (WNA), (NEI) Nuclear Energy Institute, and the International Atomic Energy Agency (IAEA) are the most frequently cited reports.  The US Energy Information Administration (EIA) and the International Energy Agency also produce outlooks however these seem to rely on data similar to others. 

Additionally, many countries publish their own national outlooks and in many cases, these forecasts reflect official energy policies – policies that can be as much about intent as national pride.  These forecasts are part and parcel to the IAEA and WNA forecasts.

Given the sources of these outlooks and forecasts (with the possible exception of the EIA), they represent what appeared to me as extreme optimism and worthy of considerably deeper analysis.  In addition, some of the drivers that impact such a forecasts are rather uncertain themselves, especially global and national economic conditions.  I therefore chose to use three different forecasting methodologies: a compendium of all the official forecasts to establish the “as announced” case using the announced operation dates and capital costs where available; a scenario forecast that incorporated some assumptions about economic futures where the as announced case reflected the highest growth; and a probabilistic analysis that sought to capture the uncertainties in the scenario analysis and provide a likely range of installed capacity and capital investment.

Other findings

In addition to the conclusion that considerably less construction of plants than announced plans is likely, and at a higher unit cost, the analysis showed that:

  • Utilities installing nuclear power plants in the US and Europe are paying a premium for this technology over other electricity generation sources. Assuming the capital costs experienced by the units currently in construction in the US and Europe, the levelized all in cost for nuclear is just about 16 cents per kilowatt-hour. That’s more than any other alternative base loaded generation.  In fact a recent Congressional Budget Office report cited a range of costs for a new coal plant with carbon capture and storage between 9 and 15 cents/kWh.  GTM cites a number of solar power purchase agreements at between 7 and 9 cents.  Since these metrics are quite public, the question is not whether nuclear power is economic, but rather how much of a premium over alternatives are utilities paying to make use of the technology.
  • Outside the US nuclear appears to be marginally economic however official capital cost estimates from those countries are not transparent – the extent of subsidization not included in those estimates cannot be determined.  China just announced that all nuclear power generation should meet a cost target of about 7 cents/kWh – this is not a number that is likely to be achieved in Western countries.
  • Small nuclear plants may be coming, but none will be commercial before 2020 and they are likely to be as expensive as their big brothers. A number of countries and private firms are working on several different concepts for small modular nuclear plants.  These designs integrate all major components in a single encapsulated system that will shut itself down without human intervention under certain circumstances.  They are small relative to the gigawatt plus sized units in construction, but are all in at least the tens if not hundreds of megawatt size.   The hope is that the smaller units might be less capital intensive and easier to construct and license but it is not clear at al they will be any less expensive (on a per unit basis) than their large cousins.  I generated probability distributions for the installed capital costs for some of the front runners and compared them with the big units- any differences were too small to be meaningful.  Over the next two or three years a few demonstration units will begin operation, but this new class of reactors will not be commercial at any time through the period of the report – 2020.
  • Decommissioning expenditure forecasts for nuclear, however substantial now, probably understate what will actually happen.  Plans are being put in place to decommission all the plants in Europe, some at the end of their license lives and some much earlier than that.  Some reactors in the US may receive license extension, but others will proceed to decommissioning.  In Japan the new licensing process is just about complete and it is not certain which of the currently shut down reactors can comply with the new regulations.  Recently a number of plants have begun decommissioning and the actual costs are considerably higher than previously considered.  The range of costs are between $1,000 and $4,000 per kWe, much higher than the originally installed costs of these units.  Long term expenditures for decommissioning are likely to exceed investments in new nuclear installations.
  • Several major issues associated with nuclear power from its inception remain unsolved and unresolved over the last 20 years.

Probabilistic risk analysis proved to be a far better means to assess a market with the inherent uncertainties of this one.

Fuel Cells and 7-Eleven

About 8 years or so ago, some friends and I were trying to get traction with a fuel cell implementation scheme.  We knew of suppliers who had 1 to 5 kW fuel cell power supplies and we knew at least one supplier of a proven hydride canister storage system.  The concept was pretty straightforward: find customers in a town or city that were interested in standby power and serve their hydrogen needs with a central hydride canister fueling station where they could just swap out empties for full cylinders, ala Blue Rhino and other propane suppliers in the US.  We wanted to demonstrate that there were ways to eliminate several perceived market barriers: safety of high pressure hydrogen; high cost hydrogen; infrastructure; and customer acceptance.  For a variety of reasons we were not successful.  A firm here in Taiwan has fully realized this notion recently, although in their case the fuel cells power motor scooters.


The fleet at dedication ceremony

Asia Pacific Fuel Cell Technologies (APFCT) has been running a demonstration program in the city of Kenting, a popular beach resort at the southern end of Taiwan, since last November.  The 80 scooters each use 2 metal hydride canisters- enough to give each of them about 80 km of range (with all of the caveats about maintaining 30 kph and no hills).  Twenty are in use by the county government, with the remainder free for use by anyone who stays at 17 B&B’s. So far the fleet has logged over 200,000 km.  Empty canisters can be swapped at police stations, scooter repair shops, the B&B’s and 7-Elevens.  Taiwan is an ideal market, with the highest concentration of gas powered scooters – over 50 million – and the highest concentration of convenience stores in the world.


