Blog Archives - Focused Sun Microgrid

Microgrid Module & Solar Farms

4/28/2015

I keep getting more questions on our Microgrid module. One question is how does a small Microgrid solar array with cement storage and ORC engine compare to Concentrated Solar Power (CSP) installations. See my previous blogs on what such a system is like and its economics.

The question comes from a simple fact: CSP solar farm electricity generation is 40% efficient. Solar energy heats steam to as high as 550C, then using the common Rankine Cycle steam turbine. “Carnot” thermodynamics (an ideal turbogenerator) shows that efficiency at these temperatures should be 40% or more.

Indeed, the U.S. Dept. of Energy says, “the current lowest-cost state-of-the-art commercial standard is estimated to be a central receiver configuration which utilizes a molten salt HTF [Heat Transfer Fluid], coupled with 10 hours of thermal storage, to deliver heat at ~550°C to a steam Rankine power cycle with a designed thermal-to-electric conversion efficiency of ~41%. As of 2013, this configuration was estimated to deliver an LCOE of approximately 13 ¢/kWhe without subsidies.”

Meanwhile an Organic Rankine Cycle (ORC) engine running at 300C gets only 20% efficiency: 20% of the sun’s heat is converted into electricity. How can a 20% efficient ORC engine compete with a steam turbine having twice the efficiency?

The answer is heat. We use the heat and they don’t. Either turbine – steam Rankine or Organic Rankine – produces heat as a result of the thermodynamic process of generating electricity. It’s called “low grade” heat because it has less value than the “high grade” heat the turbine uses. But low grade heat (heat with a temperature below boiling, 100C) is valuable for many processes. District heating (space heating of homes and businesses in cold climates), desalination (purifying salty or brackish water), absorption chilling (cool air for homes and buildings or make ice for transporting farm produce to market) and process heat (industrial heat for laundries, fabric processing and food processing) all use low grade heat.

The problem is that those efficient steam turbines can’t usually use the low grade heat they produce. With power in the 10 MW to 1000 MW range, these large installations typically need over 25 m2/kW of land area. The 100 MW Shams Solar Power plant (UAE) needed a site a mile on a side (2.5 km2) for its solar collectors. The 392 MW Ivanpah plant (USA) needed a site over 6 square miles (16 km2).

Most large CSP installations are in the desert where low diffuse radiation favors concentration with mirrors. These usually aren’t installations that are near a city. Large solar farms are sited far from population centers that could use the low grade heat they produce. Their low-grade heat is simply discarded. One method dumps the heat into rivers or the sea. Another uses cooling towers to dump low grade heat to the environment. If you’ve spent so much effort to capture solar energy, why throw most of it away?

By contrast, a 100 kW installation requires a half acre (0.2 ha) site for its solar collectors. Such a site can be near population centers that can use the low grade heat produced. For instance, a gated community in the U.S. can provide electricity to its homes while heating the same homes in the winter. An agricultural microgrid can run fans with its electricity while heating greenhouses at night. A rural microgrid in Egypt can pump water and power local villages day and night while desalinating salt water with its heat.

If the microgrid’s heat is used locally, its economics become more compelling. In my earlier economics blog, I showed that the heat energy revenue stream is about equal to its electrical energy revenue stream. Both energy forms deliver the same revenue stream. By using the heat, we have twice the savings with our microgrid compared to a desert solar farm. Doubling the savings makes up for the 2:1 efficiency difference between high temperature steam turbines and our lower temperature ORC turbine.

As I showed in the section on Microgrid Economics, the cost forecast of our system (Microgrid modules, concrete heat storage, ORC turbogenerator) is between $2.7/W (low cost labor) and $3.2/W (high cost labor) for only the electricity. By comparison, the Ivanpah solar farm cost $5.6/W for its electricity. The large solar farms can cost nearly twice what our Microgrid Module plant costs.

Capital cost is one comparison. Another is energy costs: the Levelized Cost of Energy or LCOE. By using the heat locally, we forecast an LCOE of electricity at $0.075/kW-hr. We get this low value by using the heat. We allocate half of the plant to electricity and half to heat – they both have the same revenue streams. As noted in my earlier blog, this is well below the $0.18/kW-hr LCOE of electricity from Concentrated Solar Power (see IRENA_RE_Power_Costs_2014_report at www.Irena.org), the category into which large solar farms fall.

