DETERMINATION OF TDC IN INTERNAL COMBUSTION ENGINES BY A NEWLY DEVELOPED THERMODYNAMIC APPROACH PART 1: BASE THEORY

Emiliano Pipitone, Alberto Beccari Dipartimento di Meccanica University of Palermo

ABSTRACT It is known to internal combustion engine researcher that in-cylinder pressure analysis represents an indispensable tool in the research and development process, and must be carried out with maximum accuracy. A key factor is the precise knowledge of the Top Dead Centre (TDC) position, since a phase error of just 1 degree between in-cylinder pressure and volume can cause large errors in the estimation of the most important performance related parameters. It is also known that the TDC position does not coincide with the location of the peak pressure (LPP) and their angular lag is known as the loss angle, being related to the two losses: heat transfer with wall and mass leakage. To gain an acceptable accuracy, the TDC position should be known within the precision of 0.1 degrees: this can be accomplished by means of an expensive dedicated sensor or employing a thermodynamic method, i.e. an algorithm whose input is the in-cylinder pressure measured under motored condition. In this paper the authors, starting from the study of the compression-expansion process in a motored cylinder, developed an original thermodynamic method, based on the evaluation of an expressly defined loss function, i.e. a function whose derivative although combines both losses, can be simply evaluated once the engine geometry and the motored incylinder pressure are known. The determination of the loss angle, and hence of the TDC position, can be performed through the evaluation of the loss function variation at the peak pressure point. The key point of the new method relies on its self-referencing procedure, which allows an adequate evaluation of the TDC position without a direct estimation of both mass leakage and heat transfer: this has been demonstrated, in the appendix, by means of models commonly used to describe the two losses. Keywords: Top dead centre determination, Spark ignition engine, Compression ignition engine

INTRODUCTION In-cylinder pressure analysis is nowadays an indispensable tool in internal combustion engine research and development. It allows the measure of some important performance related parameters, such as indicated mean effective pressure (IMEP), mean friction pressure, indicated fuel consumption, heat release rate, mass fraction burned, etc.. Moreover, future automotive engine will probably be equipped with in-cylinder pressure sensors for continuous combustion monitoring and control, in order to fulfil the more and more strict emission limits. For these reasons, in-cylinder pressure analysis must be carried out with maximum accuracy, in order to minimize the effects of its characteristic measurement errors. The exact determination of crank position when the piston is at top dead centre (TDC) is of vital importance, since a 1 degree error can cause up to a 10% evaluation error on IMEP and 25% error on the heat released by the combustion: the position of the crankshaft (and hence the volume inside the cylinder) should be known with the precision of at least 0.1 crank angle degrees, which is not an easy task, even if the engine dimensions are well known: it corresponds to a piston movement in the order of one tenth of micron, which is very difficult to estimate. A good determination of the TDC position can be pursued by means of a dedicated capacitive TDC sensor, which allows a dynamic measurement (i.e. while engine is running) within the required 0.1 degrees precision. Such a sensor has a substantial cost and its use is not really fast,

since it must be fitted in the spark plug or injector hole of the cylinder. A different approach can be followed using a thermodynamic method, whose input is in-cylinder pressure sampled during the compression and expansion strokes: some of these methods, more or less valid, can be found in literature [1,2,3,4,5,6,7]. This paper will discuss a new thermodynamic approach to the problem of the right determination of the TDC position.

BASE THEORY The compression and expansion processes in a motored (i.e. without combustion) engine can be described by observing the energy transformations regarding the unity mass which remains in the cylinder. The first law of thermodynamics states that:

q p v = u

(1)

where q represents the specific heat received by the gas from the cylinder walls during the crank rotation , p and v represent the gas pressure and specific volume, and u the specific internal energy variation. The gas involved in the process can be assumed to be a perfect gas, thus the following equations are also valid:

p v = R' T

dp dv dT + = p v T u = cV T

(2)

being R and T the gas constant and temperature respectively, while cV and cP represent the constant volume and constant pressure specific heat, which are functions of the gas temperature. The influence of flow-friction during the compression-expansion process in a reciprocating machine is not significant, since the fluid is macroscopically motionless. Then heat transfers are largely more important than flow-friction losses, hence, according to the second law of thermodynamics, the specific entropy variation S of the in-cylinder gas during the crank rotation is: q S = (3) T from equation (1) and (2) the specific entropy variation thus results:

