|
|
|
|
||||||||||||||||||||||||||||
|
Blog for 2010 |
Sunoba Renewable Energy Systems |
|||||||||||||||||||||||||||||
|
27 December: As the economists would say, seasonally adjusted
greetings to all. I’m currently having
a short break from my simulations.
This will also be my last post on this blog site. From the start of the New Year, I’ll be
blogging at www.sunoba.blogspot.com. 11 December: Well, I really enjoyed the
AuSES conference. Congratulations to
the organisers and thanks for all the hard work. It was fascinating to hear about the
explosive growth in solar power generation (PV and solar thermal) and the
associated technical, political and economic considerations. The chart below is due to Paula Mints, Renewable Energy World, Aug 2010, and
shows the annual global PV market.
You can see what I mean about explosive
growth. The market went from 300 MW to
13 GW in the decade up to 2010. Recent
annual growth rates are around 60%, which would mean that the market will
increase by a factor of 10 in five years.
My general understanding is that solar thermal technologies are about
2-4 years behind PV, but are if anything growing even more quickly. This is going to be a very exciting
industry! I’m currently thinking about integration of
heat storage into my concept so that electricity can be generated for several
hours after dark. My goal would be to
do this cheaply – of course! – so my starting point will be to look at
storage in a bed of rock chips/pebbles.
Initial calculations are going well, but it’s too early to report
anything specific here. 29 November: I’ll be attending the annual conference of
the Australian Solar Energy Society in Canberra this week (www.auses.org.au/conference). My oral presentation on the Wellington
simulations is scheduled for 2 pm on Friday, and I’m hoping to meet people
interested in new approaches to solar thermal power generation. Late last week, I finalised my paper for
the conference proceedings. The two
referees had suggested minor changes to my original version, particularly the
inclusion of a paragraph to describe the operation of the evaporation
engine. I particularly appreciated the
comments of Referee 2: “The
author presents an interesting study of relevance to the conference. The objectives and outcomes are clear, the
methodology is appropriate. The
modelling process and assumptions are not unreasonable and there is sufficient
detail provided for others to check the calculations.” 17 November (lightly revised, 27 December
2010): I’m happy to report that I have finished my
Wellington simulations. If you have followed this blog, you’ll know
that I have been simulating the output of a new solar heat engine, as based
in Wellington in inland New South Wales.
In this concept, heat energy provided by the sun is collected
passively under a transparent insulated canopy. The resultant hot air is fed to the inlet
valves of my evaporation engine and power
is generated. I have now simulated the
output of this system over the year from March 2009 to February 2010, for
both horizontal and sloping canopies.
(In the latter case, the canopy slopes at the latitude angle, 32.6
degrees.) |
||||||||||||||||||||||||||||||
|
The figure
at right shows the simulated daily power output for a horizontal canopy. These results are inclusive of canopy
losses but not engine losses, and are for a canopy with area 1,000 m^2. Results for the sloping canopy will be
posted after the work has been published.
(But I can say that the sloping
canopy gives about 25% more power output, more evenly throughout the year.) [Acknowledgement
– weather data was purchased from Bureau of Meteorology.] |
|
|||||||||||||||||||||||||||||
|
Subsequently I assessed engine losses, as
separate from thermal losses in the canopy, and estimated the power per peak
Watt and the Levelised Energy Cost.
The table below contains the key results. The figures are based on: collector area
1,000 m^2; cost of land/collectors/earthworks/construction AUD 25/m^2; cost
of engine and balance of plant AUD 1,000/kW; interest rate 8% and payback
period 20 years; annual maintenance cost 4% of total cost.
Details of this work (for horizontal
canopies) will be presented at the 2010 conference of the Australian Solar
Energy Society in Canberra, 1-3 December (www.auses.org.au/conference). My intention is to present results for the
sloping canopy at the ISES
Solar World Congress 2011. References:
Technical Reports 2010-4 and 2010-5;
scientific article [7]. 11 November: By popular demand from the discussion at Air Vent …
The photograph shows my evaporation engine
in April 2008. Some modifications were
later made before I finally got it to work in August 2008. Features to note are the transparent cylinder,
the piston mounted on the chrome-plated steel rod, valve assemblies (inlet
and outlet) at each end, the air heater and the hot air ducts to the inlet
valve “bellows” (grey material at each end).
The water pump to drive the spray injectors is mounted at the front of
the engine and the injection nozzles are embedded in the inlet valves. Valves were moved by compressed air. The whole device is controlled by a
computer (out of shot) connected via the black cable at the front. The dog played no role other than to sleep
in the sun and occasionally wag his tail in a friendly way. This video clip (MPG, 2.81 MB) shows
four strokes of the engine in operation.
