Solar Steam Engine Selection Process

{{This article needs editing. First, list all the candidates in an introductory paragraph. Second, perhaps organize this article by all types of practical implementations: (1) piston engines, (2) bladed turbines, (3) bladeless turbines, (4) slotted disk turbines, (5) Stirlings, (6), Rotary engines, (7) modified IC engines, (8) true flash piston steam engines, (9) Brayton cycle engines, plus many others.}}

Our solar steam engine prototype has several criteria to meet to be true to both OSE philosophy and to become a competitive economic option. It must be readily replicable, using common, environmentally friendly principles and accomplish its function in the simplest manner possible. Let it also be said that successful implementation of an economical solar energy prime mover is holy-grail energy type stuff, countless hobbyists, venture capitalists and many others more qualified than us have “broken” their teeth trying to make these project happen. Where many have failed though is in chasing highly efficient, highly optimized, large scale centralized complex solutions for our complex centralized industrial model, we are after rugged, simple, replicable post-industrial technology, technology for the global village. Using the OSE selection criteria, the steam engine appears to be our most desirable engine for converting solar energy to electricity. The remainder of this document outlines the reasoning that led to this conclusion.

In choosing the solar powered prime mover I reviewed two main heat engine cycles and multiple engines within these cycles as candidates for OSE’s first attempted solar heat engine. This comparison is intended to show my reasoning for selecting a Rankine cycle steam engine as OSE’s solar powered prime mover. It also should illustrate that as machining capabilities evolve, there will certainly be more engines to choose from depending on the criteria that we want to optimize (i.e. efficiency, reliability, simplicity, etc.).

The first cycle that was reviewed is referred to as a Rankine cycle and in its most common form is a H20 steam cycle. Approximately 80% of the world's electric energy is produced via this cycle. As with all heat engine cycles the engine extracts work by expanding a hot pressurized fluid to its low rejection state. The ultimate achievable efficiency of any heat engine cycle is given by Carnot's rule where: eff = (T_high - T_low) / T_high. However, in a Rankine cycle, this upper limit is never realized. The limitations preventing obtaining ideal efficiencies are inherent in the working medium, in order to get the steam back to the heat source either fresh liquid or condensed liquid need to be reintroduced into the cycle. This requires pumping energy to push into the pressure of the boiler, this process consumes energy and is not isentropic like the ideal Carnot cycle. There are other fluids used for this cycle as well, recent interest in harnessing industrial waste streams has brought attention to organic fluids, which have low boiling points. The organic Rankine cycle however would not be of interest to the OSE solar engine project as the fluid would offer nothing in terms of added efficiency and would compromise our principle of using readily available materials.

There are two main prime mover options within the Rankine H20 steam cycle. Currently the most common is the steam turbine. The steam turbine is attractive from both an efficiency and reliability standpoint. Since steam is converted directly to rotational motion, stress is minimized on the blading and shaft and energy density is maximized. Thermally too, the steam turbine is advantageous, as multiple stages can be added so that the steam can be efficiently expanded to its lowest point. However, this steam mover did not become an option in the industrial revolution until the 1920’s in large part because of precision and rigorous engineering that are required to make them. The reliability of this machine also has a few caveats, mainly that there is a lot of difficult engineering involved in a successful design. To avoid self-destructing vibrations, a successful turbine needs to operate outside of its natural frequencies. This requires accurate calculation of the blades' natural frequencies and rotor natural frequencies. The turbine must be successfully balanced, both statically and dynamically. A minor imbalance or unexpected natural frequency excitation will result in quick damage to the turbine bearings and blading. Additionally, turbine blade manufacturing techniques require CNC milling machines. For these reasons, the construction and engineering of a steam turbine is not yet possible at OSE, However, in the future, as OSE’s machining and engineering capabilities progress, this prime mover should be revisited.