S =

q
T

= cP

v
v

+ cV

p
p

(4)

Due to mass leakage through valve seats and piston rings, the available volume V for the in-cylinder gas increases, hence its specific volume changes:

V = vm

dv dV dm = v V m

(5)

where m represents the in-cylinder mass. Hence, considering the finite increment due to a crank rotation , the specific entropy variation in equation (4) will now result:

S = cP

V
V

+ cV

p
p

cP

m
m

(6)

being m the mass entering the cylinder (hence m0) during the crank rotation ; hence the incylinder pressure changes is:

p =

1 [Q(k 1 ) kpV ] + k p m V m

(7)

where Q=m q represents the heat received by the gas and k the isentropic coefficient =cP/cV. In an ideal adiabatic motored engine both Q and m would be zero, and pressure would reach its maximum (p=0) when the volume is minimum (V=0): the compression and the expansion strokes would cause in-cylinder pressure variations symmetric with respect to TDC and the position of the peak pressure (which can be easily determined with 0.1 precision by means of polynomial interpolation of the pressure curve sampled with 1 crank angle degree resolution) would coincide with the position of the TDC.
Vmin pmax Tmax

Twall

Vmax

Fig. 1 Temperature-Entropy diagram of the compression-expansion process in a motored cylinder: ideal engine (segments AB and BA) and non-adiabatic engine (dashed curve)

Fig. 2 Temperature-Entropy diagram of the non-adiabatic compression-expansion process: peak pressure (point D) occurs before the TDC (point E)

In a Temperature-Entropy (T, S) diagram these two evolutions would be represented by two coincident segments (AB and BA in Fig. 1). In a real motored engine the pressure variation is caused both by volume changes and by two phenomena related to the real machine, i.e. the heat received by the gas from the cylinder walls Q (which is negative when the gas temperature is higher than wall temperature, i.e. Q(TwallT)) and the mass leakage m (which is negative whenever in-cylinder pressure is higher than outer pressure): equation (7) clearly shows that both of them cause pressure rise to be zero when the volume changes are still negative (i.e. during compression). Both heat transfer and gas leakage cause the pressure curve to be asymmetric with respect to the TDC, shifting the Location of the Peak Pressure (LPP) in advance with respect to the TDC position (see the non-adiabatic evolution in Fig. 2 or a real pressure curve in Fig. 3): the angular distance between LPP and the TDC position is called loss angle, being related to the energy and mass losses, and usually assumes values between 0.4 and 1 CA degree, depending on the entity of the heat transferred and the escaped mass:

loss= LPPLTDC

(8)

Equation (6) shows that two easily measurable quantities, the in-cylinder pressure and volume, allow the evaluation of the entropy variation (i.e. heat transfer) together with the mass leakage by means of the functions V/V and p/p, which are plotted as example in Fig. 5.; defining the loss function F so that:
9.5 8.5 Pressure Volume 70 65

F = cP
Volume [cm 3]

V
V

+ cV

p
p

(9)

60

Pressure [bar]

7.5 6.5 5.5 4.5 [CAD ATDC] 3.5

55

LOSS

it will result:

50 45 40 35 30

F = S + c P

m
m

(10)

