* Cost is defined here as the cost of equipment and installation per output of energy.

 Copyright © 2017 by David M. Anderson


Concentrated Solar Power (CSP) was chosen as an example of how these principles could apply to solar power because (a) CSP can store energy most economically as hear, which can then generate power at night and (b) widespread utilization can be scaled up the fastest if following scalability principles.

s..  See the article on Scalability with a special example on scaling up solar power production: http://www.design4manufacturability.com/scalability.htm

What is keeping solar cost high now? (This section is followed by solutions in corresponding bracketed numbers).

A cost reduction expert analyzes why solar power equipment is more expensive than it could be.

1) Old sub-optimal product architectures
 Some CSP builders still use decades old heliostat guidance designs, even from companies no longer in that business, which precludes improvements.  Further, the common axis ordering is sub-optimal with an elevation axis on top of a azimuth (compass angle) axis, which forces compromised mirror shapes that lessens efficiency (as discussed in the example below). Conclusion: suboptimal architecture can miss efficiency opportunities and lock in high cost and, thus,  make it hard to improve the  manufacturability and cost;

2) No Commercialization. Dr. Anderson wrote the article on commercialization after touring the US Government solar test facility at Sandia, NM, and saw mostly what he considers prototypes being tested for function and longevity. In his learned opinion as a manufacturability expert, most of the equipment there needed to be commercialized based on the logic in that linked article, especially the sections on “What Happens Without Commercialization,” which is what he saw there and his warnings in the section, How Not to Do Commercialization, which are briefly summarized next.

3) Not Designed for Manufacturability by following the “usual scenario” of designing only for function, which is the opening section in the new white paper on Concurrent Engineering for Challenging Products, at http://www.design4manufacturability.com/concurrent-engineering.htm  . The article then continues by outlining the myriad of problems encountered.

4) Naive assumptions that cost can be reduced later (see http://www.halfcostproducts.com/how_not_to_lower_cost.htm ) or the naive assumptions that mass production volume will be able to reduce the cost of unmanufacturable designs. In reality, designing for automation has the strictest design guidelines.  A more dangerous corollary of mass production thinking is taking on a big job, assuming that  shear volume alone will automatically get the cost down. And if it doesn’t, which is highly likely, the company may be contractually obligated to a low price bid for high cost hardware

            Actively trying to remove cost later with desperate measures may have counterproductive results – meaning they actually cost more than they are expected to save! This is discussed in the web article: “Seven Reasons Why “Cost Reduction” after Design Doesn’t Work.” at

             Toyota says that late changes will “always degrade both product and process performance” (from , The Toyota Product Development System: Chapter 4: Front-Load the Product Development Process).

5) Thinking offshoring will lower cost, which it will not do because

Offshoring generates many “hidden costs” that are only known – and avoided – if all costs are quantified using total cost. See the whole truth about offshoring at http://www.halfcostproducts.com/offshore_manufacturing.htm

Offshoring will prevent real cost reduction by thwarting concurrent engineering and lean production and supply chain simplification and standardization and designing for quality,

Much effort and calendar time will be consumed transferring products offshore, expatriate travel expenses, and, worse, usually includes converting the parts to “local sources of supply” for quicker shipping (which makes sense) and to try to save more cost, which is another cost fallacy because

Lower product quality will result from the cheap parts made in “low cost” manufacturing areas

Part availability will be in months, not years because most parts in "low-cost" manufacturing are made for short-lived consumer products.

• Parts made for consumer electronics will not be available for the lifespans need for solar power components, which will halt solar component production or induce change orders, and re-testing, to change to more available parts, thus causing change-induced problems cited at the end of point (4)

Changing parts induces variables, whose problems rise exponentially when many parts are changed. The results are that

(a) proven, working design may not work any more thus requiring
(b) changes to make products work will cause so many more problems that cost much more then any expected cost savings, as discussed in the section “Difficulties trying to reduce cost later” in the white paper at http://www.design4manufacturability.com/concurrent-engineering.htm

6) Tempted by foreign sales “opportunities” that lock you into offshoring and all the problems and limitations cited above and in the cited links. These enticing sales incentives will force you to work through a local “partner” to whom you bring your intellectual property and teach your partner how to build your products.