Other advantages stem from a smaller installation. As already noted, small 100 kW plants are easier to site near population centers where their heat can be used. But these same plants can be installed by far smaller engineering firms. The 392 MW Ivanpah plant in Nevada USA was installed by Bechtel, one of the largest engineering firms in the US. Even a small engineering firm can install a 100 kW system. For smaller projects, costs are easier to finance, environmental impacts are less and approval times are shorter.

Our production methodology of sandwich fabrication means captive factory economics reign. Instead of buying solar collectors fabricated by a central factory, each solar factory makes its own solar collectors. Raw materials bought on the global market can save as much as half the costs of the system’s collectors compared to purchasing already-built collectors. Of course that’s why our LCOE is low: we take into account these savings.

Last is jobs. We bring jobs to a local community. Once a microgrid system is installed, those same factory people can produce FourFold modules that deliver heat and PV electricity at smaller scales. Solarize your community in the best way possible. Capture 70% of the sun’s energy instead of only 20%.

100 kW Electric Microgrid

4/14/2015

Some of you have asked about our Microgrid module and how it could be used to power a microgrid. Here’s the details. I envision the smallest microgrid plant to produce 100kW of electrical power. Larger plants would use multiples of 100 kW, for example a one MW plant would be made of ten 100 kW plants. This is on the small size for current turbogenerators because efficiency drops quickly if they are even smaller. A 100 kW plant can power 20 US homes at 5 kW per home.

The big reasons for picking a small plant is the land area needed and the engine size needed for efficient mass production. Since we are partnering with the Xiang Yang Institute in China, we want the size to be consistent with the area needed for the solar array. A 100 kW electric plant would require 300 of our Microgrid collectors, each requiring 4 square meters of land area. That’s 1200 square meters for the solar array and another 400 square meters around the array for the turbogenerator facility, a periphery walkway and a security fence. At 1600 square meters, the land required is 40 m on a side. That’s about a half acre or 4 tennis courts. Smaller plots are easier to find than larger ones.

A second reason for a small plant is the turbogenerator size. A 100 kW unit – about 130 horsepower -- is the size of a truck engine. That’s something that can be made easily on an assembly line. More important, its turbine can be made on machine tools of reasonable size. China has trained tens of thousands of NC (Numerical Control) machinists in the 500 vocational training centers set up in small cities throughout China. It's these NC machinists that could mass-produce the 100 kW turbogenerator.

Back to the microgrid solar array, each collector covers a length of 2 m (79”) x 2 m (79”) where the width includes 1.2 m (47”) of module. Each row of modules requires a 0.8 m (30”) service access walkway between rows. That’s where the 4 square meters per module comes from. Each Microgrid module is raised 2 m (79”) above grade level. Beneath the module is thermal storage. While we are also looking into various types of phase change storage, the simplest heat storage is concrete where heat is stored as “sensible” heat. Sensible heat is the heat required to raise the temperature of the concrete; no heat of fusion or molten salt is involved. Here’s a schematic of how concrete could be used as a storage material for the Microgrid module.

Below each module is a horizontal cylinder of concrete 600 mm (24”) in diameter and 1.8 m (72”) long. Heat transfer pipes pass through each cylinder to both add and withdraw heat from it. Each cylinder is supported on concrete blocks to reduce its conduction heat loss to the ground. The cylinder is surrounded by fiberglass batting (glass wool) insulation to prevent convective and radiation heat loss from the cylinder. A cover protects the storage and insulation from the weather. Essentially each cylinder is thermally isolated from its surrounding.

Microgrid modules are “daisy-chained” together in long rows in the North South direction. The output of one module’s absorber flows directly into the next module’s absorber. Flexing unions keep thermal expansion stresses low. In a similar way, the storage cylinders are daisy-chained together. The heat transfer pipes of one cylinder flow into the heat transfer pipes of the adjacent cylinder. Again, flexing unions reduce thermal expansion stresses.