A canister exchange station

What’s interesting here is that APFCT has done this entirely from the ground up with their own portfolio of technologies: the fuel cell; its control system; the hydride canister system; and even the scooter.  They even obtained a road worthiness certificate for the scooters from the government and, along the way, promulgated a national fuel cell scooter standard. 


The two canister solution

Once the Kenting demonstration is concluded APFCT is eyeing the potential market on the mainland, where over 50 million electric scooters were produced last year.  They intend to secure relationships with appropriate manufacturers to mass produce the scooters as well as fueling distribution relationships.  They are also looking at small personal vehicles as another potential product line. 



The ‘micro-car” uses the same power system as the scooter

Taiwan’s Nuclear Spring

Politicians and political parties in most countries occasionally take baffling decisions and positions, especially to outsiders. Taiwan’s ruling Kuomintang’s (or Chinese Nationalist Party) nuclear power policy demonstrated their version of this phenomenon this spring. 

The KMT, through Taiwan Power (Taipower, 台電), has presided over the construction and operation of three nuclear power plants.  Chin San began operation in 1978/79 with two GE BWR-4’s; Kuosheng in 1981/83 with two BWR-6 units; and Maanshan in 1984/85 with two Westinghouse PWR’s.  Combined they represent a little over 5 GW in net electrical capacity.  Lungmen units 1 and 2, GE ABWRs, began construction in 1999, are about 96% complete and would add another 2.6 GW of capacity.  The Lungmen plant (usually referred to as the “Fourth Nuclear Plant” in Taiwan) became a major political issue after Fukushima and presented a dilemma for the KMT in the 2012 elections.  The KMT has ruled Taiwan since 1945, except for 2000- 2008 when the opposition party, the Democratic Progressive Party (DPP) had control.  The DPP has been very critical of the KMT’s nuclear energy policies. 

On February 27 the KMT announced that it would conduct a national referendum whether to stop construction of the controversial Fourth Nuclear Power Station.  Ostensibly the referendum will “provide a platform for the public to decide whether to continue to build the fourth nuclear power plant,” according to KMT legislator Lee Ching-hua. The real reason why the KMT would pursue a referendum is pretty transparent: no public referendums in Taiwan have ever passed and if such a referendum is rejected it cannot be proposed again for 8 years, well after the 2016 Presidential elections and when the Fourth Plant would be complete.  The KMT would appear to be co-opting the opposition party’s ability to frame and control the issue.

March 3, the Democratic Republic of North Korea announced that it was suing Taipower for US $10 million for breach of a contract to store nuclear waste in that country.  Waste from the operating plants is currently stored on Orchid Island.

On March 7, the KMT released its draft of the referendum and helpful reasons to support or reject the proposal.  This is the wording as translated in the local press.

Proposed Referendum Wording:

Do you agree that the construction of the Fourth Nuclear Power Plant should be halted and that it not become operational?

Why to vote yes:

Why to vote no:

Operating a nuclear plant is not the safest way to generate energy and carries the risk of causing irreparable consequences.

Generating nuclear power is a relatively clean process in terms of carbon dioxide emissions and helps cut the nation’s greenhouse gas emissions so it can honor the environmental pledges it has made.

Nuclear power is not the cheapest source of energy, considering the cost of disposing of nuclear waste, decommissioning a plant and cleaning up the construction site.

A nuclear-free homeland cannot be achieved in one step. Nuclear power plants are a key element in the nation’s gradual progress toward that goal, as they provide a stable supply of energy that can allow people to change their lifestyles and the government to adjust the industrial structure to set the nation on a path toward becoming nuclear-free and having a near-zero-emission economy.

There are many safety issues have been discovered at the Fourth Nuclear Power Plant during its construction, which are compounded by the fact that Taiwan Power Company is not experienced in integrating components for the plant made by different companies, concerns that the operator has withheld information about safety violations and a general lack of confidence in the government’s regulatory mechanisms.

Terminating the construction of the power plant could lead to power shortages because all renewable energy technologies, such as natural gas, are still undeveloped, extremely expensive and vulnerable to fluctuations in the prices of raw materials in the global market.

Taiwan is frequently hit by earthquakes and typhoons and the power plant is in a vast metropolitan area. If there was a threat of a radiation leak, the government does not have the capability to evacuate the entire area that would be at risk.

Halting the plant’s construction could send prices of electricity soaring, severely impacting the economy and people’s livelihoods, resulting in a decline in GDP, driving industries overseas and raising unemployment.

After the accident at the Fukushima Dai-ichi nuclear power plant in 2011, Japan temporarily shut down all its nuclear power plants and some other countries starting working toward becoming nuclear-free. Taiwan can also adopt such a policy and develop alternative sources of energy.

If the Fourth Nuclear Power Plant cannot be put into operation, it might be necessary to extend operating licenses of the nation’s three operational nuclear plants, which could carry serious risks because of the three plants’ aging reactors.


Imagine such a presentation of pros and cons like these for a US nuclear plant.

