letnik/volume 53 - {t./no. 3/07 - str./pp. 157-214 Ljubljana, mar./Mar. 2007, zvezek/issue 503 STROJNIŠKI VESTNIK JOURNAL OF MECHANICAL ENGINEERING 19300 cena 3,34 EUR 9 770039 248001 ISSN 0039-2480 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 157 Vsebina - Contents Vsebina - Contents Strojniški vestnik - Journal of Mechanical Engineering letnik - volume 53, (2007), številka - number 3 Ljubljana, marec - March 2007 ISSN 0039-2480 Izhaja mesečno - Published monthly Razprave Papers Galovič, A., Živič, M., Can, A.: Energijska in eksergij- Galovič, A., Živič, M., Can, A.: Energy and exergy ska analiza sotočnih in protitočnih prenosnikov analysis of a parallel and counter-flow heat toplote z uporabo merilnih podatkov 158 exchangers using measured data Švaič, S., Boras, L, Andrassy, M.: Numerični Švaič, S., Boras, L, Andrassy, M.: A Numerical pristop toplotnih neporušnih preiskav Approach to Hidden Defects in Thermal Non- skritih okvar 165 Destructive Testing Popovič, P., Ivanovič, G.: Metodologija načrtovanja Popovič, P., Ivanovič, G: A Methodology for the Design zanesljivosti vozil med zasnovo 173 of Reliable Vehicles in the Concept Stage Kahaei, M. H., Torbatian, M., Poshtan, J.: Iskanje Kahaei, M. H., Torbatian, M., Poshtan, J.: Bearing- okvar ležajev z uporabo Meyerjevih Fault Detection Using the Meyer-Wavelet- algoritmov 186 Packets Algorithm Semolič, B., Sostar, A.: Mrežne organizacije - novi Semolič, B., Sostar, A.: Network organizations - a vzorec 21. stoletja 193 new paradigm of the 21st century Osebne vesti Personal Events Diplome 212 Diploma Degrees Navodila avtorjem 213 Instructions for Authors Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 158-164 UDK - UDC 536.27 Izvirni znanstveni članek - Original scientific paper (1.01) Energijska in eksergijska analiza sotočnih in protitočnih prenosnikov toplote z uporabo merilnih podatkov Energy and exergy analysis of a parallel and counter-flow heat exchangers using measured data Antun Galovič1 - Marija Živič2- Ahmet Can3 ('University of Zagreb, Croatia; 2JJ. Strossmayer University of Osijek, Croatia; 3Trakya University, Turkey) Prispevek podaja energijsko in eksergijsko analizo sotočnih in protitočnih ločilnih prenosnikov toplote z uporabo marilnih podatkov. Sestavili smo merilno progo, na kateri smo merili vstopne in izstopne temperature in masni tok vode. Za določitev brezrazsežnega razmerja izgubljene eksergije in prenesenega toplotnega toka v odvisnosti od brezrazsežnih parametrov prenosnika toplote: razmerja absolutnih temperatur na vstopu nT razmerja toplotnih moči n3 in števila enot prenosnika toplote n2, smo razvili analitični model. Za vse primere smo izračunali tudi učinkovitost prenosa toplote. Rezultate smo prikazali v ustreznih brezrazsežnih diagramih. © 2007 Strojniški vestnik. Vse pravice pridržane. (Ključne besede: prenosniki toplote, energijske analize, eksergijske analize, izmerjene vrednosti) This paper presents an energy and exergy analysis of a parallel and counter-flow recuperative heat exchangers using experimental data. An experimental rig was constructed to measure the inlet and outlet temperatures and the mass flow rates of streams. The analytical model was developed to obtain a non-dimensional relationship between the destroyed exergy and exchanged heat-flow rate as a function of the non-dimensional parameters of a heat exchanger: the ratio of inlet absolute temperatures, nT the ratio oj the heat-capacity rates, nv and the number of heat-transfer units, nT The effectiveness of the heat exchange is also calculated for each case. The results are shown in appropriate non-dimensional diagrams. © 2007 Journal of Mechanical Engineering. All rights reserved. (Keywords: heat exchangers, energy analysis, exergy analysis, measured values) 0 INTRODUCTION Energy-exergy analyses of heat exchangers have been the subject of much research over the past few decades, [1] to [7]. These analyses are based on the first and the second laws of thermodynamics. From such analyses the parameters for the improved operation of a heat exchanger can be obtained. The destroyed exergy or the lost available work of a heat exchanger is due to two factors: the transfer of heat across the stream-to-stream temperature difference and the frictional pressure drop that accompanies the circulation of fluid through the apparatus. Taking into account these two parameters it is possible to optimize a heat exchanger on an entropy-generation minimization or on a minimum destruction of exergy ([8] and [9]). In this work the exergy destruction due to the pressure drop is neglected, because from experimental data it was clear that the pressure drop was very small. 1 DESCRIPTION OF THE MEASURING RIG Figure 1. shows the experimental setup of the heat exchanger. The studied heat exchanger was “double pipe” type with only one passage of every stream. The streams were a hot-water stream (as the stronger stream) and a cold-water stream (as the weaker one). The mass flow rate of the weaker stream was kept constant and equal to 0.002 kg/s. The values of the mass flow rate of the stronger stream were 0.004, 0.006, 0.008 or 0.01 kg/s. As can be seen from Figure 158 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 158-164 Heating Tank h He N ter High Flow Meter Control Va lv e s Fig. 1. The experimental rig 1, the inlet and outlet temperatures of the streams were measured. In each of four cases (for the each value of the mass flow rate), the inlet temperature of the stronger stream was kept constant and the inlet temperature of the weaker stream was varied four times. From the obtained four sets of the measured data, for each of the four cases, both for a parallel and a counter-flow heat exchanger, two non dimen- 7l1 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 ¦ ^3=0.50 n3=04> * ^3=0.33 • ^3=0.25 /+ ¦ ^3=0.20 /^* f V3=0.50 sional parameters (p1 and p3) were calculated. The parameters are presented in the appropriate diagrams, as shown in Figures 2 and 3. The total heat exchanger area was 0.08 m2. For each point the exchanged-heat flow rate and the exergy destruction (entropy generation) are calculated by using the equations of the following mathematical model. 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 7l1 0.2 0.4 0.6 0.8 712 1.2 0.2 0.4 0.6 0.8 7l2 1.2 Fig. 2. The positions of the measured operation points for a parallel-flow heat exchanger Fig. 3. The positions of the measured operation points for a counter-flow heat exchanger N 0 0 Energijska in eksergijska analiza - Energy and exergy analysis 159 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 158-164 2 MATHEMATICAL MODEL The exchanged-heat flow rate between two streams can be calculated using the following equation ([10] and [11]): ß = C1 (^-^) = C2 (5'-4) (1). The heat exchanger effectiveness e is usually defined as: ¦$-f1' '_ Q 1 2 Ä^max (2). The exchanged-heat flow rate Q& can be rewritten as: ß = ^fimax=^1 (51'-4) (3). Since the pressure drop of the two streams is neglected, the entropy generation can be calculated from the following equation [11]: gen 1 T T C ln + Cln 2' ' (4). Using Equations (1) and (2), it is easy to transform Equation (4) into the form: ' ^ln 1-e 1 C ln Qe\T ' c r, 1+1M1-1|(5). If the following terms for two non-dimensional parameters c (6) are introduced into Equation (5), the equation for the entropy generation assumes the following form: 5gen=Qln(1-e(1-^T)) + C2ln 