The sugar industry has long been a prime example of co-generation but this has usually been on some balanced basis with all of the power produced being used on site. Export co-generation is becoming more common however, frequently driven by government influence. This will be discussed in greater detail later in this paper. The statement is true of both the cane and beet sectors.

This paper examines some of the technical and commercial factors which influence the success or otherwise of cane sugar industry co-generation for power export, utilising bagasse and some auxiliary fuel to drive the business. To simplify the discussion, the word 'co-generation' will be used to cover the expression unless it is necessary to distinguish between co-generation with and without export. Similarly, 'power' is used to mean electricity.

TECHNICAL

Most, if not all, cane sugar factories were originally conceived on the basis of balancing the site energy demand with bagasse energy supply. This can be very easily represented by a single block diagram showing cane in and sugar plus molasses out but the difficulties of achieving balance when factors such as fibre %cane are so variable do not show up. In some factories this resulted in a permanent burn of auxiliary fuel while in other parts of the world there was a bagasse disposal problem.

It is of course possible to re-draw this picture to differentiate between the sugar operations and the boiler / power house operations. Note however that the concept is still the same: to balance the energy across the site. What then has to happen before one can export power as shown in Figure 1? Before answering that question it is important to understand what the power customer - usually the local utility - requires in the way of power.

There are potentially three types of export power:

• ad hoc - sold as and when available
• continuous - sold throughout crop
• firm - sold throughout the year

Therefore the following are required to achieve power export:

• release energy by improving efficiencies in power and factory blocks
• secure the power export by including bagasse storage
• secure the power export by improving site-wide reliability
• ensure firm power by arranging auxiliary fuel and a low grade heat dump

Steam pressures and temperatures at sugar factories have been rising steadily over the years with the average probably now about 30 bar g. As the steam pressure and temperature rise, the thermodynamic efficiency of the turbine part of the cycle also rises. In part, this is a direct result of Carnot's work which became incorporated in the second law of thermodynamics: the potential efficiency of a heat engine increase as the temperature difference between the high and low temperature reservoirs increases. However, more energy is required to raise the same amount of steam at higher conditions - or less steam is raised if fuel energy is fixed. This leads to less residual energy available for evaporation.

Figure 2 presents the results of changing the HP steam conditions of the simplest possible thermodynamic cycle in the form of a three dimensional histogram. In each case factors like available fuel energy, exhaust temperature and turbine barrel efficiency were all kept constant to show only the effect of the cycle conditions. Work energy and hence power produced is given on the vertical scale as a percentage of the total energy released from the HP steam. As the pressure increases, so does the work energy and similarly, as the temperature increases so does the work energy. Some of the results are replicated in full in Table 1 so that the effect on the evaporators can be examined. The results are for a notional 100 MWh of bagasse fuel energy.

The table clearly shows that the extra electrical power is produced at the expense of evaporation capacity even though exhaust enthalpy increases. This must equate to improving evaporator efficiency or suffering process problems. However, when one re-expresses the figures as percentage improvements over the base case condition of 31 bar / 380 °C, a gearing effect can be seen which shows the benefit of higher cycle conditions despite the drawbacks.

The above was an academic exercise in one sense: a potential solution to avoid the need to improve evaporator efficiency is to burn more fuel in order to deliver the same total energy to the evaporator and hence produce even more power:

One problem which should be addressed at this point is that of the perceived danger of higher pressure cycles. When a boiler explodes, whether at 31 bar or 61 bar, it will severely damage the surrounding area. It is true that higher pressures and temperatures require better operational control than lower conditions but modern materials of construction and our understanding of the design principles embodied in the current codes make the dangers probably more remote. Provided that the code is followed and that appropriate controls are adopted, there is no more danger in a higher pressure boiler than a lower pressure one.

The one temperature effect of significance to the boiler is the need for ever more exotic materials of construction, particularly for the superheater, as the HP steam temperature rises. Of greater significance is the effect of higher pressures and temperatures on the turbine. As the HP steam conditions increase a point is reached at which silica in the boiler water vaporises and carries over in the steam to deposit in the turbine as the steam cools. The only solution to this phenomenon is to drastically cut back on the permitted silica which means more sophisticated water treatment.