You can hear three separate noises – the “click” of valves as the
water injection is turned on/off, the “clunk” of the brakes that are applied
to hold the rod at the end of each stroke, and a “buzzsaw” sound of the air
compressor (that powers the valve actuators and the water pump). Yes, I know this a weird engine
mechanism. Let me assure you that the
prototype I intend to build has a far superior mechanical principle (still
currently a commercial secret).
Experimental results have been written up in scientific
article [5]. This video clip (MPG, 2.55 MB) is
the evaporation/condensation sequence I mentioned in the Air Vent
discussion. The air in the cylinder
was initially warm and moist. All
valves were closed and the rod was manually pushed-pulled-pushed-pulled etc as quickly as possible. The pressure increases in one chamber and
decreases in the other, and then reverses.
You’ll see a condensation mist forming and disappearing about 10 times
in 2 seconds. The reason that the
piston moves to the right is that some air escapes through an outlet valve
during the “push” part of the process. The video clips were taken on my cheap
camera. I’m not expecting to win
awards for cinematography. 9 November: There’s been a long discussion on Air Vent about condensation during expansion of moist
air. A few years ago, when working out
the theory for the condensation heat pump, I
developed codes to simulate this process.
To help with the discussion on Air Vent, I ran two simulations as
described below. The key assumptions
underlying the simulations are ideal gas laws for air and vapour, constant
specific heat capacities for air and vapour, adiabatic expansion, and
saturation vapour pressure and latent heat given as tabulated functions of
temperature. Case (1) is the adiabatic expansion
(moist adiabat) for a parcel of gas with initial conditions V = 1 m^3, T = 25°C = 298.15 K, Pair
= 98,131 Pa, Pvap = 3,169 Pa. The expansion is continued until the
temperature reaches freezing. Case (2) is the adiabatic expansion
(dry adiabat) for a parcel of gas with initial conditions V = 1 m^3, T = 25°C = 298.15 K, Pair
= 100,800 Pa, Pvap = 500 Pa. The expansion is continued until the
temperature reaches freezing. I think my moist adiabatic expansion
results agree well with a completely separate calculation by Nick Stokes. I’m still working on the details of the
comparison. In the meantime, here are the plots that
accompany the post I made today on Air Vent. |
||||||||||||||||||||||||||||||
|
We start with a volume of 1
m^3 at ambient pressure, Pair + Pvap = 101,300 Pa. As the total pressure (Pair + Pvap) drops, the volume of each parcel of gases increases, but at
different rates for the two cases.
Condensation occurs in Case (1) but not Case (2). |
|
|||||||||||||||||||||||||||||
|
This plot shows temperature T as a function of the total pressure
(Pair + Pvap). As the total pressure (Pair + Pvap) drops, the temperature drops more rapidly for the dry
adiabat than for the moist adiabat. Condensation
occurs in Case (1) but not Case (2). In the atmospheric context, we
would say the dry adiabatic lapse rate is faster than the moist adiabatic
rate. |
|
|||||||||||||||||||||||||||||
|
This plot shows the amount of
water that has been condensed (delmv,
expressed in kg per kg of dry air in the parcel) as a function of the total
pressure (Pair + Pvap).
The plot is only applicable for Case (1) since condensation does not
occur in Case (2). The negative sign indicates
that condensation is occurring. |
|
|||||||||||||||||||||||||||||
|
Finally, the density (ρ_air + ρ_vap) is shown as a
function of the total pressure (Pair
+ Pvap). For a given pressure, the moist parcel of
air is less dense than the dry parcel. |
|
|||||||||||||||||||||||||||||
|
Reference: Air
Vent 7 November: It was confirmed this week that my
Wellington simulations have been accepted for presentation at the annual
conference of the Australian Solar Energy Society (www.auses.org.au/conference). Those simulations were for a horizontal
canopy. Meanwhile, I’m close to
finishing the simulations for a canopy that slopes at the latitude
angle. I’ll post some results in a few
days, but to give a glimpse of them – I found that the power output for the
sloping canopy was superior to that for the horizontal canopy except in
summer. The cost metrics will be
better for the sloping canopy. 20 October: I have resumed technical simulations. The current activity is to simulate the
performance of the Wellington canopy/engine system for the case when the
canopy slopes at the latitude angle (32.6 degrees). I expect the power output will be decreased
somewhat in summer, but increased during the rest of the year. Overall, this should give better cost
metrics than for the case when the canopy is horizontal. I expect to finish the simulations by the
first week in November. 30 September: I’ll be attending the All-Energy Conference
in Melbourne next week (www.all-energy.com.au).
