Miniature HCCI Free-Piston Engine Compressor. For Orthosis Application /

Transcription

1 / Miniature HCCI Free-Piston Engine Compressor Copyright 009 SAE Japan and Copyright 009 SAE International For Orthosis Application Lei Tian, David B. Kittelson, William K. Durfee Mechanical Engineering, University of Minnesota, USA ABSTRACT A miniature homogenous charge compression ignition (HCCI) free-piston engine compressor aimed at an ankle-foot orthosis application is described. Analysis of the human ankle shows that a fluid power source in the neighborhood of 10 W is needed. To account for compressor and actuator inefficiencies, the power output at the engine cylinder is designed to be 30 W. A compact engine compressor package has been designed and mathematically modeled. Experiments using existing engine components characterized the leakage model. Through the dynamic simulation of the engine, major parameters of the engine have been specified. Simulations indicate that the HCCI freepiston engine compressor, designed in a prototype package scale of about 80x40x0 mm is a viable compact and efficient fluid power supply. Simulation results demonstrate that the overall efficiency of the engine compressor is expected to be 5.9% and that the package should have a higher energy density than batteries. INTRODUCTION The Center for Compact and Efficient Fluid Power (CCEFP), a seven university research consortium headquartered at the University of Minnesota, Minneapolis USA, is developing a compact, untethered ankle foot orthosis (AFO) as a test bed for new technologies in tiny fluid power (Figure 1) [1]. Analysis of the human ankle during normal walking shows that to completely replace the function of the normal ankle with an active AFO, a peak torque of 75 Nm and average power of 10 W are required. Each step requires 14 J are needed for each gait cycle of the ankle which means about 70 kj per day for one side for a 10,000 step day []. Assuming a 50% mechanical efficiency from the output of the power supply to ankle power, the power supply needs to produce about 140 kj per day. A Battery and electric motor solution for the untethered power source was rejected due to low power density of batteries (90 kj/kg) and the size of a battery-motor package [3]. Hydrocarbon fuel has a power density of about 40,000 kj/kg. If an internal combustion engine coupled with compressor and pneumatic actuator are used for the power source of an AFO, an overall system efficiency of only 1.1% is needed to realize a higher power density than battery-motor package [3]. The proposed active AFO contains a free-piston engine compressor, an accumulator and pneumatic actuator. This illustrates new fluid power opportunities in medical and assisting devices. Free-piston engine compressor Accumulator Fig. 1 CCEFP Ankle Foot Orthosis. Image from CCEFP. As the scale of an engine gets smaller, surface effects such as friction, heat loss and leakage dominate [3]. In order to mitigate these losses, a free-piston engine compressor configuration was chosen, with the engine running at high speed to mitigate leakage losses. Combustion in small spaces is complex with ignition quenching and leakage problems [5]. HCCI combustion is proposed to address those problems. Chemical kinetics analysis was conducted for the engine, and an ignition model was constructed. DESIGN OF THE ENGINE Rotary actuator FREE-PISTON ENGINE A free-piston engine is a type of internal combustion (IC) engine that has no crankshaft. Without the kinematic constraint of a crankshaft, the movement of the piston is dynamically driven by pressures in the combustion, rebound and compressor chambers. A rebound device, such as a gas spring, metal spring or hydraulic accumulator, is used to store energy for combustion chamber gas compression, and two-cycle combustion is used.

