With the need to reduce CO2 emissions, there has been a downsizing trend in recent years, and engine cylinder capacity has been reduced. [309008]

INTRODUCTION

With the need to reduce CO2 emissions, there has been a [anonimizat].

[anonimizat]. [anonimizat].

RENAULT makes a [anonimizat] 90, [anonimizat], supercharged gasoline engine with 899 cc.

Figure.1. H4Bt 408

The name Energy was created to symbolize excellence in Renault's [anonimizat] 34 years of experience in F1 and 10 World Champion titles.

The ENERGY TCe 90 engine is available in two versions:

1. ENERGY TCe 90 (EURO 5) [anonimizat].

2. ENERGY TCe 90 ECOLEADER, a high-[anonimizat] a computer-controlled thermostat to optimize CO2 emissions.

MARKETING

Figure.2.1. Renault Clio IV

Renault ENERGY TCe 90 compared to the competition.

Figure.2.2. Comparation with the competition

RENAULT QUALITY MADE

Expertise Formula 1: Experience over 30 [anonimizat]: 14,000 hours

Road test: 800,000 km

Test engines: At the end of the manufacturing process

ENGINE PRESENTATION

Particularities of the H4Bt engine

The arhitecture of the three cylinder Oil pump

Figure.4. Particularities of the engine

Figure.4.1 Thermal management

Engine identification

Figure.4.2 Engine identification

SYSTEMS AND SPECIFICATIONS

Figure.5.1

Table.1.[anonimizat]. It also increases torque to full load by increasing the overlap between intake and exhaust valves. Camshafts are empty tubes and have the target for the position sensor. The bearings are covered with a special layer that reduces the frictional force. [anonimizat] (Diamond Like Carbon) is a method of surface treatment where friction is significantly reduced and energy efficiency improves.

Figure.5.2 Cams and cam follower

Aluminum cylinder head

The production process of the intake channels has been improved to get better combustion. This material reduces engine weight as well as consumption and CO2 emissions.

Engine block

The engine block is made of aluminum. This material reduces engine weight as well as consumption and CO2 emissions

Figure.5.3. Carter of cylinders

Pistons with graphite coating

Piston graphite protection layer reduces friction. This technology used on Formula 1 vehicles reduces consumption and CO2 emissions.

Figure.5.4. Pistons with graphite coating

Indirect injection

The H4Bt is a multipoint injection engine with one cylinder injector. Injectors with solenoid have 8 nozzles that provide better spraying.

Figure.5.5.Indirect injection

Thermal management

Figure.5.6.[anonimizat] < 80°C

(1)Expansion vessel; (2)Thermal management valve; (3)Thermostat valve; (4) Exchanger (water / oil); (5) Air heaters; (6) The radiator.

[anonimizat] (2) closes the water circuit around the combustion chambers.

Result:

– Engine temperature increases faster;

– The cooling fluid present around the combustion chambers heats up faster (because it is not still circulating through the system);

– Consumption is reduced (better combustion when the engine is hot, low friction as the oil becomes more fluid in a shorter time).

Figure.5.7.Cooling system – temperature < 80-95°C

Open Thermal Management Valve – Once the engine has reached the right temperature (80°C), the valve opens for a standard cooling system operation.

The coolant circulates through the entire closed circuit (except for the radiator when the thermostat (3) is not open):

– Return to normal standard operation.

– The thermostat takes over the valve.

Figure.5.8.Cooling system – temperature > 95°C

Open thermostat valve – Once the coolant has reached the optimum temperature, the thermostat (3) opens.

The coolant circulates through the system, including through the radiator.

The coolant temperature is adjusted by opening and closing this valve.

Oil pump

Figure.5.9.Oil pump – low speed operation

(1) To the engine; (2) Exhaust valve: The control valve is not to be confused with the main discharge valve located upstream of the pump and not represented in the picture; (3) To the oil crankcase; (4) Oil inlet; (5) Engine oil return: It comes from a connection port from the oil distribution ramp. This connection port is located after the oil filter.

Low speed operation: 1000 rpm.

The system tries to reach an output pressure of 1.8 bar.

At lower speeds, the electrical is open to allow the oil to escape from the chamber (a). Output pressure is set to 1.8 bar.

Cubic capacity is high because the "rotor" is offset from the "stator".

At low revs, the oil pump works in the standard mode. The fluid flows through the pump due to the rotor rotation that drives this movement.

Figure.5.10.Oil pump – Intermediate speed operation: 4000 rpm.

Intermediate speed operation: 4000 rpm.

The system tries to reach an output pressure of 1.8 bar.

The electrodes are still open to allow the oil to escape from the chamber (a). Output pressure is set to 1.8 bar.

The cubic capacity of the pump decreases to adjust the outlet pressure despite the increase in engine speed.

Figure.5.11.Oil pump – Intermediate speed operation: 5000 rpm

Intermediate speed operation: 5000 rpm.

The system tries to reach a 4 bar outlet pressure.

The electrodes are closed, the oil fills the chamber (a). Output pressure is set to 4 bar.

The cubic capacity of the pump is again high to adjust the outlet pressure.

Introduction

Knocking combustion research is crucially important because it determines engine durability, fuel consumption, and power density, as well as noise and emission performance.

Current spark ignition (SI) engines suffer from both conventional knock and super-knock. Conventional knock limits raising the compression ratio to improve thermal efficiency due to end-gas auto-ignition, while super knock limits the desired boost to improve the power density of modern gasoline engines due to detonation.

Conventional combustion has been widely studied for many years. Although the basic characteristics are clear, the correlation between the knock index and fuel chemistry, pressure oscillations and heat transfer, and auto-ignition front propagation, are still linearly stages of understanding. Super-knock combustion in highly boosted spark ignition engines with random pre-ignition events has been intensively studied in the past decade in both academia and industry.

These work shave mainly focused on the relationship between pre-ignition and super-knock, source analyses of pre-ignition, and the effects of oil/fuel properties on super-knock. The mechanism of super-knock has been recently revealed in rapid compression machines (RCM) under engine-like conditions. It was found that detonation can occur in modern internal combustion engines under high energy density conditions.

Thermodynamic conditions and shockwaves influence the combustion wave and detonation initiation modes.

Three combustion wave modes in the end gas have been visualized as deflagration, sequential auto-ignition and detonation. The most frequently observed detonation initiation mode is shockwave reflection induced detonation (SWRID).

Compared to the effect of shock compression and negative temperature coefficient (NTC) combustion on ignition delay, shockwave reflection is the main cause of near-wall auto-ignition/detonation. Finally, suppression methods for conventional knock and super-knocking engines are reviewed, including use of exhaust gas recirculation (EGR), the injection strategy, and the integration of a high tumble-high EGR.

Definition of knock

The detonation occurring in spark ignition engines is an abnormal phenomenon of combustion of the fuel mixture in cylinders, characterized by the fact that, although the burning starts correctly, when the spark occurs, it does not only continue by advancing the flame front but in the rest of the combustion chamber are areas where the fuel mixture self-ignites, causing abnormal engine operation.

How the knock appear

Detonation is determined by the chemical reactions that take place in front of the flame front. Organic and aldehyde peroxides are formed which, due to thermal and chemical inhomogeneity in the combustion chamber, favor the emergence of cold flame cores, initially propagating at a rate of tens of meters per second.

Cold flame nuclei cause the occurrence of blue flame nuclei that propagate faster. Finally, a succession of local auto springs of the fuel mixture in neighboring volumes appears, with the flame front apparently moving at speeds of 1-2 km / s. It is considered that the detonation in the SI is a low temperature polystyrene ignition process.