Delegates wishing to discuss the use of the evaporation engine for
passive solar thermal power generation are invited to contact me by e-mail. 20 September: I’ve been in The Netherlands visiting my
older son who is an aerospace engineer.
(Out of the office, he works on a team competing for the Google Lunar
X Prize; see www.whitelabelspace.com
for details.) As planned, I gave a seminar at the
Queensland University of Technology before my departure. A slightly modified version of the seminar
is available here (PDF, 831 KB). This version of the seminar includes an
estimate for the Levelised Energy Cost for power produced from the
engine/canopy system at Wellington. In
addition to the assumptions described in my post for 29 August, I used the
additional assumptions: interest rate 8%, payback period 25 years and annual
O&M expenses 4% of capital cost.
The LEC comes out as AUD 163/MWhr.
This estimate must be regarded as preliminary and does not include
taxation considerations or sensitivity to the method of financing. 25 August:
I’ll be presenting a seminar on the Wellington simulations at the
Queensland University of Technology in Brisbane next Tuesday. Details can be downloaded from https://wiki.qut.edu.au/x/XiJEBQ. Afterwards, I intend to put a PDF copy of
the seminar on this page. 19 August (revised 29 August): The conference paper I mentioned in my last
post had a strict page limitation, so various aspects of the simulations had
to be omitted. I’ve now written a
Technical Report to provide these additional details. I’ve also made a preliminary estimate for
cost per peak Watt. # Note added 29 August: A few words on the
methodology for financial estimates might be useful. The output of the combined canopy-engine
system was considered in detail over many months, leading to estimates
mentioned in the post for 9 August.
These estimates aren’t infallible, but at least a lot of work went
into them. On the other hand, my
current estimates for financial metrics are quick and dirty. Based on the 9 August
estimates, suppose we have a 1,000 m^2 canopy-engine system that would
produce 65 kW peak and 74 MWhr/yr, all losses considered. Suppose the cost of land, earthworks, frame
and canopy is AUD 25/m^2 and the cost of the engine and balance of plant
(including any water treatment process) is AUD 1,000/kW. Then the capital cost is AUD 90,000 and the
cost per peak Watt is AUD 90,000/65,000 W = AUD 1.38/Wp. Another common financial
metric is the LEC (Levelised Energy Cost).
This requires additional assumptions on interest rates, payback
period, cost of maintenance, and cost of water. Further, the LEC must take into account the
debt/equity ratio and the tax deductibility of depreciation, interest and
operational costs. At present, my
estimate for the LEC is too preliminary to give here. I’m working on a spreadsheet that will
capture the factors mentioned above (and any others that I discover to be
necessary). Reference: Technical
Report 2010-4. 9 August:
Over the last two weeks, I re-worked the Wellington simulations to
account for the dependence of transmission and reflection coefficients on the
angle of incidence. (I used the
Mitalas & Stephenson correction, as mentioned in the post for 22
July.) I’ve also submitted the paper
to the peer reviewed section of the 2010 Australian Solar Energy Society
(AuSES) Conference. Yes, the
correction reduced the power output that I mentioned in my last post. The latest estimates are as follows (from
the abstract of the AuSES paper): “With canopy losses included but not engine
losses, the peak power output from the sample of 51 days that was simulated
was 89 W.m^2. The annual output was
estimated to be 104 kWhr/(m^2.yr).
Engine losses are expected to reduce these estimates, which are for a
horizontal collection surface, by 25-30%.” The AuSES paper was restricted to 10 pages
and many details of the simulations had to be suppressed. I’ll now prepare a Technical Report to
document the work thoroughly. Reference: Scientific
article [6]. 22 July:
I’ve now completed a full year’s simulations for the Wellington
data. For most of the year, the
simulations were based on four days per month at 30-minute intervals. A few months at the end of the simulations
were restricted to two days per month, since by that stage I was confident that
there was a strong relationship between power output (in kWhr/(1000m^2.day))
and insolation received (in MJ/(m^2.day)).
As anticipated at the half-way point (see post for 12 July), the
annual power output from the 1,000 m^2 canopy was estimated at 112 MWhr
(again noting that this estimate does not include engine losses). But I’m not completely happy with these
simulations, which I think are too good to be true. Specifically, I think the transmission of
sunlight through the canopy should include a dependence on the angle of
incidence, which I have so far not included.
