Napa winery achieves sparkling results with combined cooling, heating, and power.
Mix two seasoned engineers together—one a self-described startup junkie for 25 years, the other a former avionics techie who has headed an engineering firm since 1991—and the desire to cook up the ultimate combined cooling, heating, and power (CCHP) system is irresistible. The result, commissioned in mid-2003, turned out to be a "very complicated, very sophisticated" system, says Ray Cole (the second of the two). It's one that, as the following story shows, is highly innovative, quasi-experimental, cutting-edge, remote-controllable, profitable—and kind of fun to operate.
Back to Nature in Napa Valley
In 2000 Chuck McMinn, the future owner of the system (i.e., the startup specialist), and his wife, Anne, visited picturesque St. Helena, CA, in Napa Valley; they ended up purchasing a lovely vineyard there called Vineyard 29. Their goal, Chuck says, was "to make the best wine we can" and market their high-end product from within a new, architecturally pleasing winery and hospitality center he planned to build.
 |
Photos: Axiom Engineers |
| Vineyard 29, a new winery in Napa Valley, gets 120 kW of onsite power from two microturbines. |
For the building systems, McMinn contacted Ray Cole of Axiom Engineers in Monterey, CA—specialists in heating, ventilation, and air conditioning (HVAC). Cole's firm had done work in food processing, education, health care, and commercial buildings for three decades.
Coincidentally, at about this time Cole was discovering microturbines. He'd done multimegawatt onsite power before, but what he'd seen in small generators had never seemed sufficiently durable for long-term heavy service. (Compared to his usual HVAC jobs, he notes, "a cogen plant is a whole different challenge due to the constancy of operation.") But Cole's doubts vanished when he watched the operation of a small, integrated, and self-contained microturbine. Standing no taller than head-high, it could output an impressive 60 kW; an integrated high-efficiency heat exchanger meant lots of usable thermal energy too. "I thought, ‘Wow, a little turbine—that's pretty cool!' " he recalls, and he considered all the production processes and activities that might easily use cogen power. Moreover, in its design efficiency the turbine (a Capstone C60) reminded him of a jet engine from his aeronautical days, "and it just struck some chords with me." It exuded maximized performance in a compact size. "The concept of single-shaft design, with air bearings—with no mechanical maintenance, no oil change, no friction, and no steel bearings—appealed to me personally," he says. Cole thought the Capstone designers had also done a good job with the start and stop functions, load following, and built-in grid connectivity. He knew that as a power train, "a turbine likes to be turned on and left on," and this design could do much more work, with less maintenance, than a reciprocating generator. "We saw opportunities here," he notes.
Just as Cole was scratching his head wondering where he might use a microturbine, along came the proposed winery. A Capstone seemed perfect. Cole quickly outlined to McMinn, a fellow engineer, the concept of CCHP trigeneration, and after going to see the relevant hardware at a Silicon Valley trade show, McMinn, intrigued, asked Cole for a cost-benefit study.
Greenfield Analysis Challenges
Small combined heating and power and CCHP plants are overwhelmingly designed and built as retrofits, in which a building system already exists, and the money outlay for onsite power can be justified by the expected savings on grid energy costs. A developer can simply total up recent and estimated future electric and gas bills to project a payback curve. In this job, though, with no operational history or real cost data (and in fact no winery yet), Cole had to develop spreadsheet models based on his engineering sketches, likely equipment specs, and guesstimates of usage, load, and utility charges from Pacific Gas and Electric (PG&E).
 |
| The winery cave is cooled to optimum temperature by the first US installation of a Nishiyodo adsorption chiller, which gets its heat from the winery's two Capstone C60 microturbines. |
 |
| Vineyard 29's fermentation process gets heating, power, and cooling from onsite trigeneration. |
In the new winery, multiple loads would arise from several sources. Cole made a list of everything to be powered: grape conveyors, 30 temperature-controlled tanks with cooling and sometimes heating for fermentation, piping systems, small pumps, refrigeration, elevators, and even a wine-storage cave in an adjacent hillside needing lights and cooling. The load curve would fluctuate from a high point during the "crush" (i.e., harvest and fermentation time from Labor Day to Thanksgiving) to lower non-crush times. Next, adjacent to the winery were new support facilities: a hospitality and administrative area, a laboratory, conference rooms, and kitchen appliances—all needing heat, lighting, air conditioning, and power.