At each end of the module string, the absorber pipe is connected to its associated storage cylinder. A heat transfer loop is formed where mineral oil pumped through the Microgrid absorbers collect solar energy. Each module in the string adds solar energy to the oil, increasing its temperature. At the end of the string, the oil is hottest. There it flows down into the storage cylinders where it transfers heat to the concrete. As hot oil flows through each successive cylinder in the string, it loses its heat to the concrete. Arriving at the beginning of the string, the oil is pumped once again through the modules’ absorbers.

The entire loop is called the “solar loop” because it stores solar heat. Oil flowing through successive absorbers gets hotter and hotter until the end of the string. There it reverses direction and flows through the concrete cylinders. Heat is lost to each cylinder in succession until the oil is at its coolest at the beginning of the loop.

Heat is removed from the concrete by oil flowing through a second set of heat transfer tubes called the “user loop”. Mineral oil pumped through this second loop starts at the same module as the solar loop. As its oil passes through each storage cylinder in turn, it gets hotter and hotter. It is hottest leaving the last cylinder where it flows to the turbogenerator.

There the oil flows through a heat exchanger to heat the turbogenerator’s working fluid, converting its heat to electricity in a thermodynamic cycle. After leaving the turbogenerator, the oil is still hot, on the order of 100C. This “low grade” heat can be used locally to double the plant’s return on investment. For each kW-hour of electrical energy produced by the plant, 3 to 4 kW-hours of low grade heat is available from the turbogenerator. The heat can be used commercially for heating hotels and restaurants, district heating, laundries and air-cooling. Industrially it can be used for thermal desalination, absorptive refrigeration, food processing, fabric processing and other process heating applications.

The two loops – solar loop and user loop – act like a counter-flow heat exchanger. The module absorbers have their highest temperature at the last module in the string. The storage cylinders have the same arrangement: the temperature is hottest at the last module. The arrangement assures that the turbogenerator receives the hottest oil available.

While our Chinese partners are considering various turbogenerators to generate the Microgrid’s electricity, the simplest one available is an Organic Rankine Cycle engine or ORC engine. An Organic Rankine Cycle differs from its more common cousin, the Rankine Cycle, by the working fluid used. Steam (gas phase water) is the working fluid of the Rankine Cycle. It is the standard for generating electricity in most of the world. Organic Rankine Cycles use other molecules than water as its working fluid. ORC engines are available from many sources including German manufacturer Siemens and Japanese manufacturer Mitsubishi through their Italian subsidiary Turboden.

For you thermodynamics folks, we expect the temperature range of the storage will vary between 200C and 300C. The hotter the better, of course, if we want the highest Carnot efficiency. We think the Microgrid modules can deliver 300C (572F) heat at useful flow rates using vacuum jacketed tubes (see http://www.focused-sun.com/fs/technology/hybrid_absorber). As the storage temperature is depleted, the delivery temperature drops to perhaps 200C (392F) during normal operation. Most Microgrid arrays will have a backup generator to handle the possibility of a week of cloudy weather. Backups can be diesel generators, biomass generators or even boilers that add heat to storage when little solar energy is available.

Microgrid Economics

4/14/2015

My earlier blog (Apr. 14, 2015) showed a preliminary design for a Microgrid module system with concrete cylinder storage. What capital costs ($/W) could be expected for a complete system (collectors, storage and ORC engine) if it were produced in China? What will be the cost of energy over the plant's lifetime?

I think the answer to the first question is about the same capital cost as a current natural gas fired generator if you include the distribution costs. The big difference is that the Microgrid system would provide electricity day and night for no fuel cost.

Most microgrid systems divide into 3 cost categories: collection, storage and conversion. For Focused Sun, collectors are the array of Microgrid modules, storage is the concrete cylinders and conversion is the ORC engine. Understand that the costs I’m presenting are forecast costs: the costs you could expect after a few dozen of these systems have been installed.

First consider collection. Here we assume the system is NOT built in the West. According to our BOM (Bill of Materials), the collector cost assuming pilot production quantities is $400 for raw materials, labor and the Focused Sun royalty. An array of 300 Microgrid modules would cost $120,000 or $1.2/W for the 100 kW system. In the West, labor costs are typically 10X higher. The same Microgrid modules made in the West modules would cost $170,000 giving a cost per Watt of $1.7/W.