The entity of the variation of the loss function F, which gathers the -30 -20 -10 0 10 20 30 sum of the two losses, is then Fig. 3 In-cylinder pressure and volume near TDC determined by the capability of the (MAP = 0.4 bar, n=1000 rpm, =10) cylinder walls to exchange heat with the gas and by the amount of gas 0.2 escaping from the cylinder. The 0.0 qualitative progress of the loss -0.2 function variation in a real cylinder during a compression-expansion -0.4 process, together with its two -0.6 dF F constitutive terms S and cP m/m, is dS = dQ / mT q/T -0.8 S= shown for example in Fig. 4 for a cp cpdm/m m/m -1.0 compression-expansion process in a ideal adiabatic real cylinder: the entropy variation -1.2 -180 -120 -60 0 60 120 180 starts with a positive value (being CAD ATDC T<Twall) and decreases, crossing the Fig. 4 Loss function variation F and its constitutive terms zero line when T=Twall, and reaching ( =1 CAD) a minimum near the TDC position (here the heat flux from the gas to the wall is maximum), then starts to 0.05 0.04 dp/p increase becoming positive before p/p 0.03 v/v the Bottom Dead Centre (BDC); the dv/v 0.02 relative mass leakage m/m, being 0.01 related to the difference between in0.00 cylinder pressure and outer pressure, -0.01 follows a similar trend, reaching a -0.02 minimum near the TDC: it follows -0.03 -0.04 that, in this position, the loss function [CAD ATDC] -0.05 value is the sum of the two loss angle -180 -120 -60 0 60 120 180 causes. Following this concept the authors tried to draw information on Fig. 5 Progress of V/V and p/p as function of crank rotation the loss angle entity directly from the ( =1 CAD) loss function. When the gas pressure reaches the peak value (i.e. at LPP), the ratio p/p is zero, and equation (9) becomes:
[J/kg K]

 V FLPP = c p V LPP

(11)

The latter equation shows that at the peak pressure position the knowledge of the loss function increment F allows to determine the value of V/V which, depending only on engine geometry (see Fig. 5 and equation (12)), is a known function of the crank shaft position, and hence of the loss angle. For centred crank mechanism, the function V/V can be expressed as:

V
V

cos ( ) sin ( ) 1 + 2 sin 2 ( )

2 + + 1 cos ( ) 2 sin 2 ( ) 1

(12)

where is the volumetric compression ratio and expresses the rod to crank ratio (i.e. the ratio between connecting rod length and crank radius). Since the loss angle is normally around 1 degree (= 0.017 radians), further approximations can be made:

sin(loss)loss

cos(loss)1

It follows that at the peak pressure position equation (12) becomes:

1 V = V LPP

loss 1 +

2 loss 2

2 2 + 2 loss 1

(13)

Hence, being loss2<<2, equation (11) and (13) yield:

loss =

2 1 F 1 +1 c p LPP

(14)

0.5 0.0

-0.5
loss angle (LPP=0) 1/2 loss angle 0 -1/2 loss angle -loss angle

-1.0 -1.5 -2.0 -180

[CAD ATDC] 0 90 180

-90

Fig. 6 F as function of crank position for different phase errors ( =1 CAD)

This demonstrates that, at the peak pressure position, the loss angle can be easily correlated to the loss function increment F. Unfortunately this function undergoes great distortions even with small phase errors between p/p and V/V, as shown in Fig. 6: the curves here represented have been calculated assuming different phase errors (expressed as fraction of the loss angle). As can be seen, phasing the pressure diagrams with the loss angle error (i.e. setting

F [J/kg K]

LPP=0) a considerable error on the evaluation of the function F would be introduced. This fact, without a reliable way to evaluate the F at peak pressure position, would make equation (14) useless. The same Fig. 6 however shows the existence of two zones common to each of the curves: in these two crank positions the two fundamental functions for the calculus of the loss function variation, p/p and V/V, reach their extreme values (at about 30 CAD ATDC in Fig. 5), and hence are poorly influenced by a phase error, such as the loss angle; for this reason, according to equation (9), in these two crank positions the loss function increment remains almost unchanged. This fact implies that assuming a TDC position error equal to the loss angle (easily achievable setting LPP=0), the values of the loss function variation F1 and F2 in the two points relative to the minimum and maximum of the function V/V will be nearly correct. Hence, in order to determine the loss angle from equation (14), a correlation between F1 and FLPP has been searched, and, as shown in Appendix A, it has been found that, for a given engine, the ratio between FLPP and F1 is almost constant, i.e.:

FLPP EP F1= EP Fmin dV/V

(15)

