ufc2000 Posted April 1, 2008 Share Posted April 1, 2008 Auto Engines for Aircraft KITPLANES Magazine, August 2001 Auto Engines for Aircraft By Terry Edwards Many builders of homebuilt aircraft look at the cost of certified aircraft engines and believe they can save money by using a converted automobile engine. There have been many successful installations of such engines, but the builder must be aware of the engineering aspects of engine characteristics and systems. The history of engineering is full of examples of experimenters confounding the theorists and vice versa, but in general, successful engineering depends on a combination of knowledge, analysis, experimentation and test. A lack of knowledge and analysis means more experimentation and test. In aircraft powerplant design, this can lead to forced landings or worse. This two-part article will try to impart some basic engineering knowledge and analysis for those who wish to experiment with converted automobile engines. An engine's end use determines its characteristics. Over the last century, the internal combustion aircraft and automobile gasoline engines have evolved to have the characteristics found in Table 1. Aircraft Engines Automobile Engine Reliability Reliability Light weight Low maintenance Small bulk Low first cost Good power-to-weight ratio Low noise and vibration Table 1. Engine Characteristics. The obvious differences between these two types of engines are weight and first cost. These characteristics drive the design features and compromises. To compare these in more detail, we need some definitions: Maximum Rated Power: The highest power that the engine is allowed to develop for short periods of operation. Normal Rated Power: The highest power that an engine is allowed to develop in continuous operation. The rate at which the engine does work. Units are horsepower (33,000 foot-pounds per minute). This is normally measured on a dynamometer or brake, resulting in the term brake horsepower. Torque: The ability of an engine to do work. It is a product of the rotational force produced by the crankshaft and a moment arm. Units are pound-feet. TBO: Time between overhaul. Provided the engine is operated and maintained per the manufacturer's recommendations, TBO is the expected life of the engine measured in hours. The difference between maximum rated power and normal rated power is extremely important to this discussion as it directly affects reliability and TBO. Automobile engine power is quoted as maximum rated power. The typical road-load power (rolling resistance plus aerodynamic drag) required of an automobile engine is only about 20 hp at highway speeds. The maximum rated power gives an indication of the ultimate performance of the vehicle. However, when automobile manufacturers offer the same engines as industrial engines, they are rated at normal rated power. Contrasting this, in an aircraft, we expect TBO of 1500-2000 hours in an engine that delivers 55%-100% of rated power, averaging about 75% of rated power. To illustrate the point, consider the ratings of the same V8 engine shown in Table 2. Displacement (cu.in.) Hp @ rpm Torque @ rpm Automobile sports version (road car) 350 345 @ 5600 350 @ 4400 Aircraft Conversion 350 275 @ 3500 410 @ 3500 (engine only) 350 350 @ 4800 Industrial version 350 198 @ 3000 326 @ 2400 Table 2. V-8 engine ratings depending on application. Note that the automobile sports engine power rating is 75% higher than the industrial version. It is interesting that the aircraft conversion lists two power ratings. As this engine is sold without a reduction drive, it falls on the builder to choose which rating is most appropriate. However, the aircraft conversion is still rated much higher than the engine manufacturer's industrial version. Consider that the industrial version power curves show increasing power and torque at engine speeds greater than shown in the table, but the engine is not rated at the higher speeds. That is because operating at these higher speeds would reduce engine TBO. While an engine's power rating is the most widely used rating, it is an engine's ability to do work - its torque rating - that is most important. Torque (T), power (P) and engine rotational speed (N) are related by: T (lb.-ft.) = P(hp) x 5252/N(rev./min.) If we choose the aircraft conversion figures as an example: T=275 x 5252/3500=412 lb.-ft. Interestingly, if we check the torque at the maximum power rating we find: T=350 x 5252/4800=383 lb.