        Further, the incentives will expire if you don’t keep “upping the ante” with escalating investments and commitments in that country, for instance, building all your products there for “export” back to your home country or to other projects around the world.

        Unfortunately, many companies may still think that, even without the incentives, cheap parts and “low cost manufacturing” is still worth it  For anyone who thinks that, you need a better cost system! See http://www.halfcostproducts.com/total_cost.htm

7) CSP Plants are Unnecessarily Too Large because of the following reasons. Solar power plants that are too big cause the following problems and inhibitions:

       First, big plants are harder to acquire funding for and are harder find big enough sites and get approvals for them.

       Second, all this may force plant location farther from users instead of many smaller plants located closer to all their users. And smaller mirror fields have more location opportunities for siting and approvals, and can avoid prime farm land or environmentally sensitive land.

       Third, large fields need high towers, which can raise objections about visual glare, disruption to aviation, and risks to birds. But lowering the towers will reduce the sun shinning on the most distance mirrors, thus resulting in smaller mirrors that will produce less power for the same mirror guidance cost.

On the other hand, smaller fields, along with clever design, can enable the mirror field to be build in a bowl shape that will enable lower towers and bigger mirrors on fewer heliostat controllers for a given power. This is especially important for solar fields in northern latitudes, which will need to be exploited for capacity and to be near those users.

       Fourth, the large plant size and more remote siting may produce too much power for existing electrical grids to get all this power distributed to users.

Here are some of the reasons that cause unnecessarily large plant size:

a) Mass Production thinking that leads people to the unrealistic conclusion that the only way to get cost down is get procurement volumes up, with even prestigious magazines showing pictures of Henry Ford in front of a Model T. However, this is irrelevant because CSP doesn’t have Detroit volumes nor does it have the points (1) through (4) above, which were meticulously perfected by Henry Ford.

b) Turbines that are too big because they were designed for large power plants (see corresponding solutions below).

8) Expensive mirror guidance. The biggest cost penalty in CSP comes from the wrong premise for guiding the mirrors with each mirror needing two motors, two gearboxes, two sets of sensors, and a computer to constantly position both axes all day.
.     The largest CSP plant in the US has 347,000 mirrors, which needs almost 700 thousand of these expensive closed-loop servo mechanisms.  Dubai is planning a solar field over 2.5 times that size which would need well over a million axis drives.

In the next major section, an example will show how to eliminate these unnecessary costs.

9) Depending on subsidies and over-supply discounts
. Much of the solar industry doesn’t correct the above causes of high cost, but, instead, just accepts them and depends on government subsidies and, for Photo-Voltaic panels, temporary discounts available from their government subsidies and overproduction “deals.” However, such strategies can not be counted on to save the planet.


This section follows the same numbering as the above section on “What is keeping solar cost high now?”

1] When Cost is Committed. Understand that at least 80% of a product’s cost is determined by the design (as pointed out in the DFM article at
http://www.design4manufacturability.com/DFM_article.htm  ..
The first graph in that article shows that 60% of cost is determined by the concept/architecture and achieving major cost reduction will require concept breakthroughs as shown for electronics and structures at http://www.design4manufacturability.com/designing_low_cost_products.htm 

2] Commercialization
. The science of solar power is adequate and has been proven over the years. Commercialization emphasizes identifying the “crown jewels,” preserve them, and designing manufacturable stuff around them. Thus, there is no new risk and the proven science is preserved, so the only new testing would be for part and material durability and survivability, for which there is a wealth of data for most parts and materials. The longer version of this is delineated the section, “How to Commercialize Prototypes & Research.” on the article on commercialization.