Heat storage is the cost of concrete cylinders and their heat transfer piping at $70,000 or $.7/W. Each Microgrid module has 30 kW-hr of heat storage. The concrete has a cost of less than $5/kW-hr and lasts for decades. Compare this with batteries at $400/kW-hr that last only a few years.

We think the 100 kW ORC engine can be mass produced in China for about $80,000 or $.8/W. I’ll discuss why I think the Chinese can make ORC engines this size in more detail later.

All told, the plant cost is $270,000 or $2.7/W. In the West with our higher labor costs, the estimated costs are $320,000 or $3.2/W. If we were only producing electricity for the microgrid, these costs are more than the going price for a PV solar farm.

But wait. We have leftover low-grade (less than 100C) heat from the ORC engine. That heat can double its return on investment. That's the same as cutting its payback in half. It’s the combination of heat and power that makes this system economical. Applications should use both electricity and heat. Heat uses in the commercial/industrial market are hotels, district heating, desalination, air-cooling, laundries, refrigeration, food processing and fabric processing.

But how to apportion the capital cost of the heat versus the electricity? Focused Sun uses a computer analysis to forecast the heat and electrical energy we could expect from a 300 module Microgrid system. Back in the day, we were among the first to use monthly weather data to estimate performance of our MIT solar module. We could only get weather data from 10 cities. Today it’s much easier: the U.S. National Renewable Energy Laboratory (Golden CO) does it for nearly 200 US cities (http://rredc.nrel.gov/solar/old_data/nsrdb/1961-1990/redbook/).

Where I live Las Cruces, NM is closest to El Paso, TX. Using NREL data for “DIRECT BEAM SOLAR RADIATION FOR CONCENTRATING COLLECTORS (kWh/m2/day)” for El Paso TX gives the monthly averages of solar radiation of various types of reflecting solar collectors including our single N-S rotational axis with horizontal collectors (http://rredc.nrel.gov/solar/old_data/nsrdb/1961-1990/redbook/sum2/23044.txt).

Using this data, our computer analysis based on NREL collector type says the 100 kW system will deliver 260,000 kW-hr of electricity a year plus 870,000 kW-hr of low grade heat (less than 100C) each year. Assumptions we use include a reflection efficiency of our mirrors at 88%, collector heat loss at 10%, ORC electrical generation efficiency at 20% and heat delivered to it and ORC engine heat loss at 10%.

In the US, heat from natural gas typically costs $0.04/kW-hr; electricity costs $0.12/kW-hr. At these energy costs, the system’s electricity savings are $31,000/yr and its heat savings are $35,000/yr. The combined savings total $66,000/yr. Then 47% of the total savings come from electricity ($31K/$60K) and 53% come from heat ($35K/$66K). Note that heat savings double the total savings; heat produces as much savings as the electricity.

That means 47% of the plant’s $270,000 capital costs are apportioned to electricity ($127,000) and 53% apportioned to heat ($143,000). The cost per Watt of electricity is $127,000 for 100 kW of electricity or $1.3/Watt. This is a little less than the going cost of a PV solar farm that doesn’t have energy storage. Note that the economics of the entire system requires that the leftover heat from ORC engine is used locally. If the heat isn’t used, then the cost per Watt is much more: $270,000 for 100 kW of electricity is $2.7/W.

ORC engine pricing is based on Mitsubishi’s Turboden Division that have sold ORC engines since 1980. Small engines (100 kW) are priced at $2.5/W, bigger engines (1 MW) at $1.6/W and their largest engines (10 MW) at $0.8/W. When you look at the Turboden website, most of their installations are in the 1 MW to 10 MW range. Far fewer are as small as 100 kW. Clearly, they are not set up to mass produce the smaller engines that we need.

China manufactures products less expensively than Western production because China does not have the high fixed costs (mostly professional salaries) of Western companies. In my experience – 5 years running a sourcing company in China – China can price a product at half the Western price. My choice of $0.8/W for a 100 kW ORC engine reflects lower Chinese pricing as well as the economies of scale of mass production. If a Western company like Turboden can produce a 10 MW engine for $.8/W, Chinese manufacturers should be able to match that price/Watt for mass produced 100 kW engines.