The practical limits for industrial boilers in the cane sugar industry seem to be around 100 bar and 500 °C although the wisdom of going above 80 bar is debatable.

Energy Release - Factory Efficiency

As discussed at the beginning of this paper, many factories were deliberately designed to be inefficient and offer several opportunities to substantially improve energy consumption. The two key ones are replacing inefficient, often single wheel turbines with electric drives and reducing the evaporator steam consumption, these being the main users of steam in the factory. In addition, changing from milling to diffusion also reduces the electrical power consumption of the factory although it also increases the evaporative duty.

Modern co-generation turbines have specific steam consumptions approaching half that of typical mechanical drive turbines. As an example of that, a recent project was optimised at 45 bar g and 465 °C at the boiler crown valve, assumed to be 45 bar a and 460 °C at the turbine. Exhaust pressure was 2.5 bar a. The turbines on that project consumed 6.7 tons/MWe, perhaps 6.5 tons/MW of mechanical power.

It can be seen that replacing perhaps 2 MW of drive turbines might release another 2 MWe for sale.

Increasing the number of effects on the evaporator can also have a dramatic impact on the availability of power for export although it does create other points to be considered. Using the same example project as above, a nominal 450 tch diffusion plant, the evaporative load was estimated to be 467t/hr. With a triple effect evaporator this would have probably been over 50 % steam on cane, for a quad it was 45 % steam on cane but with a quintuple effect unit only 41%.

Given that the bagasse is a constant, what does one do with the 'surplus' steam arising from adding effects to the evaporator station? There is a spectrum of possibilities to answer this question ranging from condensing it all to not generating it but storing the bagasse instead. The correct solution needs to consider the last of the four targets established in the introduction: 'ensure firm power by arranging auxiliary fuel and a low grade heat dump'. In high fibre cane areas with low power prices the correct solution may be to not make the factory too efficient and thereby hold down capital cost.

Clearly there is one other way of improving the overall site efficiency and that is by improving the quality of the bagasse presented to the boilers. It is important when considering bagasse quality to look at the total non-combustibles - ash and moisture - combined. There are therefore two possibilities, improving the moisture content and reducing the ash. Neither is particularly easy but the effect, predicted by the SMRI formula, can be important:

In the typical value ranges seen in cane factories, reducing the sum of the moisture and ash by just one percent increases the calorific value of the fuel by over two percent.

Energy Release - Power Station Efficiency

Just as balanced factories were designed to dump energy with inefficient evaporators and drive turbines, many of the boilers were also designed for energy dumping by leaving out any form of heat recovery equipment after the main bank. In addition, as modern turbines have lower specific steam consumptions than older units, so modern boilers are more efficient at converting fuel energy to steam. The fundamentals of physics still apply of course, what has changed is a reduction in energy loss, primarily in the flue gas.

One way of improving the power station efficiency is therefore to re-equip with modern equipment optimised for efficiency. This does not necessarily mean replacing the boilers but certainly means retro-fitting airheaters and economisers. A ready way of judging boiler efficiency is the final gas temperature. Adding a simple airheater will bring that temperature down to 180 to 200 °C and adding an economiser in addition to an airheater will bring it down to about 160 °C but at a higher capital cost. Even more expensive heat recovery arrangements can bring the gas temperature down to 140 °C but one is in the area of diminishing returns. Most modern co-generation stations operate at about 160 °C.

Many older boiler installations are not optimised and waste energy by not controlling the excess air used in combustion. Excess air, rather than the stoichiometric requirement, is used to ensure that all the fuel is burnt despite the less than perfect mixing in the furnace. However, excess air [and its accompanying moisture content] has to be heated and carries some of that heat energy up the stack. The extent of excess air is measured as the oxygen content of the final gas. Modern co-generation stations operate at less then 3% oxygen in the flue. The specific consumption of the turbine was discussed in the previous section. The specific production of older boilers was typically 1.8 tons steam per ton of standardised bagasse at 52% moisture content, 2% ash. A modern boiler with full heat recovery burning the same fuel will produce over 2.2 tons/ton of the same quality steam, a >20% improvement in efficiency.

