Power Control Systems

Pressure level modulation

Variation in the mean pressure level of the working fluid is the most widely used and best-known control system for power regulation in Stirling engines. It was used to some extent on the more sophisticated air engines of (he nineteenth century and was early adopted as the principal power control system for Philips engines (van Wecnan 1947).

Constant speed system (Philips) Meijer (1959a) gave an excellent description of the system used for power regulation on the first rhombic-drive engines using hydrogen or helium as the working fluid. The objective was to maintain a constant speed of operation of the engine and was achieved by increasing or decreasing the pressure level of the working fluid in the cylinder as the engine speed changed, causing the engine to accelerate or decelerate to the specified speed.

The system is shown schematically in Fig. 10.5. It operated as follows. The governor (1), driven by the engine shaft ensured that, at the nominal speed of operation, a certain oil pressure was maintained in pipes (2) and (3). If the load torque increased, causing the engine to slow down, the governor raised the oil pressure in the pipes with the result that valve (4) in the feed device (5) opened. High pressure hydrogen then passed from the storage reservoir (6) through the feed device (5) and valve (4) to the engine cylinder (8) via the non-return valve (7).

Increase in the pressure level of the working fluid caused an increase in the power output from the engine and consequently an acceleration in speed. Injection of additional working fluid to the engine continued until the engine speed had climbed back lo the original value. At the set speed

pressure level ol Ihe working fluid (after Mcijcr I959n).

the oil pressure in pipe (2) controlled by governor (I) was sufficiently low for the valve (4) to close and admission of additional gas was therefore curtailed.

If a decrease in the load torque occurred it resulted in an increase in the speed of the engine. As a consequence the governor decreased the oil pressure in pipes (2) and (3) with the result that valve (9) of the regulator (10) opened. This permitted gas to flow from the engine cylinder via the non-return valve (I 1) through valve (9) to the auxiliary compressor (12) and hence back to the storage reservoir (6). If the maximum cycle pressure exceeded the pressure of the storage vessel sufficient gas would flow intermittently from (8) to (G). A sharp decrease in the load however would require the auxiliary compressor to function.

Release of gas from the cylinder reduced the power produced thereby resulting in a slow-down of the engine. This continued until the power output had declined to a compatible level with the new value of load torque at which time valve (9) closed and furthei release of gas from the engine ceased.

The need to compress the fluid back to the higher pressure of the storage reservoir restrained the flow of gas from the cylinder to the amount that could be handled by the compressor. As a consequence the response of the engine to sudden decreases in the load was substantially longer than the response to sudden increases in load.

Short-circuiting (Philips)

To improve the engine response to sudden load decrease a supplementary power control system was incorporated in the speed regulator. The new control was remarkably simple. Two or more spaces in the engine were interconnected so as to 'spoil' the pressure characteristic of the individual spaces operating separately. This supplementary system called 'loss regulation' or 'short-circuiting' caused both a change in the phasing and a reduction in the amplitude of the pressure variation in the engine cylinder and clfected a reduction in power output.

In the unit described by Meijer, the two spaces interconnected were the single cylinder and the 'buffer space' below the piston. In a rhombic-drive engine, the space below the piston (the buffer space) is normally charged with working fluid to the same mean pressure as the engine. This relieves the gas dynamic forces acting on the piston and adjusts the hermetic seal problem to one of scaling the small diameter piston rod rather than the large diameter piston. The buffer space experiences a cyclic pressure variation that is approximately inverse to the cylinder pressure variation.

In addition to opening and closing valve (9) as described above the regulator (10) also actuated a slide throttle (16) which offered the gas a direct connection between the cylinder and the buffer space. This was more or less equivalent to a leak past the piston and resulted in a virtually instantaneous reduction in the amplitude and a change in the phasing of the pressure variation and hence in the power output of the engine. The system resulted in a decrease in engine efficiency and foi this reason was termed by Meijcr (1959a) 'loss-regulation' but more recently has come to be called 'short-circuit control'. It was a useful device for it made the engine response virtually instantaneous to sudden changes in load. Moreover, it permitted the use of a very small auxiliary compressor to pump the gas back to the storage compartment.