Curiously the announcement of the referendum and its wording came on the Thursday prior to the 2nd anniversary of the Fukushima incident, which fell on the following Monday, allowing for a full weekend of protests, which had already been planned but became much more popular after the referendum announcements.  That Saturday, an estimated 200,000 people took to the streets to demand construction of the Fourth Nuclear Plant to be halted. In addition, they sought the early decommissioning of the other three plants.

The New Taipei City Council passes a resolution to halt construction on March 20.  The Fourth Nuclear Plant is within New Taipei City.

On the 21st the Taiwan EPA announced that without the Fourth Nuclear Plant it cannot meet its CO2 reduction goals, and that “reaching the goal of 100-percent renewable energy is not realistic,” according to Environmental Protection Administration (EPA) Minister Stephen Shen.  The Taiwan official goal for renewable energy is 8% by 2025; in 2012 renewables accounted for 1.5% of total generation. 

On March 28 (the 34th anniversary of the Three Mile Island accident) Taipower claims not completing the Fourth Nuclear Plant will harm US relations.  The Ministry of Foreign Affairs strongly rebukes the claim the next day.

KMT legislators have starting coming out against completing the plant, including the Mayor of Taipei.

The China Post recently cited a Taiwan saying used to describe the fickle weather in spring on the island. “A spring day is like a stepmother’s face,” her face changing color many times a day.  Politics in Taiwan seems just as unpredictable as the stepmother this March.  

I suspect the DPP at this point is quite happy they were co-opted.


Avoiding Another Smart Grid False Start

Utilities now looking to automate their distribution grids ought not repeat the mistakes made by some AMI deployments by designing it as one more hardware overlay. Instead, future programs and especially communications should be designed within the context of a network using state-of-the-art information and data management technologies. That’s one of the key messages in the recent GTM Research report “Distribution Automation Communication Networks: Strategies and Market Outlook, 2012-2016.” In fact, the report finds that: ”Implementing and obtaining the benefits of DA programs requires access to new communications networks that do not now exist within most distribution grids. In addition, the design, engineering, implementation and operation of these systems require intellectual resources and competencies that are usually associated with IT operations, not electric utilities.”

Thus far, we’ve seen a number of situations where AMI systems have been installed by utilities using purpose-built communications systems, systems that are not going to be adequate to support DA and other more sophisticated technologies over the long term.

Last week a discussion on a LinkedIn group had a topic listed as “The Case Against the ‘Smart Grid’.”  What had been posted for comment was a YouTube presentation by Bruce Nordman at Lawrence Berkeley Laboratory. Although three years old, the presentation makes a number of points that build upon the findings in the GTM Research report. One of Bruce’s arguments is that semantics are important because they frame the thinking behind system design. Defining the “smart grid” as encompassing everything from power plants to end-use devices drives thinking to a mix of networking concepts with hardware concepts.

Without a clear separation between the two, such thinking can distort network design and allow ancient control paradigms to flourish. This hardware-centric focus distracts attention from the real grid and limits the understanding of its broad potential. The focus ultimately was on building systems, but its observations were quite prescient when we look at what happened to AMI.

In many AMI systems deployed to date, the meter was regarded as the end-use device, sometimes connected to a home area network. AMI-unique communications were installed to periodically talk to the meter and deployed to meet the requirements of that metering system, most likely using the least cost option. What the GTM report identifies and which Nordman amplifies is the fact that the distribution system constitutes one domain, the home network another domain and the meter the interface between them.  Further upstream in the grid, the distribution domain interfaces with a number of other domains that include substations, transmission, business operations, customer data and the overall enterprise network.  Unfortunately, this longer-term perspective has been the exception rather than the rule in many programs.

It’s not hard to understand why we ended up with meter-centric “smart grid” programs. Meters are easy to describe, customers can see them and billions were paid out to implement meter programs. Presumably that’s one of the reasons why the Department of Energy chose to invest billions in meter programs, rather than gain the larger and more immediate returns from invisible distribution and transmission infrastructure automation investments. Nonetheless, this hardware focus distorted the transformation of the grid to a network and, because of its difficult business case, has made it harder in some regulatory jurisdictions to gain rate recognition of DA programs.

Regardless of whether or not a utility has an AMI network, the opportunity exists to design its next steps within the proper context and with the longer-term view. At present, that’s an IP-based networking system that connects the application and physical layers in a distributed, universally interoperable network. The GTM report noted as much in its recommendation that utilities adopt the OSI Layer model in network design. Key to the flow among layers and interoperability is the common layer: Internet Protocol, as shown in the figure below.

FIGURE: DA Communication Article Figure

Source: GTM Research

As the report notes: “ …Internet Protocol (IP) networking frameworks are becoming the baseline for smart grid communication networks and are likely to be the only realistic path to achieving interoperability within the system.”

Note: This is an article I submitted to and was published by Greentech Media:

India’s Outages: What Can We Learn?

India’s recent national outages on July 30 and 31 have received a great deal of attention in the press.  Nearly 800 million people were without power and India suffered hits to its economy and its global reputation as a result.  Years from now some definitive report will outline the details, but enough is known now to glean some important considerations for energy policy makers.    While some may quickly dismiss the event as endemic to uniquely Indian conditions, the event highlights a number of important considerations for grid operators, engineers and policy makers in other countries.  Here’s a very quick summary of India’s grid status and what we know about the events that ultimately tripped much of the national grid.

