852070 An Overview of Zero-Dimensional Thermodynamic Models for IC ‘AnoTRACT The heat release type models have been classified aa zero-dimensional because they have to spatial resolution and therefore don't contain any information on the fluid aechanics In them, They nave been a elgnifleant contri~ bution "because they Incorporate the rate processes occurring in the engine and are an ald In the analysis of the cate, Because they are thermodynamic models it 1a necessary to define ‘the control volumes on vhich the theraodmanic analysis 1s to be performed. The different. ‘control volume dsaoriptions, called one, tvo and three zone nodels, and the analysis of the conbustion event” using these models is, Finally a description of second law zero- @imensional models 1s given, These nocels have pimlar Liaitations as the first baw models; no Spatial resclution and a control rolune definition 1s required. These models are useful Decause they enapie one to analyze tne magnituae of the leases” that ocaur in the different, Processes vhich comprise the engine cycle. SINCE “THE LATE 1950'S there have been many detailed simulations of engine cycles. In such detailed treatments, the equations of change for a fluid are applied to Ue engine cole and solved step-by-step on a computer. These Siaulations can de classified aa to vhather they are predictive or dlagnostic end to the extent Engine Data Analysis David E. Foster University of Wisconsin- Madson Department of Medhancal Enginesing ‘The purpose of this paper Ls to present an overview "of heat release type oonbustion nodels. These models hava deen olasaified as Zero-dinensional thermodyienio modela because they ave n0 spatial resolution In deseribing the combustion phenonera. Furthermore the heat Pelease type models discussed merein are Glagnostic and not predictive. The discussion centers on using these type models to facilitate Gata interpretation and on the handling of the fata in éoing the analysis, ‘The presentation wili follow the chrono= Logical sequence in vhion the heat release nodele appeared in the literature. The First Law analysis will be derived and discussed; tne ane heat release stens from the First Law Interpretation, This Will be followed by what the author belleves to be a natural extension, that 1s. @ Second Lav analyals of tne sane control volume, These Second Law mocels, orven referred to as availability motels, are useful for quantifying the losses associated with the attrerent engine processes. DERIVATION Heat release analyses nave toon cited in the Literature for the past twenty five years. The earliest reference that the author 13 fomiLiar with ie the work of Lyn (1) and Austen etal. (2), The foundation cf all of the models Is the Firat Lax of Thernogynanics with attention being focused on uhe combustion that apatial resolution 1s incluced in the * Numtans in parenthesis daaignate referances at nodel. tthe end of paper. 196 ‘0096.7363/069400-0496512.50 Ccopycgtt 1908 Soclaty of Automotive Engineers, Ine.process. The derivation given here is not Original to tne author, 1% follows very closely the derivation given in the paper by Krieger and Boraan (3) and is reproduced here to act 2s the cornerstone for the discussion which follows. mihi Control Surface Arbitrary control volume for First Law Analysis (zero shaft work 19 assumed) Figure 1 represents an arbitrary control volue “at a specific instant in time. The analysis assunes that at any instant in tine tne control velune ia in theenodyramie aquilib= Plum, Thus it is assumed that’ the entire control volume containa a homogeneous equilib= ium @ixture of air and conbustion preducte at each instant. Phenomena such as teaperature gradients, pressure waves, ron-equilibrium Soxpeoitions, fusl vaporization, aixing ete. are ignored in the medeis desorived here, Following these assuaptions an energy balance, written as @ rate equation, can be witten lanediately amusdt = rate of change of total internal enersy, mu, of the control yolune of P(aV/dt) = rate’ of mechentcal work dore by the _ System on its boundaries. G= rate of heat transfer into the control volune through the control velune boundary. ‘energy convected into or ovt of the control volume at f1ow area 1osavion i by mass flow rate #, For flow out” of the control? volume the enthalpy, ny is the control volume total enthalpy. For Flow into the control volume the stagnation absolute enthalpy of the flow stream fat the boundary Le to be used. any = 852070 ‘A masa belance on the control volume and tne appropriate Tlow equations for une intel and exit Movs Is used to relate the rate of change of mags of the control volume and the mass flovs orossing the boundaries: ea, web here a sass flow in is taken as positive. In order to solve equation I additional relationships are required. An equation of state and an equation for the internal energy as 2 Tunction of pressure, tenperature and coapos!