Another causes:

The temperature and the pressure at the end of the compression is to high;

The lower the octane rating, the lower the resistance to self-ignition is;

Ignition spark to high, etc.

The effects of knocks in internal combustion engines

The SI engine can be damaged by knocking combustion in different ways: piston crown melting, piston ring sticking, cylinder bore scuffing, piston ring-land cracking, cylinder head gasket leakage and cylinder head erosion. To enhance power density and reduce fuel consumption, high boost with direct injection has become the main stream technology in SI engines in recent years and a new knocking mode, called super-knock has become a challenge for engine designers, especially with respect to direct injection engines in the low-speed, high-load operating regime. Fig.1 shows typical damage resulting from conventional knock and super-knock.

Figure.1. Typical damage caused by engine knock

Basic characteristics of engine knocking combustion

Fig.2.Combustion parameters of engine knock cycle

Engine knock arises from auto-ignition of the end gas ahead of the propagating flame.

Fig.2 presents the pressure trace, pressure oscillation, heat release rate (HRR) and unburned gas temperature (T) of a typical knocking case. The combustion process of the knocking case has two stages: flame propagation induced by spark ignition and end-gas auto-ignition causing pressure oscillation.

The flame propagation stage is from spark ignition to the crank angle of CAko (the onset of pressure oscillation), reaching the value of Pko.

The heat release rate shows an increasing tendency generally during this stage, with possibly a short term drop caused by the down ward movement of the piston and heat transfer. The unburned gas temperature rises stably, caused by the compression heating effect of the burned gas and propagating flame, plus the compression or expansion due to the moving piston. The knock sensor signal shows no oscillations, since only flame propagation exists and there is no pressure oscillation in this stage. In the auto-ignition stage, the pressure trace at first increases dramatically, reaching a peak value, and then it oscillates with decaying amplitude. The knock sensor signal also oscillates from CAko. As for the combustion process, at CAko, the pressure and temperature of the end gas reaches a high level, causing the auto-ignition combustion phenomenon in the combustion chamber. Once auto-ignition is induced, a pressure wave can propagate into the chamber, reflecting back and forth from the walls, causing pressure oscillations.

Four important parameters that characterized knock behavior

Heat Release Rate (HRR), Temperature of the unburned gas (T unburned), maximum pressure rise (Λp) and Knock Intensity (KI) are four important parameters that characterized knock behavior. HRR is calculated according to the first law of thermodynamics. In the calculation of T unburned, the end-gas compression process is considered to be adiabatic. KI is integrated using the pressure oscillation obtained by high-pass filtering (HPF), as follows:

Heat release rate (HRR)

Temperature of the unburned gas (T_unburned)

Knock Intensity (KI)

The effect of oil and fuel properties on knock

Research on knock has indicated that the composition and physicochemical properties of the oil and fuel also affect seriously the frequency of knock.

Oil properties

Base stocks

Pre-ignitions were thought to arise from local auto-ignition of areas in the cylinder which are rich in low ignition delay “contaminants”, such as engine oil and/or heavy ends of gasoline. These contaminants are introduced in to the combustion chamber at various points in the engine cycle. In general, less reactive base stocks (i.e. those with along ignition delay) appeared to have the lowest ignition tendency. The calculated ignition index, which is an estimate of heavy hydrocarbon reactivity based on empirical measurements from bunker fuels, was found to predict the ignition tendency of group I-IV lubricant base stocks (Fig.3).

Fig.3.The effect of base stock on LSPI

Oil additives

Many researchers have noted that oil additives correlated directly with the frequency of knock due to possible catalytic reactions. The frequency of LSPI increased with the content of additives.

A new engine oil formulation has been developed, which reduced LSPI frequency to less than 10% of that of conventional ILSAC certified gasoline engine oils. High quality base oils and optimized additive components were formulated in which the amount of calcium-based detergent was reduced to levels lower than that in general ILSAC oils, and anti-oxidants were added.

Based on these findings, a correlation to estimate LSPI frequency (relative) was obtained:

In an assessment of engine oil degradation effects, the influence of wear metals and engine oil degradation was also investigated. It was found that addition of Fe and Cu compounds clearly showed contributory effects on LSPI frequency.

In addition, the observed auto-ignition frequency decreased with increasing viscosity and density, which is related to the physical property of the additives.

However, so far there ported work does not provide an explicit conclusion about the effect of oil additives. For example, there is no consensus on whether Zn and Mo have effects on pre-ignition.

This is because pre-ignition in engine production is random and oil additive effects are only one of the possible pre-ignition sources. To isolate the effect of oil additive on pre-ignition, a single factor experimental method needs to be developed to reproduce pre-ignition and knocking are search engine.

Fuel properties

The fuel's chemical composition, octane number and volatility have effects on the frequency of knock. On one hand, the quality of the fuel directly influences the ignition properties of the mixture. On the other hand, fuel droplets may play a role in the origin and progress of pre-ignition.

Composition

The effect of fuel composition on knock was experimentally investigated in a boosted direct-injection gasoline engine. Despite similar RON and MON ratings, the knock characteristics of the test fuels indirect injection SI engines were different. Fuel blends with high levels of aromatics increase knock frequency, where as a low aromatic fuel and E10 (10% mass of ethanol) reduced knock frequency under different test conditions of A/F ratios and EGR levels, as shown in Fig.4.

Fig.4. The effect of fuel composition on knock

In essence, the fact that a high aromatic content increases the knock frequency can be explained by two main reasons. Aromatic combustion leads to soot formation and enhanced deposits, and these solid particles are possible sources of pre-ignition for knock.

Moreover, soot and deposits also coat the surface of the combustion chamber, which resists heat transfer from walls. The reduced heat transfer leads to higher temperatures and pressures in cylinder before spark timing, which also promotes pre-ignition. For ethanol addition, this influences the latent heat and flame propagation and will be further discussed in the following section: octane number.

Octane number

Fuel octane number has comprehensive effects on pre-ignition and super-knock. Most experiments in the literature show that there is no correlation between pre-ignition propensity and RON or MON.

Some experiments show that pre-ignition frequency decreases with increased octane number. Pre-ignition temperature tends to be higher with increasing RON, and pre-ignition temperature is more related to RON than MON.

However, it was found that ethanol had a relatively high pre-ignition tendency with high intake manifold pressure, although its octane numbers (RON and MON) are particularly high. This can be explained by the fact that fuels like ethanol and hydrogen are very susceptible to pre-ignition because of their high laminar burning velocities and smaller laminar flame thickness.

A stable flame can be established if the kernel radius is sufficiently larger than the laminar flame thickness. On the other hand, lower knock tendency with ethanol fuel than gasoline fuel may be thanks to the higher heat of evaporation of ethanol. To sum up, pre-ignition with a high octane number fuel might not lead to knock due to its higher resistance to auto-ignition.

Volatility

A higher frequency of pre-ignition was observed for a fuel featuring a higher fraction of low volatility compounds. Fuels that had a distillation point of T50 below 103°C showed a lower number of stochastic pre-ignition (SPI) events.

Fuel distillation points (T90 and T95) also have effects on LSPI frequency. It was concluded that a heavier fuel could lead to an increase in LSPI tendency.

This conclusion is consistent with the theory that high boiling point fuel components can cause more liner wetting due to poorer atomization of the fuel spray and slower evaporation after adhesion on the wall, potentially leading to more fuel accumulation in crevices and higher LSPI frequency.