Regional Electricity Supply and Demand Imbalances

India has five regional grids: Northern, Western, Southern, Eastern and North-Eastern.  The Northern, Eastern and North-Eastern grids were affected by the outages. 


Source: BBC and Power Grid Corporation of India

India has an abundance of generation in the Eastern and North-Eastern grids, primarily coal, but the other three grids are in deficit at peak periods, regardless of season.





Source: Power Grid Corporation of India

It doesn’t take a transmission planner to see a looming problem here.  While the summer is theoretically in net surplus, potentially significant swings in demand in the Northern or Southern grids could quickly upset the balance. The national grid company, Power Grid Corporation of India, Ltd. (PGCIL) has been constructing a number of HDVC and HVAC lines to link the regions. Most transmission links are AC; at present there are 3 HVDC lines in operation, with three more under construction. 


Regulatory and Operational Framework

Regional Load Despatch Centres (RLDCs) operate the regional grids.  The RLDCs are theoretically under the control of the National Load Despatch Centre (NLDC).  The NLDC and the RLDCs are wholly owned subsidiaries of PGCIL. 

Like the US, India has a patchwork quilt of regulatory jurisdictions.  Each of the 28 states has its own regulatory body run by elected officials, the State Electricity Regulatory Commissions (SERC).  Each SERC is autonomous.  Federal regulation is provided by the Central Electricity Regulatory Commission (CERC), however all real control lies with the SERCs.   

There is far more political maneuvering within India’s regulatory bodies than in the US, however.  Rate increases only occur immediately after elections, for example, and while the US is not immune to parochial decision making, it is rife in India.

What We Know So Far

In both disturbances, heavy power flow on the Bina -Agra line exceeded limits.  This particular line connects the Northern and Western grids through two 765 kV circuits that have been operating at 400 kV and is one of four major corridors between the two regions. 



At the time of the outage, one of the circuits was being upgraded to 765 kV and was out of service.  The operating line had a Surge Impedance Loading (SIL) of 691 MW but was operating above 1,000 MW.  Apparently several circuits in the Eastern grid were operating above SIL as well.   

In addition, frequency regulation deteriorated to 47.69 Hz in the 50 Hz system, reportedly because of a refusal on the part of some states to install frequency regulation which would have initiated automated demand reductions.  As in the case of the SIL standards, states continued to draw on generation at system frequencies that were below regulatory minimums.

Although some reading between the lines is necessary from subsequent corrective orders, the NLDC in a memo to the regions noted that, in addition to the protective technology “mis-operations” the outage was exacerbated by the fact that there was sustained high loading during a period of high ambient temperatures and the absence of dynamic reactive power compensation resulted in voltage dips in the system. The NLDC’s primary corrective action was to demand that operational protocols and limits be enforced by the RLDC’s throughout the system.

The Indian press has speculated that many of the seemingly technical problems were in fact man made.  In India, access to reliable electricity directly correlates to economic growth.  Those regions that have reliable power are prospering; the have-nots are not.  Electricity is becoming seen by some as a right.  If the electricity is less reliable, or unavailable, it’s the politicians that are blamed.  The Indian press alleges that before this event, controls and technology such as under frequency relays were not enforced or installed because of political intervention; during the outage some politicians in the states of Uttar Pradesh, Haryana, Punjab and Rajasthan demanded that no power reduction measures be taken so that their state would continue to be served.

The inevitable cascade then occurred.  Twice.  With one exception:  the Southern grid is not synchronized with the others and is linked via an HVDC line, providing that grid with a firewall against cascades in the other systems.

We also know that the system had a warning the day before.  On July 29 the Bina – Agra line had a “near miss” due to the same set of circumstances. 

Strategic Implications

Clearly there are several very obvious lessons learned here: never take a major line out of service during a peak power period; and protective technologies; operating limits are worthless if they are not engaged or ignored; and HVDC interconnections of asynchronous grids do, indeed, isolate grids from cascading disturbances.  More important, however, are the longer term implications.

Transmission investment and demand growth mismatch.  This issue is perhaps best highlighted with a comparison of India and China.  The two countries share some similarities: huge populations; high economic growth; and growing electricity demand.   India and China are in the process of building High Voltage DC lines to both link regional grids and to import large amounts of power that happen to be located in regions of low demand to the high demand areas.  But that’s where the similarities end.  See below.




Percent of population without access to electricity (WEO2011)



Population without access to electricity (WEO2011)

289 million

8 million

Forecasted growth in electricity demand, 2012 – 2020 (quads: IEO2011)



Forecasted annual growth in electricity demand, 2012 – 2020 (IEO2011)



Km of HVDC lines planned through 2020



Planned investment in HVDC lines through 2020

$ 9.4 billion

$ 84.2 billion

India’s planned investments in transmission do not seem to keep pace with its demand. 

Over reliance on central generation.  About 55% of Indian generation comes from coal fired thermal plants.  At present 87 GW of coal units are under construction and another 380 GW are in some stage of planning and permitting.

