Photovoltaics, Batteries, Tractors, Horses, and Biofuel

 
by Andrew R.B. Ferguson This is a draft paper prepared for the Optimum Population Trust Journal, published in the UK. For more information see www.optimumpopulation.org Abstract. As fossil fuels dwindle, there will have to be a change to different ways of farming, and of providing transport for essential commercial goods such as food. Engineer John Howe’s experiments show up the limitations that apply to using photovoltaics and batteries. Using this combination, it might be possible for a farmer to farm 20 hectares. However, equally as likely, once fossil fuels become scarce, this technique may prove too expensive. An alternative is for farmers to make use of some of their productive land to grow biofuels, using which they could carry out cultivation using typical internal combustion engine tractors. It is shown that 20 hectares could be farmed using soy oil in place of diesel, but there would be a need for about an additional 5 hectares to be farmed to produce the biofuel needed. In other words, 20% of the total land used would have to be utilized for producing the fuel used on the farm. Difficulties of expense in running small-scale biofuel processing plants might be a barrier there too. The task of cultivating the 20 hectares with horses is also considered. The major problem with that is the additional land needed. The extra land needed is estimated at 7 hectares, about a third of the land which is being cultivated. In previous issues of the OPT Journal, it has become apparent that photovoltaics are not suitable for contributing more than a small proportion of power to an electrical grid, and that their low load factor makes them unsuitable for producing hydrogen by electrolysis (in both cases unless storage becomes feasible). However, some families — mainly those which are unable to connect to the grid — use only batteries in conjunction with photovoltaic (PV) modules to provide themselves with power at home. This suggests the possible use of batteries, recharged by PV, for other vital requirements. At least it is worth investigating whether batteries combined with PV modules could do some useful work on the farm, because when fossil fuels become scarce, maintaining farm output is of paramount importance. One clear advantage is that the batteries, when not being used for the relatively short periods of time that plowing, harrowing, and harvesting are being carried out, could be used to carry out transport tasks and to power the farmhouse throughout the 24 hours. Those are points to return to later. Engineer John Howe has done excellent ground work on the subject, with the aim of determining the practicability of this approach. () It is his work that — with his guidance — I am about to recount, with special reference to the experiments he has made with the fourth of a series of test vehicles that he has built, namely an 8N Ford Tractor. It is a tractor of substantial size. The original version, before modification, weighed 1130 kg. With batteries, the modified version weighs 1680 kg. The original liquid fuel version was rated at 22 horsepower. In the modified tractor, the battery pack consists of ten DEKA 9C12 220 Ah batteries. Each battery is rated at 75 amperes with a capacity of 150 ampere hours (Ah). At the 12 volts provided by each battery, each battery can thus usefully store 150 x 12 = 1.8 kilowatt hours (kWh). The batteries are connected together in series to produce 120 volts. Operating at the same 75 amperes, i.e. maximum power output, they can operate for 2 hours, thus providing a total capacity of 18 kWh. The power output is 75 amps x 120 volts = 9 kW (electrical), equal to 12 horsepower. About 5 horsepower is needed just so the tractor can move itself in soft ground; the remaining 7 horsepower being needed to do the required work. That sounds like a lot of horses to have in hand, but of course the electrical energy does not convert with perfect efficiency into propulsion power. And we may note that the original tractor was apparently able to offer a surplus pulling power of 17 horses. The energy value of the total 18 kWh battery capacity is equivalent to 0.5 gallons of gasoline. However electric motors are more efficient than internal combustion engines, so in terms of ability to do work, the 18 kWh is about equal to 1.5 US gallons, 5.7 litres, of gasoline. Even doing fairly heavy work, like 16-inch plowing (this means plowing over a lateral distance of 16 inches, 40 cm), or double-disk harrowing in new ground, at 2 mph (3.2 km/h), the tractor will have sufficient energy stored in the batteries