2 Compared to a crankshaft IC engine, the free-piston configuration is more compact and simpler, having fewer moving parts, and no side-thrust between piston and cylinder wall. Large free-piston engine air compressor had been developed, such those by Pescara [6], Junker [7], and Braun [8]. The biggest challenge for free-piston engine is that the two-stroke cycle requires efficient scavenging but the piston motion is undefined [9]. More recently, attention has been paid to developing small free-piston engines. Riofrio et al at Vanderbilt University designed and prototyped a free liquid-piston engine compressor in the power range of 100 W [3]. Aerodyne Research Inc. designed and manufactured 10 W and 500 W miniature engine-generators, which are two-stroke free-piston engines coupled to linear alternators [10]. Their test showed 16% thermal efficiency. Aichlmayr et al proposed miniature freepiston engine coupled with Homogenous Charge Compression Ignition (HCCI) combustion, and experimentally demonstrated that HCCI can occur in a 3 mm bore, circumventing the flame quenching problem in small space [5]. HCCI COMBUSTION The proposed engine proposed here incorporates HCCI combustion, for several reasons. First, in tiny dimensions, spark plugs or a fuel injector are problematic because of their size and timing is challenging without a crankshaft. Second, the -piston engine is suited to HCCI since there is no crankshaft and the compression ratio can adapt to the onset of HCCI combustion. Third, as the engine dimension goes down, the flame is more likely to quench due to higher surface area to volume ratio. HCCI circumvents this problem [5]. ENGINE DESIGN CONCEPT The design concept for the new engine is shown in Figure and uses HCCI combustion and a metal spring for rebounding the piston. Silencer Engine piston Carburetor OPERATION RANGE FOR MINIATURE HCCI ENGINE The miniature HCCI engine is not simply a scaled down full-scale engine. There are unique features which restrict the operation range for the engine. In this section, overall parameters such as cylinder bore and speed are specified. PERFORMANCE ESTIMATION Performance estimation following the approach used by Aichlmayr [5] was used to find the major engine parameters. In this estimation, several rough approximations were made, including the scavenging efficiencies model in Taylor et al [11] r 1 Λ = r 1+ Λ e η = ch where Λ is delivery ratio, r is the compression ratio and charging efficiency η is the arithmetic mean of ch the completely displacement and completely mixing charging efficiencies. The engine power can be determined by 1 3 ( ) 3 π r 1 Vt = ρ BR i Φ η η η S c fc, i m ch P SN F e 4 rr where P is the power of the engine, ρ is the inlet air BR i density, S is the stroke, N is the engine speed (Hz), V is the volume of cylinder, R is the stroke to bore t aspect ratio, F is the stoichiometric fuel air ratio, Φ is S the equivalence ratio, e is the lower heating value of c η the fuel, is the indicated fuel conversion efficiency fc, i and η m is the mechanical efficiency. Using this equation, Figure 3 shows the relation between speed, compression ratio and cylinder bore and demonstrates that the engine should be chosen to be big and slow, or small and fast. Λ Starting valves Rebound Compressor piston Compressed air output Bore (mm) Fig. Design concept for miniature HCCI free-piston engine. Speed (Hz) Comp ratio Fig.3 Relation between speed, compression ratio and cylinder bore, when aspect ratio = 1.

3 LEAKAGE SIMULATION As the engine gets smaller the charge leakage through the cylinder-piston gap becomes the dominant factor affecting the efficiency of the engine [1]. This is because the leakage gap stays the same so that with smaller engines, a higher percentage of the mass is leaked. Simulation Approach An analytical solution can be derived to solve the gap flow using the Navier-Stokes equation [1,13]. Sher et al. [1] found that with a mechanically attainable gap width of 0 um, an engine of displacement smaller than 0. cc, cannot run at rpm. However, the COX.010 model airplane engine can run at this speed. This is because the Sher model neglects the sealing effect of the lubricating oil inside the cylinder gap that reduces the leakage. We improved Sher s model by including a parameter that models this effect. Experimental Validation The leakage model was validated by comparing the compression stroke pressure trajectory of the simulation with the pressure recorded in an experiment. An AP Hornet.09 engine was used with a 1.5 mm bore and aluminum piston and brass and chrome coated cylinder liner, the current state-of-the art in small engine machining technology. After a running in period, the model engine was motored at constant speed of 4900 rpm, and an Optrand D55-Q pressure sensor was used to measure the cylinder pressure. Lubricant oil is added before motoring to simulate the lubricated running situation of the engines. Fig. 5 Actual versus geometric compression ratio for several engine speeds CHEMICAL KINETICS SIMULATION Approach - HCCI is a different combustion mode because the onset of combustion is determined solely by chemical kinetics during the compression process instead of being triggered by a spark as in a SI gasoline engine or by high pressure fuel injection as in a diesel engine. The compression process must be simulated properly to determine the operation characteristics of the HCCI engine. The simulation package CHEMKIN is capable of simulating the chemical kinetics during engine compression based on chemical kinetics. The heptane mechanism developed by Lawrence Livermore National Laboratory was used. Fig. 4 Experimental validation of leakage model. Results - The Chemkin simulation showed that the required compression ratio to ignite the fuel increases with the engine speed. This can be understood from the ignition delay theory for HCCI combustion. As the engine speed increases, the ignition delay must be decreased to onset the HCCI combustion, thus a higher compression ratio is needed. For different hydrocarbon fuels, the higher the number of carbon atoms in the molecule, the easier the fuel is to ignite through compression. Because the reaction mechanism for complex mixed fuels such as kerosene is not readily available and therefore cannot be simulated, the ignition curve for kerosene based model diesel fuel was based on experimental observations of model engines operating using this fuel. This experimental result shown in Figure 4 demonstrates that the leakage inside the cylinder is exaggerated at high pressure, and under-estimated at low pressure. This is partially cause by the fact that the actual engine cylinder is tapered, which means the gap is larger around BDC. The simulation model will need to be further improved to model this effect. Simulation results - The simulation results shown in Figure 5 reveals that the engine should be run at high speed to minimize the effect of leakage. While the geometric compression ratio is limited by engine geometry the actual compression ratio depends on the engine speed with the ratio increasing with speed because of the reduced leakage as speed increases.