Status (through 2020)

# of Plants

Capacity, GW




Early Development



Advanced Development









Commission since 1/1/2010















It is not at all clear that an already fragile transmission network with limited forecasted investment can accommodate all this new generation.  In addition, the outage occurred because of overloading at a peak period.  And of course, base loaded, multi-GW sized coal plants cannot solve the peaking problem inherent in the Indian system.

Perhaps the more important consideration is whether or not complete reliance on the central generation paradigm continues to make sense.  Perhaps some portion of this massive investment in large scale generation plants might be better directed to increasing transmission capacity, especially using multi-link HVDC, placing a much higher reliance on distributed generation; and optimizing the network as it currently exists.

Many of the most prosperous economic zones in locations like Mumbai are prosperous simply because they have installed their own generation sources, shielding them from the poor reliability of the public network.  Distributed generation can add considerably to balancing load disparities across the various grids and act as an effective supply (when aggregated) during peak periods.

In addition, India has considerable solar and wind resources that are not being utilized in the Western and Southern regions.

Absence of a true national grid.  Indian economic planning, including its energy infrastructure planning, is driven by a 5 year planning process.  The 11th 5 Year Plan ends in 2012; the 12th 5 Year Plan is in development.   The ability of state entities to override RDC’s and the NRDC effectively renders India’s “national” grid impotent.  As the 12th Plan is completed, investments in infrastructure need to be coupled with strong means to implement national policy and national control. 

VSC vs. CSC HVDC technology.  All of the currently planned HVDC lines in India are expected to use current sourced converters (CSC).  CSC systems are tried and true, and have the ability to cope with the very large capacities intended by Indian system planners.  CSC systems do not offer several advantages of voltage source converters (VSC), which include independent continuous control of active and reactive power, dynamic voltage control, and multiple stations.  VSC technology is currently limited to about 1,200 MW at 500 kV, whereas CSC can reach 16,000 MW at 800 kV, however VSC is making rapid advances.  India’s currently planned lines may need to be CSC technology, however new lines might be best suited for VSC technology.

Smart Grid technology. Putting aside the jurisdictional control issues, clearly the Indian system has no ability to visualize the state of its system on a real time basis and does not rely on any automated protective systems.  Any number of technologies could have made a difference: dynamic loading; synchrophasors; Flexible AC Transmission (FACT) systems; etc.

Final Thoughts

This was a rather breezy explanation of what is known about the Indian outages.  It would be easy to dismiss the events as par for the course in India and ignore some of the implications for other national systems.  The US is by no means immune from the potentiality of becoming something that looks like the Indian system.  Certainly the US networks are not as fragile as India’s, but we suffer from the same:

  • Inability to enact national policies, especially as they relate to smart grid technologies and distributed generation;
  • Mismatch in transmission investments and electricity demand growth;
  • Patchwork quilt of regulatory policies that prevents the implementation of a national grid;
  • Insufficient HVDC regional interconnections;
  • National legislation overly influenced by fossil fuel providers; and
  • Over reliance on the central generation paradigm.

Difficult as it may seem, these are real issues for which there are no easy solutions.  We may not be as fortunate as the Indians to have a wake up call like having half our population in the dark for two days. 

Taiwan Power: Quietly Getting the Smart Grid Right?

Taiwan Power – Quietly Getting the Smart Grid Right?

Imagine a utility that is vertically integrated and operates the entire grid through which it provides service.  It has various components of substation automation in place, as well as a Fault Isolation and Restoration (FLISR) system.  It faces the same challenges that many other utilities face as it develops its strategic plans for the next 5 to 10 years: increasing and substantial interconnection of renewable generation; strong pressure on the part of its regulator to minimize and defer new capital investment in generation and transmission; and a need to plan and implement its own version of an intelligent grid operation and management system, including smart meters.  Unlike other utilities, though, it is just beginning its smart grid planning now and has the opportunityof designing its smart grid program components and communications network from high voltage systems down to individual customers and can do so with the benefit of lessons learned around the world.  The utility is Taiwan Power Corporation (TPC) – and its story offers an interesting and instructive case study of how smart grid can be successfully implemented.

A little background: the TPC system operates the generation, transmission and distribution of electricity inTaiwan.  Peak load (summer) is not quite 34 GW; total installed capacity is nearly 41 GW; and annual sales at about 208 Billion kWh to about 12.7 million customers.  InUSterms, Taipower is the rough equivalent of PG&E, SCE, SDG&E, LADWP and SMUD combined.  Its generation mix is similar to theUSas a whole: 40 % coal; 19% nuclear; and 28% LNG.  TPC’s System Average Interruption Duration Index (SAIDI) and System Average Interruption Frequency Index (SAIFI) are 18.224 minutes/customer-year and 0.204 frequency per customer-year, respectively. These data indicate a system reliability much higher than theUSaverages (244 for SAIDI and 1.49 for SAIFI) and placeTaiwanin the top 5 most reliable national systems. Taipower system wide line losses are about 4.6%; in theUSthe average is about 7 %.  This despite the fact thatTaiwanis subject to earthquakes and typhoons (nearly 40 % of all their feeders are underground).

