Extending Lean Limit with Mass-Timed Compression Ignition Using a Catalytic Plasma Torch Mark A. Cherry and Robert J. Morrisset, Automotive Resources Inc. N. John Beck, BKM, Inc. ABSTRACT Research on the Catalytic Plasma Torch (CPT) ignition system was conducted this last year at BKM, Inc. in San Diego. The results showed that under certain conditions CPT can not only time ignition properly, but also extend the lean stability limit. This concept is based upon compression ignition of the charge in the CPT's integral pre-chamber. Compression ignition is induced by timed catalytic reduction of the pre-chamber's activation energy. This produces almost instantaneous combustion in the pre-chamber and is divided into multiple high velocity torches to rapidly ignite the main chamber charge. The timing of the ignition event is based on the location of the heated catalyst in the prechamber and the mass of the charge inducted into the cylinder. The base timing curve can be modified via current control which effects the catalyst activity. Dynamic modification of the timing event is accomplished by using the catalyst as an in-cylinder hot wire anemometer. In this mode, the activity of the catalyst, and thus the timing, is inversely proportional to the mass of the charge. This system works with multiple fuels in any homogeneous charge engine and differs from hot surface ignition in that it is not timed via fuel injection. This device replaces conventional ignition components such as the distributor, breaker system, high-tension wires and spark plugs. In addition, this system contains no moving parts or sophisticated electronic controls and can run under load without external energy and is entirely water proof. Resources Inc. (ARI) at BKM's facilities in San Diego, California. The purpose of this research was to define what advantages the Catalytic Plasma Torch (CPT) ignition has over conventional spark ignition systems. CPT was shown to be equal to or better than spark ignition in both increased power and lower fuel consumption. CPT proved to have the same effect as a spark-ignited, fuel-rich pre-chamber without the added complexity of pre-chamber fuel supply. Thus, CPT can have a extended lean limit which will result in better fuel efficiency and lower emissions than spark ignition. CPT timing demonstrated the ability to be modified via catalytic current adjustment with simple electronic controls. CPT demonstrated that has the ability to lower the ignition energy requirement prior to combustion via timed selective ionization of reactants in the pre-chamber. Finally, an improved version of CPT was demonstrated to have the ability to inherently retard timing with load as normally required. Spark ignition requires many pails and sophisticated controls to accomplish this whereas CPT has only one part per cylinder and no moving parts. These preliminary test results demonstrate that CPT ignition already has many advantages over spark ignition even in this early stage of development. REASONS FOR DIFFERENCES IN FORCED IGNITION METHODS SPARK The main advantage of spark ignition is that it can be externally timed. State of the art timing methods cannot anticipate anomalies in the individual cylinder therefore spark timing is based upon a theoretical optimum. This event can be modified by closed loop feed back through 02 sensors and detonation sensors. Yet to date, this is done on an averaged basis which cannot anticipate the needs of an individual cylinder. Furthermore, of the total energy released through the spark gap, only 30% is useful for ignition. Since a spark must ionize all of the gasses in its path and since normal air only contains 20.8% oxygen, then approximately 20-25% of the 30% (or 6.0-7.5% of the total energy) is available to the charge for ignition. Therefore, Spark ignition requires a relatively large amount of discharge energy to achieve satisfactory results. This high-voltage, high-energy discharge is difficult to insulate and is moisture sensitive and also leads to increased erosion of the spark plug electrodes Another well recognized weakness of spark ignition is that the flame front development is passive and uncontrolled. The direction that the flame travels is dependent on fluctuating variables in the cylinder. This leads to a random burning process, causing undesirable cyclic variations and loss of efficiency. This is improved with a pre-chamber but this introduces other design problems such as decrease electrode durability and dual mixture delivery. HOT SURFACE IGNITION Hot surface ignition systems, such as glow plugs used in diesels during warm up and in "compression" ignition methanol engines, differ from CPT in several ways. First, all surface ignitions require timed injection of fuel under high pressure near TDC in order to control ignition timing. Secondly, research has shown that in an identical prechamber, steel wire took three times as long to ignite the charge as catalytic wire under identical conditions (1). This is due to the fact that hot surface ignition must rely 66 solely on thermal activation mechanisms to achieve ignition. There are no known chemical surface interactions that either add or subtract from the ignition process. Only radiant and convective heating add energy to the reactants until they reach their given activation energy level and spontaneously ignite. Also, the hot surface ignition must be actively heated by a continuous external energy source. No heat can be added to the igniter until after combustion has begun. Because of the relatively high amount of heat thermally shunted to the cylinder head water jacket, hot surface igniters of current design could never achieve self-sustaining ignition when only heated by thermal energy from combustion. Finally, prior to ignition, the hot surface igniter develops no active radicals that can lower the activation energy of the surrounding charge and therefore must have a stoichiometric mixture present in order to achieve reliable combustion. CATALYTIC PLASMA TORCH By