4 Fig. 6 Ignition curves for various rules showing the compression ratio required to ignite as a function of engine speed. Fuel for Miniature HCCI Engine - The viable engine operating speeds for different fuels can be determined from the data shown in Figures 5 and 6. While higher speed is needed for reducing leakage, the higher speed makes it harder to ignite the fuel in HCCI mode. Figure 7 combines the data for n-heptane from Figures 5 and 6 and shows the fuel retention efficiency, which is the percentage of fuel not leaked out from the combustion chamber before combustion. When operating on n-heptane fuel, the engine only ignites at rpm, while efficiency is negative due to severe leakage. At low speeds the leakage is too high and at high speed the required compression ratio is too high. Fig. 8 Operating range for model diesel fuel. Based on the curve of model diesel fuel shown in Figure 6, the operating range of model diesel fuel for a miniature HCCI engine is shown in Figure 8. Viable operating speeds using this fuel range from about 10,000 to about 40,000 rpm. DETERMINATION OF ENGINE SPEED Although the previous section demonstrates that the engine should operate over a range of speeds using a kerosene based fuel, additional simulations are needed to specify a rated nominal operating point at a specific speed. REED VALVE RESPONSE - Because the free-piston engine lacks a rotating crank shaft, the rotary valves commonly used in two-stroke engines cannot be used. Instead a reed valve was chosen to trap the fuel air mixture inside the crankcase chamber. The reed valve designed here is essentially the same as the valve used in COX reed valve model engines, which is a check valve working on a pressure difference (Figure 9). The valve needs time to open and close and its response is also affected by vibration. At higher speeds, the reed lags when opening and closing and reduces what otherwise would be an increase in power with speed. Intake channel Fig. 7 Operating range for n-heptane fuel. Reed Reed retainer This leads to the conclusion that an easily ignitable fuel must be used for the miniature HCCI engine so that the speed can be high enough to minimize leakage but the fuel can still ignite. Our engine will use model diesel engine fuel, the fuel commonly used for two-stroke model airplane diesel engines, which are basically HCCI engines. This fuel is based on kerosene, a large molecule hydrocarbon mixture that is easy to ignite. The fuel has two percent additive of ignition improver amyl nitrate to further facilitate ignition. Fig. 9 Reed valve used in COX model engine The valve vibration was simulated as a cantilever beam with an equivalent spring-mass-damper system [14]. For simplicity only the first vibration mode was considered. Pressure Model for Reed Valves - A model for pressure on the reed must be employed to determine the load on the reed valve due to pressure differences. For example, Blair et al. assume linearly changing pressure from inlet tract pressure to crankcase pressure [15] and Fleck et al. assume linear relationship fitted with a reed-lift associated pressure reduction factor [16]. In our simulation, the FLUENT computational fluid dynamics software was used to simulate the pressure distribution on the reed valve and a model that relates pressure distribution on reed surfaces to reed lift was constructed for a onedimensional simulation. During the inflow to the crankcase, the pressure force is

5 F = (0.33x ) A P p reed and during the backflow from the crankcase the pressure is where F = (0.899x ) A P p reed P is pressure difference in Pa, reed A is reed surface area, and x is the reed lift in mm. Part of simulation results are shown in Figure 10. Fig.1 Engine performance with speed. ONE DIMENSIONAL DYNAMIC SIMULATION Once parameters such as the target speed and engine dimensions are specified, a one dimensional dynamic model can be constructed to simulate the performance of the entire engine. Fig. 10 CFD results on the pressure distribution on the inlet side faces of reed for a 0.3 mm valve lift. The left image shows the inflow to the crankcase and the right shows the backflow from the crankcase. Results - This simulation results shown in Figure 11 reveal that the