Taiwanenergy policy, regulation and rate regulatory matters are developed and administered through the Bureau of Energy (BOE), part of the Ministry of Economic Affairs.  How these policies are implemented is ultimately a matter of negotiation between the BOE and TPC.  BOE, with TPC input,  just promulgated a 20 year US $ 4 billion smart grid investment program. The program’s objectives are not much different from most other programs: 1) ensure continued high reliability; 2) encourage conservation and emissions reduction; enhance the use of green energy by improving interconnection capacity to 30% by 2030; and 4) develop low carbon smart grid industry that ultimately generates US $30 billion in value.  The following table lays out the phased goals for the program:



Phase I
2012 – 2015

Phase II
2016 – 2020

Phase III
2021 – 2030






System line loss targets (%)





Smart substations




Distribution Automation System





Renewable energy integrated
(% system capacity)

< 10 %




AMI (meters)

1,200 High Voltage (HV)

2,300 HV;
1 million Low Voltage (LV)

6 millionLV

National deployment

Emissions  reductions
(millions MT/year)




Revenue from newly developed Industry
(USD millions)






About $2.74 of the $ 4 billion investment is targeted for AMI; with about $ 800 k devoted to distribution automation and smart substations.  The remainder will be used for emissions reduction programs and for economic development.  Prior to this program, TPC had already invested considerable amounts in its distribution automation systems.  Seventy percent of its distribution system is already automated.  Its fault isolation and restoration system is well established throughout the island.  Some preliminary testing of meters has already occurred.   

TPC has just begun to implement Phase 1 for all but the last goal, which is not its responsibility.  As in theUS, virtually all the public attention and much of the investment is focused on smart meters.  Of most interest, however, is how TPC s approaching the communications network necessary for the program and the distribution level activities that are planned.

Phase 1 is really an intensive period of technology verification testing, the results of which will guide detailed planning for the future.  Taiwan includes a number of islands in addition to the main island- one of those, Peng-Hu (澎湖), will be the test bed for smart grid testing.  Peng-Hu already has a small and large scale wind and solar generation to supplement its diesel generators.  TPC will be installing a total of 30,000 meters on Peng Hu, along with a few smart substations and a demand response program.  During the later stage of the phase electric vehicle charging stations will be installed. 

TPC will be testing both PLC and RF mesh systems in the Peng-Hu trial.  They are out for bids on the initial meters, one requirement of which is the ability to upgrade firmware to accommodate future technology enhancements or changes in communications networks.  PLC is probably the more likely near term choice for the pilot and the initial AMI, as they have fewer concerns about outage disruption of the distribution automation system than other utilities. Longer term, and with its anticipated future growth, RF mesh or other radio options are more probable.  TPC has the ability to obtain licensed frequency spectrum from the National Communications Commission (NCC).

TPC intends to maximize the use of its extensive fiber optic network to support much of the substation and distribution automation program.

Peng-Hu and later phases will also be used to determine the best method to integrate their AMI, demand response, and existing SCADA with the distribution automation control system through their Common Information Model.

TPC, in some respects, is a much larger version of someUSutilities and faces many of the same issues regarding capital investment, cost reduction, efficiency improvements, enhanced customer engagement and integration of distributed energy resources.  Unlike US utilities, it benefits from a fully vertically integrated business structure and answers to only one regulator in a country with a national energy policy. Regardless TPC offers a few interesting lessons for other utilities:

  • Craft a long term vision and work tactical planning accordingly.
  • Define the implementation of intelligent grid management and automation in networking terms and utilize the OSI model to guide the network architecture.
  • Give transmission and distribution grid improvements higher priority than metering as the plan is rolled out.
  • Build in flexibility for both customer growth and step function improvements in technology.

TPC appears to be well positioned to further modernize a grid system that is already far more reliable than many systems.  Careful examination of TPC’s approach, as well as monitoring their future decisions and results could be quite valuable to other utilities and vendors, regardless of where they might be in their own smart grid program or product development plans.

(Also posted at Greentech Media:





Biofuels Retrospective: What Lessons Apply to Future Due Diligence?

And now for something completely different…  Last week I gave a presentation at Renewable Energy World 2012 as a member of a panel entitled “Biomass Due Diligence: How Your Project Can Learn from Failure to Better Ensure Success.”  I provided the following retrospective on the biofuels experience and a few recommendations on improving due diligence.

The Boom

In mid-2006 87 biodiesel plants were operating (33 since the beginning of the year) and 77 were under construction. Similar activity was happening in ethanol.  E85 was selling at 40 cents a gallon less than regular unleaded, and B20 was selling at par with diesel.  Interest in new plant construction (and investment) was so high that plant process providers and E&C firms had over a year backlog – some 2 years.

The reason for this boom period was not difficult to determine: profits in biofuels, especially ethanol, made these new plants seem like money machines.  Margins for both are shown below:

A year later, ethanol margins dropped, in part because by the end of the year there was more ethanol produced than there was distribution capacity to handle it.

The Bubble Bursts

Early in 2007, the relationship between feedstock costs and sale prices for biodiesel flipped to upside down even with federal subsidies.

By the end of 2009:

  • 47 Biodiesel plants and 37 ethanol plants were idled
  • Subsidies expired
  • Soy based biodiesel, except in states with additional subsides or for export (“splash and dash” phenomenon) never returned to profitability since 2007
  • Plants were being sold for cents on the dollar

So what happened?