definition, a catalyst accelerates a reaction but is not consumed by it. In a practical catalytic internal combustion igniter, the catalyst accelerates the reaction rate by lowering the activation energy through the selective generation and the release of active radicals to the surrounding mixture prior to ignition. Unlike spark ignition which ionizes everything in its path, the catalyst selectively ionizes only oxygen and fuel molecules. The catalytic ignition is the only ignition that can selectively ionize fuel and oxygen prior to gaseous combustion, thus lowering net activation energy of charge in prechamber (1). No nitrogen radicals are generated until gaseous combustion takes place. There is preliminary evidence that catalytic initiation causes a substantially different chain-branching reactions. These reactions not only lead to more complete combustion but also provide in-cylinder reduction of NOx. Furthermore, the catalytic prechamber needs no secondary mixture supply and delivers the torch evenly throughout the main charge in a predetermined controlled fashion. Unlike hot surface ignition, CPT is not dependent on sophisticated fuel injection to control timing, but is dynamically timed by the conditions in the cylinder. Thus, CPT can work in a homogeneous-charge engine. Also, CPT can be actively heated by the chemical energy in the charge, which enables self-sustaining operation without any electrical energy input. There are three specific ways in which a catalyst can react with the charge. 1. SURFACE REACTION At the first energy state, the temperature of the catalyst is such that fuel and oxygen are ionized on its surface and react with each other forming completed products and transfer most of their reaction energy to the catalytic wire with no net heat gain to the surrounding charge. The boundary layer on the catalytic surface contains no active radicals, only pre or post-reaction products- Given enough time, all reactants would be consumed and there would be no gaseous phase reaction (2). In this mode the catalyst actually removes energy from the charge in order to maintain its own temperature. This reaction is isothermal. 2. PARTIAL REACTION (Radical Generation) In this second energization level, the catalyst is not only hot enough to ionize the reactants on its surface but also transfers enough thermal energy to the reactant to eject it from its Surface prior to a Surface reaction. In other words, the mass transfer rate is high enough to force radicals from its surface prior to a surface reaction (although some percentage of surface reactions do occur). These active radicals mix with the surrounding charge and pre-condition it by reducing the ignition energy in inverse proportion to their concentration. A point will come where the ignition energy requirement will reach zero and gaseous ignition occurs (1). This is primarily species activation with vary little thermal activation occurring other than that which is inherent in species activation. This process can take place at bulk -gas temperatures considerably below the thermal activation temperature of the mixture. This enables the prechamber charge to have a considerably lower ignition energy requirement than the main chamber which is in effect the same as a richer mixture in the pre-chamber. This is something that neither spark nor hot surface ignitions can achieve. This process is exothermic due in part to the deposition of radicals at a higher energy state into the pre-chamber charge. 3. HYBRID REACTION This is where both thermal and species activation happen simultaneously at a lowered activation energy level. In this third mode of operation the catalyst is operating just below the thermal activation temperature of the charge. Between the contact angle and gaseous ignition, approximately 40' Of mixture passes the catalyst. A sizable percentage of this mixture is transformed into active radicals and steam. This charge is now preconditioned and has a lower ignition activation requirement (1). Gaseous ignition is initiated by two mechanisms. One is thermal/catalytic activation of the charge in the immediate vicinity of the catalyst d.ue to its increasing temperature from the surface reaction . The second is compression ignition in the pre-conditioned zone of the pre-chamber resulting from the rapid temperature increase near the end of the compression stroke in combination with the lower activation temperature. This near - instantaneous combustion happens when the lowered activation temperature of the pre-chamber charge and bulk gas temperature are equal. This produces near constant volume combustion within the pre- chamber, leading to a rapid pressure rise and a powerful torch of ignition products being distributed to the periphery of the main chamber. The distribution of the torch is accomplished by a torch distributor or "showerhead" TOP AND SIDE VIEW OF TORCH DISTRIBUTION The flame their propagates toward the center of the cylinder similar to Mazda's surround combustion concept. In this way CPT combines the desirable features of both Otto and Diesel cycles. THEORY OF CPT CONCEPTS 1. TIMING CHAMBER THEORY U. S. Pat. ft 5,109,817(3) The timing chamber principle is based upon the concept of a gaseous spring which responds in equilibrium with the applied pressure (4). These gases consist of noncombustible gases either from the previous cycle or pure air. The chamber is assumed to be a cylinder of uniform diameter and a length chosen to match the desired characteristics for the engine in question. An interface between the non-combustible gases and the fresh charge is formed at the mouth of this chamber during the intake stroke. The diameter of the nozzles should be relatively small to minimize the mixing effect of turbulence within the cylinder during the intake stroke. Some mixing due to gaseous diffusion will occur. To compensate for this effect the chamber needs to be lengthened to increase the timing resolution or the distance the interface travels per degree CA. Since