to plow, or harrow, 0.2 hectares (2000 square metres), which it will do in 2 hours. So far that is moderately encouraging. Heavy ground might slow them up, but in moderate ground, a team of two horses would plow a hectare in about 8 hours; so in 2 hours they would manage 0.25 ha. Thus over a two hour period the rate of plowing with 2 horses is comparable to the battery powered tractor. The problems arise in the recharging of the batteries. If 14 square metres of module were used, then at a typical peak output rating of 150 watts per square metre (W/m2), the peak output would be 14 x 150 = 2.1 kW. At the 120 volts used, that is 2100 / 120 = 17.5 amps. That is below the maximum charging rate, but, as will become apparent later, when several sets of batteries are used sequentially, there is no point in recharging faster, so we will consider using only 14 square metres of module with one set of 10 batteries. How long will it take to recharge the 10 batteries? Even if we assume summer in the mid west of the USA, where insolation is fairly high — about 270 W/m2 — the 14 m2 of panel will provide only 399 watts,{} or 399 x 24 = 9.58 kWh per day. Thus to recharge the batteries with a full 18 kWh would take 18 / 9.58 = 1.9 days. Recognizing that some days will be less sunny than average, it is clear that 2 days should be allowed for recharging. To plow only 0.2 hectares in 2 days (i.e. at a rate of 0.1 hectares per day) is obviously inadequate. It is hard to see anything like modern life continuing if a farmer cannot cultivate as much as 20 hectares (a square of 450 metre sides), so let us use that area as a basis for analysis. With this one set of batteries and modules, it would take 200 days to plow 20 hectares. That is excessive. So as to be able to work continuously for 8 hours a day, we need to consider solving the recharging delay problem by using a sufficient number of batteries and PV modules. If we had 8 sets of batteries and 8 x 14= 112 m2 PV modules that would allow 8 x 2 hours = 16 hours of work to be done, without recharging. The important point is that the two days of work would be covered by charged batteries, and those two days would allow sufficient time for the first set of batteries to have been fully recharged, so work could continue at the same rate, day in day out, indefinitely. Working 8 hours a day, the tractor would achieve 0.8 hectares per day. Thus the 20 hectare field could be finished in 20 / 0.8 = 25 days. That seems like an acceptable proposition. Let’s just do a rough cross-check on that, so as to see that the energy captured equals or exceeds the energy used. In two days, 8 periods of plowing will be done, each of 2 hours. Since each two hours requires 18 kWh That is a total of 8 x 18 = 144 kWh. In two days, the 112 m2 of modules will capture 153 kWh. That is slightly above 144 kWh, but then one cannot expect every two-day period to have average or above average insolation, so some margin for lower insolation is advisable. Now to look in more detail at the comparison with horses. David Pimentel has provided me with background information about this, supporting his own experience with data from Morrison (1946). We have noted that a team of two horses might plow at a rate of 0.125 ha/hour, so 20 hectares would take 160 hours. A team of horses can plow for 8 hours a day according to Morrison (1946), so that would take 20 days of work. That is actually better than the battery powered tractor; but there is a rule of thumb, which says that a team of two horses is adequate for contributing to the management of 10 hectares. Thus we should really consider two teams to be required, with two plowmen of course, cutting the time to 10 days. Obviously a substantial improvement over the tractor. In fact since oxen plough at about two-thirds of the speed of horses, with a similar set up, they would achieve the task in 15 days, which is still faster than the tractor. Then there is the question of the area of land needed for the horses. Using data from Morrison (1946), a team of two horses would require 0.8 ha of pasture; 450 kg of corn, needing 0.06 ha; and 1.2 ha of hay land. That works out at just about 1 hectare per horse. So for the four horses that we are considering, 4 hectares would be needed. But another factor to account for is that horses need time to mature into working animals, so allowance needs to be made for replacements for the 4 horses; ideally there should be reserve horses too. With a small group like this, John Howe suggests that the land needed for replacement and reserves might be as much as a hundred percent, but for a larger group of