Variable speed system (Philips)

A similar control system was described by Neelen et al. 11971) foi the four-cylinder rhombic-drive Philips Type 4-235 traction motor. In this case the engine must operate at variable speeds as well as variable loads.

A simplified diagram of the system described by Neelen <*f al. (1971) is shown in Fig. 10.6. The valve S corresponds to the valve 4 in Fig. 10.5 and the valve D to valve 9. Similarly the short-circuit valve SC corresponds to the bypass valve 16. The complete functional diagram of the system described by Neelen el al. for an automotive engine is shown in Fig. 10.7.

Depression of the accelerator pedal caused the supply valve S to open and for the valves D ami SC to remain closed. The cylinder pressure was

variation in the pressure level of llic working fluid (after Neelen ri al. I 1 ).
Philips Stirling Engine Cooler
Fig 10 7. Functional diagram foi torque (pressure) control for Philips Type 4-235 four-cylinder rhombic-drive Stilling engine.

connected in a closed feed back loop so that the valve S closed when the cylinder pressure attained a value proportional to the displacement of the accelerator pedal. Release of the accelerator pedal opened valve D and working lluid passed from the engine back to the storage reservoir.

Depression of the brake pedal caused the dump valve D to be closed and the supply valve ,S' and the short-circuit valve SC to be opened. Hie degree of opening of the SC valve was determined by the position of the accelerator pedal.

The 'loss regulation' feature of this system was so effective that the engine could be used for braking purposes. Neelen el al. (1971) gave the work diagrams reproduced in Fig. 10.8. These show the output work diagram of the engine under full load conditions and the corresponding diagrams with the 'loss-regulation' short-circuit valve open. When partly open the engine produces some power but at a reduced level. When the short-circuit valve was fully open the engine absorbed power and could therefore be used in braking. Neelen el al. (1971) forecast that a maximum braking torque of 60 per cent of the rated engine torque would be possible. In one test an engine output of 7.44 N m (176 fl lb) changed to a braking torque of 140 N m (100 ft lb).

Other test results achieved a response time from no load to full load within 0.3 seconds. With the short-circuit valve operating a similar response lime (0.3 seconds) was obtained for sudden release of the load. Without the short-circuit valve operating it took the compressor 30 seconds to reduce the pressure from the full load to the zero load value.

Stirling Engine Philips
FiO. 10.8. Work diagrams lor Philips I'ype 4-235 four •cylinder rhombic-drive Stirling engine illustrating the control achieved by the loss-regulation short-circuit feature.

The above system, developed for multiple-cylinder rhombic-drive engines, has been carried over to the double-acting Siemens engines now universally in favour, and van Bcukering and Fokker (1973) have briefly discussed relevant control systems. Postma et al. (1973) have indicated that the power control system for the Philips/Ford type D.A. 4-215 swashplate engine is essentially identical to the system described above, A similar system was incorporated in the conceptual study for a small engine carried out by Phi lips./ Ford for the Department of Energy (Kitzner 1977a).

Variable speed system (United Stirling)

A schematic diagram of the power control system used on United Stirling engines is reproduced in Fig. 10.9. It can be recognized as virtually identical to the above system. To increase power, the control valve is moved to the right so that gas (lows directly from the reservoir to the engine. Hallare and Rosenqvist (1977). in a discussion of the control system, have revealed that a timed supply system is used which admits additional hydrogen to the engine cylinders only when the cycle pressure is near the maximum value. Gas flow into the cylinders without a timed system resulted in an undesirable torque drop during increase in the pressure.