diffusion is directly linked to time, a slower engine will require a longer chamber than a faster one. The interface moves in direct response to the applied pressure in order to maintain equilibrium with the cylinder. Since the pressure in the chamber and the cylinder are equal and the change in pressure is proportional to the change in volume, then the change in volume of the gas spring will be proportional to the change in volume of the cylinder. Therefore, if the chamber and stroke were of equal length, the interface location at any given time would be identical to the location of the piston in the bore. If the chamber is shorter than the stroke then interface location is equal to the chamber/stroke ratio times the piston position. Since piston position determines interface location, the contact angle of the charge with the catalyst can be determined and the catalyst can be positioned accordingly. This interface response is constant regardless of load or speed and thus the contact angle of the charge with the catalyst remains constant. 2. DENSITY SENSITIVITY The timing effect of the catalyst is based oil the density of the charge. As the density of the charge decreases, the number of charge molecules that can contact the surface per unit time also decreases. The energy density, or kJ/unit volume, obviously varies directly with charge density, assuming a constant air/fuel ratio. The reduction in surface reaction energy lengthens the delay between contact and gaseous phase combustion. In the mass controlled model, the contact angle is advanced for proper light load operation and the increasing charge density cools the catalyst causing ignition delay to increase with increasing load. Another important consideration is the inherent EGR of a given engine. For example, if an engine had negative valve overlap, the ratio of exhaust gas to fresh charge becomes proportionally greater as the intake density decreases. This is due to the fact that the atmospheric pressure of the trapped exhaust gas may be as much as four times higher than the pressure of the intake manifold. On a 101 compression ratio engine this would mean that the exhaust gas could comprise as much as 50% of total cylinder volume at high vacuum conditions. At full load this would only be 10% of total charge. This alone would severely reduce the energy available to the catalyst surface at lighter loads. The above factors when combined with the lower compression temperatures make it difficult to achieve compression ignition below 75% load. To compensate for this the mass controlled unit uses much greater catalyst surface area combined with vortex entry to ensure uniform mixing of the incoming charge. This ensures the lowest activation energy possible so as to achieve compression ignition in the pre-chamber at all loads- Ideally this system would operate in an unthrottled, stratified charge engine where pre-chamber energy and compression temperature would favor Compression ignition under all conditions. 3. ELEMENT TEMPERATURE FEEDBACK It has been shown that platinum wire, with its high resistivity and linear relationship between resistivity and temperature, makes it an excellent candidate for temperature sensing. This has been done with Resistance Temperature Detectors (RTD) and hot wire anemometers as well as mass flow sensors. By monitoring the resistance of any given element, one can derive that element's temperature if the temperature coefficient of resistance for that wire is known. Since we are using platinum as the heated catalyst and heating it via a constant current power source, the voltage across the element will vary in direct proportion to the resistance, which varies in direct proportion to the temperature of the element- This method can be used to monitor the chemical reaction on the catalyst surface and the temperature of the surrounding gases (2). Under operating conditions, the platinum wire will reach a thermal balance where the electrical energy in will equal the heat energy out. This will be determined by the heat capacity and mass transfer rate of the Surrounding gas and the heat sink effect of the supporting structure. When thermal balance is attained, the temperature of the gas surrounding the catalyst will be approximately equal to that of the catalyst's temperature. As long as this gas temperature is higher than bulk cylinder gas temperature, changes in cylinder temperature will not effect catalyst temperature. This being true, the only phenomena that will change the catalyst temperature will be a chemical reaction on its Surface or an increase in the bulk gas temperature above that of the catalyst's temperature or a change in the mass transfer rate due to turbulence in the vicinity of the catalyst. Therefore, when a deviation from steady state temperature occurs, it must be because of one of the above reasons. By using crank angle and cylinder pressure traces one can determine which is occurring. If catalyst temperature deviates during the compression curve, this must be due to Surface reaction heating or turbulent cooling. This then becomes the indicator of the contact angle when the interface reaches the catalyst's location in the timing chamber. In a non-mass controlled CPT, the rate of heat input into the catalyst after fuel air contact is determined by charge density and the Surface area available to the charge that is above the reactive temperature. Therefore, since the Surface area of the catalyst does not change, the rate of heat released to the wire from the given charge will be approximately constant and thus the temperature change will be linear at a given density until the temperature difference between the Surface reaction and catalyst becomes small. Since the cycle time between contact angle and ignition is relatively short (120 ms or less), one can assume that the heat balance between surface reaction and catalyst temperature will [lot Occur. This implies that any change in the rate of temperature change must be due to gaseous combustion of the