There are also more subtle ways to improve the power station by making the thermodynamic cycle more complex. One of these was discussed in the paper by Magasiner presented at the 1995 AGM of BSSCT : to take an uncontrolled, medium pressure steam bleed from the turbine in order to pre-heat the feedwater and combustion air of the boiler. Initially this approach seems counter-intuitive because the turbine specific consumption increases. However, treating the boiler as a 'black box' shows that there is benefit gained because flow through at least part of the turbine is increased without increasing the large latent heat loss through the condensing part of the cycle.

Security of Production - Bagasse Storage

High availability is every factory manager's dream. With export power it becomes an absolute requirement, at least on the export side of the scheme.

The one specific point in this section relates to the bagasse supply. The bagasse flow from the mills [tandem or post diffuser drying mills] is subject to short term and longer term fluctuations and disruptions, sometimes completely outside of management control. Many sugar factories have lived with this problem, accepting that a mill trip will disrupt the steam supply until auxiliary fuel can be switched in. Export power co-generation cannot accept this situation if it is to be profitable so the mills and the boilers have to be de-coupled.

Converting the knives, shredder and mills to electric drives de-couples any MP steam consumption [which otherwise could have a dramatic effect on the turbine of the TA set when the mill trips] and the bagasse has to be de-coupled with a bagasse store. The capacity of the store and the mode of operation depends on factors such as the auxiliary fuel selected as much as on the expected frequency and duration of stoppages. If either oil or gas is selected as the fuel then switching is a relatively simple and quick operation requiring only a short term bagasse supply. Solid fuels, whether coal or some other form of biomass such as waste wood chips, require a much greater store of bagasse.

Another part of the bagasse storage equation flows from the exhaust steam ratio of the factory and the solution adopted to the 'surplus' bagasse mentioned in the previous section.

A steam efficient factory operating in a high fibre cane area has the potential to generate a large quantity of surplus bagasse if it is designed around the steam consumption of the process. Skeldon II will be a good example of this where one can expect about 40% steam on cane in a country with cane in the 16 to 17% fibre range. Keeping with the 450 tch model used previously and the same HP steam conditions, such a factory could generate a gross ± 27 MW from the steam requirement of the factory but would still have a surplus of ± 70 tons of bagasse per hour! For comparison, in condensing mode that is equivalent to another 35 MW of power.

One solution is to raise more steam than is required by the process and hence generate more power. The surplus exhaust [or lower pressure] steam has to be condensed which can be done by deliberately backing off the steam efficiency of the process or by installing a separate condenser. The correct condenser solution will depend on the off-crop power generation requirements and the selection of turbine type. How far one goes with this approach will depend on the price and availability of the auxiliary fuel and the cost of storing the bagasse. The optimum might be to partially increase the steam flow and extend the bagasse generation into off-crop before switching to the auxiliary fuel but solutions involving only burning bagasse in off-crop have been proposed.

This is clearly a problem which does not arise in low fibre cane areas.

Security of Production - Plant Reliability

Because the power station and the remainder of the factory are so closely integrated it is not adequate to consider the availability of the power generation plant in isolation: any aspect of the operations has the potential to disrupt power export. As we will see later, this can carry severe penalties. It is the nature of these penalties which will dictate the available capital expenditure and hence scope of measures to assure security of production.

Because improvements in the thermodynamic cycle are likely and power will probably be required in off-crop as well as crop, the power station side of the equation is often resolved because different and therefore new, more reliable equipment is installed. There are examples - notably in Central America - where this is not necessarily true because second hand plant has been used in upgrading the power station.

On the factory side of the equation, the de-coupling from the station in terms of bagasse supply and MP steam demand has already been discussed earlier. This leaves the exhaust steam, the corresponding condensate return and the power demand as the major connections for short term disruption. In the longer term the reliability of bagasse supply must also still be a consideration.