Control strategy for conventional knock

Conventional knock is a race between the flame in the engine and the thermal auto-ignition of the unburned “end-gas”. The principle for avoiding engine knock is that the time off lame propagation to the end-gas is less than the time of the end-gas auto ignition. Possible control strategies presenting SI engines are summarized next:

Retarding spark timing, improving octane number and enriching mixture

Retarding spark timing, improving octane number and enriching the mixture are commonly applied approaches for knock suppression in production engines without component modifications. Retarding spark timing is the most effective method to simultaneously reduce the end-gas temperature and pressure. Lower end-gas temperature and pressure prolong the ignition delay. However, late spark timing usually leads to an un-optimized combustion phase with lower thermal efficiency. It also may deteriorate engine performance due to the decrease in mean combustion chamber temperature and pressure. Increasing the fuel octane number can be achieved by introducing anti knock additives to the fuel, like ethers, etc.

In addition, adding alcohol even water in to the combustion chamber could also increase the time for the end-gas auto-ignition and the suppress knock. Injecting excessive fuel could increase the effect of charge cooling and decrease the temperature of the mixture, which could prolong the time to end-gas auto-ignition and finally suppress knock. However, under the condition of fuel enrichment, HC and CO emissions increase, and since the TWC can only work under stoichiometric conditions, this causes problems of fuel economy and emissions.

Exhaust gas recirculation

Exhaust gas recirculation (EGR) is regarded as an effective method for suppressing knock in advanced SI engines due to the resulting increased. Furthermore, cooled EGR has more potential in knock mitigation without a loss of output power. Some demonstrated that by increasing the EGR ratio from 7% to 13%, the maximum BMEP increases with advanced spark timing. But, with further advanced spark timing, BMEP does not increase, since it is limited by knocking combustion. Increasing the EGR ratio reduces this limitation. Because an increase of cooled EGR lowers the gas temperature, and more fuel is consumed before reaching the knock temperature limit, therefore, it can further widen the ignition advance angle.

Stratified mixtures

Use of stratified mixture is a flexible approach to suppress knock and is usually achieved using direct injection. Fuel directly injected in the cylinder can lower the combustion temperatures due to the cooling associated with fuel vaporization. Therefore, the cooling can lower knock sensitivity and the compression ratio can be increased.

A two-stage injection concept was proposed by Kuwahara and Yang for knock suppression and low speed torque improvement in DISI engines, but the proposed rich mixtures (A/F ratio=12»12.5) would deteriorate fuel economy and exhaust emissions, which limited the application of this method. Use of stratified stoichiometric mixture has been demonstrated.

Using a two-stage injection strategy in gasoline direct injection engine, Stratified Stoichiometric Mixture (SSM) could suppress knocking effectively compared with a Homogeneous Stoichiometric Mixture (HSM) case, as shown by the smooth pressure trace and lack of pressure oscillations in Fig.4.

Investigation of at two-stage strategy indicated that locally slightly rich mixtures due to a late injection helps to suppress auto-ignition by providing a local cooling effect and by enhancing the local flame propagation speed. The assessment concludes that two-stage injection is an effective method for sup-pressing knock under high load operating conditions. However, the injection strategies and combustion systems need to be carefully designed and optimized to avoid soot emissions.

Fig.4. Experimental results of Stratified Stoichiometric Mixture (SSM) and Homogeneous Stoichiometric Mixture (HSM) for knock suppression

Other methods

Turbulence

Increasing turbulence intensity usually suppresses engine knock. Once combustion has started, increased turbulence leads to fast combustion and decreases the tendency to knock. However, at the start of combustion, increased turbulence also increases heat transfer from the spark electrode. To solve this problem, modern engines adopt high energy ignition (more than100mJ) to support strong turbulence.

Cooling

Cooling of the combustion chamber walls is also an effective approach to reduce end-gas temperatures. An effective method to suppress knock is changing the coolant flow patterns to improve the distribution of wall temperatures. However, increased heat transfer lowers the thermal efficiency. To suppress knock without a considerable loss of thermal efficiency, only the upper portion of the exhaust side bore should be cooled, and foam rubber can be utilized to provide heat insulation to maintain the temperature of the center and lower portions of the bore, as shown in Fig.5.

Cooling the intake charge is also useful for knock suppression. The initial condition of the end gas during a cycle is the intake air temperature. A water-cooled exhaust manifold (WCEM) with high cooling capacity was adopted to reduce the temperature of the exhaust valve surface and the EGR so as to suppress knock.

Fig.5. High tumble intake-port and cooling water jacket for knock suppression

Lowering effective compression ratio

Variable valve timing (VVT) is a practical way to change effective compression ratio at relatively low cost for the different engine operating regions. Late intake valve closure (LIVC) is commonly used at high load to achieve a lower effective compression ratio to avoid knock.

Using a variable compression ratio (VCR) is a more ideal method, but the complex structure makes the cost increase significantly However, FEV has developed a VCR engine to suppress knock and improve fuel economy effectively.

Description of the knock sensor system

The detonation sensor system allows the Engine Control Module (ECM) to control the ignition timing to achieve the best performance while protecting the engine against possible detonation levels that damage the engine. The ECM uses the detonation sensor system to test abnormal engine noise that could indicate detonation.

The detonation sensor is of the linear response type. The sensor uses piezoelectric quartz technology that produces an alternating amplitude and frequency variable voltage signal depending on the vibration or noise level of the motor. Amplitude and frequency depend on the detonation level detected by the detonation sensor. The detonation sensor is connected to the engine control module (ECM) via a signal circuit and a low reference circuit. Both of the detonation sensor circuits are protected by electromagnetic interference by a shielded grounding circuit. The shielded circuit is grounded by the ECM.

The ECM stores a minimum sound or background noise level at idling speed and uses calibrated values ​​for the engine speed line. The command module uses the minimum noise level to calculate a noise channel. A normal signal from the detonation sensor falls into the noise channel. As the engine speed and load change, the upper and lower noise channel parameter changes to accommodate the normal signal of the detonation sensor, keeping the signal within the channel. To determine which of the cylinders produces detonation, the ECM only uses the information from the detonation sensor signal when each cylinder is near the upper dead end of the ignition time. When there is a detonation, the ECM detects that the signal does not fall into the noise channel.

If the ECM detects the presence of detonation, delay the ignition timing to eliminate the detonation. The ECM will always try to adjust zero compensation or a sparkle delay. An abnormal signal from the detonator sensor remains outside the noise channel or will be missing. Diagnostics of the detonation sensors are calibrated to detect the detonation sensor circuit faults inside the ECM, the detonation sensor harness, or the detonation sensor output voltage. Some diagnoses are also calibrated to detect constant noise emitted by an external influential factor such as an appropriate / damaged non-fixed component or excess mechanical noise emitted by the engine.

Inside the Knock Sensor there is a piezoelectric ring with a metal contact in each of their faces and perfectly isolated from the body and the seismic mass. The piezoelectric sensor is a ceramic ring which polarizes when exposed to an external electric field, so that, when subjected to compressive forces can generate a potential difference. The seismic mass is a metal ring perfectly calibrated to achieve the required sensitivity, so that when placed next to the sensor and compressed by a spring washer and a threaded nut (due to the inertial force) transmits the vibrations to the sensor element. The sensor metal part (body) is responsible for transmitting the vibrations from the engine block, so that, before fitting it we need to ensure that the area is clean and in a good condition, otherwise it would not be possible to guarantee the proper functioning of the sensor.

Fig.6. Elements of knock sensor

CAPITOLUL 3 – Equipment used for experimental research

3.1 Equipment’s used engine tests

An engine test facility is a complex of machinery, instrumentation and support services, housed in a building adapted or built for its purpose. For such a facility to function correctly and cost-effectively, its many parts must be matched to each other while meeting the operational requirements of the user and being compliant with various regulations. Engine and vehicle developers now need to measure improvements in engine performance that are frequently so small as to require the best available instrumentation in order for fine comparative changes in performance to be observed. This level of measurement requires that instrumentation is integrated within the total facility such that their performance and data are not compromised by the environment in which they operate and services to which they are connected.