-ft. which is 7% less than the peak torque rating. This is typical of all piston engines. Propellers The function of the propeller is to convert the engine power into thrust. The power absorbed by the propeller is given by: P =& (sigma(rpm/1000)3 x D5 x Cp)/50 and the thrust is given by: Th = (sigma(rpm/1000)2 x D4 x Cth)/1.515 where: sigma = relative air density D = propeller diameter (ft.) Cp = propeller coefficient of power Cth = propeller coefficient of thrust Propeller design and the power and thrust coefficients are beyond the scope of this article. But note that the power absorbed increases as the cube of the rpm and the fifth power of the diameter. If we plot a typical engine power curve against a propeller demand curve, we see the information in Chart 1. Engine power available vs. propeller power required. To minimize the tip losses and noise, the propeller blade tip speed should be less than the local sonic velocity. SV = sonic velocity SVsea level = 1116.5 ft./sec. x 60= 66,990 ft./min. SV10,000 ft. = 1077.4 ft./sec. x 60 = 64,644 ft./min. A typical 75-inch (6.25-foot) propeller should be limited to about 0.85 Mach or, at sea level: rpm=0.85M x SVsea level/¹D = 0.85 x 66,990 ft./min./(¹ x 6.25ft=2900 rpm.) Thus the thrust required by the aircraft determines the propeller diameter (D4 in the thrust equation) and the sonic velocity determines the maximum rpm. Therefore, if our engine is rated at a higher rpm than the optimum propeller speed, we need a propeller speed reduction unit (PSRU). This fact is rather obvious, but the engine rated speed and the optimum propeller speed determine the actual ratio. Provided the correct power and torque ratings are observed, any core automobile engine should provide reliable service. Core in this case means the basic block, crank, rods, pistons, heads and valve train. The difference between a reliable and unreliable installation is in the systems engineering, which means all of the support systems, PSRU, fuel, oil, cooling and controls that the engine needs. The engineering in these support systems can be considerable. All systems should be examined to determine if a single-point failure can cause engine failure. PSRU types include gear, toothed belt and chain drive. A gear drive is the most reliable and the most costly. Gears are precision-machined components whose contact surfaces have been hardened. Gears are manufactured in sets and can be manufactured to a variety of accuracy standards. The American Gear Manufacturers Association (AGMA) lists these standards in their publications. AGMA Class 6 to 8 are used in automobiles, and this is the minimum class that should be used on a PSRU. Spur gears are more efficient because they have no sliding friction and no end thrust. Helical gears are quieter due to the sliding action of the teeth, but they are slightly less efficient and require the bearings to absorb the end thrust. Toothed-belt manufacturers do not endorse PSRU applications, so users are on their own establishing application factors and service life. Although they are light, relatively inexpensive and quiet, the reinforcing fibers have a finite life, and the molded teeth need to be inspected on a regular basis for wear. Many housings have open sides to facilitate inspection, but consideration should be given to enclosing them to prevent debris from damaging the teeth or breaking the belt. Chain drives may be either roller chain or Morse Hy-Vo chain. All chain drives require an enclosed housing and continuous oil lubrication. They are compact but require accurate alignment. As there are hundreds of parts in the links and pins, many more parts can fail compared to gears. Chains are prone to galling (tearing away of metal from load- bearing surfaces), tension failures and fatigue failures. In addition to the tension force required to transmit power, additional tension force is due to the centrifugal force caused by the rotational speed of the drive. All PSRUs require bearings. Rolling-element bearings are preferred for both the thrust bearing for the propeller and the radial bearings for the input and output shafts. Once the loads are accurately known, bearing manufacturers' handbooks provide excellent formulas to rate the life of the bearings. The L10 life is generally used, which means that for the load in question, 10% of a large sample of bearings will fail before reaching the calculated life (either hours or revolutions). Failure means that the races will start to exhibit rough surfaces causing noisy operation. It does not mean that they fall apart. If left in service long enough, all bearings will fail due to fatigue failure of the rolling surfaces. The propeller shaft bearings should be sized to support the propeller thrust loads plus the gyroscopic loads caused by maneuvering. As a minimum, the L10 life calculated should be at least 1.5-2 times the TBO of the engine: 3500-5000 hours. If sealed bearings are used, they may not reach this life if the hours per year are low because the grease sealed inside the bearings can dry out. Oil-lubricated bearings should reach the calculated life. Only one bearing on a shaft can be fixed axially; the other must float to allow for thermal expansion. All PSRUs should isolate the side loads from the gears, chain or belt from the engine crankshaft bearings. Neither the crankshaft nor the crankshaft bearings was designed for side loads. Keeping It Cool No engine will function