surrounding gasses- As gaseous combustion proceeds from prechamber to the cylinder, both the bulk gas temperature and cylinder pressure rise until combustion rates begin to fall and cylinder expansion begins. The change in element temperature will follow the same pattern until the bulk gas temperature falls below the thermal equilibrium of the catalyst. This their implies that the peak rate of temperature change will coincide with peak temperature in the cylinder. In mass controlled CPT the temperature of the element is directly related to the instantaneous density of the cylinder. Since the element is exposed to the gas stream in such a way as to be cooled by it, the primary phenomena governing element temperature is mass transfer due to turbulent flow. When no fuel is in the mixture, the element losses temperature in proportion to charge density during the compression stroke. With ail air/fuel mixture, the element either gains or losses heat based on which phenomena is more dominate, catalytic surface reaction or turbulent cooling. In this way energy content and density of the mixture can simultaneously be measured for each cylinder. The principle of element temperature feed back gives valuable insight into the operation and relationships of mass flow and catalyst reactivity in the pre-chamber. With further development this may prove to be a viable way to measure in real time charge mass and air fuel ratio on a cylinder to cylinder basis 4. THERMAL MASS RESPONSE TIME The thermal mass response time is an important design parameter. This factor determines the time it, takes between turning the power on and cranking tire engine. Also, if forced ignition is desired, it allows one to determine how far in advance and how much voltage must be applied to achieve the desired results. It will also tell how many cycles will be needed to achieve a change in ignition timing for a given change in current. Finally, it will aid in the proper design of a mass controlled timing igniter. The platinum element has a thermal inertia based on its specific heat and mass. If this element were in a zero heat-loss (adiabatic) environment, a constant energy input would cause the temperature to rise in a linear fashion until the element melted. If the same element were placed in a infinite heat loss (isothermal) environment, the temperature of the wire would stay constant regardless of the energy input. In real life, the environmental conditions are somewhere in between. Since the thermal inertia of the wire is linked to physical constants such as mass and specific heat, this factor also will remain constant. The other factor which comes into play is the heat loss characteristics of the environment. This is governed by the specific heat of the substance surrounding the catalytic element and the mass transport rate due to thermal convection and radiant heat loss. Assuming that radiant heat loss is directly proportional to the temperature, the only non-linear heat loss factor is convective losses. This could be described as a thermal RC circuit where thermal balance (heat in = heat out) is a fully charged capacitor for a given energy level. The "R" factor would be the thermal inertia or resistance to change of state o f the wire. The "C" factor would be the mass transfer rate and specific heat of the gas. This system then should produce a time constant (RC) that would be characteristic for its construction, 5. TIMING ADJUSTMENT BASED UPON MASS The final concept, called mass controlled timing, is an improvement on the basic timing chamber concept for which ARI has applied for patent protection. As mentioned earlier, the ignition delay in the first CPT prototypes increased when density decreased, producing a timing curve proportional to load but with a negative slope . This design has a timing delay that is also proportional to load but has a positive slope that can be optimized for any given engine. The design is based on the concept of hot wire anemometry. The heated catalytic element is placed in a fixed location that will provide optimum light load operation, i.e. most advanced. The catalytic element is exposed to the gas stream in such a way as to be cooled by it. Since the velocity of the interface mimics piston velocity times the pre-chamber/stroke ratio and since the heat removal ability of the passing gas is directly proportional to its mass, then the final temperature of the catalytic element will be inversely proportional to the mass or density of the charge. This will cause the ignition delay to increase with increasing load, giving the timing curve a positive slope. As engine speed increases so will gas velocity. But at the same load, the net mass passing the element remains constant but the time to remove heat decreases. This should cause the catalyst temperature to increase with faster engine speed thus providing dynamic advance. Therefore, for a given load, the ignition delay decreases, in terms of crank angle, as rpm increase. Detonation and intake air temperature also can be easily compensated for by dropping current as the signal from the detonation sensor or temperature sensor increases. Hence, this ignition system dynamically responds to load specifically tailored to the need of each cylinder. Very simple and reliable controls can compensate for temperature and detonation as necessary. Also note that as atmospheric pressure decreases, this system will inherently advance thus compensating for changes in altitude and weather. Further, as the engine wears, thus lowering charge density, this system will also compensate by advancing the ignition event as needed for the weak cylinder. MATHEMATICAL MODELS This section give the mathematics to Support the foregoing conceptual models. 1. TIMING CHAMBER In order to determine the function of a particular CPT/engine combination, one must know four variables. These are rod length, stroke length, chamber length, and compression ratio. Using the following equations one can predict the contact angle of the system.