With itemized loads in hand, Cole summed the total at between 60 and 70 kW, including peak, off-peak, and fully autonomous operation, if the grid should ever fail. He thus specified two Capstone C60s, for a total output of 120 kW—allowing ample reserve for the future, and plenty of backup. As initially planned, one or both of the turbines would run eight hours a day, shutting down after-hours.
Turbine exhaust heat would be captured and utilized in various ways (see below), and Cole's design called for lots of plumbing, wiring, and two integrated chillers.
Another critical issue here was energy reliability. Being at the valley's north end meant isolation from PG&E substations, and perhaps more frequent and longer outages. Opening a winery without providing emergency backup power was unthinkable. McMinn realized he might have to spend about $80,000 for a backup generator, and perhaps nearly that amount again for its delivery, installation, and wiring. The usual type used for this is a diesel reciprocating engine; it sits idle nearly all the time, costing six figures but serving no purpose except in emergencies. On the rare occasions when it must run (usually to see that it still works), there's noise and noxious fumes, which, in this environmentally sensitive neighborhood, would be most unwelcome.
So, McMinn thought, for about the same price it seemed smarter to invest in a clean-burning, quiet microturbine that would run all day.
Also appealing was the realization that, as McMinn recalls, "we would have control of our own destiny when it came to the power grid," which, in 2001, looked shaky. Power was unreliable, and the rates were soaring. All in all, then, the turbine concept "was a win-win-win," says McMinn—"for reliability, for the environment, and for saving money." And it offered an estimated payback time frame, Cole told him, of three to five years.
Cole and McMinn were also raring to go for another reason. With McMinn's skills and connections as an Internet access provider (he was one of the founders of Covad Communications in Silicon Valley), and Cole's experience in facility control systems, they wanted to pool their knowledge and create an innovative Web-based interface for facility monitoring and management. As they envisioned it, literally scores of data elements in the winery and support facility could be remote-controlled from a Web browser. Nothing like this had been done before, to their knowledge; certainly there was nothing affordably off-the-shelf in this line. The two thought the technical challenge might be interesting, and the resulting innovation important. McMinn gave the OK on Axiom's proposal, and late in 2001 the vineyard's contractors started building.
Really Cool Design
In Cole's view, one "magical" aspect of the design was the complementary, integrated operation of the two chillers. One unit was a first-of-its-kind US application (although with a nearly 20-year track record in Japan) of a Nishiyodo adsorption chiller; the other was an efficient electric chiller made by RTI. Both are rated at about 22 to 25 tons of chilling at standard conditions, Cole notes.
 |
| A Capstone C60 microturbine produces up to 60 kW of continuous energy, with automatic load-following adjustment. |
 |
| The natural gas fuel compressor. |
 |
| The power system's integrated control panel. |
What makes the Nishiyodo unique and optimal for this job, he says, is its adsorption technology, surpassing the more standard absorption type. An adsorption chiller produces cooling by adsorbing water at a high vacuum onto silica gel beds. The water thus adsorbed boils in the vacuum chamber. This boiling transfers heat from a heat exchanger, through which the winery process coolant circulates, yielding a stream of 30% cold propylene glycol/water mixture at 40°F.
The Nishiyodo chiller has two adsorbing beds; one provides the cooling effect while the other is regenerated, and thus the cooling remains continuous. Regeneration (or evaporation) occurs by the influx of hot water at 192°F, which has been heated by the exhaust streaming out of a at 500°F to 550°F. The Nishiyodo's input need here, Cole notes, "matches up well with the Capstones' heat output." Integrated heat exchangers turn this searing exhaust into near-boiling water of 190°F to 195°F. Even if this drops to as low as 120°F, the glycol still evaporates; this generously broad inlet temperature range, Cole finds, "is one of the nice things about the Nishiyodo."
Because its energy comes from exhaust heat, the Nishiyodo is naturally cheaper to run than the electric chiller. It also demands less maintenance and offers greater reliability. So, most of the time, the adsorption unit is relied on to produce adequate cooling, but starting every July, the RTI chiller takes the lead. It pipes out lithium bromide (LiBr) refrigerant at a chilly 25°F, an optimal temperature for stabilizing the wine prior to bottling. Now the Nishiyodo becomes secondary, since it is limited to a supply temperature of 37°F.