1.0 0.5 comp. ratio error -4% -2%

F [J/kg K]

0.0 0% -0.5 -1.0 CAD ATDC -1.5 -60 -40 -20 0 20 40 60 2% 4% dF1 F1 F2 dF2

where EP is a proportionality constant. As shown in Appendix A this constant mainly depends on the engine compression ratio and on the heat transfer law, and its mean value has been estimated to be 1.95. Thus equation (15) becomes:

FLPP 1.95 dFmin dV/V

(16)

As a result, the top dead centre can be determined phasing the pressure 1.0 cycle with the loss angle error (i.e. 0.5 setting LPP=0) and calculating the -0.05 0.0 loss function increment F1 at the 0.00 -0.5 minimum V/V position 1, which 0.05 -1.0 requires, according to equation (9), 0.10 -1.5 the estimation of the functions V/V dF1 -2.0 and p/p. Unfortunately both of F1 CAD ATDC F2 dF2 these functions can be affected by -2.5 -60 -40 -20 0 20 40 60 measurement errors: the in-cylinder pressure in fact is subjected to bias Fig. 7 Loss function variation F in presence of compression ratio error (above all if an un-cooled error (top) and pressure bias error (bottom): F1 () and F2() piezoelectric transducer is used) and are shown (pressure cycle phased with LPP=0, =1 CAD) to electric and mechanical noise, while the in-cylinder volume estimation may present inaccuracy related to the compression ratio, which is normally known with some approximation. Moreover, as shown in equation (9), the specific heat at constant pressure and volume are required, which are functions of the gas temperature; this in turn can be deduced applying the perfect gas law by means of the gas temperature at inlet valve closure, which is normally known with an approximation as high as 30C. All these uncertainties may strongly affect
1.5 pressure biar error [bar] -0.10

F [J/kg K]

the F1 evaluation, as shown for example in Fig. 7: here the loss function variation is calculated supposing both different compression ratio errors (top figure) and pressure bias errors (bottom figure). As can be seen, in presence of these measurement errors, the evaluated F1 may considerably differs from the real one thus preventing a reliable evaluation of the loss angle. However the same Fig. 7 also shows that the evaluated F1 and F2 move in different directions in consequence of the measurement errors: this effect implies that the mean entropy variation Fm, defined as

Fm =

F1 + F2 Fmin dv/v + Fmax dv/v = 2 2

(17)

remains almost constant, as shown in Table 1 and Table 2.

pressure bias error [bar] -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 Err% F 1 -332% -202% -93% 0% 82% 155% 221% err% F m 18% 11% 5% 0% -4% -8% -11% Compression err% F1 ratio error -6% -4% -2% 0% 2% 4% 6% -219% -144% -71% 0% 70% 138% 205% err% F m -3% -2% -1% 0% 1% 2% 3%

Table 1 Percentage error on both F1 and Fm for different pressure bias errors

Table 2 Percentage error on both F1 and Fm for different compression ratio errors

It follows that, in order to correctly evaluate the loss angle, the loss function increment at the peak pressure position FLPP should be correlated with the mean value Fm rather than with the F1. Thus relations (15) becomes:

FLPP = E P Fm

(18)

Therefore the method proposed by the authors reposes on the evaluation, at the minimum and maximum V/V positions (1 and 2), of loss function increment F1 and F2 which, according to equation (18), allows to evaluate the loss function variation at the peak pressure position FLPP; this, in turn, is linked to the loss angle loss, which can be calculated using equation (14) and furnishes the top dead centre position (see 32 equation (8)). The determination of the angular positions 1 and 2 at 30 2.8 3.4 4.0 which the function V/V is minimum 28 and maximum requires the derivation of equation (12), whose result is a 26 function not solvable in the variable 24 . Hence these angular positions must be evaluated using numeric 22 methods; considering compression 20 ratios ranging from 10 to 20 and 8 10 12 14 16 18 20 22 rod to crank ratios ranging from Compression ratio 2.8 to 4.0, the authors determined the Fig. 8 Position of maximum V/V with varying compression ratio angular positions 1 (=2) using a and for different rod to crank ratio
Position of maximum