Image coming soon

Image coming soon

Piston position= Piston ratio (Rp) = Sp/sl Chamber/stroke ratio (Rcs) = cl/sl Interface location (Li) = (Rcs)Sp Contact ratio (Rc) = Rp for a given piston position Contact Angle (Ca)=

Image coming soon

* Where: rl = rod length; cr = crank radius; Sp = piston position measured from top down', sl = cylinder length; cl = chamber length. In this, one can see that the compression ratio determines the final volume of the gas spring. The mathematical reasoning behind the interface position mimicking piston position is based upon the compression ration being equal throughout the cylinder. Since the pressure in the chamber is equal to the pressure in the cylinder at any point in time, the compression ratios of chamber and that of the cylinder must also be equal as well. This ratio can be expressed as C, of cylinder = Cr of chamber Cr of cylinder = Cr of chamber

Image coming soon

Where Lc2 is the inter-face position within the chamber.

Image coming soon

After this project we mathematically explored the effects of different contact ratio adjustment strategies. The three we explored were: Fixed primary volume with an adjust-able secondary similar to our first prototypes.

Image coming soon

Fixed secondary with an adjustable primary such as adding an extension to [tie mouth of the plug as we did in this test.

Image coming soon

Fixed chamber length with an adjustable catalyst location.

Image coming soon

Of the three approaches, only one proved to have a nearly linear response to a given distance of adjustment. This was the adjustable catalyst strategy (variable contact ratio). All of these graphs assume that the igniter exists on a two dimensional plane perpendicular to a point on the axis of the chamber. In actuality, this cannot be achieved although it is approximated as far as possible. This and the decreasing distance traveled by the interface towards TDC set limits on the maximum retard for contact angle that can be achieved with this system. When the effects of gaseous diffusion at the interface and the length of the catalyst are considered, practical contact angle retard cannot be less than approximately 15' BTDC. The effect of compression ratio on the before mentioned relationships is to add a fixed volume to the total length of the pre-chamber. A lower ratio requires a larger addition while a higher ratio requires a smaller addition. This addition accommodates the exhaust gas spring at TDC. When the compression ratio is relatively high and the chamber/stroke ratio is less than 60%, the additional volume can be ignored due to turbulent and diffusive mixing at the interface (since this advances the interface position) prior to the compression stroke. 2. PRE-CHAMBER ENERGY The energy in the pre-chamber at the point of ignition determines the ignition effectiveness and rate of main chamber combustion. Several factors effect the energy density of the pre-chamber. Heat density determined by air/fuel ratio, charge density, heat added by compression and exhaust gas concentration all play a part in energy density. These variables to be considered are: Fuel/air ratio. Heat content of fuel. Pre-chamber volume ratio. Percentage exhaust gas. Heat of compression. Volumetric efficiency. One can see this system favors full load conditions where high energy density along with high compression temperatures lead to compression ignition in the prechamber which produces a rapid and powerful torch to ignite the remaining charge. The relative energy density compared to the main chamber should stay constant if the charge is perfectly homogeneous. Although the amount of inherent EGR remains constant, its percentage of the fresh charge increases dramatically at light loads (Lip to 50% of the fresh charge) due to throttling. This would cause slower combustion rates Under these conditions. 3. TEMPERATURE MEASUREMENT The resistivity of pure platinum at 20'C is 10 x 10-6 omega cm and 19.0 x 10-6 omega cm for platinum containing 13% rhodium. We know that resistance times area divided by the length equals the resistivity of a given element and that there is a linear relationship between resistivity and temperature. Therefore, by monitoring the resistance of the element during operation, we can calculate the magnitude of temperature change and gain some valuable insight into the operation of CPT. The resistance measurement is based on ohm's law which states that R = V/1. If "I" remains constant then AR OC AV. In practice, the element was powered by a current regulated power supply and the voltage across the element was monitored via *an oscilloscope timed by the engine cycle. This enabled us to describe temperature changes in terms of crank angle relative to BTDC 4. RESPONSE TIME Using a current regulated supply and monitoring the voltage across the element with an oscilloscope, the typical response curve of a 1.5 volt catalytic element was captured and analyzed to determine the mathematical equation that described this line. The response curve resembles that of a partially charged capacitor that is undergoing additional charging. The derived equation which mimics a charging RC circuit is:

Image coming soon

V = resultant voltage to maintain constant current E = maximum voltage at thermal balance T = time in milliseconds K = constant (in milliseconds) M = thermal inertia H = thermal balance (energy in = energy out)

Image coming soon

The K constant adjusts the equation for its "pre-charged capacitor" state. The time constant (M - H) for this response curve proved to be 525 ins. For automotive applications this would mean that this ignition Would have instant response on start-up. For ignition timing modification via current control, a higher or lower temperature could be achieved in 13 cycles at 3000 RPM. Using this equation, one could determine the magnitude of voltage and advance time necessary to force ignition at a given time with pulse width modulation of a higher voltage. This method also could be used to achieve faster transient times to a new temperature within one or two cycles at any speed. The following factors all play a role in deriving the factors in the equation: The mass of the element times its specific heat (thermal inertia). Heat loss to supporting structure based upon AT. Heat loss to surrounding environment (radiant and convective) based upon AT. Specific heat of surrounding gas. Mass transport rate. Based on these factors, one can design an ignition system to meet the given conditions. The objectives of this project were as follows: To quantity CPT's natural timing tendencies and controllability. To verify and quantity timing adjust ability via current control. To compare CPT's prototype performance with conventional ignition. To quantify CPT's lean stability limit. To find air optimal configuration for timing characteristics and lean stability. RESEARCH EQUIPMENT The following equipment was used during the test procedure: Hercules G2300 ST natural gas engine provided by Hercules Engines, Inc. Stuska 90 HP water brake with dyno load cell BKM's test cell #2, including pressure gauges and temperature sensors . Merriam Laminar air flow element Sierra instruments electronic mass fuel flow meter PCB Pieziotronics miniature pressure transducer and charge amplifier AMI 486 33 MHz computer to run PEI's Engine Cycle Analyzer (ECA). Encoder Products Inc.'s 1800hev crank angle encoder Tektronix 5441 storage oscilloscope. Bondwell B310 286 12 MHz compuler for spreadsheet data entry and computation. RESEARCH METHODOLOGY BASELINE The data was taken for both baseline and CPT with a 4x3 matrix comprising of four different speeds at three different loads. LOAD/RPM 1300 1500 2100 2800 30% 30% 30% 30% 60% 60% 60% 60% 100% 100% 100% 100% These percentages were determined from the spark ignited baseline at 100% load. 100% means wide open throttle (WOT) using the load to maintain a given RPM. When running the CPT matrix, 100% means WOT with whatever load would maintain the set RPM. At 100% load, torque is used as the comparison. At lower CPT loads, the baseline's torque was matched and the thermal efficiency and BSFC were used for comparison. The Stuska dyno was rated at 67 kW at 7000 RPM. Since the engine produced almost 60 kW at 2800 RPM, the dyno needed to be driven at 2.5 times the engine speed in order to exert sufficient load. Even though this arrangement caused a small reduction in indicated horsepower due to belt and loaded bearing losses, it still gave comparable figures. Pressure sensor data was captured by a 500 KHz A/D converter and stored on a hard disk for later analysis by the ECA. Baseline was taken according to the 4x3 matrix. The test conditions were spark ignition at Lambda of 1.00. CPT OPTIMIZATION AND TESTING We began running the CPT igniters which had formerly been in ARI's Quad Four, demonstration vehicle These igniters had been originally optimized for a 9:1 compression ratio. By analyzing the pressure traces from these igniters, we found that the timing was too advanced for a stoichiornetric mixture as would be expected if installed into an engine with a 10:1 compression ratio. Therefore, a modification to the primary volume was done by lengthening it 2.5 cm, using a threaded extension. This retarded the ignition event to approximately the proper location, though it was still slightly advanced. With this configuration running at 1500 RPM, we found that the engine produced 2% more torque than spark ignition at MBT and exhibited slightly better fuel consumption. We ran all the rest of the data points with this configuration. Before we ran the complete matrix we spent two weeks exploring an optimal configuration for the Hercules engine. The parameters evaluated for optimization were: Igniter position. Chamber length. Torch distribution. Density sensitive timing. As mentioned previously, the first adjustment made was chamber length in order to retard contact angle of the fuel/air mixture and the ignition event. Contact angle is the moment when the interface contacts the element in terms of crank angle BTDC. It was found that approximately 2.5 cm added to the primary length retarded the peak pressure angle to within 4' before MBT spark timing. Peak pressures were about 10% higher, and there was no evidence of detonation. Mass burn rates were equivalent.