Engine test facilities vary considerably in power rating and performance; in addition there are many cells designed for specialist interests, such as production test or study of engine noise, lubrication oils or exhaust emissions.

The common product of all these cells is data that will be used to identify, modify, homologate or develop performance criteria of all or part of the tested engine. All post-test work will rely on the relevance and veracity of the test data, which in turn will rely on the instrumentation chosen to produce it and the system within which the instruments work.

An engine test bench must include:

– engine test chamber;

– control chamber;

– space needed for measure equipment’s;

– space needed for prepare workshop;

– space needed for parts;

Fig.3.1.Engine test bench presentation

Structural organization of an engine test bench

Minimum requirements for an engine test bench:

– Fuel supply system;

– Ventilation and air conditioning system;

– Engine cooling system;

– Exhaust gas evacuate system;

– Phonic isolation;

– Seismic block;

– Fire extinguishing system;

– Blower for cooling intercooler, oil and exhaust system;

– Engine charging equipment;

– Control unit.

Ventilation and air conditioning system

In dealing with air services to the engine test cell, there are two primary functions that the air provides for the engine: cooling and breathing. The control of one of them (cooling) may be critical to the effects that the other (breathing) will have on the engine.

As a general rule, internal combustion engines are considered to be air pumps, in that they breathe in air for combustion purposes. Therefore, it follows that the quality of that air with regard to its temperature, pressure, humidity, and condition will affect the performance of the engine.

Fig.3.2.Ventilation and air conditioning system

Attempting a very important engine is the air in the cell, as it is designed to evacuate the heat generated by the engine and cellular accessories during operation, but also to prevent the accumulation of dangerous amounts of gas and vapors. For this reason the air is conditioned in terms of temperature, pressure and humidity.

Combustion air is the air that actually participates in the combustion process (the suction air). Also to ensure repeatability and reproducibility conditions is temperature and humidity conditioned (25° C temperature and 40% relative humidity respectively). The air is first cooled and dry, and to have the temperature and humidity values ​​constantly conditioned by a heater and a vaporizer.

The combustion air conditioning system is separate from that of the air in the cell and allows the control of the two parameters regardless of the conditions in the cell, for this reason the air is directed to the intake manifold through a duct coming from the upper floor of the test cell, where it has already been optimized. For a more precise adjustment, the combustion air duct is equipped with a humidifier and a temperature probe.

The volume of air supplied by the boiler is kept constant at a level that assures the engine for the operating conditions and at the same time prevents the engine from aspiring air from the cell that has not been optimized for intake.

Combustion or induction air describes the air that the engine will draw in or induct, through the inlet tract into the inlet manifolds and into the cylinders for the combustion process. The source of this air usually is the ambient air that is present within the test cell itself (e.g. the same air that the technician is breathing-fumes, vapors and all).

The quality of the air is critical to the performance of the engine and will be reflected in the engine test results. Obviously, from the viewpoint of the customer who is commissioning the tests, undesirable air quality will produce undesirable test results and inevitably lead to wasted money, time and effort in conducting the tests. Where in-cell air quality is of an extremely dubious or poor condition with regard to the four points mentioned (air temperature, pressure, humidity or condition) or the test requires absolute guaranteed and consistent air quality, then the air for the combustion process may be drawn from outside the cell and via additional filtering equipment. In the main, air is drawn from the cell environment, and this is the focus of this section.

The temperature of the air entering the inlet manifold has a direct influence on the complete integration of the air and fuel as a mixture, with regard to the evenness and atomization of the mixture (e.g. the dispersal and size of the fuel droplets throughout the air stream and on the mass of the air fuel mixture). When the air temperature is too low, fuel does not mix as effectively with the air stream due to the higher density of the air; therefore, it tends to fall to the sides and floor of the inlet tract. This gives an uneven mixture and poor atomization, which may cause misfiring and so forth. In turn, this can lead to higher emissions of hydrocarbons and carbon monoxide. In the case of air temperatures that are too high, the air charge is expanded by its heat and, as a consequence, has a reduced density. This reduced density means that the fuel lair charge will have a lesser mass than preferred (lower total oxygen content).

The efficient flow of air through and around the test cell is important for the safety and reliability of the testing operation. Although airflow rate and extraction through the cell usually are controlled automatically, the flow around the assemblies and equipment within the test cell is affected by the care and thought applied by the technician when setting up the test. It is important that cells are kept tidy and clear of unnecessary equipment, particularly near the ventilating fans and so forth.

Engine cooling system

The hot and cold wells provide a volume of water required for the running of numerous engine and dynamometer rigs under test at the same time. (Dynamometers are cooled by the water or air.) The wells provide the means for an infinitely variable and controllable engine heat range by controlling the flow and temperature of the raw water through the cell manifolds.

This system is designed to provide instant engine cooling for repeated cold-start condition testing by rapidly dissipating the heat of the engine coolant at the heat exchangers.

Water is lost through evaporation due to high temperatures in the system and through deliberate leaking. This deliberate leaking is done to ensure that the cooling water is recycled with fresh water over a period of time to prevent excess buildup of impurities and water quality deterioration. The bleed-off reduces the requirement to drain and replenish the system completely on a regular basis.

Fig.3.3 Water cooling towers

It must be emphasized here that the engine coolant (with or without antifreeze mixed into the coolant) is separate from the cooling raw water.

The engine coolant is stored in a host radiator chamber that also may incorporate the heat exchanger, with separate chambers keeping the engine coolant away from the cooling raw water. The components of the engine cooling system and their relative positions and connections are critical and must conform to the requirements laid down.

The cooling raw water carries away the engine coolant heat to the hot well and cooling towers. Frequently, a customer requests that his or her engine be run on a particular coolant and antifreeze mixture for certain tests.

This mixture is set up for the coolant system of the engine, and the cell raw water services are used to maintain running temperatures as required by the test procedures. Note that all antifreeze solutions are not the same.

Therefore, care should be used to ensure that technicians select the correct antifreeze, per the instructions of the engineer and client.

This is particularly important when ultra-low emission levels are being measured, for the heat transfer rate through the mixture can affect the combustion process. In addition, the onset of nucleate boiling and effective heat transfer can be adversely affected by incorrect coolant mixtures, making repeat tests at some time in the future practically impossible.

Fig.3.4 Heat exchanger

The main function of heat exchangers is, as the name suggests, to exchange heat. It is important to remember that heat can be exchanged in both directions using the heat exchanger system.

The new generation of plate heat exchangers has been designed as a low-cost alternative to the shell tube units.

High-pressure devices are made from grade 3-16 stainless steel heat transfer plates, two outer covers, and four connections, with copper vacuum brazed together to form an integral unit.

Low-pressure plate-type heat exchangers are manufactured from aluminum and duralumin.

These types of heat exchangers are suitable for heating, cooling, evaporating, or condensing any fluids compatible with the materials of construction.

Fuel supply and control system

Fuel is delivered in engine testing cell from two locations according to fuel type used. Commercial fuel is delivered from one of the master tanks with a capacity of 5000 liters; special fuel is delivered form barrels with a capacity of 200 liters.

Fuel is delivered by pneumatic pumps (with membrane), those pumps have the advantage that doesn’t use electric power to operate eliminating fire risk and if the pump is switch off, pressure remains in the system.

Because the fuel flow required to operate the engine is smaller the fuel flow produce by the pump it is necessary to use by-bass valves along fuel lines.

All the lines that enter in the engine room must have an electronic cut-off valve connected to fire extinguish system.