The chief technical difference between adsorption and absorption, as Cole explains, is that the latter uses a LiBr solution as its coolant, while the Nishiyodo, as noted above, uses water under a vacuum. Granted, the chilling power of LiBr is greater if sufficiently hot water or steam is available, but because the coolant is a solution, the equipment is less forgiving about temperature variants in both the inflow water and on the heating side. "If you get out of spec, you can cause chemical equilibrium problems, and you get precipitation," Cole amplifies. Crystallization within the pipe may occur, which can be disastrous. Because it cycles constantly, it must be replenished periodically to maintain its balance. Also, getting rid of depleted LiBr presents a major ecological issue, particularly at a winery. On this score, McMinn adds that "the nice thing about the Nishiyodo is it's basically silica gel and water, with some valves and steel. The water in it is almost 100% recyclable and benign to your environment."
LiBr chillers have well-suited roles in certain industrial applications, but in this agrarian one, another potential drawback is that they are a corrosive salt solution refrigerant. Over time, it's more apt to damage pipes, fittings, and even operating parts, especially if portions of pipe are exposed to temperature swings. One final "chilling" fact to consider here is, at least in Cole's experience, well-trained repair technicians are as rare as a case of Vineyard 29 Cabernet. The average refrigerant system repairperson, he says, often doesn't understand how chillers work.
The Nishiyodo brand is sold and supported in the western United States by CoGen Equipment Solutions of Carmel, CA, whose CEO, Donald Pruss, also represents Capstone.
Plenty of Hot Water
The workload of that exhaust-heated water isn't finished with the Nishiyodo cycle, not by a long shot. The same loop (temperature range now 175°F to 185°F) continues to other winery locations where it is used to heat water for washing and for wintertime building heat. Between these three functions and one or two others, the hot water is thoroughly utilized all year round.
To deliver these multiple services, Cole designed a prioritized hot-water dispersion system that he's rather proud of. The highest priority always goes to the wine production and aging process, specifically, for the year-round cooling of a nearby storage cave for the wine casks, which age best at an optimal 57°F to 59°F. Chilled water is pumped in, and fans circulate the chilled air evenly "for free," Cole adds, in the sense that this chilling comes from cogenerated exhaust heat. "The winemakers are tickled about that," notes McMinn.
All of which means the Nishiyodo gets assigned the first hot water out of the tank, because the chiller works at its peak with an inflow temperature of 190°F. Wash water requires only 160°F, and so the pipe supplying this service is positioned farther down the loop. Hot water for wintertime office heating can also use 160°F water (rather than the design-spec 180°F), thanks to the slightly larger coil surface Cole installed for the offices. A final loop sends water to warm glycol to 80°F or 90°F for warming the fermentation tanks.
Rarely do the hot-water needs conflict. ("Obviously you don't need full chiller capacity at the same time you need full building heat," Cole points out.) The hottest water goes where it's most needed, to the adsorber, and the less-hot water is tapped out "where it can still do useful work for us, down there warming the wine" slightly. "And damn, it's worked!"
With all this work to perform, in order to maintain a steady loop average above 175°F, the water also circulates into a boiler for supplemental heating, if need be. If the turbine exhaust alone can't keep it hot enough, the boiler will automatically light up for an extra burst. When the inlet temperature reaches 185°F again, the gas burner turns off. Moreover, if the turbine exhaust jet should rev up and sustain 205°F or more due to high electrical demand, the exhaust gas is bypassed within the heat recovery unit. Conversely, if the turbine should conk out, then the boilers take over and keep the water temperature constant. One way or another, says Cole, "We'll always have hot water, because if the turbines don't produce it, then the boiler will." It's an extremely simple, highly efficient cogeneration setup, with a monitored thermal efficiency of about 85%. It burns very little boiler fuel, but having the redundancy brings peace of mind.
"My Check, Please"
McMinn's cost for this cornucopia of heating, cooling, and electrical power—including the installation of the two Capstone C60s, two chillers, heat recovery / heat exchange hardware, plumbing, wiring, and transfer boxes, together with all the associated engineering and professional services, taxes, and shipping—came to a total of $608,763.
 |
| The interior of a Capstone C60 microturbine. |
A breakdown of this reveals the equipment cost $270,000; engineering was about $100,000 (and included a la carte services such as elimination of PG&E as the primary energy supplier, review of PG&E overcharges, the cogen feasibility study, microturbine and cogen engineering, construction documentation and administration, rebate filing, and installation supervision). Installation by the electrical contractor and a mechanical contractor came to $220,000 (including lots of wiring and plumbing).