Before we look at what happened, a basic understanding of cost/revenue relationship for this business is important.

The basic business model for biodiesel is rather straightforward:

Cost of feedstock + cost of processing chemicals (methanol) + O&M + Fixed Costs

must be greater than or equal to:

Biodiesel sales price + blender credit rebate + glycerin sales price

In February 2012 this equation yields the following results:

  • Costs:  ($4.00/gal soybean oil+ $.31/gal O&M + $.50/gal debt) = $4.81/gal
  • Revenues: ($3.14/gal B100 + 0 rebate + $0.05/gal glycerin) = $3.19/gal
  • Net Loss/Gain: ($1.62/gallon)

Note that 83% of the Cost of Goods Sold and 100% of the revenue are subject to both commodity volatility and government subsidy.   The only real control most producers had was over operating costs and some degree of risk mitigation through hedging.

The boom was created by federal subsidies.  For all practical purposes a biodiesel “industry” did not exist prior to 2005. In 2004 the Energy Policy Act was passed containing numerous biofuel tax credits and grants.  A partial list of federal supports include:

  • Biodiesel
    • Biodiesel Excise Tax Credit
      • $ 1/gal per gallon blended
      • Began in 2004: lapsed all of 2010, revived in 2011, ended 12/31/11
  • Small Producer Tax Credit
    • $ 0.10/gal, ended 12/31/11
  • B100 Income Tax Deduction
  • $ 1/gal to dispenser of pure biodiesel to vehicles
  • Ethanol
    • Excise Tax credit
      • .45/gal per blended gallon, ended 12/31/11
  • Small Ethanol Producer Tax Credit
    • $0.10/gal, ended 12/31/11
    • Biofuels
      • Improved Energy Technology Loan Guarantee
      • Advanced Biofuel Production Grants and Loan Guarantees
      • Advanced Biofuel Production Payments.
      • Ethanol Infrastructure Grants and Loan Guarantees
      • Value Added Producer Grants
      • Commodity Credit Corporation Production Incentives

These do not include state programs, some of which were more lucrative than federal.

The impact of subsidies is readily apparent in the following charts.

Agricultural Unintended Consequences

Once various agricultural commodities were impacted by subsidies, both in the crops themselves and from anticipated biofuel related demand, a number of “unintended consequences’ occurred.  When farmers saw these kind of returns in corn, they shifted from soy beans to corn in the 2007 planting season.

The reduction in acreage devoted to soy beans, the anticipated explosive demand from biodiesel and an increase in exports to China resulted in soy oil prices shooting up, peaking mid-year 2008.  And by the way, so much glycerin was being produced that its market became so depressed that it began to be regarded as a waste with disposal costs, rather than a revenue stream.

Referring back to the cost/revenue equation for biodiesel: when soy oil prices grew much faster than heating oil (the futures commodity used to hedge diesel fuel), once costs exceeded revenues, the differential just got worse.

 As can be seen below, except for a very brief moment in 3Q08, costs have exceeded revenue from 2006 to present.

Ethanol production, however, continues, and even with the loss of subsidies, at pretty much breakeven.  The heady days of huge margins are over, but production plants can continue to operate.

So where do we stand today?

  • Approximate total investment in ethanol plants, $ 9 to $12 billion; biodiesel $2 to $4 billion (since 2005)
  • Ethanol projected to breakeven for foreseeable future
  • With the federal $1/gal blender credit no longer available, the biodiesel industry’s long term health is questionable

Lessons Learned

For future biofuels, and biomass projects, it is strongly recommended that:

Due diligence:

  • Employ risk based due diligence that:
    • Identifies the nature of and quantifies the scale of all risks involved
      • Commodity risk – in biofuels, impact is on both expense and revenue
      • Government policy risk – assume overnight loss of support!
      • Technology risk – might your economics be eclipsed?
      • Clearly assesses all project management competency limitations – do you have the depth to fully appreciate farming, commodity trading, wholesale fuels blending and distribution – and curb your appetite to what you know!
      • Never rely on commodity cost projections out beyond a year or the available hedging horizon
      • Examine and test the details behind the metrics (e.g., gallons/acre)

Feedstock Control

In addition, if the project is buying feedstock or selling product, hire a strong hedging consultant and amass sufficient working capital to maximize hedging on the expense and sales sides.

If at all possible, disconnect from commodity feedstock markets using strategies others are already employing in the market:

  • Consider new feedstocks: micro & macroalgae, jatropha, carinata, cellulosic, switchgrass, etc.
  • Unlock value in wood residues; bagasse, MSW; waste gases
  • Add value to corn starch; cane syrup


And finally, good advice from Vinod Khosla:

“Subsidies bring cash flow forward but seldom create your market or build your business. In order to succeed, your product must be price competitive without subsidies.”

From Hill Street Blues:

“And be careful out there…”

Smart Grid 2012 Predictions: Something Missing?

In the previous post 8 different sets of predictions for the smart grid in 2012 were synthesized, and one prediction that the 8 sets seemed to miss was identified: customer engagement services will be a growth opportunity. In fact, it was remarkable how limited was any discussion of customers in the prediction sets. Take a look at the word map below that highlights the frequency of words among all of the 51 predictions and see if you can find “customer.”