Fig.3.5 Tanks used for commercial fuel Fig.3.6 Special fuel barrels

A pressure regulator is used to obtain different pressures according to the engine tested.

Fig.3.7 Pressure regulator

For measuring fuel consumption and regulate fuel temperature it is used AVL system.

The combination of AVL Fuel Balance and AVL Fuel Temperature Control is a high precise fuel consumption measurement and conditioning system, which is used worldwide at almost all engine test beds where engines of a maximum consumption of 150 kg/h are tested.

The AVL Fuel Balance is mainly used where high measuring accuracies and gravimetric measurements are required.

The built in calibration device enables calibration of the system under real test bed conditions. The AVL Fuel Temperature Control is used for fuel temperature conditioning on engine and chassis dyno test beds in research, development and production.

As a controlled cooling system it allows the user to set the fuel temperature anywhere within the range of 10 … 80 °C.

This system is capable to condition fuel consumptions up to 150 kg/h with a typical temperature stability of better than 0.02 °C and thus guarantees highest measurement accuracy when determining the fuel consumption on modern combustion engines.

Reducing the fuel consumption of engines requires the measurement of increasingly small differences in fuel flow.

The AVL Fuel Balance allows measuring these slight differences with maximum reliability.

The AVL Fuel Balance with AVL Fuel Temperature Control is based on the principle of gravimetric measurement. The amount of fuel consumption is determined directly by measuring the time related weight decrease of the measuring vessel by means of a capacitive sensor. Convenient calibration and easy maintenance provide optimum ease of operation. With the FlexFuel option, up to 100% alcohol and biodiesel can be measured.

Fig.3.8 AVL fuel balance

The following measuring data and functions are available:

indication of the fuel consumption values in kg/h and g;

measurement and indication of the actual fuel consumption at a measurement frequency of 10 Hz (measurement time 0.1s);

average consumption for pre-selected measuring time or pre-selected measuring weight;

total/interval consumption for determined measuring time;

running average calculation with additional indication of standard deviation and min. /max. values (with option remote control);

fully automatic built-in accuracy check and calibration;

fast and efficient fuel change;

indication of error and status report.

Application The AVL Fuel Balance is used to measure the fuel consumption of engine- and chassis dyno test beds for transient and steady-state measurement.

Benefits

the precision measurement accuracy of 0.12% can be verified according to ISO9001 within a few minutes on the engine test by the integrated calibration unit;

test bed times are minimized due to extreme reliability and long maintenance intervals;

dynamic fuel consumption measurement on engines with air bubbles in the engine return line;

easy to integrate in different automation systems thanks to the presence of compatible interfaces;

measurement results absolutely comparable to the AVL Fuel Mass Flow Meter;

direct mass determination of the fuel;

eminently suitable for state-of-the-art high-pressure injection systems;

not sensitive to pressure pulsations from the carburetor system.

The fuel consumption is determined using an appropriate weighing vessel linked by a bending beam to a capacitive displacement sensor. Due to the fact that the weighting vessel has to be refilled for each measurement this is a discontinuous measurement principle. The mass of fuel consumed is therefore determined gravimetrically, which means that the density does not have to be determined in addition. The fuel consumption can thus be determined to an accuracy of 0.12%.

The built-in calibration unit is standard scope of supply and allows calibration and accuracy check according to ISO 9001 which helps to reduce downtimes.

Fig.3.9 AVL fuel balance operating principle

Tabel.3.1Tehnical Data

Engine charging equipment

William Froude is regarded as the father of the modern dynamometer. His first project was to design a dynamometer for the steam engine in the HMS Conquest. The unit was fitted to the propeller shaft of the HMS Conquest, and the unit was submerged to provide cooling capacity for the absorbed power. Handles located on the stern of the ship operated a complex series of bevel gears that opened and closed sluice gates. An arrangement of levers read the torque on a spring balance located on the quay; a mechanical mechanism noted the engine revolutions .These were coupled to a rotating drum, and this produced a speed-versus-load chart, the area under the graph being the power.

The dynamometer is as fundamental to the in-cell testing of engines as is the engine. In establishing the engine characteristics and performance under different “road load” conditions, it is necessary to be able to safely and effectively replicate actual on-road conditions on a consistent and repeatable basis. This is in essence what the dynamometer enables one to do when running engines are tested.

The function of the dynamometer is to impose variable loading conditions on the engine under test, across the range of engine speeds and durations, thereby enabling the accurate measurement of the torque and power output of the engine.

Many types of dynamometers are available to the industry, with each having its own distinct advantages and disadvantages compared to those of its rivals.

The main types of dynamometers considered here are as follows:

"Hydraulic" (or water brake), of which there are two types:

Constant fill: This type uses thin sluice plates inserted between the rotor and the stator, across the mouth of the pockets, to interrupt and affect the development of the toroidal (or whirlpool effect) flow patterns within the pockets. These sluice plates can be inserted to infinitely varying degrees to provide variable control of the loading of the engine crankshaft.

Variable fill: As the name suggests, this method relies on controlling the amount of water available within the dynamometer casing, thus affecting the water supply available to the rotor stator assembly. This in turn will have an effect on the developed resistance force. The use of water outlet valves in varying the water flow through the dynamometer casing replaces the sluice plate control found in the constant fill machines.

Electrical, of which there are three main types:

DC current

AC current

Eddy current

With the eddy current dynamometer, the engine turns on the driveshaft, which is mounted on the rotor within the dynamometer casing. The outer edges of the rotor disc run between electromagnetic poles of the stator. Varying the excitations of these magnets, thereby altering their effect on the spinning rotor disc, will develop a resistant force or drag to counter the torque of the engine. Electromagnetic devices such as this are infinitely variable and have the added advantage of almost instantaneous implementation, thus giving greater control for the test-bed controller. Water cooling is achieved by passing raw water into the cavities in the stator near the point where the rotor and the stator are closest (eg. where the magnets act upon the rotor plate).

Eddy current dynamometers are the most popular type used within the test cell environment. These vary in size and application, depending on the power-torque output of the engine being run. Special care must be given to the elimination of vibration when using electric dynamometers because the sensitivity of the control will be affected.

The AC or DC transient dynamometer consists of a variable speed generator or alternator, the electrical output of which is delivered outside the test cell to a controllable load bank. In certain cases, particularly where large power output and continuous operation are concerned, the output may be delivered into the mains supply.

The generator cooling air is drawn from the test cell and returned to it, contributing to the total heat release in the cell. No water cooling is likely to be involved. This type also can be used for starting the engine and is used primarily for transient testing. Here, the aim is to evaluate fuel consumption variances (among other values), during split-second operations such as gear changing and overrun situations, which also affect he1 use and so forth, via the engine management systems.

Fig.3.10.AC dynamometer

For our tests we use a AC dynamometer (Fig. 3.10) with the following specifications:

– power 227kW;

– maximum speed nn/nmax4130/8540 rpm

– power supply 380V;

– frequency fn/fmax 68/141 Hz.

This type of machine is characterized by a small inertia and is successfully use to simulate the whole vehicle run.

Because the stator is fixed, the engine torque is measured by an electromagnetic system without contact (flange torque sensor).

Equipments and softs used for data acquisition

3.2.1 Connecting box

All the sensors that we use to measure temperature and pressure are located in the connection box that does the link between sensors and network module (E-GATE01).

Fig.3.11.Connection box

The box has two modules:

– One module with 16 analog multifunctional inputs, can be configured for multiple signals;

– One module with 16 analog inputs configured for type K thermocouples.

Analog signal from sensors are converted into digital signals and transferred to E-GATE by a PROFIBUS-DP link up to 12Mbit/s.