3] Design for Manufacturability
. The web site: http://www.design4manufacturability.com  has many leading-edge seminar descriptions and articles on DFM. The main principle is that manufacturability must be designed into the product using Concurrent Engineering ( http://www.design4manufacturability.com/concurrent-engineering.htm  ) with the right mix of resources at the right time to accomplish this in half the time with half the resources (http://www.design4manufacturability.com/half-the-time.htm )

4] Design for Low Cost.
Most of the site http://www.design4manufacturability.com/  shows how to design products for low cost. The article: http://www.design4manufacturability.com/designing_low_cost_products.htm  discusses how to design low-cost products with examples for electronic products and large structures. The home page of HalfCostProducts.com  offers several more methodologies to lower total cost.

5] Avoiding offshoring enables real cost reduction, such as:

Working together in multifunctional teams in real time to use concurrent engineering to design low cost products

• Stable production can be reached in half the time
by using Concurrent Engineering teams to work together every day, interacting often instead one round of email per day. See: http://www.design4manufacturability.com/half-the-time.htm

Designing products for quality using multifunctional teams working together in real time, which will save much more in “cost of quality” than any part cost price saving for cheap parts. See: http://www.design4manufacturability.com/design_for_quality.htm
Benefit from efficient flow/lean production inside your own factory instead of offshoring to batch production in a remote contract manufacturer who just “builds to print.”  Unless you have enough production  for a dedicated line (which you could do at home), the contract manufacture will build batches that that will require the cost and delays of setting up the batch and tearing  it down (you will pay for both!), which will:

6] Be Objective about all opportunities, either for access to markets with strings attached (the temptations in # (6) above) or thinking offshoring will lower your net cost (the point# (5) above about the problems of offshoring. These risks will be minimized by:

• quantifying all costs as recommended in this sites total cost article

• Actually lowing your total costs, as recommended in all these points and links, so as not to be tempted by any magic elixirs

7] Make CSP fields the optimal size.  First, don’t fall for any fallacies that high volumes alone will lower the part cost and cut assembly costs dramatically.  Next, all components must be sized for the optimal plant size. Probably the biggest problem for big components forcing large plants would be the turbine and steam plant. Dr. Anderson has proposed to one of the turbine suppliers the need to use commercialization to retain their proven technologies to:

(a) design scaled-down versions.

b) design for lowest cost per output. Any machinery that consumes high-cost feed stock, like non-renewable fuels, must maximize efficiency However, machine cost can rise exponentially to get every last percent of efficiency.  On the other hand, solar power fuel is “free” but most of the cost is paying off the equipment. Therefore, since the economics are so different, it may be possible to design the lowest cost per output without the “efficiency at any cost” penalty.

 Given the effort, and the gain, this will need to be funded appropriately and quickly to assure this is ready when widespread implementation is needed fast.

8] Design the lowest cost mirror guidance. Given that heliostat mirrors may number in the thousands for a large solar plant, this could be the biggest opportunity to substantially slash the cost of concentrated solar power. The heliostat mirror field is half the installation cost for CSP electricity generation and constitutes almost all of the cost of heat generation installations for instance, for industrial heat needs, to replace existing burners, and augment conventional plants with solar energy whenever possible. (See next example on how ultra-low-cost mirror fields can be designed.).

9] Don’t depend on subsidies or over-supply discounts. 
Doing all above will lower the cost so much that it will eliminate the need depend on these or wait for them to appear.


The overall strategy: Instead of depending on hundreds of thousands of expensive servo-controlled motors, controllers, sensors, and gearboxes, the ultra-low-cost strategy would be to couple almost all mirrors together mechanically, all driven by a few common “master” drives..

        One might say: "Sounds good! Why hasn't this been done before."  The answer is that the design of mechanical couplings is not trivial  because each heliostat mirror must be uniquely positioned based on its location relative to the tower target, Each mirror must go through its unique daily motion to precisely reflect sunlight to a stationary tower throughout each day.

Fortunately, linkages can be designed to do this, with enough knowledge of the broad range of linkage functions possible, as shown at http://www.design4manufacturability.com/linkages.htm

Except for the few master drivers, the dependent mirrors would not need any motors, controls, gearbox, or sensors, thus savings many millions of dollars for the largest planned solar fields, one of which will need 1.8 million of these drives!.