The steam engine is the second most common Rankine cycle prime mover and is also the engine that ushered in the industrial revolution. Steam engine efficiencies are historically of low efficiency, averaging around 10%, and although they are typically much lower than the turbine, this is not due to the engine itself but the lower temperature and pressures typically used in a steam engine. In fact, towards the end of their era, numerous engines were reliably achieving Rankine/Carnot efficiencies approaching that of the steam turbine. With our projected boiler steam conditions of 150psig and 550 deg F., we could historically expect 10% overall cycle efficiency for a single-cylinder, non-condensing engine and 13 % overall cycle efficiency for the more complicated double expansion engine with a condenser. This would be 50% and 65% of Carnot cycle efficiencies, respectively. Without thermal data on a steam turbine operating at the same thermal conditions, it is hard to compare the two, however it is known that optimized turbines operating at higher steam pressures and temperatures typically achieve less than 70% of Carnot efficiency. This then places the steam engine on relatively equal thermal footing for these steam conditions. It should be noted though that at much higher temps and pressures the steam turbine runs into cylinder lubrication burn-out problems, a problem not encountered in the steam turbine. The construction of a steam engine is difficult but does not require machining beyond OSE’s current capabilities. Roughly 50 lbs of cast iron will need to be cast into a cylinder, which may pose the most difficult problem. Due to the low speed designs, overall balance of crank and flywheel are orders of magnitudes less critical than a high-speed turbine. Valve and governor design and manufacturing will require a fair amount of engineering work, but again these problems have been solved and recorded by countless engineers from the turn of the century. Overall, these workhorses are proven and well demonstrated, and though certainly not initially simple to manufacture, we can lean heavily on countless successful designs of the past if we chose this option.

The second and last cycle that was reviewed is the Stirling cycle. This cycle much more closely approximates the ideal Carnot cycle. Mainly this is due to the use of a gas as the working medium and thus reducing the non isentropic losses associated with the condensation portion of the Rankine cycle. Additionally the closed cycle use of air as a working medium eliminates both the complex valve arrangement of the steam engine and the inherent dangers of steam. Despite the ideal safety and theoretical thermal properties of this engine, its cult of hobbyists and developers have yet to succeed in a low cost entry-level Stirling engine. To achieve high efficiencies and a modest size, it is required to pressurize the cylinders and preferably with an ideal gas. This greatly adds in complexity and cost. Furthermore to obtain the inherent thermal advantages of this engine, a high performing regenerator is needed. This device acts as a heat exchanger to store energy during the expansion stroke and keep from rejecting it to the low temperature discharge side. In practice, it has proved difficult to make a compact efficient regenerator, and few Stirling engines achieve >60% of Carnot efficiencies. These problems though are mainly due to pursuing a compact highly efficient Stirling engine for mobile applications. One alternative is to look at slow speed large cylinder, low pressure devices. This approach was tried in the S5. This engine was designed for burning agricultural wastes and is 20% efficient at transforming applied heat to electricity. The engine could presumably be retrofitted to accept solar input, however the initial construction and subsequent maintenance are reported to be fairly difficult. Most concerning to us at OSE is the very large cylinder, which is a result of using low pressurized air as the working medium. Manufacturing and machining of this cylinder will not be friendly to relative novices. Therefore, this engine with the advantage of using a safe working medium (air) and the ability to accept a high input temperature, should be reviewed further when our machining capabilities are more mature.

There are two additional variants of the Stirling engine that probably deserve to be mentioned as potential candidates. Thermo-acoustic engines operate essentially as a Stirling cycle. One end of a resonant chamber is heated while the other is cooled, this creates a standing longitudinal sound wave that is linearly oscillating a diaphragm which has a magnet mounted to it. The magnet is then moved through coils to act as a generator. I know of none in commercial manufacturing and have not seen performance statistics on it. Presumably some tuning is required to achieve desired oscillation speed and amplitude. Development of this engine would then require extensive fluid study, not necessarily global village material, yet.