Image coming soon

Next we designed, built and tested several torch distribution tips, i.e. "showerhead", aimed at evenly dividing the combustion chamber with as many as 16 individual torches. Of the six designs tested, we found that an 8 hole, 120', included angle tip resulted in the best combustion characteristics. With this design mass burn rates approximately doubled, consuming the entire mixture 15' CA for CPT compared to 36' CA for spark.

Image coming soon

This design concentrates on igniting the most difficult areas around the periphery of the combustion chamber then letting the flame propagate towards the center. The eight nozzles enter the CPT's chamber tangential to its cylindrical bore thus creating high turbulence and a well mixed charge before the interface reaches the catalyst. Note that only the mixture following the interface is well mixed. All of the early CPT designs have the common feature of a peak pressure position which advances with load. This is due to the proportional increase in the energy available to the catalyst as the cylinder density increases. Therefore in order to keep full load operation near MBT it is necessary to sacrifice pail load timing. The timing of peak combustion pressure can vary by as much as 60' CA between 100% load and no load conditions. Therefore, we designed built and tested a design which would retard with load. In operation, this design operates in a similar manner to a thermal dispersion mass flow sensor. Since the gas velocity in the chamber is directly proportional to piston speed regardless of load, the only variables are the density and composition of the gas. By designing the chamber in such a way that the flow of gas into the pre-chamber cools the catalyst, the timing of the ignition event can be retarded with load. Our first design works so well in fact, that it will actually extinguished combustion when the load is increased sufficiently. Because of the limited time, the final design still needed further optimization to achieve proper thermal balance. Therefore, the rest of the test matrix was run simply with the extension added to the primary volume. This configuration performed as well or better than spark ignition at full load but performance degraded at part load conditions, as would be expected without optimized timing. Also due to lack of time, we were not able to quantify the effects of current adjustment other than to verify that it does move the timing event. All of the test data was taken with fixed current configuration. TEST RESULTS In order to determine contact angle and gaseous combustion, we monitored element resistance. The contact angle of deviation appeared to be approximately 54' BTDC regardless of load or speed. This is just what the mathematical model predicted. This phenomena appeared very stable and steady with very little cyclic variation. These preliminary results indicate that the concept of controlling the moment of contact is feasible and stable. Gaseous ignition timing also seemed stable. This was based on the assumption that the moment of the torch exiting the chamber coincided with the beginning of Measurable heat release. The fluctuations in peak pressure angle and peak pressure seem to be caused not by fluctuations in the ignition event but by large differences in the rate at which the charge burned. Peak pressure angle at full load was almost always 4' before spark ignition, yet performed equal to or better than spark ignition at these conditions.

Image coming soon

The mass burn rates of CPT (the non-optimized design) versus spark seem to be equal at full load. As load decreases, total burn time increases. Of note is the remarkable performance of the torch distributor in this area. At 30% load the CPT with torch distribution burned the mixture approximately twice as fast as the spark ignited mixture under the same conditions.