V/V

[CAD ATDC]

2nd order polynomial interpolation on the V/V curve extended to a range of 0.4 degrees around the position of the extreme values. The results, as pointed out in Fig. 8, showed that the angular positions 1 and 2 depend both on the compression ratio and on the rod to crank ratio. The data obtained allowed to trace a formula for the calculation of the minimum and maximum V/V angular positions with a precision of 0.1 degrees:

1,2 = m 76.307 0.123 0.466

(19)

Summarizing, once the motored pressure cycle has been sampled, the procedure for the TDC estimation consist of 5 steps, here resumed: 1) 2) 3) 4) pressure curve phasing setting LPP=0 (in this way the position error is exactly equal to the unknown loss angle ) numerical evaluation of the angular position 1 and 2 of the minimum and maximum V/V (for example using equation(19)) calculus of the loss function increments F1 and F2 at the angular position 1 and 2 by means of equation (9), and hence calculus of their mean value Fm (equation (17)) evaluation of the loss function increment FLPP at the peak pressure position from equation (18), where the constant EP can be estimated by means of equation (30) in Appendix A or set to the mean value 1.95 the loss angle loss, and hence the TDC location, can be then evaluated by means of
equation (14)

5)

It is worthwhile to mention that the first step is not necessary if the pressure cycle has already been phased with a small error, since the method proposed is unaffected by small phasing error (i.e. in the range of the loss angle entity). Moreover the specific heat cP and cV in equation (6) (9) (11) and (14) should be temperature dependent and evaluated according to the classical known functions valid for air:
1000 1000 1000 cP = 1403.06 360.72 [J/kg K] 10.79 + 108.24 T T T
2 3

cV = cP -R

and

R=287.1 [J/kg K]

However a satisfactory approximation is equally reached if the cP and cV are supposed to be constant. In this case the evaluation of the gas temperature is completely avoidable for the TDC determination.

Appendix A In this section a relation between the loss function variation at the peak pressure position FLPP and at the minimum V/V position F1 is derived. As first step, the in-cylinder evolution will be considered without mass leakage; hence the ratio between the two loss function increments can be expressed in terms of entropy variations:

[Q/T ] LPP FLPP S LPP = [Q/T ] F S

1 1 1

(20)

where the amount of heat received by the gas from the walls during the time interval t is:

Q = h A (Twall T) t

(21)

being h the heat transfer coefficient, A the area of the heat exchange surface, T and Twall the gas and wall temperatures. Hence the entropy variations ratio becomes:

[hA(Twall - T )/T ] LPP S LPP = [hA (Twall - T )/T ] S

1 1

(22)

The total in-cylinder wall surface area A is:

d d2 x A = d x + = + 1 2 2 d/2
where x represents the piston distance from the cylinder top (function of the crank angle ):

(23)

x=

d 2

2 sin( ) 2 + + 1 cos ( ) 2 -1

(24)

Here is the volumetric compression ratio, while is the rod to crank ratio (i.e. the ratio between the connecting rod length and the crank radius). Introducing the dimensionless variable = 2x/d, the ratio between the heat transfer surfaces become:

ALPP [ + 1] LPP = [ + 1]1 A1

(25)

According to the most used model for heat transfer between gas and internal combustion engine cylinder, the heat transfer h coefficient is related to gas pressure p, temperature T and volume V by means of three power with exponents a, b and c respectively:

h pa Tb Vc
Hence the ratio of the heat transfer coefficient becomes:

paT bV c LPP hLPP = h1 paT bV c 1

Both gas pressure and temperature are linked to in-cylinder volume by the polytropic law:

[ [

] ]

(26)

p V =constant T V -1 = constant
where is the mean polytropic index. It follows that the ratio between the heat transfer coefficient is:

hLPP V1 = h1 V LPP

(a+b) -bc

1 = LPP

(a+b) -bc

(27)

The last fundamental ratio in equation (22) regards the temperature difference between gas and wall:

[(Twall - T )/T ] LPP = [Twall T ] LPP LPP [ (Twall - T )/T ] [Twall T ]