for long without adequate cooling. As the amount of heat rejected to the cooling system is approximately the same as the heat used to produce the shaft power of the engine, a properly sized radiator with good inlet and exit ducting is essential. The cooling system is a primary engine support system, along with the fuel and ignition systems. Unfortunately, many radiator systems in homebuilt aircraft look like afterthoughts. A close examination of WW-II-era aircraft will give some guidance in how to integrate the cooling system into the airframe. The Spitfire, Hurricane and Bf109 used underwing radiators. The P-51 Mustang used an aft-fuselage-mounted radiator, while the P-40 Warhawk's was mounted beneath the engine. One of the most efficient was the Mosquito, whose radiators were buried in the wing between the fuselage and the engine nacelle. The theme of all of these was efficient inlet and exit ducting with few or no bends, with the radiator core perpendicular to the airflow direction. It is better to make a duct that increases the airframe frontal area but is well streamlined on the outside and inside, than to try mounting the radiator in the space between the engine and the firewall with several 90° bends in the duct. Sharp bends increase cooling drag. The function of the inlet duct is to increase in area gradually so the air velocity slows to the most efficient heat-transfer speed. The function of the exit duct is exactly opposite: contracting gradually to increase air velocity. Because heat (energy) is added to the air in the middle of the duct, it is possible to obtain net thrust from the duct as exit energy is greater than inlet energy. Radiator design is rather complicated because the heat rejection depends on the air and water temperatures and the air and water flow rates. Custom radiator shops use charts to determine radiator area and core thickness. Automobile radiators can be used, but it may not be possible to obtain performance charts from the manufacturer. Radiators must be mounted in vibration-isolating rubber mounts to the structure. Do not mount them to the engine. All coolant pipe ends must be beaded and should be mounted to the structure with cushioned Adel clamps. Coolant pipes should be joined with short straight pieces or elbows and can be high quality automotive silicon hoses. Hose clamps should be constant-tension type, and two clamps per hose end provide extra safety. In addition to a coolant temperature gauge, cockpit instrumentation should include a coolant pressure gauge, which will give the fastest indication of leakage. Make sure no coolant pipes or hoses enter the cockpit. If a separate radiator-type cabin heater is used, make sure this can be shut off completely in the event of a leak. The system fill/pressure cap should be a coolant-recovery type, and the pressure rating should be the same as that provided by the manufacturer in the automobile application. Higher-pressure caps will raise the boiling point of the water, but may also cause head gasket or water pump seal leaks. Keeping It Oily Engine oil systems, the sump, pump, distribution galleries and bearings are generally adequate for the normal rated power. However, the pump drive system should be checked to see if there are any pins connecting the drive gears to the shafts. In some cases, these pins are designed to break if the pump encounters debris. All pins should be examined to determine if the shear feature provides a margin of safety or if it is a single-point failure that will disable the entire engine. Filters must be lockwired so they don't unscrew. Lubricating oil should be the type and grade recommended by the engine manufacturer. Note that the internal clearances in the journal bearings and cam drives are designed with the recommended grade in mind and that using a heavier grade may lead to increased wear. Also, different types of oils have various additive packages, which may affect the engine materials. The valves should be those recommended by the manufacturer. Aftermarket valves made from titanium or stainless steel are made specifically for reduced valve inertia at high rotational speeds or for special fuels such as alcohol. They do not have the surface hardness and wear characteristics that the special valve and valve seat alloys used by the engine manufacturer have. Next month we will continue with a discussion of ignition, balancing, and using the right fuel. Last month we covered some of the differences between engines designed for aircraft and those that might be converted from auto use. Topics included propellers and speed-reduction units, cooling and lubrication systems. This time we conclude with other systems that deserve careful attention when considering flight with an auto engine. Sparking Automobile ignition systems have progressed from points, condenser coil and distributor types, through solid-state switch systems, through integrated computer-controlled engine management