Finally the Stirling fluidyne pump perhaps deserves mention. This pump is typically not thought of as having electrical generating capabilities but the utter simplicity of it deserves some attention. Alternatively, fluidynes are referred to as a liquid piston heat engine. The displacement of the water column can then be used to operate a check valve pump or possibly be fitted with a cylinder and crank for rotational motion. The demonstrated efficiencies of this pump are low, with the highest reported value being 7%. To my knowledge it does not seem that people have pushed this engine towards high temperatures and therefore higher efficiencies. If this was done there would be some obvious technical challenges that would need to be overcome. Perhaps some of the most advanced research on the fluidyne in this area is being conducted by Thermofluidics ltd. It is certainly interesting as it presents the ultimate in simplicity, but it remains to be seen whether they can be directly coupled to produce mechanical power and therefore is not an option at all, yet.

Therefore, in conclusion the steam engine appears to be our best compromise of thermal efficiency, ruggedness, reliability and ease of manufacturing.

Various Internet Discussions Types of Stirling Engines The Stirling cycle can be implemented in practice through various types of engines. A breakdown of the key parts of a typical engine follows. The key parts of the engine are the pistons, connection rods, the crankshaft assembly, the heat source and heat sink, and the regenerator. All engines contain at least one power piston, which must have very low friction and near-perfect sealing for satisfactory engine operation. The engine output per cycle is the net work that the power piston does on the working gas over a cycle (i.e. one rotation of the engine shaft). The linear motion of the power piston is transformed into rotational motion by a drive mechanism. Usually the drive mechanism consists of connection rods. However, other configurations such as Ross yokes are possible1. The three basic types of Stirling engines are alpha, beta and gamma. Figure 3.1 contains basic schematics of each type. Alpha engines have two power pistons, while beta and gamma engines have a power piston and a displacer. Beta engines have a power piston with a coaxial displacer; while gamma engines consist of a power piston and displacer in separate cylinders. The three basic types are well described in the literature [1,4], and we will outline only their main characteristics here. Figure 3.1.a) A typical alpha-type Stirling engine. b) A typical beta-type Stirling engine. 1 The Ross yoke drive mechanism is discussed at http://www.ent.ohiou.edu/~urieli/stirling/engines/engines.html 11 source: http://www.ent.ohiou.edu/~urieli/stirling/engines/engines.html Figure 3.1. c) A typical gamma-type engine. Each of the three engine types has certain advantages and disadvantages. The primary disadvantage of alpha type is a requirement for perfect sealing for two pistons. Also, the configuration of the two pistons at an angle to each other can be cumbersome for construction, especially for demonstration engines. Beta engines offer potential significant advantages, especially in terms of efficiency and size, but are technically complex. The requirement of a coaxial displacer and power piston makes for difficult machining of the engine. This requirement also increases friction in the engine because the displacer necessarily slides in and out of the power piston. Moreover, the mechanical drive mechanism used to convert translation of the pistons into rotation of the output shaft is a difficult design problem to say the least. Finally, while a beta demonstration engine would be interesting, the coaxial piston motion and regenerator position make it difficult to see Stirling cycle principles at work. Gamma engines are best equipped for educational purposes. With two parallel cylinders attached to a common crankshaft (see Figure 3.2 below), it is relatively easy to discern the four stages of the Stirling cycle, and to study the functions of the displacer, power piston and regenerator in each thermodynamic process. 12

There is are two additional variants of the stirling engine that probably deserves to be mentioned as a potential candidate. Thermoaccoustic engines operate essentially as a sterling cycle. One end of a resonant chamber is heated while the other is cooled, this creates a standing longitudinal sound wave that is linearly oscillating a diaphragm which has a magnet mounted to it. The magnet is then moved through coils to act as a generator. I know of none in commercial manufacturing and have not seen performance statistics on it. Presumably some tuning is required to achieve desired oscillation speed and amplitude. Development of this engine would then require extensive fluid study, not necessarily global village material. Finally the sterling fluidyne pump perhaps deserves mention. This pump is typically not thought of as having electrical generating capabilities but the utter simplicity of it deserves some attention. Alternatively fludynes are reffered to as liquid psitons heat engine, see diagram. The displacement of the water column can then be used to operate a check valve pump or be fitted with a cylinder and crank for rotational motion. The demonstrated efficiencies of this pump are low, with the highest reported value being 7%. However I am only aware of this unfinished high temperature approach. If temperatures of 550 deg can be used as heat input this pump may convert to mechanical motion at efficiencies close to our other options.