~ ton are needed. These are written a9 rate equations for substitution into equation T. Teal Goes PV = eT As 8 rote equation: a Pa v vat Bean SE cath a The internal energy 1s a function of vemerature, pressure aid composition; the pressure ‘dependence arises because the equlliorium coaposition changes with pressure. wea r where F = fuel/air equivalence ratio. ‘Ae 2 rate equation this becomes: au au, aT, ayy oP, au, oF rt ae By Ty Pe cry For products formed directly froa the reactants (fuel plus air) the value of F is the equivalence ratio of tne reactants. — For products of such a reaction mixec with alr or additional profucts of the same fuel the value of F may be computed from a hypothetical, Peactant F which would give the same atomic Properties as the resulting mixture. The equivalence ratio at any instant 1s given by reference 3: Fees where Fg = initial fuel-air equivalence ratio = instantaneous miss in the control volune mg = initial mass in the control velure e202 “initial fuel-air ratio prior to fuel addition fg + stoichiometric fuel-air ratio. quation this becomes:an Because the composition of the control volume is changing with tine, shifting equilib- ium, one must reelize that the apeciric gas constant, %, used in the ideal gas equation of state 1s changing. That is, the specific gas constant will te @ function of tenperature, proseure and composition. 8 _ aR at, aR, aF a” SY at * Pat BE w Equations I thru V are the equations tat fare solved simultaneously for the First Law analysis. They represent five equations and Sever unknowns. The wiknowa are alt, oT/3t, aP/at, aF/at, | eR/ct, du/dt and a¥/a:. The partial derivatives of u and R with respect to T, P ane F (used in equations 111 and \) ean be obtained from a ohamloal equilibrium progran. That 1s, under the assumption of shifting equitibriue they are not unknowns. The equilib= Plum program ia discussed lator. In addition to equations I thru ¥ the follewing supplenental Information is required: 1) A set of initial conditions; known composition at a krown temperature, pressure and volume. 2) an experimental pressure-time aLagran. 3) Expressions for the Internal energy, enthalpy and gas constant for the products and unburnes mixture. 4) A eorrelation fon the gaavatde heat= tranafer coefficient. 5) in estimate of metal temperatures tor ‘the exgoaed cylinder eurfacoa. 6) ata wnich describe the geonetry of the engine From the geomstry of the engine and ite operating speed one can readily obtain dV/dt. The rate of change of pressure, dP/dt, 18 obtainec foe the ‘experimental pressure dat For these type models It 1s accurate to describe then as First Law nodele driven by the expert: mental pressure trace. This 1s wnat aakes the models diagnostic. If one were to Input an ‘analytical expresaion for the greseure trace, oF ‘@ mass burning curve, the models would have Slosure and could be eonsidered predictive to tthe extent that the input analytical expression is accurate, ‘Aa ore alght guoes, the resulte of the model are extrenely sensitive to the shape of tthe input pressure trace. Secause the pressure Serivative Lo required slight perturbations in the data will cause wild osetllations in the Pressure derivative vith attendant wild Oscillations in the calculates neat relesse; a fone percent. error in tha pressure measurement, 426 can result in approximately a 50 percent error In the rate of heat release (2). In order to analyze the pressure data in more detail Pessaroners often vill ensemble average many Spline smoothing routine to obtain a well behaved average pressure trace, A subroutine TSScU from the International sathematical and Statiatical (IMSL), meference i, is comonly lused for this purpose. The expressions for the internal energy, ontnalpy and gaa constant oan be odtained aa nixture properties from the equilibrium Composition at that instant in time. The equilioriun thernodynamic properties ‘of the products of ccmbustion can be obtained several Gitrerent ways. yn (5) used the data of Giionrise (5), Krieger and forean (3) used tne Newnali-Starkman charts (7), MASA fas made a chemical equillbrius program availabie (8) and Oiieara and Borman (9) have publisned a special program for this explicit purpose. Furthermore, in examining equations 11 ané V oe quickly observes that the partial derivatives of tie internal energy and the specific gas constant with respect to temperature, pressure and Squivalence ratio are required. The usual practica 1a to expreaa sacn property in torma of the mole fractions of the constituents, Then the partial derivatives becone a function of the Perthel derivatives of the individual mole fractions with respect to teaperature, pressure and equivalence ratio. The program by Olikara ana Berman (9) gives these derivatives expiicit= ly 1n the output In caleulating the entnalpy and internat energy it 13 necessary to. use absolute quantities. That isthe enthalpy on internal energy is the sum of the sensible energy and the respective formation energy. Therefore one also needa relationship for the apecific neat af the mixtures. In addition, the neat transfer correlation chosen may require iaformation on ‘such transport properties as viscooity and thermal conductivity. Many relationships for these transport properties are available in the Literature, Rererences 10, 11 and 12 are but a few of these. Similarly there 1s latitude in ‘ones choice of a heat transfer comelation. In our laboratory we usually use the heat transfer cormalation by Amnand (13) ubleh haa the form: a-ak (nor acre, ‘ way? wo Finally one need to eatinate the total mass of fluld in the cylinder. This will include the mass of alr and fuel which enter through the Intake plus the wave of exaust geser remaining after the blow down and exhaust. Typically the tase flow into the cylinder 1s measured and the exhaust residual is estimated, A common technique fon estinating the residual is by