At oblique angles, the amount of sunlight reflected is a greater
fraction of the inbound radiation than is the case for nearly normal
incidence. I suspect a sinusoidal
dependence on the sun angle, which I need to confirm. (# Note added 25 July: I’ve found some useful data and the effect
is NOT sinusoidal. An old technical
note from Mitalas & Stephenson, 1962, refers to original work by Fresnel
and gives good data.) The overall
effect will be to reduce the annual power output. My deadline for the conference publication
is ominously close, but I hope to be able to tweak the model to include the
features mentioned above and to meet the deadline of 8 August. I’d better get started! 12 July:
Today, I completed the first half of the Wellington simulations. General features of these simulations have
been described in previous posts. My
evaporation engine has now been simulated for the months of March-June 2009
and January-February 2010, with energy input coming from passive solar heat
collection under a 1,000 m^2 transparent insulated canopy. The simulations are for 30-minute intervals
for four days per month. The following
two conclusions are noteworthy: · The results show a clear relationship between power output (in kWhr/(1000m^2.day)) and insolation received (in MJ/(m^2.day)). This enables prediction of the overall performance over a full year. · Based on simulations so far, the annual power output from the 1,000 m^2 canopy would be about 112 MWhr. (Once again, please note that the simulations include canopy losses but not engine losses.) Reference:
This work will be presented at the 2010 Conference of the Australian
Solar Energy Society. My deadline to
complete simulations for the remaining six months and write up the paper is 8
August 2010. 22 June:
The Wellington simulations continue …
As described in earlier posts, I’m
simulating the performance of the evaporation engine over a full year at a
suitable inland location, namely the city of Wellington in inland New South
Wales. The engine is powered by
passive solar heat collection under a 1,000 m^2 transparent canopy with
insulation and low-emissivity coating.
Weather and insolation data has been obtained from the Bureau of
Meteorology. At regular intervals
during the day, I calculate the flow-rate through the system (comprised of
canopy plus engine) to optimise the power output. This takes time, and a huge number of
mouse-clicks, and my original goal of 15-minute intervals has been replaced
by 30-minute intervals. Also, I’m not
going to simulate every day; that would be just too much. Instead, I’m carrying out simulations on
the basis of one day per week. That
should give a representative sample. Full results will be presented at a conference
later this year. For the moment, let
me mention the best daily output so far.
This was 22 January 2010, a lovely day for solar energy. The maximum temperature was 38.6°C, with
12.5 hours of sunshine and 32.94 MJ/m^2 insolation on a horizontal surface. As an advantage for my engine, the relative
humidity was low. Shown below at
30-minute intervals are the inlet temperature (given by a spline
approximation to the data) and the optimal power output. The maximum power output predicted was 89
kW and the total output for the day was 657 kWhr. The temperature under the canopy peaked at
122°C, and typical canopy losses amounted to about 40% of the inbound radiation in the middle
of the day and 50% in the morning and late afternoon. A reminder – these results include canopy
losses but not engine losses, which are a topic of ongoing investigation.
Reference:
This is work in progress.
Results will be presented at the annual conference of the Australian
Solar Energy Society in December 2010. 11 June: I
have now revised my canopy heating model and applied it to simulate the daily
performance of the BEE. The model now
incorporates transmission, reflection and absorption coefficients that are
based on published data. These apply
for both the visible and infrared spectrum.
Also included in the model are molecular heat diffusion (both through
the glass cover and a diffusion zone immediately below it) and convective
heat transfer from the top of the glass cover. The simulations are based on the actual
angle of the sun above the horizon and ambient weather conditions, with the
latter provided by the Bureau of Meteorology. The figures below show results from 1 March
2009 at Wellington, 305 m above sea level in inland New South Wales. This, the first day of autumn, was a fine
and sunny day with 968 W/m^2 direct irradiance averaged over the day and a
total of 25.2 MJ/m^2 insolation on a horizontal surface. The humidity was also low, with the average
of the 0900 and 1500 vapour pressures being 732 Pa. At each quarter-hour interval throughout
the day, the flow-rate under the canopy was calculated to optimise the power
output of the evaporation engine. Other features of the simulation: · The canopy area is 1,000 m^2. The glass cover is 3 mm thick, below which is a 25 mm transparent convection suppression zone. Of the inbound solar radiation, 12% is reflected, 18% absorbed in the glass cover and 70% transmitted. The low-emissivity coating reflects 85% of infrared radiation emitted by the ground under the glass. The glass cover emits infrared radiation both up and down, and both are included in the heat transfer model. · The BEE expansion ratio is 1.9. · The output in the period from 0815 until 1800 hours is 441 kWhr. · 4.4 m^3 of water is evaporated to produce this power. (Comment: note that desalination of seawater by reverse osmosis requires approximately 4 kWhr per m^3; less energy than this is required to treat waste or brackish water. Also the average annual rainfall in Wellington is 650 mm, so that the average rainfall run-off collected per 1,000 m^2 is approximately 1.8 m^3/day.) · These results include canopy losses but not engine losses, which are the subject of ongoing investigation.