Decrease in power is accomplished by moving the control slide to the left, thus releasing fluid from the cylinder back to the gas reservoir and

1 Hydrogen storage

2 Hydrogen compressor

3 Control valve

<t Compressor short-circuiting valve

Fig. L<J.l». Schcmntic diagram of power control system on United Stirling engines.

Short circuiting

1 Hydrogen storage

2 Hydrogen compressor

3 Control valve

<t Compressor short-circuiting valve

Fig. L<J.l». Schcmntic diagram of power control system on United Stirling engines.

Short circuiting also short-circuiting the various cylinders lor rapid response. The hydrogen auxiliary compressor of the United Stirling system is described (Hallare and Rosenqvist 1977) as an oil-free single-stage double-acting compressor with piston rings acting as suction valves. The displacement is 10 cm3 (0.61 in3) and the pressure ratio is 10: I. To unload the compressor during the increase of power and steady state conditions the suction and pressure sides of the compressor are connected to each other by a compressor short-circuit valve.

General Motors system

General Motors invested considerable effort in the development of control systems during their decade of intensive work on Stirling engines. Percival (1974) has indicated that on General Motors engines, temperature and torque were always independently controlled. Pressure modulation was principally used for torque or power control.

The small generator sets made by General Motors for the U.S. Army. GM Type GPUI, 2 or 3, required control of the speed to close limits. Stability was to be maintained at 3600 revolutions per minute to within ±10 revolutions per minute. Speed droop was not to exceed 90 revolutions per minute and the surge limit for sudden changes in load was set at

216 revolutions per minute. The recovery time for 100 per cent load change was limited to 6 seconds.

By 1967 the GPU-3 system was capable of holding the stability at ±5 revolutions per minute, the droop did not exceed 10 revolutions per minute, the surge limit was met and the recovery time was reduced to two seconds. Despite this achievement Percival writes feelingly that:

'From the standpoint of reliability, however, the entire speed governing system was a constant source of trouble until nearly the end of the programme. The hydrogen compressor was perhaps the major problem in the beginning. Il was incorporated into the base of the crankcuse of the GPU-3 as an extension of the displacer piston rod in Ihc form of a hydraulic plunger. Hydraulic pressure activated a diaphragm compressor which eliminated the need of a sliding or rotating seal. This made servicing more difficult. The hydraulic plunger required precision machining and was subject to binding. In retrospect, it would have been better to mount an experimental compressor outside the engine and drive it front a gear, with a break-away coupling, or from a belt. On the other hand, an outside compressor requires a good seal to prevent hydrogen leakage.

Another item which often stopped endurance tests was failure of the small (fin (1.27cm) dia.) hydrogen check valves and main control valve -usually the seats were pounded out or distorted sufficiently to leak. The hydrogen control valve was actuated by hydraulic pressure delivered by the speed sensing governor which was mounted on the crankcasc and gear driven*.

In 1964. the governor system had 5 separate valving units and 10 adjusting screws; bv the end of 1965. it had 2 valving units and one adjusting screw.

Results of endurance testing of the GPU-3 at GMR in 1967 showed that the hydrogen compressor had failed twice in the 1537 hour run and the governor hydrogen valve had failed four times.

In 1969, the GPU-3 at GMR was operated on a more rigorous 500-hour test, equalling a military qualification lest. In order to meet military requirements, a 'certified parts list * for the package was established so that all parts were like the engineering drawings. This defined exactly what was being tested and prevented casual substitution of components which would have caused the test to lose significance. At the conclusion, the maximum overhaul life was extended slightly but under more rigorous conditions to 560 hours from the previous 553 hours.

The longest run with no service was extended to 525 hours from the previous high of 196 hours. The longest run without stopping was extended from 159 hours in 1967 to 235 hours. In all. the engine was stopped four times, all caused by building safety interlocks and in no way connected with the GPU operation. The limit of 560 hours was caused by the hydrogen compressor—a small valve assembly failed lo function properly. However, the hydrogen check valves and governor control valve were in excellent condition.