1 1 1

(28)

If TIVC represents the gas temperature at inlet valve closure, then the ratio between the temperature differences becomes:

[Twall - T ] LPP [Twall - T ]

1

IVC Twall TIVC LPP = 1 IVC Twall TIVC 1

(29)

Hence, from equations (22)(25)(27)(28) and (29), the entropy variations ratio can be evaluated by means of:

S LPP = S LPP
1 1

[ + 1] LPP [ + 1]
1

1 IVC Twall TIVC LPP 1 IVC Twall TIVC 1

(30)

being the exponent = (a+b) (b+c) ( 1). As can be noted, this ratio mainly depends on the engine geometry and on the heat transfer law, then for a given engine, it can be considered a constant:

S LPP = EP S
1

(31)

Assuming the values in Table 3 and taking into consideration three different heat transfer models (Woschni [8, 9], Eichelberg [8, 10] and 10 to 20 Hohenberg [8]) it has been found that compression ratios the values assumed by the ratio of 2.8 to 4.0 connecting rod to crank ratios equation (30) ranges from 1.81 to 2.05 Twall 70 C according to the compression ratio and TIVC 40 C the engine heat transfer law, as shown 1.32 in Fig. 9. A negligible dependence has -1 LPP been found with respect to the rod to Table 3 crank ratio . The mean results obtained by each heat transfer model 2.10 Woschni Eichelberg are resumed in Table 4, and, as can be Hoehnberg 2.05 noted, for the constant EP a mean value equal to 1.92 could be adopted. 2.00 Thus the following relation can be 1.95 assumed to calculate the loss function 1.90 increment FLPP at the peak pressure position, once the F1 at the minimum 1.85 V/V position has been evaluated: compression ratio
Entropy variation ratio . 1.80 8 10 12 14 16 18 20 22

FLPP 1.92 F1

(32)

Fig. 9 Entropy variation ratio as function of compression ratio for three different heat transfer models

Heat transfer model Woschni Eichelberg Hohenberg

a 0.8 0.5 0.8

b -0.53 0.50 -0.40

c 0 0 -0.06

EP (mean value) 1.91 1.83 2.03

This relation however has been derived for a constant mass process; it will be now shown that a similar relation can be derived in presence of mass leakage. As shown in equation (10), for a real adiabatic evolution the loss function increment is:

F = cP

m
m

(33)

Table 4 Mean entropy variation ratio using three different heat transfer models

It follows that for an adiabatic process in presence of mass leakage (neglecting the specific heat change) the ratio of the loss function increment is:

FLPP F
1

m cP m LPP mLPP m1 = m1 mLPP m cP m 1

(34)

The mass escaping from cylinder through valve seats and piston rings during the crank rotation (i.e. during the time interval t) can be evaluated by means of the equation for the mass flow through a convergent nozzle. Once the gas pressure is above the critical pressure (which is about 2 times the outer pressure), the leakage mass is:

m = Gnozzlet = AN

p 2 k 1 k m t V k + 1

k +1

(35)

where AN is the constant equivalent flow area. It follows that the ratio in equation (34) becomes:
k +1 p 2 k 1 k m V k + 1 LPP m1 = k +1 mLPP p 2 k 1 k mV k + 1 1

FLPP F1

(36)

Assuming that during the rotation arc from 1 to TDC the isentropic coefficient k remains constant, the loss function ratio becomes:

FLPP p = F m V LPP
1

p m V 1

(37)

The mass escaped in the considered crank rotation arc can amount to few percentage points of the total mass, hence:

m1 mLPP

(38)

Thus by means of the polytropic law pV=constant and of the already introduced dimensionless variable = 2x/d, the ratio in equation (37) becomes:

FLPP F
1

1 = LPP

+1
2

(39)

Using the same values of Table 3 it was found that this ratio moves from 1.94 to 2.07, with a mean value of 2, which is not too far from the result obtained in the case of heat transfers and no mass leakage (see equation 32), i.e. 1.92. Hence, considering a real non-adiabatic process, the constant Ep should assume a value between 1.9 and 2, i.e. 1.95