systems. Modern engines will not run without all the sensors and controls, so when selecting a candidate engine, the ignition system should be included. Factors that are not considered by automobile engine designers are radio frequency (RF) shielding and high-altitude operation. Modern engines are known to emit RF interference that can be picked up with a VHF radio. GPS systems are also sensitive to stray interference. The high-voltage portion of the system is the most susceptible to dielectric breakdown at high altitudes due to the lower air pressure. Aircraft engines have used sealed or pressurized magnetos to overcome this. Careful testing should reveal whether this is a problem. Some redundancy should be incorporated into the ignition system. FAR 33.37 for certified engines requires dual spark plugs per cylinder, a dual ignition system, and two separate electric circuits with separate sources of electrical energy, or a system of equivalent in-flight reliability. While it may not be possible to have two spark plugs per cylinder due to cylinder-head design, other portions of the system can be made redundant. The choice of fuel requires careful consideration. Table 1 illustrates the major differences between aviation and automotive gasolines. Fuel Grade Vapor Pressure (kPa) Lead (mg/L) Sulfer % Gum (mg/100mL) Octane Performance Number Mogas regular All grades (max): 5 0.15 5 87 Mogas midrange summer 79 spring 86 5 0.15 5 89 Mogas premium fall 97 winter 107 5 0.15 5 91 Mogas super premium 5 0.15 5 93 Avgas 100LL 38-49 560 0.05 6 100 lean/130 rich Table 1. Comparing fuels. Avgas 100LL has a low and fixed vapor pressure all year to minimize vapor lock at all altitudes. Contrasting this, all grades of mogas vary the vapor pressure throughout the year to provide easy starting, but this leads to increased susceptibility to vapor locking. In addition, mogas absorbs more heat in the carburetor, leading to a greater susceptibility to carburetor icing. The lead content of avgas is required to attain the octane (or anti-detonation) rating of 100. However, on lower-compression engines that do not need the octane rating, the lead content can cause lead fouling and erosion of spark plugs, valve erosion, and valve and ring sticking. In addition, lead is incompatible with catalytic converters and oxygen sensors used on modern automobile engines. While catalysts probably wouldn't be used in an aircraft conversion, the oxygen sensors in the exhaust system are required for correct operation of the engine control computer. In many areas of the country, alcohol is allowed in mogas. Alcohol may separate from the gasoline, and it will absorb water, which may freeze. Alcohol also attacks the elastomers used as seals in the fuel system. If using mogas, test for the presence of alcohol before filling the tanks. When designing the fuel system, conduct tests of the seals using the fuel chosen to ensure compatibility. This should include the fuel tank sealant, O-rings, face seals, carburetor jets and floats, fuel pump components, and fuel injection components. With up to four mogas octane ratings available, it is important to use only the grade specified by the engine manufacturer. The use of lower octane ratings can cause detonation and engine failure. Ice Induction-system icing must be considered. With a carburetor, some form of carburetor heat must be used. This may be a diverter valve to direct induction air over the exhaust manifold similar to aircraft engines. If the engine is fuel injected, an alternate source of air is required in case the primary inlet becomes iced over. Mounts Engine mounts must be designed to support all of the flight loads encountered. The weight of the engine and propeller system, the load factors (positive and negative) that the airframe is designed for, and thrust provided by the propeller must be supported by the engine mount. When analyzing the mounts, remember that the torque reaction of the engine may cause some members of a seemingly symmetrical mount to be in compression while the opposite member is in tension. Aircraft stressed for aerobatic use must consider propeller gyroscopic forces on the propeller shaft as well as the engine mounts as these may dominate the normal forces, particularly in snap maneuvers. Steel tubing is the preferred material as it can be welded to the full strength of the tube and because it is more heat- and fatigue-resistant than aluminum-sheet structures. The mount should be designed on the space frame principle: Each tube is axially loaded in compression or tension. Each compression tube must be checked for buckling and have an adequate margin of buckling load to actual load. Transverse loads in the middle of tubes such as brackets for oil coolers and ECU modules must be avoided. If components must be mounted on the engine mount, the structure should be triangulated at that point and analyzed carefully. Accessories Automobile engines usually do not have