horses fifty percent would suffice. Too much precision is obviously not possible, so let us pencil in an additional 3 hectares, making a total of 7 hectares for the two teams with backup. So about a third as much land is needed to maintain the horses as is being cultivated. Some confirmation for that estimate is given by Britain’s experience in the 1900s, for Britain only has about 18 million hectares of ecologically productive land, and the following information from Clive Ponting (2007) is therefore broadly in line: In the early twentieth century Britain had a horse population of about three and a half million…. In 1900 Britain’s horses consumed 4 million tonnes of oats and hay every year (taking up the production of about 6 million hectares of land). The land needed by horses provides one good reason for using PV and batteries rather than horses. Another is that it might take a long time to revive the skills of a sufficient number of plowmen. Yet another is that the PV system offers several advantages in addition to plowing and harrowing, as will be mentioned later. But before counting the benefits, we need to look at the costs inherent in installing 112 square metres of module and 80 batteries. The 112 square metres of module, rated at 150 W/m2, would have a peak capacity of 16,800 watts. At a slightly optimistic fully installed cost of $5 per peak watt, that would cost $84,000. The batteries cost about $200 each, so 80 would cost $16,000. Assessed on a 20 years basis, and assuming the PV modules to last 20 years, and including four lots of batteries, the total cost would be $148,000, or $370 per hectare per year (although that might be somewhat reduced by the trade-in value of the used batteries). While that cost looks bearable (at about 5 cents per kg of corn), it is based on the present cost of energy. The cost of lead acid batteries, PV modules, their installation, and the cost of the tractor, are all related to the cost of energy. Without the benefit of fossil fuels, energy would be much more expensive, and the increased cost of energy would impinge on raw materials (which need energy to mine and refine), and energy is needed for manufacturing the product and delivering it to where it is wanted. So the whole concept, even if possible while fossil fuels are available, looks like being too expensive to bear when fossil fuels are no longer available. Thus we hardly need to pause to consider other problems, like the cost of the lead used in the batteries when such use become extensive (thieves are already stealing lead off roofs in the UK), or note that the farmer would already be impoverished by the high cost of nitrogen fertilizer, which needs a lot of energy, or pause to reiterate the positive sides, such as that 112 batteries each of 1.8 kWh capacity could store 202 kWh, which would be a wonderful thing for the farmhouse during the winter months. Moreover even when the tractor was working 8-hour shifts, there would always be one set of modules that were not in use (because the batteries are in the tractor) and the output from the PV modules could be used to supply current directly to the farmhouse. It is a good prospect, but the big question is whether it will be possible in a world without fossil fuels. Without doubt, that world will be a materially impoverished one, and perhaps the most telling question to ask is whether this system is one that could be adopted in a poor country such as Cuba, because without the benefit of fossil fuels, that is the sort of situation that humans will be experiencing. The answer to the question must surely be that it is quite likely that it will not be possible, so let us now turn to biofuels, to see if they could offer a more attractive alternative than returning to the use of horses. The biofuels alternative The task of fermenting maize (corn), and distilling it to remove about 95% of the water to produce ethanol suitable for use in internal combustion engines, is a task that could not be undertaken at small-scale farm size. However, it is possible to envisage the necessary processing to produce a suitable biodiesel from an oil crop being done at a small scale. Of all the oil crops, soybeans are probably the best choice for a renewable energy world, as they need little or no additional nitrogen (the most energy intensive fertilizer). Thus it would seem sensible to choose soy oil as the biodiesel to evaluate. What we need to see is whether it offers a better alternative than either horses, or batteries recharged by PV (which as mentioned may not be economically viable). As Table 1 (see end of article) shows, the gross output