Aii alternative system for compressing hydrogen was investigated briefly in 1961. It was based on electrolytic generation of hydrogen and diffusion through palladium tubes. Pressures to 7.9 MN/m" (1150 lbs per sq in) were maintained inside the tubes; but the concept was abandoned when piston rod seals were found to seal hydrogen better than expected.

Pressure-amplitude variation

An alternative power control system for Stirling engines was described by Aim el al. (1973) in an account of developments at United Stirling. Power control was achieved by means of variation in the amplitude of the engine cylinder pressure excursion. The system is illustrated in Fig. 10.10. A number of different gas bottles cast into the engine crankcase could be put into direct communication with the working space of the engine by a series of valves operated by the cycle pressure amplitude.

To reduce engine power, one or more of the valves was opened so that the volume of the gas bottle associated with it became part of the dead volume of the working space. Increase in the dead volume decreased the volume compression ratio V,,,.,,/ Vnill, and also reduced the amplitudes of the pressure variation as shown in the pressure/time diagram given in Fig. 10.11.

The gas bottles were of different sizes and the valves could be sequenced so that the power level could be progressively reduced in small

Fio, Id II). United Stirling system for power eonlrol by variation in the amplitude of the cyclic pressure curve.

Fio, Id II). United Stirling system for power eonlrol by variation in the amplitude of the cyclic pressure curve.

Solar Stirling Engine
Fig. 10.11. Cyclic pressure characteristic of a Stirling engine with power control by variation in the engine dead space volume (pressure amplitude variation).

incremental steps approximating to the smoothed curve shown in Pig. 10.12.

The system of power control by pressure-amplitude variation was developed by United Stirling as an alternative to the system preferred by Philips of control by pressure-level adjustment. Aim el al. (1973) pointed out that the Philips system required a rather complicated and expensive compressor to pump the hydrogen working fluid back to the pressure vessel.

Furthermore, because the compression process required a considerable time, it was necessary to resort to the complementary short-circuiting system and thereby sustain a decrease in engine thermal efficiency. With repeated unloading of the engine, as for example, in a city bus, a substantial efficiency loss would accrue.

To support this contention the figure reproduced in Fig. 10.13 was given by Aim el al. (ll>73). This shows the thermal efficiency for Stirling engine as a function of speed at full load and half load with three different control systems. At half-load the engine with short-circuit power control has only half the thermal efficiency it has at full load. With mean-pressure or pressure-amplitude control the decrease in efficiency is much less pronounced. No details of the engine or any numerical data were given in support of this figure.

It was interesting subsequently to observe thai United Stirling appeared to abandon the pressure-amplitude control system in favour of a return to the Philips pressure-level control (Hallare and Rosenqvist 1977). No

Fig. III.12. Power ouipui as n (unction oi dead spacc in :i Stirling engine with power control by pressure amplitude variation.

reasons for this change were given and the shortcomings of the pressure-amplitude system are unknown.

A possible explanation may be that, in practice, the decline in engine efficiency with the pressure-amplitude system was greater than anticipated. Some support for 1 his may be found in data presented in the important paper by Neelen et al. (1971) following experiments by MAN/MWM on a 7 kW (111 hp) Stirling engine to establish the effect of

Engine speed

Fid. I0.L< Comparison of Stirling engine efficiency as a iunclion of speed nl full load and half load conditions with thtcc ditferenl conliol systems.

Engine speed

Fid. I0.L< Comparison of Stirling engine efficiency as a iunclion of speed nl full load and half load conditions with thtcc ditferenl conliol systems.


Dead space (per cent)

FIG. 10.14. Effect of dead space on the power and thermal cdlcicncy of a Stilling engine thermal. The calculated and measured values aie compared as obtained in tests on a 10 hp Stirling engine at MAN/MWM (alter Neclen et al. 1971).