sufficient accessory drives for aircraft applications. While some automobile engines have mechanically driven fuel pumps, none has a vacuum pump drive or propeller governor drive. For aircraft that require these drives, the method of driving requires careful consideration. Probably the easiest way is to drive these from the propeller speed-reduction unit (PSRU), but most commercially available units have no provision for extra drive pads. These accessories may be driven by gears, belts or chains. V-belt drives are probably the easiest to add, similar to automotive alternator (and Lycoming alternator) drives. If this is done, the accessory manufacturer must be consulted to ensure the component will withstand the side load on the input shaft imposed by the V-belt. Modern automobile alternators are very powerful, compact, light and reliable. However, an electrical load analysis should be done by adding all the aircraft equipment and engine ECU computer loads. Total load should be at least 20% less than capacity. For redundancy, it is worth considering using two smaller alternators instead of one large one, especially if the engine is fitted with an ECU. Of course, they should have separate drives. Also, it is worth considering fitting dual batteries with isolation between the two electrical systems so a single malfunction does not cause an engine shutdown. All the major electrical components, alternator, voltage regulator (if separate), and the ECU must be supplied with cooling air. Nothing destroys electrical components faster than heat. Brackets for the accessories are just as important as the accessory itself. An engine that fails because a bracket falls off fails just as surely as if the crankshaft broke. Brackets must be strong enough to manage the loads, stiff enough to not vibrate, and light enough so they do not impose a weight penalty. All the usual engineering principles apply. Before Flying Getting back to our engineering principles from the opening paragraph in Part 1 last month, the analysis and ground test of the finished engine installation is probably the most important element before committing to flight test. Strangely, this seems to be the most poorly done, but its importance cannot be overstated. There are two main aspects to this: power and reliability analysis and test, and vibration analysis and test. Power and reliability should be verified using similar procedures to the certification requirements for aircraft engines. These are detailed in FAR 33.49 and prescribe a 150-hour endurance test differing in detail whether the engine is supercharged or not. This endurance test may be used as a guideline for automotive conversions, and the total time and detail tests may be shortened as desired. If the engine passes this test, it should be disassembled and checked for wear and crack-tested. The valves should be inspected and relapped or reground. The valve guides should also be inspected. Vibration analysis and measurement is probably the least understood and most important part of testing. Most builders understand that dynamic balancing of the rotating components, especially the pistons, rods, crank and propeller, results in smoother running. But realize that there are different grades of dynamic balancing. ISO 1940 lists 11 grades from crankshaft drives of large marine diesel engines to gyroscopes. The choice of grade should be discussed with the balancing shop and depends on the balancing equipment and the budget. A Demonstration Even an engine that is perfectly balanced dynamically will have torsional vibration due to the intermittent nature of the power production and transmission. A simple model to visualize torsional vibration can be made using a small-diameter steel wire (such as a piece of coat hanger wire about 12 inches long) suspended vertically with a weight (such as a gallon paint can) firmly attached. Rotate the weight about a quarter turn, release it, and watch the weight oscillate around the axis of the wire. Now, with the weight oscillating, turn the suspended end of the wire at a constant rotational speed. This simulates a piston and connecting rod causing torsional vibrations in the rotating crankshaft. Of course, due to multiple cylinders, the frequency is much higher in an engine, and because the crankshaft, drive and propeller shaft are stiffer, the amplitude or amount of rotation is much lower. If the amount of weight is kept the same but the shape is changed (a 2x4 piece of wood weighing the same as the paint can) and the experiment repeated, the wire would oscillate slower due to the higher rotational inertia. This simulates the propeller causing torsional vibrations in the propeller shaft. Notice that once excited, the torsional vibration continues for a long time because the wire has very little internal damping. Engine Vibration Many automobile engines have a torsional damper on the accessory drive end of the crankshaft. However, the entire drive train