Without a condenser in the cycle, be this the factory process or a stand alone unit, it is not possible to run the power station. Even allowing for the off-crop condenser, if the evaporator stops drawing exhaust there is likely to be an imbalance as the factory continues to draw power. Luckily there is usually sufficient buffering capacity in juice and syrup storage [or it can be added at relatively low cost] to allow some notice of the problem to the station. This is not true of many milling trips of course, hence the desire to de-couple bagasse and MP steam.

One of the problems which has been encountered in various parts of the world is condensate return, both in terms of quality and quantity. Earlier discussion highlighted the problems of feedwater quality as steam conditions are increased. It has been found that if the cycle does not go over about 46 bar a and 420 °C then quality is not a problem provided that strict separation of the first effect condensate is practised and only this is returned to the station. Above this level a better solution is probably the installation of a 'steam transformer' as described by Magasiner, completely de-coupling the station from the factory. The transformer is a single effect evaporator which, although it carries a small pressure drop penalty, guarantees the safety of the station in terms of feedwater quality and minimises the need for blow down.

Of course, limiting the return to first effect condensate does give a quantity problem to the station which has to blow down in order to keep the system within limits on tds and other quality parameters. Neither does it recognise the clean water generating capacity of the factory: cane is, after all, more water than anything. Sometimes, at least at lower steam conditions, this can be resolved by using V1 condensate which has been kept in isolation pending quality testing.

Ultimately, however, the project will only be successful if the whole site is running reliably and not being forced to switch frequently from crop to off-crop conditions. One cannot build in precautions for rain days and other natural events which disrupt cane supply but otherwise the whole system from harvest planning to maintenance QA will probably have to be upgraded to cope with profitable power export.

Firm Power

In most locations it is only when the facility is able to offer export power throughout the year that a reasonable sales price can be obtained for the power exported. Because this is directly related to the utility's ability to avoid installing its own generating capacity, the optimum solution is to export the same amount of power out of crop as in crop. A careful balance is therefore required to replace the fuel and condensing provided by the factory and to reduce power output by the amount consumed by the factory in crop.

The latter point also reinforces the argument for converting any turbine drives to electrical as this increases the factory power consumption and reduces off-crop power generation when one is paying for the fuel.

The factors to be considered are therefore the auxiliary fuel, its handling and storage, the off-crop cycle and the appropriate turbine system for that and the alternator(s) to achieve the differing conditions.

The auxiliary fuel can also serve as a standby fuel during crop so needs to be selected with care although price will usually dictate the right solution. Many of the existing stations have been built around the use of a solid fuel: coal or wood but some have used fuel oil in one form or another. Solid fuels involve more complex handling and storage facilities than fluid fuels.

One hears of several proposed power stations based on the use of green wood as a renewable fuel source. The problem with any green biomass is the high alkali metal ash content which leads to severe fouling and special forms of corrosion, particularly around the superheaters. This forces up the capital cost of the boiler in order to manage fouling and in terms of materials of construction. The one advantage of such fuels is that they can be claimed to have low environmental impact when compared to fossil fuels.

In high fibre cane areas one could make an argument for not using auxiliary fuels but spreading the bagasse burn over the full year. The practicalities of doing this have been discussed, as has the possibility of partly doing this in order to optimise the export power. A practical example of this is discussed in the next section.

Some of the existing stations have been built around pass-out turbines with full condensing capability at off-crop conditions. However, condensing turbines with controlled pass-out are expensive and can only be designed for maximum efficiency at one particular set of operating conditions. The conceptual design of the system therefore has to consider whether to take advantage of the extra power from full condensing or whether to dump the energy in the exhaust out of crop, perhaps operating at atmospheric exhaust rather than some higher value such as 2.5 bar a. One way of achieving this, at least in part, is to use the heat transfer surface of the evaporator out of crop.