This is all the more remarkable in the face of headlines like:

“Smart Grid Backlash: Michigan Opens Smart Meter Investigation”

“Smart Grid Consumer Pushback Spreads to Florida, Returns to Maine”

“Signatures Gathered to Put Smart Meter Issue to Naperville Voters”

And there are plenty more stories about customer concerns and pushback to smart grid. Further, a Zpryme customer survey indicated that 69.9% of customers were not knowledgeable about the Smart Grid and 85.8 % had not received any information about Smart Grid from their utility or didn’t know if they had (“HTU” stands for High Tech User):

The Consumers Electronics Association found very similar results in its survey last year.

So it’s very clear that end users need education, and that without that education greater resistance to smart grid implementation is likely.

And then there is the knee-jerk attitude found among some utility managers and smart grid vendors, best summed up in the following quote:
“… most (residential) customers …have such tepid interest in what smart grid can do for them. They don’t care because they have never cared. They’ve never had to think about their electricity supply, and asking them to engage with their utility via demand response, rooftop solar or time-of-use rates presupposes that they have an interest in power in the first place. They simply do not perceive a need to change.”

They don’t believe customers are interested in participating in better management of their energy, and this belief unfortunately gets reinforced when the utility approach to engaging the customer is to overwhelm them with statistics and meter readings. When properly approached, however, several studies have indicated that customers do, indeed, want to have a direct role in managing their electricity and respond to demand management programs such as time of use pricing. Tendril uses an interesting segmentation of residential customers as they relate to the smart grid:

Another segmentation done by the Smart Grid Consumer Collaborative developed 5 categories:

In summary:

  • Customer engagement does not seem to be high (or at least as high as it should be) on the list of priorities in smart grid implementation.
  • Some utility and vendor managers believe that customers don’t care, and that biases their approach.
  • Failure to make the compelling argument guarantees customer pushback.
  • Properly conducted customer engagement using behavioral techniques generates meaningful results. (Opower has coined the term “Information-based energy efficiency.”)
  • One size does not fit all when it comes to communicating with customers and segmentation is important.

So what do we conclude from all of this? My opinion is that unless careful adoption of customer engagement approaches become a trend for 2012, more customer, and by extension, more regulatory pushback will occur. In addition, since the smart grid industry is plagued by a variety of confusing messages to consumers, and since “smart meters” have become synonymous with “smart grid” because of these confusing messages, problems with smart meter programs will negatively impact other smart grid programs.

Smart Grid 2012- Making Sense of All the Predictions

Over the last few days I collected 8 different sets of predictions for the smart grid market in 2012.  No doubt there are more out there.  These sets incorporate 51 different trends, some similar and some unique.  I’ve identified the various authors at the bottom of this piece.  Distilling these entries into more meaningful groups results in: trends with some consensus; trends common to more than one expert; unique perspectives; and conflicting opinions. And while my population sample may not be statistically significant, none of them picked up on a big one, discussed below.

Trends with some consensus

  1. Participants will begin to identify new benefits from focused data analytic programs and analysis: the sheer volume and potential value will drive significant utility investments in this area. (5)
  2. Distribution Automation investments will dominate. (4)

Trends common to more than one expert

  1. Demand response programs accelerate. (3)
  2. Grid scale battery systems will see significant cost reductions and commercial deployments. (3)
  3. Smart Buildings will see greater investment and utility interest. (3)
  4. Market consolidations, mergers and acquisitions continue. (3)
  5. Attention on cybersecurity and risk will increase by policymakers and market participants. (3)
  6. Prepaid Electricity will emerge as a new service. (2)

Unique perspectives

  • More progress on smart grid standards.
  • Excess renewable generation will become a problem.
  • Demand will flatten or even fall.
  • Federal government cuts back R&D and funding.
  • Constant reinvention of business models will occur as smart grid market matures.
  • Municipals and coops will drive new AMI deployments.
  • Cellular systems are coming on strong.

Conflicted opinions:

  • Electric vehicles will gain traction (2); year to make or break a market (1); backburnered (1).
  • Solar PV growth will moderate (1); solar PV growth will be greater than 25%

A missing trend: Growth of customer engagement services

What is interesting to me is the fact that none of these sets of predictions flagged a very important trend: the emergence of customer engagement services.  There is a pervasive attitude that the average customer does not care about energy management, an attitude that is probably correct if all a customer is given is a lot of statistics.  If meaningful insight is provided, insight that the average customer can understand and use, the picture is quite different.  And some utilities have begun to take notice.

The origins of the trends I paraphrased above were:

Jesse Berst, “Top 9 predictions for 2012”,

Heather Clancy, “3 smart-grid trends to watch in 2012” ZDNet,

Katie Fehrenbacher, “The top 10 trends from the year’s big smart grid show” gigacom,

Chris King, “2012 Smart Grid Predictions” AolEnergy,

Donald Rickey, “Smart Grid Trends to Watch for in 2012” The Daily Energy Report,

Rob Wilhite, “Three Smart Grid predictions for 2012” KEMA Blog,

Electric Light & Power “2012 to bring more smart grid, less government support”

IDC Energy Insights, “North America Utility Industry Top 10 Predictions 2012” webinar December 6, 2011