per hectare is 524 litres of soy oil per hectare. However liquid inputs are required for the cultivation of soybeans. For reasons that will be amplified later, we need to budget for 23% of the soy oil output to be used as inputs, so this leaves 524 * 0.77 = 403 litres (106 US gallons) per hectare as “useful” output for cultivating other crops. If we allocate 4.5 hectares of land to growing soybeans, then the “useful” output of biodiesel would be 403 * 4.5 = 1814 litres per year. Turning now to how that “useful” oil is to be used, it depends on the crop exactly how much liquid fuel is required for cultivating and harvesting, but to get a general idea let us take the figures that are appropriate to soybean production, namely about 90 litres per hectare. Therefore aiming to cultivate the same 20 hectares that were previously considered, the biofuel requirement would be 1800 litres. So the 4.5 hectares of soybeans provide sufficient “useful” biofuel for the 20 hectare task, but more land is required for non-liquid fuels needed in the production of the soy beans. So far we have only considered the liquid inputs needed to produce the soybeans. In addition to the liquid input, there is a substantial non-liquid energy input required. The liquid input amounts, as a first step, to about 17% of the gross biofuel output,4 while the remaining inputs are about equal in energy content to 117% of the gross biofuel output. For these reasons, the whole process is an energy loser, insofar as it requires in total, as inputs, 134% of the energy contained in the gross output of biofuel (Pimentel and Pimentel, 2007). We can calculate that additional energy requirement in the following manner. The gross biofuel yield was mentioned above as 524 litres per hectare, so the 4.5 hectares we are using to produce biofuel would yield a gross energy equivalent to 524 * 4.5 = 2358 litres of soy oil. The energy density of soy oil is about 34 MJ/litre, so at 117%, the total additional (additional to liquid) energy demand for producing the soy oil from 4.5 hectares is 2358 * 34 * 1.17 = 94 gigajoules (GJ). The gross yield of an energy crop like short-rotation woody perennials might optimistically be put in the region of 160 GJ per hectare per year. Thus we need about 0.6 hectares to supply the 94 GJ of non-liquid energy. But harvesting this area and transporting the wood to where it is needed will also use some liquid energy, probably at least 5% of the gross wood output from the 0.6 hectares, i.e. 94 * 0.05 = 4.7 GJ. That is equal to about 138 litres of biofuel. As a proportion of the total gross output of liquid, that is 138 / (524 * 4.5) = 6%. That 6% liquid output needs to be added to the 17% used in providing the straight forward liquid inputs for cultivation of the 4.5 hectares of soybean, and thus a total of 23% of the output will need to be used as liquid input (the figure drawn out of the air earlier without explanation!). Conclusion We can now sum up the position like this. Using a tractor and biofuel, 5.1 hectares of land is needed to provide the inputs to farm 20 hectares. Of this, 4.5 hectares needs to be arable land and the remaining 0.6 hectares needs only to be suitable for growing an energy crop such as short-rotation woody perennials. This is much less than the 18 hectares needed to provide horsepower. Moreover the biofuel powered tractor would plow at a rate of about 0.15 hectares per hour, so working an 8-hour day, it could cover the 20 hectares in about 17 days (an improvement over the 25 days taken by the battery powered tractor). However, even the biofuel operation might run into cost problems, depending on how simple and cheap can be made the process of producing the soy oil from the soy beans. The battery plus PV option has the advantage of not using ecologically productive land, but as already mentioned it may not be viable economically in a world that is impoverished by shortage of fossil fuels. The above theoretical calculations probably give a fairly realistic idea, but they cannot be held with any certainty without trying them out in practice. It should be a priority to explore every alternative, as John Howe is doing with batteries and PV, so as to get a good idea of whether the theoretical calculations are in the right region, because we need to know ahead of time what situation we will be in as fossil fuels become scarce, so that we can make useful estimations of the population that could be supported without their benefit. It is likely to be a small fraction of the present world population, not only because of these limitations on food supply but because of the need to provide energy for keeping warm, and for manufacturing processes. References Giampietro, M., S. Ulgiati, D. Pimentel 1997. Feasibility of Large-Scale Biofuel Production: Does an enlargement of scale change the picture. Bioscience Vol. 47, No. 9, October 1997, pp. 587-600. OPTJ 4/2. 