Another solution is to adopt multiple TA sets with a mix of exhaust and pass-out turbines. There is a further capital cost penalty to doing this however.

Technical Summary

Technically, the problem is more complex than appears at first sight. Perhaps the simplest way of summarising is by way of a worked example, the project used earlier in this paper.

The location is a high fibre cane area, typically 16%, and the factory is a 10,000 tcd [450 tch] unit with a diffuser and hence high imbibition [350% on fibre]. Internal power consumption of the site was 15 MW when the power station was included, the drying mills being electrically driven.

The conceptual design started with a relatively simple model of the site operations with crop length, holidays, planned maintenance and unplanned stoppages all included. This, plus the bagasse analysis, the energy balance of the process and similar factors allowed a power export model to be created. The model predicted export power for whatever HP and exhaust conditions were selected and had provision to add turbines, make the cycle more complex and vary the auxiliary fuel.

As it happens, the auxiliary fuel decision was simple because the factory was built almost on top of a gas field.

In the end a cycle operating at 45 bar g and 465 °C at the boiler crown valve was selected as being a compromise between efficiency and lifetime cost with a nominal steam flow of 207 tons/hr or 46% steam on cane, some 11% more exhaust than the process demands.

The excess was to be condensed in a dump condenser and would provide the primary control for absorbing the inevitable fluctuations in exhaust demand. A secondary let-down station was also provided but was envisaged primarily for start-up and shut down and to allow for some process activity in the event of the grid being unable to accept the export power.

The high fibre cane meant that no economiser was required on the boilers and the specific steam production could be kept down to about 1.9 tons/ton. Even so, at the conditions selected the bagasse was accumulating at over 30 tons/hr and, after allowing for start-ups, stoppages and similar, reached 75,000 tons for use during off-crop.

Two TA sets were to be used, each being driven by an exhaust turbine and rated at 16 MW. During crop 31 MW would be generated, of which 16 would be exported while in off-crop nearer 17 would be generated and 16 exported because the turbine would operate at atmospheric exhaust pressure. In order to operate during the off-crop the dump condenser had to be augmented by using the first effect of the process evaporator as a condenser. This required carefully planning around the number of evaporator bodies to ensure that maintenance could take place without disrupting power sales.

The overall result from this factory was therefore as shown on Figure 5. The bagasse accumulated during crop and was then used up in the following 8 weeks before power generation switched to burning gas, at which time the net revenue from the power sales fell off as the cost of the auxiliary fuel was taken into account.

Was it the correct solution? Some of the technical possibilities considered and their consequences included the following:

Of course, none of this is worth doing unless the project is profitable and that depends as much on the commercial aspects as it does on the technical ones.

COMMERCIAL

We have already seen that the commercial aspects of the project have a direct impact on the technical aspects. This section will look at the project from an engineer's commercial point of view.

There are three factors required for the commercial success of any project: market, market and market. In the case of a co-generation project this means that there has to be a sufficient demand for the power at the right price. The right price is, in turn, a function of the operating costs which primarily comes down to the cost of the auxiliary fuel. The implication of the whole subject is that there is a willing - or at least cooperative - power purchaser, usually the local utility company.

The three types of export power have already been identified:

• ad hoc - sold as and when available
• continuous - sold throughout crop
• firm - sold throughout the year

There are exceptions to this generalisation of course. One such is where the utility uses hydroelectric power for some of its supply. When the rains stop this may no longer be available but continuous power supplied through the dry period sugar crop could avoid the utility having to install an alternative source of power.

Demand

Power demand may constrain output more than the technical considerations, particularly where the local grid is unsophisticated as might well be the case for many sugar mills. In these situations the demand could be limited to that in the local area.

Provided that the utility is a willing partner in co-generation, absolute demand is likely to be less of a key parameter than the ratio of the station's supply to total demand. The French DOM's of Reunion and Guadeloupe are prime examples of locations where bagasse fired co-generation provides a substantial portion if not all of the power for the islands. In these situations there needs to be greater security of supply than locations where the station is a small part of the total supply system.