Sorry to be a party pooper, but forget Stirling engines for generating power from biomass in developing countries, unless your target market is a small group of enthusiasts with lots of time, fuel and technical expertise. We started our company in the early 1980s with the same purpose in mind as the correspondents to this blog. We saw the Sunpower ST-5 in action in Bangladesh, and although it was a brave effort, it was clearly not going to provide the desired solution – too large, inefficient and expensive, difficult to maintain, and very difficult to get the heat in without sooting up. I later visited the factory in Chennai (ex-Madras) where about 100 ST-5 engines were produced, and talked to the Indian Government agency who put them into villages. A couple of years later, 1 was still working, run by a real enthusiast – otherwise, they just gave too many technical and economic problems, and the factory was shut down. Resources were not the only problem. So we decided to develop our own engine. We built a 1 kW and a 7.5 kW engine in the UK, and the latter in particular ran very well (it had 5 internationally granted patents) with a Beale number of 0.15 straightaway (highest recorded after many years development is 0.22). However, we could see that there were still too many problems to be solved, particularly in the heat exchangers and regenerator, for the engine to be commercialised, whether in developing or industrialised countries. We have come up with heat exchanger solutions using new technology from outside the Stirling industry, and are working on a new engine with a different cycle that is simpler than the Stirling cycle. A few fundamentals will put everyone’s ambitions in context. The Stirling cycle is at first sight very simple, but putting it into practice it is extremely complex. Companies have been trying to commercialise the modern Stirling since the late 1930s, including the US Navy, Ford, General Motors, Philips, the major Japanese industrial groups and many, many others. None has succeeded: Whispertech has come closest to it in the niche market of auxiliary CHP systems for yachts. The UK, Germany and the Netherlands are field testing Stirling-based domestic CHP systems: the UK’s Carbon Trust has just published a very thorough study of these trials and shown that with present technology efficiencies (electricity out/fuel in) are mainly in the 4-8% region. This is admittedly with fairly high cold end temperatures, but even so, these efficiencies are extremely low. Do not be fooled into thinking efficiency is not important in developing countries. It is important if you are going to use wastes and other biomass fuels. Our experience in South and South East Asia is that what may be a very low or even zero-cost waste suddenly becomes valuable once somebody finds a use for it. If you are going to use a fuel grown specifically for power generation, then in most parts of the world you will want to use as little land as possible – again, efficiency is important. The highest recorded Stirling engine efficiency is about 60-65% of Carnot, from an extremely large and sophisticated United Stirling engine in the 1980s. More modern small engines achieve way below 50%. To expect 80% of Carnot from an engine to be built and used in developing countries, as suggested by one correspondent, is, I fear, unrealistic. Another major constraint on the rural biomass Stirling is getting the heat into the engine. Direct combustion of the fuel will, even if it is suitably sized and with low moisture content, tend to foul the heater tubes. Technical University of Denmark minimise this problem with very large tubes, but the penalty is high weight and costs, and high internal volume, which reduces power output. In our view gasification is a better option, but again, this is difficult and expensive on a small scale, although there are reports of successful biomass gasification for diesel engines in India. Which brings me to the question: what to do instead. The simple although often unpalatable answer is to use a diesel engine, even if fuel is expensive. Diesel can be obtained wherever a truck can get through, which is nearly everywhere. Why foist technical ideas on poor people in developing countries that have not been proved and used commercially in our own developed countries? If we want to help with local energy problems in this sort of way, then liquid biofuels used in a diesel engine are a better bet, and will continue to be so unless and until something better is developed. We are working on it, but without any illusions. The same applies to Solar dish Stirlings. A US report in 1989 (Dish Stirling Commercialisation Workshop, Electric Power Research Institute, Atlanta) noted that all that the technology required for commercialisation was an efficient, cost effective Stirling engine: it was right and it is still waiting. Drummond

-- Drummond Hislop, December 10, 2007