engine power density, indicated as Delivery ratio * rpm, actually decreases with speeds higher than rpm. CHEMICAL KINETICS FOR ONE DIMENSIONAL MODEL The CHEMKIN simulation is too detailed for the one dimensional simulation. Thus a model adapted from that of Gregory et al. [17] was used for onedimensional simulation of the miniature HCCI engine. The free-piston engine does not have a crankshaft angle so instead a time-base integral was used to determine the onset of HCCI combustion as shown by 0 [ ] [ ] t a b C T exp( C / T ) fuel O dt > RRth n 1 where T is the combustion chamber temperature, C, 1 C, a, b are coefficients, and RRth is the threshold value for onset of combustion. The coefficients and threshold values were fitted to the CHEMKIN simulation results of Figure 6 to match the simplified model to the detailed chemical kinetics. After ignition, the combustion process is modeled by a Vibe function Fig. 11 Reed valve response simulation results. ENGINE PERFORMANCE WITH SPEED Combining the simulation results from the previous sections, the engine power density and efficiency can be related to engine speed taking into account leakage, chemical kinetics and reed valve dynamics with speed. The results are shown in Figure 1. The simulation shows that the optimal speed is between 0,000 and 40,000 rpm. Because the higher speed results in much higher audible noise, 0,000 rpm was specified as the target speed for the engine. This speed corresponds to an engine bore of about 7 mm. t t x = 1 exp a o t m+ 1 where x is fuel consumption percentage, t o is ignition time, t is combustion duration and a and m are coefficients for the Vibe function. THERMODYNAMIC MODEL FOR EACH CHAMBER The energy balance of combustion, crankcase and compressor chambers are determined by the first law of thermodynamics ( )... dt m c = Q W + m h u dt ( ) c. v v in in c. v.. m h u out out c. v.

6 .. where m is the mass, Q is heat transfer and W is work transfer. The sub-scripts in and out mean flow in and out of the control volume. HEAT TRANSFER MODEL The heat transfer in the combustion chamber was calculated using the methods of Annand et al. [18]. The heat transfer equations are where p a K ρu B p h = B µ. b 4 4 ( ) Q = A h( T T ) cσ T T cc cc wall cc wall h is the convection heat transfer coefficient, U is piston speed, K is conductivity, µ is viscosity, σ is the Boltzmann constant, area, and a, b, c are coefficients. A is the heat transfer cc FREE PISTON DYNAMICS The free-piston shown in Figure is subject to inertia dynamics defined by d x piston = piston ( ) cc c engine comp comp spring m P P A P A F dt where P cc, P and P are combustion chamber, c comp crankcase and compressor chamber pressures, x is the piston position, and F is the force spring piston exerted by the rebound spring. Fig. 14 Piston dynamics model in Simulink. DYNAMIC SIMULATION RESULTS CHOOSING PARAMETERS - The Simulink simulation was then used to select remaining engine design parameters such as the rebound spring constant and the compressor dimensions. The procedures to specify the major parameters are first to chose a rebound spring constant so that HCCI combustion will readily occur from the rebound energy of the spring. Next a piston mass is specified to match the target speed because the rebound spring constant and the spring mass are two of the major factors that determine the engine speed. A compressor piston size is chosen so that the energy of fuel combustion is partially absorbed by the compressor and the piston ends up in a position that scavenging can occur and with sufficient stored energy in the rebound spring to drive compression for the next cycle. Based on those procedures, a spring constant of 1800 N/m and a piston mass of 5 gm are calculated for 300 Hz (18,000 rpm) operation. A 5 gm piston is possible if it is fabricated from aluminum as a typical piston in a model airplane engine weighs 1.3 gms. ONE DIMENSION MODEL OF THE ENGINE All the models discussed in this paper were put into a Matlab Simulink application to simulate the overall engine dynamics (Figures 13-14). Fig. 15 Simulated pressure trace of one cycle at 300 Hz operation simulation. Fig. 13 Simulink model for the entire engine. SIMULATED EFFICIENCIES Engine Indicated Efficiency - The engine indicated efficiency is the work done on the engine divided by the energy contained in the fuel flowing into the engine. By