Pressure sensors: is used PR 824 sensor with a precision of 0.25 % with different measure domain according to the engine parameter that is measured. The sensor can operate from -20 ˚C up to +90°C.

Fig.3.12 Pressure sensor and type K thermocouple

For temperature acquisition are use type K thermocouples with Ni Cr-Ni, precision is 1% of measurement field from -20 ˚C up to +1300°C.

The knock sensor contains a piezoelectric crystal (3) and a seismic mass (1).

When the detonation occurs, strong vibrations in the cylinder are produced that are propagated through the block motor and captured by the sensor. Vibrations transmit the seismic measurement of the skin to the piezoelectric element and produce an electrical voltage.

Fig.3.13.Knock sensor

Components of a knock sensor:

seismic mass;

the housing;

piezoelectric element;

electrodes;

electrical contacts.

3.2.2 Instrumented spark plugs

The pressure sensing instruments (fig.3.14) have the ability to measure the pressure in the cylinder according to the piezoelectric principle.

Fig.3.14.Spark plugs with pressure sensor

Since the electric charge produced by the crystal is low, it is transmitted through a high impedance cable and converted to a voltage signal by an amplifier (fig.3.15).

Fig.3.15. Signal amplifier AVL MICRO IFEM

After conversion, the signal is sent to the data acquisition system (fig.3.16)

Fig.3.16. Data acquisition system AVL INDIMODE

3.2.3 Encoder

To determine the position of the crankshaft and its angular speed, a transducer for the AVL 365C encoder crankshaft position is used.

The operating principle is based on the reflection phenomenon of light, having a slotted disc, which gives a very high accuracy under extreme operating conditions.

Electronic components are mounted separately from the sensor that is mounted on the crankshaft flange to minimize the influence of electrical interference, temperature and vibration.

The information is transmitted by light pulses from the encoder through a 2 meter long optical cable to an AVL INDIMODUL data acquisition unit that synchronizes the information received from the crankshaft position transducer with the engine cylinder pressure measured by the sensors piezoelectric spark plugs.

The encoder (figure 4) has the following features:

– Maximum speed: 20000 rpm;

– temperature range: from -400 to 700 ° C for electronic components;

– temperature range: from -400 to 1200 ° C for mechanical parts.

Fig.3.17. AVL 365Ce Crankshaft Position Transducer

Software used

DDT 2000.

It is a software that can perform diagnostics (feather reading and deletion) as well as uploading new calibrations into the ECU.

Fig.3.18.DDT 2000

INCA

In today's automotive industry, the standard is the INCA (Integrated Calibration and Acquisition system). The system allows a real-time calibration of hundreds of variables and tables while performing simultaneous recordings.

INCA tools are used to develop and test the ECU as well as validate and calibrate electronically controlled systems in the vehicle, on the test bench, or in a virtual environment on the PC.

The tools offer a wide variety of functions, including pre-calibration of functional models on the PC, programming Flash ECU, measurement data analysis, calibration data management, and automatic optimization of ECU parameters.

Generated data can be continually processed and evaluated.

Fig.3.19.INCA

INDICOM

IndiCom is a combustion analysis software that combines data acquisition with professional data evaluation for a clear graphical presentation.

AVL provides a complete solution for measuring combustion from sensors for data processing to post-processing solutions.

Benefits:

– a powerful and flexible data acquisition system for a wide variety of applications;

– extensive capabilities of on-line computing;

– easy integration into test cell automation with personalized and generic interfaces;

– maximum measurement accuracy.

Flexible data acquisition system that studies the combustion process for a wide range of engines. IndiLod AVL is the high-performance measuring device.

Fig.3.20.INDICOM

MORPHEE

It is a program used to control and monitor engine test stands.

As a result of this, Morphee performs the following real-time operations:

acquisition and storage of gross (unworked) or point-to-point measurements (ready processed);

continuous data acquisition for post-mortem examination – in the event of a failure in one of the stand equipment, all parameters of the stand can be examined the length of time before the fault occurred; the length of time can be set according to the size of the allocated memory;

control of all external instruments of the stand (both analogue and digital);

surveillance of all parameters, with different stop levels in case of repeating alarms or warnings.

Fig.3.21 Morphee

STRATEGY ANTI – KNOCK

THEORY ABOUT STRATEGY ANTI – K

The ECU diagram which describe the process of anti-k

Fig.4.1. The ECU diagram which describe the process of anti-k

A piezoelectric sensor ( knock sensor ) mounted on the engine block produces an output voltage in proportion to the engine vibrations. Engine vibrations are caused by various sources (combustion noise, moving parts noise (valves, pulleys…)).

An input stage serves the following purposes:

a capacitor decouples the sensor from the following stages allowing only AC signal which contains information about engine vibrations.

an op-amp circuit provides gain to the sensor signal (K= R2/R1) and a DC shift of value ‘V mid’. This makes the signal uni-polar and hence adapted for the following stages. V mid is usually the middle voltage value between V cc and ground (typically V mid = 2.5V). V mid is used as a noise free DC reference for the circuit.

An anti-aliasing filter limits the input frequencies and adapts the signal to be sampled by the following analog to digital converter.

LDB – basic software

LDA – application software

PSL – Formatting the noise of the knock sensor

LDB – basic software

This basic software is processing the signal from the sensor and is used to filter the signal from the knock sensor and to calculate the energy of the signal.

The LDB have three components:

during an observation window, determined by software, the sensor signal is acquired by the analog to digital converter (ADC);

a band-pass filter (digital implementation) extracts frequency components specific to knocking combustion. This filtering improves knocking/non-knocking signal ratio and diminishes risk of false detection due to perturbations other than knocking combustion;

summation module whose calculation window (or detection window) must also be calibrated.

The measured signal energy within the observation window is then calculated.

From the measured energy and using look-up tables, the noise energy is calculated.

The noise energy together with the measured energy is delivered to the application layer ‘LDA’.

Fig.4.2. LDB

The test consists of acquiring on an Indicom PC the accelerometer signal (taken at the level of the anti-aliasing filter in the case of an instrumented computer (or anti-aliasing filter), see fig.4.1, or directly at the sensor for an un-instrumented computer) and the cylinder pressures for each operating point defined in the test program.

These operating points must be performed first with a high octane fuel and a second time with a low octane fuel. The purpose is to be able to distinguish the normal combustion noise from the noise caused by rattling. For this we use the signal-to-noise ratio (SNR), that is to say the ratio between the signal of the noise caused by the rattling and the average background noise without knock.

The LDB is a part of the Anti-K strategy that allows to determination the knock detection window and the frequency of knocks.

The purpose of this step is to identify the area representative of the knock of the engine (when the knock interving in the angular window in ° CRK of each cylinder after the PMS) and the knock noise frequency.

As can be seen from Figure 4.1, in the window calibration step and the knock frequency, the analogue signal from the accelerometer (piezoelectric) cap is amplified in the "Input Stage" section and sampled in the "Anti-Aliasing filter" side, then converted in a digital signal on the "ADC" side, it is further filtered from the frequency point of view (only those with knock frequency (2-20KHz) in the "Frequency Filter" part are preserved, and finally summed up by integration to quantify its energy.

In order to calibrate the window and the frequency of knock measurement we need:

– ECU (Computer Control Engine) instrumentation with knock output (knock capture signal);

– Knock sensor mounted on the motor and connected to the motor wiring harness;

– AVL INDIMODE equipment equipped with detonation function + 3 spark plug operated instruments with cylindrical pressure gauge;

– BNC cable for connection between the detonation of the ECU and the AVL INDIMOD;

– a PC where the INDICOM software is installed.