Dead space (per cent)

FIG. 10.14. Effect of dead space on the power and thermal cdlcicncy of a Stilling engine thermal. The calculated and measured values aie compared as obtained in tests on a 10 hp Stirling engine at MAN/MWM (alter Neclen et al. 1971).

Measured dead space volume. The dead space of the engine was enlarged artificially and the result, shown in Fig. 10.14, was given as the percentage decrease in power output and thermal efficiency of the engine as a function of the dead space. According to Ncelen el al. (1971) the power curve declined as predicted by calculation but the efficiency declined at a much greater rate than predicted for reasons that had lnot yet been clarified'.

Phuse-angle variation

In a Stirling engine the volume variations of the expansion space lead those of the compression space by the phase angle or, one of the principal design parameters of the engine. Variation in the phase angle is one possibility for engine power control. The power output of a Stirling engine as a function of phase angle is approximately sinusoidal in form, as shown in Fig. 10.15.

At zero phase angle (point A in the figure) the volumes ol decompression and expansion spaces vary exactly in phase. The change in total system volume is a maximum and the range of the cyclic pressure excursion is a maximum. However, if we assume isothermal, or adiabatic, processes in the compression and expansion spaces there is no work output from the engine because the pressure—volume diagram is the single line a-b shown on the work diagrams at A and li in Fig. 10.15. The pressure in the engine cylinder is the same on the downstroke of the v '—l JS3_fc- t



Fig. 10.15. Power output of a Stirling engine as a function of the phase angle between volume variations in the expansion ami compression spaces.

piston as il is on the upstroke of the piston. There is no cyclic flow of working fluid through the system.

At point C, with a phase angle of 3.14 radians (180°) the converse is true. Here the volumes of the compression and expansion spaces are exactly out of phase so that the variation in the total system volume is at its minimum possible value. II the swept volume in the two spaces is the same, the variation in total system volume is then zero. The flow of working fluid between one space and the other is maximum. The range of the cyclic pressure variation is small because it is due solely to the change in the mean temperature as the fluid moves between the hot and cold spaces at constant volume. Again the work output is zero because the pressure volume diagram is the single line 'c-rf' as shown at point C on Fig. 10.15.

At any phase angle between 0 and 3.14 radians (0 and 180°) the volume variations lead those in the compression space and the cyclic pressure and volume changes will result in a work diagram as shown at point IJ in Fig. 10.15. If the expansion space is hotter than the compression space the work diagram will be positive and a surplus of work will be available al the shaft to drive the engine and an external load. The net work output will reach a maximum value at a phase angle of about 1.57 radians (90°),

If the phase angle is between 3.14 radians and 6.28 radians (180° and

360°), points C and If in Fig. 10.15, a similar situation will prevail but now the work diagram will be negative as shown at F. An input of work will be required to drive the engine and the maximum value occurs at a phase angle of about 4.71 radians (270°). The direction of heat flow will be reversed. Heat will be transferred from the cooler to the working Huid and from the working fluid to the heater. In this situation the engine is operating as a heat pump, taking in heat at a low temperature and rejecting it at a higher temperature.

Of course if there is no power input available at the shaft to drive the engine as a heat pump it will simply stop and then run in the reverse direction as a prime mover, taking in heat at high temperatures and rejecting it at a low temperature with a positive power output.

Power control by phase-angle variation is characterized by instant response and represents an extremely convenient way to provide a rapid reversing engine. The system was adopted by the Electromotive Division of General Motors for an 590 kW (800 hp) Vee 8 Stirling engine intended as the propulsion motor for coastal surface vessels where good manoeuv-ering capability was required. The engine had a piston and displacer in each cylinder but they were connected to separate drive shafts. The separate shafts were coupled by sun-and-planet gears so that movement of the sun wheel caused a change in the phase angle between the piston and displacer motion and. hence, a change in the phase angle between the expansion and compression spaces.