2004. Optimum Population Trust Journal, Vol. 4, No 2, October 2004. Manchester (U.K.): Optimum Population Trust. 32 pp. Archived on the web at www.members.aol.com/optjournal2/optj42.doc OPTJ 7/2. 2007. Optimum Population Trust Journal, Vol. 7, No 2, October 2007. Manchester (U.K.): Optimum Population Trust. 32 pp. . Archived on the web at www.members.aol.com/optjournal2/optj72.doc Morrison, F.B. 1946. Feeds and Feeding. Ithaca, NY: Morrison Publishing Co. Pimentel, D, Pimentel, M. 2007. Food, Energy, and Society. Third edition. Boca Raton, Fl; London; New York: CRC Press. ISBN 978-1-4200-4667-0. CRC Press website http://www.crcpress.com Ponting, C. 2007. A Green History of the World: the Environment and the Collapse of Great Civilizations. London: Vintage books. 452 pp. Table 1. Inputs to produce and transform 2668 kg of soybeans (the average yield of 1 hectare) into 480 kg, 524 litres, of soyoil, as a suitable substitute for diesel fuela Quantity/ha/yr MJ/ha/yr Inputs Soybeans 2,668 kg 15,679 Electricityb 130 kWh 1,401 Steam 2,714 MJ 2,714 Cleanup water 322 MJ 322 Space heat 306 MJ 306 Direct heat 884 MJ 884 Losses 603 MJ 603 Stainless steel 5 kg 318 Steel 10 kg 494 Cement 27 kg 213 Methanol (needed for transesterification)c 105 L 927 Total 23,861 Gross output of soyoil = 524 litres = 17,816 Liquid inputs = 17,816 x 0.23 = 4,098 So output of ‘useful’ ethanol 13,718 = 403litres soy oil/ha/yr. So power densityd = at 34 MJ/litre = 403 x 34 MJ = 13.702 GJ/ha/yr = 0.43 kW/hae Notes to Tables a. This table is derived from two tables (pp. 323-325) in Food, Energy, and Society, 3rd edition, by David and Marcia Pimentel (2007). The first of those two tables shows the energy analysis for the cultivation and harvesting of the soybeans, but only the result is shown here (first row of the table). b. The original tables explain and reference the derivation of the energy inputs, but that information is not repeated here. The original second table dealt with 1000 kg of soybeans. The figures here have been adjusted to deal with the 2668 yield from one hectare in one year, and also for one missing item (later advised to me by David Pimentel), namely the 200 ml of methanol needed per litre of soy oil for transesterification of the soy oil (see next Note). c. The energy for the 200 ml of ethanol needed per litre of soy oil was calculated on the basis of the inputs needed to provide the methanol from wood, using the data in Giampietro et al (1997). 1.73 kg of wood is needed per litre of methanol, so here the amount of land for growing wood is negligible. d. This 0.43 kW/ha in only a “partially net” power density, because only the liquid inputs have been subtracted. To allow for the non-liquid inputs, an additional (134 - 23) = 111% of the input would be needed, i.e. 1.11 x 17.816 GJ = 20 GJ/ha. Using the rather optimistically high figure of 160 GJ/ha/yr to provide that non-liquid energy from woody crops would require an additional 0.13 ha. The effect of needing 0.13 hectares more land would be to decrease the power density to = 13.702 [GJ/ha/yr] / 1.13 = 12.1 GJ/ha/yr = 0.38 kW/ha. Even that is not the end of the story, because some liquid inputs will be needed to harvest and transport the wood. Even if we put that as low as 5% of the 20 GJ yield, providing that required liquid fuel will reduce the “useful” output by 29 litres (= 0.03 GJ/ha), and thereby reduce the 0.43 kW/ha figure to 0.40 and the 0.38 figure to 0.35 kW/ha. The last figure is the fully net power density, and amply indicates the basic problem with biofuels: low power density. e. The equivalent power density for ethanol from Brazilian sugarcane is 2.9 kW/ha (OPTJ 7/2, p. 21), but there are many reasons why farmers cannot grow sugarcane and produce ethanol from it. One important reason for choosing soybeans is that they need little or no nitrogen, and nitrogen is the most energy intensive fertilizer. BACK TO HOMEhttp://www.optimumpopulation.orghttp://www.members.aol.com/optjournal2/optj42.dochttp://www.members.aol.com/optjournal2/optj42.dochttp://www.members.aol.com/optjournal2/optj72.dochttp://www.members.aol.com/optjournal2/optj72.dochttp://www.crcpress.comhttp://www.crcpress.comHome.htmlshapeimage_3_link_0shapeimage_3_link_1shapeimage_3_link_2shapeimage_3_link_3shapeimage_3_link_4shapeimage_3_link_5shapeimage_3_link_6shapeimage_3_link_7