So the signal of the knock sensor, the three signal of pressure in the cylinder and the angular encoder signal will be transmitted to AVL INDIMOD and synchronized for processing on a PC on which the INDICOM software is installed.

In order to distinquish the signal of the knock sensor when is knock and when is not knock it is require to do two recordings with two type of fuels RON 87 and RON 105.

It is intended that the knock signal / non-knock signal (red zone / blue zone – figure 4.3) should have the highest value for each of the three engine cylinders.

The window depicting the engine's detonation is located between the angles for which the ratio of the detonating signal to the engine's mechanical noise is the highest.

Fig.4.3.Example of a the detection window

Criteria followed in window selection:

the maximum window length is 20 °CRK;

best RSB;

number of detonation cycles 10%.

Note: To select the detection window, select an area where the red surface is above the blue. The goal is to find the best compromise for all cylinders.

As for the frequency selection, proceed likewise: selecting a frequency band (central frequency +/- 1.4kHz) for which the ratio of the detonating signal to the non-detonating signal (red zone / blue zone – figure 4.4) is the highest for each of the 3 cylinders and where the background noise is as stable as possible (eg not being influenced by closing the exhaust valve from the neighboring cylinder). It will avoid the frequencies that background noise has a "bump", even if the ratio of the detonating signal to the undefined signal is good.

Fig.4.4 Example of a frequency

LDA – application software

part of the Anti-K strategy that collectively includes calibration of formatting tables and calibration of detection lines ( gain + offset )

The LDA determines whether there is knocking or not.

At the end of the window calibration step and the frequency of the detonation measure, the gross engine noise remains unchanged because it did not intervene on it.

Calibrating the formatting tables is intended to modify the gross engine noise to produce for each cylinder a formatted and centered noise signal at a specific value (generally 50 when there is no detonation).

This is due to the fact that the noise of the motor varies greatly from one operating point to another, especially depending on the speed. The objective is to return the noise at the same operating speed (speed-load) of the engine, when we have normal combustion, so calibration of the BMS format tables is made with high octane fuel – RON105.

The formatting table calibration principle applies to each individual cylinder and is the following:

Formated noise =

On the test bench, calibration of the format tables requires manual editing for each point of operation of the 3 format tables corresponding to the four cylinders to retrieve the noise at a specific value (50 for example for the Nominal Captor).

Samples are performed with a high octane fuel of RON105 and a nominal detonation cap. After the calibration has been completed, formatted noise measurements will be performed to validate the calibration table calibration on the entire engine operating range.

But from a constructive point of view, the parameters and characteristics of the detonation caps are not identical, there is practically a dispersion, which results in the instantaneous raw noise values acquired from the detonators being dispersed.

Therefore, it is necessary to define a new parameter, namely filtered noise.

Filtered noise is the "average" of the instantly formatted noise.

In a graphic, where the filtered noise is on the abscissa and the formated noise is one the ordinat is established a threshold from which the noise is considered a detonation.

The threshold is in the form of a straight line that is called the right of detection.

The difficulty in determining the position of this line is the compromise between detection / false detection. The Anti-Detonation strategy must protect the engine without degrading its undue performance.

The detection right is defined by a GAIN and an OFFSET. The basic principle consists in delimiting the noise considered to be detonations from those resulting from normal operation of the engine (including mechanical noises).

Fig.4.5 Right of a detection

Noise energy calculation for knock detection

The main contains 3 blocks

Read basic layer: reading of the knock energy.

Reference noise energy: based on Look-up tables.

Noise energy calculation: using an adaptive strategy.

A piezoelectric sensor mounted on the engine block produces an output voltage signal in proportion to the engine vibrations. The energy of this signal measured for each cylinder is read in the block read basic layer.

The measured signal energy information about engine vibrations caused by different sources: normal combustion, moving parts (valves, pulleys…), knocking and other sources.

The aim is to detect the knock. Hence the energy due to all vibration sources other than knock is considered as ”noise”, while the energy corresponding to knock is considered as the ”desired signal”.

In order to obtain the desired signal, the noise is subtracted from the measured signal. In the figure below, the continuous waveform represents the measured energy, while the discontinuous waveform represents the noise energy. By comparing the two waveforms the knocking can be detected.

Fig.4.6.Measured signal with knock versus noise energy without knock

If the sensor malfunctions or is disconnected, the obtained noise energy will be too high exceeding a threshold.

This threshold is a tuning parameter used by knock sensor diagnosis to check whether the sensor is working properly or not. The following figure depicts the noise threshold used for knock sensor diagnosis.

Fig.4.7.Noise threshold used for diagnosis

The noise depends on many factors among which are the engine speed and the manifold pressure. By measuring the instantaneous noise of several combustions, when knock is not present, for the same cylinder, a look-up table can be built. This look-up table is used to construct the reference noise energy. The reference noise energy is constructed in block reference noise energy.

This analysis may not be so accurate due to the fact that the noise is affected by other several parameters, like engine ageing, engine dispersion, fuel dispersion, variable climatic conditions and road irregularities. Hence more sophisticated methods are developed to obtain more accurate results. The figure below illustrates the reference noise energy obtained using look-up tables versus the measured energy with knock and the actual noise energy for vibration sources other than knock.

Fig.4.8.Reference noise versus measured signal and actual noise signal

Knock sensor diagnosis

Knock sensor diagnosis is activated under certain engine operating conditions (coolant temperature, engine speed). In this mode, only the two inner cylinders are investigated. The noise levels are compared with a threshold.

It is require to diagnosis the knock sensor because we may have failures (open circuit, short circuit to ground).

Knock sensor is checked regularly, the knock sensor diagnosis is scheduled based on a number of TDC counts.

Validation Anti-K

Validation of anti-knock calibration takes place at the end of the development and certification loop in accordance with standard calibration process of the activities of the "Basic Tuning & Torque Management" team.

Anti-knock validation can also become necessary in other cases: transparency of anti-knock setting during a technical definition upgrade of the engine, a change in the Reliability Specifications, etc.

Anti-knock calibration should be entirely calibrated in accordance with the standard process as well as all of the complementary strategies:

Preventive anti-knock strategy as a function of the manifold air temperature,

Correction of enriching threshold and richness corrections as a function of ignition timing correction,

Calibration of engine protection against knocking for turbocharged engines.

The anti-knock calibration validation criteria are linked to criteria relative to:

Reliability / Durability: compliance of Reliability Specifications note on basic tuning and detection quality

Performance / Consumption: compliance with nominal performances note on false detection.

Fig.4.6. General operation of anti-knock strategy – stabilized point

Along the abscissa: actual knock amplitude measured by cylinder pressure

Along the ordinate: formatted engine noise determined by Engine Control Unit.

The actual knock limit is materialized on this charge by a vertical straight line. This limit varies according to the speed and/or the cylinder pressure measurement system (signal processing).

The Engine Control knock limit is materialized on this charge by a horizontal straight line. This limit is relative and varies according to the average noise from the operating point.

These two limits are used to divide the charge into four zones:

Nominal operation: no knocking is present and the strategy does not detect any OK

Operation with knocking detected: knocking is present and the strategy detects it OK

Operation with false detection: no knocking is present but the strategy detects it risk for performance and consumption.

Operation with no detection: knocking is present but the strategy does not detect it risk for reliability and durability.

During validation of the anti-knock strategy, the detection Booleans per cylinder should be measured and returned into the cylinder pressure measurement system (INDIMODUL) allowing to sort the cycles that were detected and those that were not detected by the Engine Control.

To validate the strategy Anti-K we need to do two tests:

In the first part, a fuel with a lower octane rating than the one specified in the DT is used to verify that the computer injection is facing correct backflows when the detonation phenomenon occurs (quality detection);

In the second part, an engine with an octane number above DT is used to check if the injection computer is experiencing inappropriate feed backs (false detection).