The calculated power and efficiency curves for this engine were given by Percival (1974) and are reproduced in Fig. 13.9. Only one bank of four cylinders of this engine was in fact constructed for development work before General Motors abandoned piston-displacer engines in favour of Siemens double-acting engines. In 1967 the author witnessed the operation of this four-cylinder half-engine at the Electromotive Division works at La Grange, Illinois. 'Hie response of the engine was remarkable to change in the phase angle by simple adjustment of the sun wheel. It was said at the time that the engine could be reversed in less than a revolution (but only when the stress office engineers were not present!).

Power control by phase-angle variation is obviously not applicable to double-acting engines where the phase angle is limited to the value 6.28/N radians (3607N) where iV is the number of cylinders. Double-acting engines can be conveniently reversed by switching the interconnected cylinders so that in effect a phase change of 3.14 radians (180°) is introduced. Reversal of the cylinder connections can be done by a simple slide valve fitted to the cold side of the engine. It is not known if the system has, in fact, been used in double-acting engines but the potential for reversing engines was early recognized by van Weenan (1947).

Stroke variation

Power output from a Stil ling engine may be controlled by variation in the stroke of the reciprocating elements. This may be cither or both the piston and displacer of single-acting engines or the piston-displacer of double-acting engines. This method of power control is more applicable to free-piston or free-displacer machines than to engines with conventional drive mechanisms. In the single-cylinder Beale free-piston Stirling engine, an adjustment of piston stroke to the load condition does occur naturally in the engine. When the load resistance to movement is light the piston stroke extends to the maximum value permitted by the stroke limiting controls oí the engine. As the load resistance to motion increases the piston stroke decreases but the force exerted by the piston increases and attains it maximum value when the piston motion is completely restrained. It is an eerie sensation to grasp the pump rod attached to the piston of a Beale engine and to feel the engine respond instantly by an aggressive increase in the driving force as one attempts to restrain the piston movement.

A hybrid Stirling engine has a free displacer and a crank-controlled piston for driving a rotating shaft. The concept was first reduced to practice at the University of Calgary and is being further developed at the University of Bath. Control of engine output by adjustment of the displacer stroke was investigated briefly. By limiting the displacer stroke, the mass flow of gas in and out of the hoi and cold spaces was reduced, and so the range of the cyclic pressure amplitude was reduced. The effect was somewhat analogous to the system for pressure-amplitude control described above where the system dead space was increased by opening valves communicating with dead volume gas bottles in the engine crank-case. Limitation of the displacer stroke increases the effective clearance space in either or both the expansion or compression volumes. Clearance space in the cylinders may be accounted as dead volume to correspond with the above system. However, a change in the displacer stroke also changes the ratio of swept volumes in the expansion and compression spaces. This swept volume ratio »i is another principal design parameter that may be used for power control independent of I he dead space effect.

John Malone (1930) in his liquid engines (sec Chapter 8) used a system of displacer stroke limitation as his principal mode of power control. In his paper lie shows a rack-and-pinion device on the displacer rod for varying the displacer stroke but the details of the actual mechanism were not disclosed.

In the Thermoelectron Tidal regenerator engine for artificial hearts (sec Chapter 17). power control was achieved by variation in the stroke of the piston. A flexible metal bellows acting as the piston was caused to reciprocatc by the action of a ball nut on a screw shaft driven by an electric motor. Rotation in one direction caused the nut and hence the piston lo ascend. Rotation in the other direction caused it to descend. An electronic unit controlled the direction and duration of rotation in one direction or the other and so regulated the rise and fall of the piston and hence the output of the engine. The output in this case was. in fact, hydraulic power to operate a blood pump.

Solar Stirling Engine Basics Explained

Solar Stirling Engine Basics Explained

The solar Stirling engine is progressively becoming a viable alternative to solar panels for its higher efficiency. Stirling engines might be the best way to harvest the power provided by the sun. This is an easy-to-understand explanation of how Stirling engines work, the different types, and why they are more efficient than steam engines.

Get My Free Ebook

Post a comment