REFERENCES 1. M. Tazerout, O. Le Corre, P. Stouffs, "Compression Ratio and Tdc Calibrations Using Temperature- Entropy Diagram", SAE Paper 1999-01-3509 2. M. Tazerout, O. Le Corre, S. Rousseau, "Tdc Determination in IC Engines Based on the Thermodynamic Analysis of the Temperature-Entropy Diagram", SAE Paper 1999-01-1489 3. Ylva Nilsson, Lars Eriksson, "Determining Tdc Position Using Symmetry and Other Methods", SAE Paper 2004-01-1458 4. Mitsue Morishita, Kushiyama Tadashi, "An Improved Method for Determining the Tdc Position in a PV Diagram", SAE Paper 980625 5. A. Hribernik, "Statistical Determination of Correlation Between Pressure and Crankshaft Angle During Indication of Combustion Engines", SAE Paper 982541 6. Marek J. Stas, "Thermodynamic Determination of T.D.C. in Piston Combustion Engines", SAE Paper 960610 7. Wang Xi-Bo, Deng Kang-Yao, He Fang-Zheng, Zhou Zhen-Hua, A thermodynamics model for the compression and expansion process during the engine's motoring and a new method for the determination of TDC with simulation technique, Applied Thermal Engineering, 2007, 27 (11-12): 2003-2010 8. C. A. Finol and K. Robinson, "Thermal modelling of modern engines: a review of empirical correlations to estimate the in-cylinder heat transfer coefficient", Journal of Automobile Engineering, Proc. IMechE Part D, Vol. 220, 2006 9. J.I. Ramos, Internal combustion engine modeling, Hemisphere Publishing Corporation, 1989 10. J.H. Horlock and D.E. Winterbone, The Thermodynamics and gas dynamics of Internal Combustion Engines, Volume II, Clarendon Press, Oxford 1986 SYMBOLS AND ABBREVIATIONS A = in-cylinder heat exchange surface area Ac = cylinder section area AN = equivalent nozzle flow area for mass leakage calculation c = engine stroke cP = in-cylinder gas constant pressure specific heat cV = in-cylinder gas constant volume specific heat d = piston bore dY = differential of the generic function Y EP = proportionality constant errorT = gas temperature uncertainty at inlet valve closure error = engine compression ratio uncertainty F = loss function h = heat exchange coefficient k = gas isentropic coefficient = cP/cV l = connecting rod length m = in-cylinder gas mass p = in-cylinder gas pressure

p = in-cylinder gas pressure affected by measurement errors Q = heat received by the gas from the cylinder walls q = specific heat received by the gas from the cylinder walls r = crank radius R = in-cylinder gas constant S = in-cylinder gas specific entropy T = in-cylinder gas temperature t = time Twall = cylinder walls temperature u = in-cylinder gas specific internal energy um = mean piston speed v = in-cylinder gas specific volume V = in-cylinder volume x = piston distance from the cylinder top

= crank position 1 = crank position for the minimum dV/V 2 = crank position for the maximum dV/V B = BDC crank position measured with respect to cylinder axis (non centred crank mechanism) loss = loss angle T = TDC crank position measured with respect to cylinder axis (non centred crank mechanism)

= adimensional piston position = 2x/d

F1 = Fmin dV/V = loss function increment at the minimum dV/V angle F2 = Fmax dV/V = loss function increment at the maximum dV/V angle FLPP = loss function increment at the peak pressure position Fm = mean loss function increment = 1/2 (F1+F2) Y = finite increment of the generic function Y during the elemental crank rotation

= exponent of the polytropic evolution = adimensional crank pin offset = z/l

= engine connecting rod to crank radius ratio = engine compression ratio

ATDC = after top dead centre BDC = bottom dead centre BTDC = before top dead centre CA = crank angle CAD = crank angle degree(s) IMEP = indicated mean effective pressure IVC = inlet valve closure LPP = location of pressure peak LTDC = location of top dead centre MAP = manifold absolute pressure TDC = top dead centre