Given that, solar power planners should seriously consider major cost reduction opportunities, like clever linkage couplings that would orient each mirror to be guided in a unique way to reflect sunlight to the tower all day. It is also possible to design such couplings to be stiff enough to enable many mirrors to be driven by the same master driver with little deflection or error The field would thus consist of low-cost mass-produced mirror assemblies which would be driven and controlled by one or more master driver in each array.

Example # 3 at http://www.design4manufacturability.com/linkages.htm  shows a CAD layout of to-scale mirrors (without showing the connecting linkages) that shows the unique orientations of connected mirrors. 

The description on the connecting linkage also discusses the value of optimizing the joint ordering – the first step in robot design. Conventional practice ignores this opportunity by arbitrarily placing the elevation axis on top of the azimuth (compass angle) axis. The trouble is that this actually turns the mirror out of the sun-mirror-tower plane (the tangential plane), which results in a compromised contour that precludes more optimized contours. (see that discussion for more details).

Simple-to-build heliostat couplings would represents a huge opportunity because:

a) The heliostat mirror field is half the installation cost for building CSP electricity generation plants

b) The mirror field constitutes almost all of the cost of heat generation and food processing installations for instance, for industrial heat needs and for solar power fields.

(c) Compared to trying to "cost reduce" or redesign today's complex servo drives, developing simple mechanical mechanisms would be faster, considering simple linkage parts could be automatically fabricated to high tolerances from ordinary materials that have already been proven to survive a long time outdoors. 

(c) Then, ordinary machine shops and factories could knock out large volumes of easily manufacturable mirror mechanisms.  . Since most of the heliostat mirror field would be manufactured by ordinary machine tools from ready available materials, this alone would easily satisfy local content requirements, without having to lose control of the crown jewels or outsource anything too hard to build or too proprietary.

This could also be scaled up much faster than photovoltaic panels, which need multi-billion “fabs” (semiconductor factories), which take years to build, compared to the 21,000 machine shops already in the United States.  See the Scalability article at: http://www.design4manufacturability.com/scalability.htm  

This is one example of how half the cost can be designed out of solar power plants,  All potential cost reduction breakthroughs, like this one, should be developed now so implementation can be able to quickly  commercialized and be designed to be scaled up very quickly.

About the Author

Dr. David M. Anderson
has been providing customized seminars and webinars on DFM and Concurrent Engineering for 25 years. He has unique expertise in both commercialization and scalability, which gives him unique expertise that enables him to create strategies and implementation plans to rapidly commercialize complex systems for optimize manufacturability so that they that they can be rapidly be scaled up as many times as needed.

Notable seminar/workshop engagements include eight at Hewlett-Packard, five at GE, four at Boeing, four at BAE Systems, four at Korea's LG Electronics, two at Emerson Electric. Advanced Energy Industries (power plant scale PV Inverters), Itron (smart meters), and five at GE, including GE Nuclear, GE Power (distributed power plants), and GE Energy (power plant scale fuel cells). See the complete list of Clients of Dr. David M. Anderson, P.E., CMC. .

Since 1990, he has published books on DFM and Concurrent Engineering, with updated editions published every couple of years, based on his seminars, workshops, consulting. His current 2014 DFM book is now being translated into Mandarin.

In 1993 he twice taught the Product Development course at the Haas Graduate School of Business at U.C., Berkeley.
Dr. Anderson is a Fellow of the American Society of Mechanical Engineers and a Life Member in SME. He has been certified as a Certified Management Consultant (CMC) by the Institute of Management Consultants. His credentials include professional engineering (P.E.) registrations in Mechanical, Industrial, and Manufacturing Engineering and a Doctorate in Mechanical Engineering from the University of California, Berkeley. .

He can be reached at 805-924-0100 or anderson@build-to-order-consulting.com 
He has published dozens of articles that are posted at www.design4manufacturability.com , www.HalfCostProducts.com, and www.build-to-order-consulting.com


 Seminars   Consulting    Credentials  Client List  Articles    Books     Site Map