Reference: the methodology and full results
will be presented at the 2010 annual conference of the Australian Solar
Energy Society. 11 May:
I’ve finally discovered on the internet some experimental data for
transmission, absorption and reflection coefficients for glass, either clear
or with a low-emissivity coating.
These coefficients are shown as a function of wavelength, so the
results are applicable to both the sun’s spectrum and infrared
emissions from heated surfaces. This
will enable me to refine the heating models described in the last two
posts. That’s work in progress;
results soon. 21 April: I
have now simulated the optimal performance of the BEE based on passive solar
heating. The key point is the air
flow-rate. Increasing the flow-rate
decreases the temperature of the air, and so reduces the work available per
cycle. On the other hand, increased
flow-rate means less canopy losses and greater mass flow of air to the
BEE. Where is the optimum? These simulations involve a 1,000 m^2
canopy made of transparent material like ‘bubble wrap’, 20 mm thick with
thermal conductivity equal to that of air.
It is coated with a low-emissivity coating, ε = 0.2. The ambient air is 25°C at relative
humidity 53%. The temperature of the
injected water is 20°C. The expansion
ratio of the BEE is r = 1.8 or
2.0. The solar radiation is Σ = 350, 500, 650 or 800 W/m^2. The figure below shows the power output as a
function of flow-rate for the various values of r and Σ. At optimal flow-rate, the temperature at
the outlet of the canopy is less than 137°C for all the insolation
values. The simulations include canopy
losses but not thermal or mechanical losses in the BEE. At optimal flow-rate for the highest
insolation (800 W/m^2), the overall system efficiency is just under 10% and
the power output is 78 kW/1,000 m^2.
Actual losses in the BEE will reduce these estimates; by how much is a
matter of ongoing investigation. So,
you see the philosophy of the BEE in a nutshell – gather the heat cheaply and
passively, and then convert that heat energy directly to power without any
heat exchangers. It is all a matter of
how well the job can be done, and at what cost. And that’s why I need investors to help me
build a prototype.
Reference: Technical
Report 2010-3. 29 March: Over the past two weeks, I’ve been looking
again at heating of air under a transparent insulated canopy. My original motivation for invention of the
BEE in 2004 was based on the desire to convert heat energy in hot dry air
directly into power. The energy
conversion was to be achieved without heat exchangers, and the hot dry air
was to be obtained by passive solar heat collection. About five years ago, I developed a
simulation model for solar heating that included airflow, incident solar
radiation, and insulation of the canopy, but did not include the effect of
radiation losses. I have now included
the radiation losses. I think it
should be technically possible to design and build a canopy that will produce
air temperatures up to 140°C. Under
ideal conditions (strong insolation, good insulation, low-emissivity coating
on the canopy), theoretical power outputs of up to 75 kW should be achievable
from a canopy of area 1,000 m^2.
After allowing for inevitable losses, the actual power output from a
canopy of area 1,000 m^2 might be up to 40 kW. These matters have been written up in Technical Report 2010-2. 15 March: I’ve now finished Technical
Report 2010-1 on the effects of incomplete evaporation during
re-compression. This involved several
weeks of computer simulations. In
brief, the work reveals the trade-off between energy gained as an excess of
water injected versus the extra
energy required to purify and inject the water. I’m heartened by the results, which
indicate that the BEE can be run a bit faster than I expected. In other things: (1) Work
on the Business Plan has been suspended at a reasonably advanced stage of
preparation. To make further progress,
I need input from potential investors and collaborators. It’s a current priority to identify these
people. (2) I
still have some theoretical results from last year that need to be properly
documented – another Tech Report is required. 20 January: Thanks to all you folks who keep contact
with this web site. Based on the web
stats for the month so far, January might produce a record number of hits on
the site. Yes, I’m still working on my
inventions, both in a technical sense and commercialisation. As mentioned in the page for investors, the plan remains to
modify the existing experimental engine to test new concepts, and then to
proceed to a full-sized prototype.
Details will depend on preferences of industrial collaborators, still
to be identified. I’m currently
working on drafts for three Technical Reports and the Business Plan. In February, I’ll be presenting some
results to my mathematical colleagues in New
Zealand and then recharging my batteries with a short holiday. |
||||||||||||||||||||||||||||||
|
|
||||||||||||||||||||||||||||||
|
|
© Sunoba Pty Ltd 27 December 2010 |
|||||||||||||||||||||||||||||