must be considered as a system in controlling torsional vibrations. An automobile drive train includes the engine, a clutch with damping springs for a manual transmission or a hydraulic torque converter, a transmission, driveshaft and wheels and tires. Being made of rubber, the tires damp a lot of the torsional vibration. When the engine is converted for aircraft use, the entire drive train is removed, and a new one is substituted. The PSRU and the propeller both contribute torsional vibrations to the new system, and engine vibrations can be fed into the propeller and vice versa. Once the engine, PSRU and propeller have been chosen, the mass elastic system is finalized except for torsional dampers. A mass elastic drawing is prepared that lists the mass moment of inertia (a property with the mass of the paint can in the simple model) of each major part such as pistons, crank and propeller - plus the torsional spring rate of the elements connecting the masses (a property of the coat hanger in the simple model). The mass elastic system is then analyzed by a specialized computer program to compute the natural frequencies of the elements in the system and the forcing frequencies (the input). The critical frequency occurs when the forcing frequency coincides with the natural frequency. Critical frequencies should be computed with all cylinders firing and with one cylinder misfiring. If the system is operated at its critical frequency, it will fail in short order. If the system has inadequate margins between forcing and natural frequencies, it must be modified to include extra mass, extra elasticity or a torsional damper, or a barred speed range must be observed. When the analysis proves the system has adequate margins, the engine should be run and tested to prove the analysis with electronic torsional-vibration recording equipment. Failure to perform either the torsional vibration analysis or to test the engine system can result in spectacular failures of parts such as crankshafts or PSRU drives - or unspectacular but annoying failures such as jets backing out of carburetors, flywheel bolts loosening, and cam drives wearing prematurely. The world of certified engines is full of examples of newly designed engines or new propellers on old engines having torsional vibration failures. These are normally corrected during the certification process by the methods noted here. Certified vs. Auto Engines 360 cu.in. Aircraft Engine 4.3L V-6 Auto/industrial Engine 5.7L V-8 Auto/industrial Engine Displacement (cu. in.) (road car) 36 262 350 Bore (in.) 5.1 4.0 4.0 Stroke (in.) 4.3 3.5 3.5 Number of cylinders 4 6 8 Cylinder arrangement horizontal V V Rated power (hp @ urp) 180 @ 2700 120@ 3000 198 @ 3000 Engine weight (lb.) 295 388 582 Including cooling system (lb. estimate) 295 425 622 Speific weight (lb./hp) including cooling system 1.6 3.6 3.1 Table 2. Comparing a 180-hp aircraft engine to two automobile-derived engines. Finally, we will directly compare a 180-hp aircraft engine to several automobile-derived industrial engines. In Table 2, notice the differences in weight and the specific weight between the engine types. Naturally aspirated aircraft engines typically weigh 2 pounds per horsepower. The auto/industrial engines in the table weigh 3-3.5 pounds per horsepower including the cooling system. Most aircraft weigh 10 pounds per horsepower, so the substitution of a heavier engine means the airframe must be lighter, or the payload must be reduced. The normal rated power of the auto/industrial engine occurs at 3000 rpm, which is only about 10% above the aircraft engine rated speed. Therefore a PSRU is not really needed to drive the propeller. The V-6 is the lightest but has inadequate power at its normal rated power. If we compromise on longevity (lower TBO) by increasing the power and rpm, we can re-rate the engine to 190 hp at 5400 rpm. Then we need a PSRU of 2:1 ratio to reduce the propeller speed to 2700 rpm. But the PSRU adds another 50 pounds, increasing the total weight to 478 pounds, and it is only 95% efficient, thereby consuming about 10 hp. Now we need to reduce the engine weight by 183 pounds to get it to the same weight as the aircraft engine. We can substitute an aluminum block and heads, but this adds cost. When the cost of all the components is known, a cost/ weight/power analysis should be performed to ensure the engine will have adequate performance for the intended airframe. What Does It Mean? In summary, an aircraft engine must be reliable and light with small bulk and good power-to-weight ratio. Successfully converting an automobile engine for aircraft use requires a good deal of engineering design, analysis and test. The hope is that the information in this two-part series serves as a guide for those who wish to embark on the quest of lower-cost aircraft engines. 1 Quote Link to comment Share on other sites More sharing options...
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.