Gratuity was not included, but the California Energy Commission (CEC) graciously helped McMinn pick up this tab, providing a rebate of $1,000 per kilowatt yielded by his equipment and installation investment. Hence, two Capstone C60s providing 120 kW made the CEC's total contribution $120,000
.Parenthetically, one "slightly bitter aftertaste" regarding the rebate is notable: Approval of this sum was overseen, in this case, by the utility, PG&E; it readily OK'd $113,000, but made life tougher for the winery by requiring rigorous documentation for the $7,000 balance. This meant coming through expense records and gathering up canceled checks. Complicating matters was the fact that several subcontractors had lumped together their charges for cogen work with non-related and non-rebatable services. The moral here, says Cole, is "be sure to segregate how you do the billing, pay the bills, and have contractors submit invoices" so that rebatable items are explicitly identified with an accounting code and by "a very clear paper trail."
At any rate, McMinn's net after-rebate expense came to $489,000.
But remember the diesel power backup, mentioned above, which McMinn didn't have to buy. The Capstones handily negated the need for this six-figure expense; and thus, taking that cost-avoidance in mind (as well as the excellent tax advantages of a cogen investment), McMinn regards his outlay for trigeneration at something well below $400,000.
Positive Savings, Good Reviews
In return for his investment, the winery has succeeded in reducing McMinn's power costs by perhaps half, compared to what the likely billings would be with PG&E. Cole estimates the Capstones' current operating costs at $0.04 to $0.045 per kilowatt. As for heat output, he calculates that "we've captured 2 billion Btus of cogen heat for useful purposes in the winery," the great bulk being expended on the Nishiyodo. McMinn adds, "We know we're saving money every day it's online. … It seems to be everything we'd hoped for." Full payback is expected within five years.
PG&E still supplies a trickle of 2–5 kW, the minimum needed to keep the grid connection going and comply with interconnect rules. A transformer feeds this into 1,000-amp main switchgear; there's a point of common coupling where the microturbine juice also comes in. Capstones carry dual-mode controllers so that, if the incoming grid power fails, disconnection is automatic and the microturbines supply all the winery power in standalone mode. (This necessity has happened, in fact, just once in the first year, prior to actual commissioning.)
What are the engineering connoisseurs' verdicts on this project? Wine tasters applaud a tasty vintage with phrases like "it has a lush, full-bodied spiciness that is pleasing to the palate." The equivalent praise for a "first-one-in-the-USA" Nishiyodo maiden voyage is this bottom-line assessment from Cole: "We fired it up in mid-August [2003] and it has worked almost flawlessly since." Initially, a small pinhole leak appeared; it was easy to spot and fix. Too, the PG&E power to it surged once, "throwing a big, nasty spike" and tripping a breaker. That episode (not the chiller's fault, of course) was its only service outage. All in all, then, its performance has been "spectacular. … From my perspective and the owner's it's been a beautiful machine," says Cole.
As for the two compact turbines, McMinn also gives a solid rave. Originally he'd planned to run the power plant eight hours a day, but after startup, he quickly found that "the economics and performance"—especially the multiuse exhaust heat—turned out to be so unexpectedly favorable that he runs the Capstones day and night. One C60 suffices most of the time, but the second springs to life during peak loads (which top out at 80 kW or so during the crush—a figure close to Cole's preconstruction estimate). During non-crush, the daytime loads average around 40 kW and drop to 25 or 30 kW after-hours. A load-following meter adjusts the power output to match demand hour by hour. Automated controllers also balance the run-time between the two Capstones so that one doesn't work harder and wear out sooner.
In sum, says McMinn, "They're online all the time, providing all our power and much of the heat. In terms of being workhorses, they're not very temperamental. They just seem to run."
Coming over the immediate horizon, the CCHP plant will soon be circulating heat, power, and cooling to McMinn's adjacent new home and pool. Neither is built yet, but their anticipated loads were factored into the initial plumbing and electrical design.
For Cole, this first microturbine foray was exhilarating, and it whet his palate for more. Since his successful completion of the winery, he's developed five more cogen jobs, including an innovative year-round pool and rec center and a county detention facility. He's looking for more applications besides. The threshold for strong payback potential, he's finding, seems to be a year-round heat load of about 6,000 hours; achieving this, a project will likely earn a good return.
Summing up, Cole says, "This first Capstone got us started, and it taught us a lot. The devil is always in the details; we've learned a lot of the details." Vineyard 29 was, he adds, "a really challenging project, because we just tried to do so much. It all worked out, but it was not easy," he concedes. "There were so many elements that had to be brought together and integrated," including new technology and the significant hurdle of dealing with the utilities. Although the job was complex and difficult, he says, "It is turning out to be fairly satisfying—just because it's all working the way we intended it to."