The samples are run with the temperature of the coolant (T_EAU) at 90 ° C and measurements are made in Morphee, INCA for 3 minutes and INDICOM for 5000 cycles up to 3000 rpm and 10000 cycles over 3000 rpm.

False detection

The analysis is conducted in the form of statistics: tool in EDES PCMAP PV Cliquetis.

Table 1: PCYL statistics on test with high octane number gasoline (RON105)

Explanations:

Point 1: Actual number of knock beats: the number of knock beats whose amplitude is above a threshold defined in column 2 is counted, threshold corresponding to the 3% knock criteria defined by a specific test. This threshold is defined on the test conducted with RON105 (maximum measured on test).

Point 2: Actual number of knock beats that were detected by Engine Control. If the number of events is quite important, the detection rate is calculated. In the same way, the actual number of knock beats not detected by Engine Control can be calculated.

Point 3: Number of False Detections. A false detection is a knock detection by Engine Control when the actual knock beat amplitude is below the threshold defined in column 2.

DEV analysis:

Point 1: The number of knock beats should be null. If this is not the case, we need to verify the connections and to modify the thresholds of knock

Point 2: The number of detections and non-detections should be null. If this is not the case, we need to modify the threshold.

Point 3: The number of False Detections should be null. If this is not the case, calibration creates False Detections. A small number of False Detections is acceptable. To know if the false detection rate is acceptable or not, the INCA file should be analyzed using the ANALYFICH tool.

Note:

This analysis allows measuring the KP_PK corresponding to the combustion noise. This value depends on the measurement system. Note that amplitudes of 1.5 to 2 bars are measured at high engine speeds without knocking.

From measurements under INCA (ANALYFICH tool):

In case of False Detection, this FD rate should be checked to be acceptable:

On each cylinder slow corrections should be less than 1 °V on average.

One (false) detection is authorized for 30s with transverse calibrations.

If this is not the case, the anti-knock calibration process should be redone. Generally, formatting is validated and only the detection lines should need to be adapted.

Quality detection

Table : PCYL statistics on test with low octane number gasoline (RON87)

DEV analysis:

Point 1: The knock beat rate is used to validate performance of the strategy:

Rate > 5% the knock level is still high, the amplitude of the noise signal should be checked

Rate < 5%  the strategy is effective and globally corrects the knock beats

Point 2: The detection rate of this test gives an indication on the performance of the anti-knock strategy but does not fully reflect the performance of the strategy: in fact, in case of detection, the setting is modified and decreases the number of knock beats – cf. point 1. The detection target is as follows:

Rate < 20%  : low detection check the False Detection margin.

20% < Rate < 40% : average detection enabling proper correction of the knock beats

Rate > 40%  : good detection enabling effective correction of the knock beats

Note:

If it is desired to evaluate the performance of the strategy, a validation test should be planned using limit fuel with the anti-knock strategy active but by inhibiting the spark advance corrections proposed in the presence of knocking. This test is not recommended because it creates a risk in engine reliability. It will only be conducted on the request of the Reliability sector if the performance of the anti-knock strategy is not considered acceptable.

EXPERIMENTAL RESEARCH ON ENERGY OPTIMIZATION OF SPARK IGNITION ENGINE BY ANTI-K REGULATION AT STATIONARY BENCH

Calibration LDB

Calibration window detection and frequency

Figure.4.7.Window detection

Figure.4.8.Frequency

Colour blue represent the average noise ( RON 105 );

Colour blue represents the gross noise ( RON 105 );

Colour red represents the average noise ( RON 87 );

Colour orange represents the gross noise ( RON 87).

In the AntiK adjustment samples, calibration of the window and detection frequency are the first samples to be performed.

In the first part of the test, a high octane fuel (RON ≥ 105) is used, and in the second part a fuel with a lower octane rating than the reference fuel used. The aim is to be able to distinguish the normal burst noise (without detonation) as compared to the noise caused by the detonation firing.

To calibrate a window and detection frequency we need to have some criteria:

the lenght of the window must have 20°CRK;

the best ratio signal to noise;

the number of cycles with knock should be maximum 10%;

the frequency must be between 2 and 20 kHz.

It is preferable to choose a frequency to allow for a more stable engine background noise. ( fig.4.8 ).

With EDES program´s help we can observ that the highest ratio for cylinder 1 is 9,78, for cylinder 2 is 14,7 and for cylinder 3 is 5.

Calibration LDA

Line of detection

In this part of the experiment, to calibrate the line of detection, it is necesary to use a low octane fuel RON87, to check if the injection computer withdrawing the advance corectly (quality detection), in the second part it is use a high octane fuel to check if the injection computer experiencing inappropriate feed backs (false detection).

The samples are run with the temperature of the coolant at 90 °C and measurements are made in Morphee, INCA for 3 minutes and INDICOM for 5000 cycles up to 3000 rpm and 10000 cycles over 3000 rpm.

Quality detection

Fig.4.7 Values of KP_PK and DET_m of cylinder 1 at 6000 rpm with full load

In the figure 4.7 we have two signals:

A signal from Morphee which represent the value of knock

A signal from INCA which represent the moment when the ECU withdraw the ignition advance.

Detection – when we have a amplitude of knock above the threshold and the withdraw of ignition advance we have a detection.

False detection – corresponding of a detection of knock by ECU when the amplitude of knock is inferior to threshold.

Non detection – taking place when the ECU doesn’t withdraw the ignition advance when the knock occurs.

Fig.4.8 Values of KP_PK and DET_m of cylinder 2 at 6000 rpm with full load

Fig.4.9 Values of KP_PK and DET_m of cylinder 3 at 6000 rpm with full load

Table.3. PCYL statistics on test with low octane number gasoline

Table.4. Reliability formatting

Analysis:

In table 3, column KNOCK, we can observe that the strategy is effective and globally corrects the knock beats (knock beat rate < 5%);

In table 3, column DETECTION, on each cylinder is confirming good detection enabling effective correction of the knock beats but at 4000 rpm on cylinder 1 and 5500 rpm on cylinder 2 we have average detection enabling proper correction of the knock beats and at 6000 rpm cylinder 3 we have a low detection.

In table 4, the first column represents the knock beat category and the next columns represents the detection and non-detection of every cylinder.

As an example in table 4, at 6000 rpm with full load in interval 3 – 4 bar we have 1 detection and 3 non-detection. At this speed we have a 2.6 bar threshold and the ECU had to withdraw the ignition advance but he hasn’t.

False detection

Fig.4.10 Values of KP_PK and DET_m of cylinder 1 at 6000 rpm with full load

Fig.4.11 Values of KP_PK and DET_m of cylinder 2 at 6000 rpm with full load

Fig.4.12 Values of KP_PK and DET_m of cylinder 3 at 6000 rpm with full load

Table.5. PCYL statistics on test with high octane number gasoline

Table.6. Reliability formatting

Table.7.INCA measurements

Analysis:

In this test the number of knocks should be null. But at 6000 rpm with full load on cylinder 1 and 2 we have 2 knocks (figure .5.). To be sure if we have knock in cylinder 1 and 2 we analyze the measurements from INCA. In INCA measurements at 6000 rpm with full load there isn’t any record of knock and a withdraw of ignition advance (table .7.).

At 4500 rpm with full load, on cylinder 3 (table .5.) we have 14 false detection and it is required that we may have maximum 6 false detections. In table 7 we have 13 false detections. At this speed it is required to be redone the calibration process. Generally, formatting is validated and only the detection lines should need to be adapted.