7 this definition, the indicated efficiency was estimated to be 4.4%. Overall Efficiency Using an analysis similar to Barth et al [19] and taking into account that the compressed air will eventually cool to ambient temperature and lose some of its energy, the energy stored in compressed air is P P comp atm Energy _ stored = m RT air P comp where subscripts comp and atm denote compressed air and atmosphere. Based on this equation, the overall efficiency of the engine compressor is defined as the energy stored in the cooled compressed air, divided by the energy of the fuel that flowed into the engine to create the compressed air. By this definition, the simulation showed that the overall efficiency would be 5.9%. CONCLUSION A compact HCCI free-piston engine compressor was conceived, designed and modeled. The operation range for miniature HCCI engine was analyzed and a target speed of about 0,000 rpm was specified. Experimental measurements were used to calibrate and validate the leakage model. Dynamic simulation shows the potential overall efficiency of the enginecompressor to be 5.9%, which would be a higher power density than batteries. Further experimental research must be conducted to characterize the fuel and to validate the simulation models. ACKNOWLEDGMENTS This research is supported by the National Science Foundation through its Engineering Research Centers program. REFERENCES 1. D. Winter, Biomechanics and Motor Control of Human Movement, 3rd edition, 005, Wiley 3. Jose A. Riofrio, Design, Modeling and Experimental Characterization of a Free Liquid- Piston Engine Compressor with Separated Combustion Chamber, 008, Ph.D. thesis, Vanderbilt University 4. Andrew Alleyne, William Durfee, Liz Hsiao- Wecklser, Eric Loth, Geza Kogler, Manak Jain, Jicheng Xia, Jason Thomas, Joel Gilmer, Alex Shorter, 009, CCEFP TB6 presentation at Univ. of Minnesota, Minneapolis 5. H. T. Aichlmayr, Design Considerations, Modeling, and Analysis of Micro-Homogeneous Charge Compression Ignition Combustion Free-Piston Engines, 00, Ph.D. thesis, University of Minnesota atm 6. R.P. Pescara, Motor Compressor Apparatus, 198, U.S. Patent 1,657, K. Neumann, Junkers free-piston compressor, 1935, American Society of Mechanical Engineers, Volume 57 issue 4 8. A.T. Braun, Paul H. Schweitzer, Braun Linear Engine, 1973, SAE Preprint R. Mikalsen, A.P. Roskilly, A review of free-piston engine history and applications, 007, Applied Thermal Engineering, Volume 7 issue Kurt D. Annen, David B. Stickler, Jim Woodroffe, Glow Plug-Assisted HCCI Combustion in a Miniature Internal Combustion Engine Generator, 006, 44 th AIAA Aerospace Science Meeting 11. C. F. Taylor. The Internal Combustion Engine in Theory and Practice: Volume I: Thermo-dynamics, Fluid Flow, Performance, 1985, The M.I.T. Press, Cambridge, MA 1. I. Sher, D. Levinzon-Sher, E. Sher, Miniaturization Limitations of HCCI Internal Combustion Engines, 009, Applied Thermal Engineering, volume S.K. Grinnel, Flow of a Compressible Fluid in a Thin Passage, 1955, American Society of Mechanical Engineers 14. E. T. Hinds and G. P. Blair, "Unsteady Gas Flow Through Reed Valve Induction Systems," 1978, SAE Paper 15. G. P. Blair and E. T. Hinds, "Predicting the Performance Characteristics of Two-Cycle Engines Fitted with Reed Induction Valves," 1979, SAE Paper 16. R. Fleck, A. Cartwright and D. Thornhill, "Mathematical Modeling of Reed Valve Behavior in High Speed Two-Stroke Engines," 1997, SAE Paper 17. Gregory M. Shaver, J. Christian Gerdes, Parag Jain, P.A. Caton, C.F. Edwards, Modeling for Control of HCCI Engines, 003, Proceeding of the 003 American Control Conference 18. W.J.D. Annand, Heat Transfer in the Cylinder of Reciprocating Internal Combustion Engines, 1963, Proceedings of the Institution of Mechanical Engineers 177 (36) 19. Eric J. Barth, Jose Riofrio, Dynamic Characteristics of a Free Piston Compressor, 004, Proceeding of ASME International Mechanical Engineering Congress and Exposition 004 CONTACT Lei Tian, Address: Dept. of Mechanical Engineering, University of Minnesota, Minneapolis, MN 55455, tianx055@umn.edu