The bottom line is that a Focused Sun microgrid system produces steady power for about the same capital cost as utility electricity if the heat produced is also used. But that’s the capital cost.

What about the plant’s cost of energy? For energy costs, the Levelized Cost of Energy (LCOE) is the cost of energy over the plant’s lifetime which we’ll assume is 20 years. As noted above, the electricity portion of the plant costs $127,000 and produces 260,000 kW-hr of electricity annually. The heat portion costs $143,000 and produces 870,000 kW-hr of heat annually. To find the LCOE, we also need to know the maintenance and operation costs. Let’s assume a two man maintenance crew working single shift at $2400 annual wages in a non-Western country. Annual maintenance expenses are estimated at 2% of the initial capital cost. The total cost of about $10,000/yr means $4700/yr goes to maintenance of the electricity portion and $5300/yr goes to maintenance of the heat portion.

Given these parameters, we can calculate the LCOE for each type of energy independently. The LCOE equation can be found on the internet, for example at http://large.stanford.edu/courses/2010/ph240/vasudev1/. Using a 10% discount rate to include the time value of money gives an LCOE of $0.075/kW-hr for the plant’s electricity. Using the same formula for heat delivery gives an LCOE is $0.023/kW-hr. Both these values are less than utility heat and electricity in most places. In fact, to calculate the solar savings of our microgrid plant, we used average US energy values of $$0.04/kW-hr and 0.12/kW-hr respectively. And since we’re storing energy, we can deliver heat and electricity from the ORC engine 24 hours a day and 7 days a week. That’s steady power for less than utilities charge. By comparison, the chart below shows LCOE electricity values for various types of energy from the International

RenewableEnergyAgency,IRENA:(http://www.irena.org/menu/index.aspxmnu=Subcat&PriMenuID=36&CatID=141&S...).

Notice that $.075/kW-hr is below all Solar Photovoltaic, Concentrated Solar (CSP) and Offshore Wind. It’s about equal to Biomass, Geothermal, Hydro and Onshore Wind. It’s also lower than the average Fossil Fuel power. Not bad for delivering steady power day and night. And that price won’t go up in the future since the capital cost has already been paid.

In many regions, electricity is not only less reliable – 6 to 8 hours of power a day are common in developing countries – but more costly. I’ve calculated the cost of electricity by diesel generators in Africa and found it costs $0.35/kW-hr. Even in America, electricity can cost $0.44/kW-hr if you are a Tier 3 consumer in northern California.

Island communities also pay high prices for electricity. We have had interest from many island nations like Malta, Indonesia and the Philippines. There the fuel to produce electricity must be imported making energy costs especially high. Renewables are a much better deal than fossil fuels for these island communities.

Another comparison is the cost of heat. At $0.023/kW-hr, our heat LCOE is very low. While America has used fracturing to tap into its bountiful gas energy, even natural gas heat in the US is $0.04/kW-hr to $0.05/kW-hr. Heat from other sources is more pricey. I recently switched from propane heat to natural gas heat when Zia Gas put a gas line out our way (we live 2 miles past the Las Cruces city limits). I was pleased to see my heat cost drop by better than half. Propane, at $0.12/kW-hr, is three times more expensive than piped natural gas. It's also tied to the cost of oil which is likely to increase as the world uses more oil.

A last concern with economics is risk: new products have a higher risk than old, reliable products. We combine three components: collector, storage and conversion. Only our collectors have not been proven in the marketplace. Cement has been used to store heat at 300C by the Europeans for decades. ORC engines have been made by substantive companies like Siemens and Mitsubishi for decades. Only our linear Fresnel collectors have not yet been proven. Yet the linear Fresnel concentration method is itself decades old, invented by Francia in Italy in the 1950s.

What we bring to the party is a low cost way of making linear Fresnel mirrors. Sandwich fabrication, which we pioneered on solar panels with Chevron in the 1980s, is the lowest cost method to make the mirrors. We have squeezed the costs out of linear Fresnel to make the most economic solar energy approach available today.