Image coming soon

Unfortunately, time didn't permit us to run the data matrix with this configuration. Also, the burn rate at the lambda = 1.5 was a noteworthy 15' of crank angle compared to 30' with spark. There is much more data that could be explored here but time constraints require us to be specific and brief. The Combustion Quality Index (CQI), as defined below from the ECA technical manual, was calculated by hand. COMBUSTION QUALITY INDEX In order to discriminate between engine configurations creating low NOX emissions due to poor combustion stability and those which consistently ignite the very lean mixture, yet yield very low emissions levels without a loss in power, Enterprise engineers developed a dimensionless parameter, the Combustion Quality Index or CQI, for evaluating combustion stability.. CQI is mathematically defined as follows:

Image coming soon

Typical values of CQ1 for standard turbocharged engines are shown in the following table 2. TABLE 2 COMBUSTION QUALITY INDEX for TYPICAL TURBOCHARGED COMBUSTION SYSTEMS SYSTEM Diesel Standard Dual Fuel Standard Spark Ignited Lean Dual Fuel Lean Spark Ignited Pre-Chambered Spark In general, a minimum CQI of 4-6 appears desirable for generator drives, a good system will have a CQ1 of 1215 and an excellent one over 20. Please note the statistical results comparing spark (MBT) of lambda 1.5 and CPT of lambda 1.5 at 1200 RPM. Spark showed a CQI of 18 while CPT showed a CQI of 40.

Image coming soon

Image coming soon

CONCLUSIONS MOST SIGNIFICANT RESULTS The most significant results of the CPT demonstration project are that CPT: Has a similar effect to a fuel-rich spark-ignited pre-chamber. Demonstrated faster combustion rates - almost two times as fast as spark. Ignition timing can be adjusted via current or voltage control using a simple electronic interface. Improved version is inherently self-timing with respect to load and self- compensating for individual cylinder and atmospheric conditions. Can lower pre-chamber ignition energy requirement via timed selective generation of radicals prior to combustion. ADVANTAGES Based on the data gathered during this test procedure, and other tests conducted by ARI, CPT has the following advantages over spark ignition: Lower cyclic variation under lean conditions. No high voltage needed. No EMI. Higher- peak pressures. No moving or consumable parts (less maintenance). * Self-sustaining (no external energy required). Waterproof. PROJECTED ADVANTAGES The following are projected advantages of CPT: With the right compression ratio, CPT has an unlimited lean limit. Less detonation prone (due to different chain - branching reactions). In-cylinder reduction of NOx DISADVANTAGES The following are observed disadvantages of CPT: More heat to water jacket. Sensitive to inherent EGR (effects rate not timing). EGR effects worsen at light loads - % EGR larger. Must have homogeneous mixture for operation. SOLUTIONS The following are successful, recently tested solutions for CPT's disadvantages-. Fabricate from materials that limit heat flow to water jacket. Use mass controlled timing with extremely advanced contact angle. Use vortex entry to ensure well mixed prechamber charge. Deliver fuel to optimize torch content at ignition (special stratified, direct injection) in combination with an unthrottled engine using fuel delivery to control load and speed. FUTURE RESEARCH Since the close of this project, ARI has further developed and studied the principles of mass controlled CPT. ARI's findings have been very encouraging. ARI plans to further perfect this concept based oil the present data base and refine it to the point of commercial application. The benefits of catalyzed ignition warrant every effort to make this concept a commercial reality since social, economic, political and environmental concerns are forcing the internal combustion engine to be clean and efficient and cost effective to meet the needs of the 21st century. Since the invention of the sparking plug, property timed catalytic compression ignition represents the first major advance in Otto cycle ignition technology in over 130 years. REFERENCES 1. Beyerlein, S. and Wojcicki. S., A Lean-Burn Catalytic Engine, SAE paper 880574, 1988. 2. Law, C. K., Catalytic Ignition and Combustion of Lean Mixtures, NTIS, 1984. 3. Cherry, M., Catalytic-Compression Timed Ignition, United States patent 5,109,817, filed November 1990. 4. Cherry, M., Timing Chamber Ignition Method and Apparatus, United States patent 4,977,873, filed June 1989 5. Daneshyar, H. and Orme, J. R., Cyclic Compression, Combustion and Expansion of a Fuel and Air Mixture in a Cylinder Containing a Catalyst, Cambridge University, 1981. 6. Dale, J. and Oppenhiem, A., A Rationale for Advances in the Technology of I.C. Engines, SAE paper 820047,1982. ACKNOWLEDGMENTS Funds for this research project were provided by Altronic, Inc. of Girard, Ohio and BKM of San Diego, California. The Hercules G2300 engine was loaned to the project by Hercules Engines, Inc. ARI expresses its gratefulness to Dr. N. John Beck for his valuable insight and advice given during this project and to the staff of BKM for their assistance in carrying out the research.