Accordingly, the stations are divided into two separate but cross-linked 30 MW systems which allow either boiler to drive either TA set. Compare that to the stations proposed in Australia where one boiler will drive one TA set but the station capacity is insignificant in the grid scheme. [The implications for the security of sugar production is another matter although correctly designed modern equipment should not present unacceptable risks.]

One of the commercial aspects of demand which is often overlooked is the potential demand from the estate itself, particularly from irrigation and social facilities. Where these demands are currently provided by the utility, the cost of reticulating factory power is often quoted as a reason for not following that route. However it is possible to set up a 'wheeling' arrangement with the utility where the utility's power lines are used to transmit the company's own power rather than the utility's power.

Price

The actual costs of power production and hence the price achieved, varies widely around the world. South Africa for instance has very large coal fired power stations on top of coal fields which are barely sub-surface and is therefore able to sell power for about 2 US¢ per unit. At the other end of the scale, countries without any fossil fuel or natural resources such as hydro-electic potential have to import fuel and generate at some considerable cost. Jamaica for instance sells power at 11 US¢ per unit.

Both countries have bagasse fired co-generation potential but clearly South Africa is unlikely to see any such stations for the foreseeable future. It is possible to take a first view of whether a project will be viable just by considering the results of previous projects. In most circumstances, if the power sales price is not at least 5 US¢ per unit then it is probably not worth studying the project. There are of course exceptions to that rule of thumb, often brought about by political interference in one form or another.

Obvious forms of political interference include special prices for renewable energy [India is a good example of that in the bagasse fired co-generation context but even the UK has such schemes] and heavily subsidised capital expenditure [such as in the French DOM's]. More subtle forms include 'carbon credits' where the government of an industrial country will grant planning permission for new boiler plant [or extend the operating permission for existing boilers] against remote investment in renewable energy projects. The Canadian government is supporting this approach strongly and there is talk of a 'Bank of Carbon Credits' under the auspices of the World Bank where such credits can be traded.

Contractual Aspects

In general both parties to a co-generation contract negotiation want a long term contract and there is little difficulty on this score. What is difficult however, is predicting inflation over the long term and hence determining a price escalation formula.

One approach which is frequently adopted is to relate the escalation to the price of an appropriate fossil fuel. In one sense this is very logical as it will reflect the auxiliary fuel cost of the project over time but the counter argument is that the primary fuel is 'free' and the utility therefore wants to share in this benefit of renewable energy.

The other key aspect of the contract is the level and application of penalties. The utility is relying on the project in order to avoid installing generation plant of its own. Whilst it can understand and plan for maintenance outages of limited duration [albeit requiring these at times which are usually most inconvenient to the factory] plus longer term annual shutdowns, it cannot accept outages with short or nil term notice without risking the grid coming down. It therefore wants to include severe penalties for such events, often with escalator clauses for extended periods of outage.

On the other hand the factory management knows from past experience [although probably with older equipment] that power generation is not reliable and does not want such penalties incorporated. Both sides need to be educated with perhaps a phasing in of the penalties.

CONCLUSIONS

A project to establish a power export co-generation business is both technically and commercially complex. The optimum solution for each will be unique and the result of much hard work.

However there is no doubt that the impetus will increase in the future with more and more bagasse fired units coming on stream. Bagasse is, after all, one of the two most widely available biomass fuels and the sugar industry has been co-generating for many decades.

What will that future hold? The ultimate goal has got to be practical gasification of the bagasse which will completely change the thermodynamics of co-generation. Gasification of bagasse will allow the use of a combined cycle where the gas is fired in a gas turbine and then the turbine exhaust is used to raise steam for a traditional steam cycle. This approach, which has a combined 'fuel to power' efficiency twice that of a steam cycle, is already practised in the beet sugar industry where fossil fuel is used rather than renewable energy and gas is often available at attractive rates.

There seem to be two main problems with gasification:

• bagasse has a low bulk density and is difficult to feed through an airlock into the pyrolisis chamber
• the blades on the gas turbine become fouled with condensates