1+*,*—-1| (7). After multiplying the above equation by the environmental temperature, T0, the equation for the irreversibility or the exergy destruction is obtained as: ^^ =r0šgen =r0kln ( 1-f( 1-^T )) + C2ln 1+^J1 -1j (8). The above equation can be written in a non-dimensional form. For that purpose, it will be divided by the product of the heat capacity rate of the weaker stream, C1, and the environmental temperature, T0, as follows: 1=^ = ln(1-f(1-^T)) + -ln T0& 1+M3L\-------1 7T (9). Because the scope of this work is to show the ratio of the irreversibility and the exchanged-heat flow rate, it is useful to write Equation (1) in a non-dimensional form, dividing it by the product C1T1': Q (10). Finally, Equation (9) is divided by Equation (10), and a relevant ratio is obtained: ln(1-f(1-a-T)) + ln 1+7lM------1 E\1-71 (11). For a parallel heat exchanger the effectiveness e is obtained using the following formula [10]: -(1+»3K 1-e{ 1 + ^3 (12). and for a counter heat-transfer heat exchanger e is estimated as [10]: 1-e (1-»3 )»2 1 - jie (1-»3)»2 where: Q (13). (14). As can be seen, this ratio is a function of the non-dimensional parameters p3 and e = p1, which represent the operation points of the heat exchanger and which are obtained by the energy analysis of heat exchangers. The additional parameters for irre-versibility or exergy destruction are the input temperature ratia, pT, and the ambient temperature, T0. 3 CALCULATION RESULTS AND DISCUSSION The diagrams in Figures 4 and 5 present the values of i1 and i1/q1 for a parallel-heat-flow heat exchanger, and the diagrams in Figures 6 and 7 for a counter-flow heat exchanger. These values are calculated using Equations (9) to (11) and Equations (12) and (13) for the parallel and the counter heat exchanger, respectively. The black points in the diagrams represent the values of the operating points, which are presented in Figures 2 and 3. 2 160 Ga/ovic A. - Živič M. - Can A. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 158-164 0.012 0.01 0.008 0.006 0.004 0.002 0 0.5 0.6 0.7 0.8 0.9 a) ;r=0.50 0.7 0.65 0.6 E 0.55 0.5 0.45 0.4 0.35 0.3 0.012 0.01 0.008 0.006 0.004 0.002 0 0.5 0.6 0.7 0.8 0.9 0.7 0.65 0.6 e 0.55 0.5 0.45 0.4 0.35 0.3 7h b) tt3=0.33 0.012 0.01 0.008 0.006 0.004 0.002 0 0.5 0.6 0.7 0.8 m" 0.7 0.65 0.6 0.55 0.5 0.45 0.4 0.35 0.3 0.012 0.01 0.008 0.006 0.004 0.002 0 cK=0.25 0.7 0.8 0.9 d) ^=0.20 0.7 0.65 0.6 0.55 0.5 0.45 0.4 0.35 0.3 33 Fig. 4. The non-dimensional irreversibility i1 and the effectiveness e as a function of p2 and the parametric curves pT for a) p3=0.50, b) p3=0.33, c)p3=0.25 and d) p3=0.20 for a parallel flow heat exchanger 0.12 0.1 0.08 0.06 0.04 0.02 0 0.5 0.6 0.7 0.8 0.9 a) ;r=0.50 0.12 0.1 i °-08 — 1\ 0.06 0.04 0.02 0 3 0.12 0.1 i, 0.08 — * 0.06 0.04 0.02 Sj=1.07 0'5 0^6 0^7 0^8 0^9 ^=1.11 m *=1M b)?r=0.33 ^=1.17 0.12 0.1 0.08 h 0.5 0.6 0.7 0.8 0.9 ih 1\ 0.06 0.04 0.02 0 cW^O.25 0.5 0.6 0.7 0.8 0.9 7I2 d) ;t=0.20 Fig. 5. Relative non-dimensional irreversibility i1/q1 as a function of p2 and the parametric curves pT for a) p = 0.50, b) p = 0.33, a) p = 0.25, a) p = 0.20 for a parallel-heat-flow heat exchanger h h 1.1 1.2 1.1 1.2 h rc=1.07 e TT = 1.11 TT =1.14 TT =1.17 h h 1.1 1.2 0.5 0.6 1.1 1.2 h 1.1 1.2 1.1 1.2 1.1 1.2 1.1 1.2 Energijska in eksergijska analiza - Energy and exergy analysis 161 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 158-164 0.012 0.01 0.008 0.006 0.004 0.002 0 0.5 0.6 0.7 a) ^=0.50 0.7 0.65 06 E 0.55 0.5 0.45 0.4 0.35 0.3 0.012 0.01 0.008 0.006 0.004 0.002 0 0.012 0.01 0.008 0.006 0.004 0.002 0 0.5 0.6 0.7 0.7 0.65 0.6 0.55 0.5 0.45 0.4 0.35 0.3 T: =1.01 K=\.\\ K=\.\A 7l=\M 0.5 E 0.8 0.9 b) tt3=0.33 0.7 0.65 0.6 0.55 0.5 0.45 0.4 0.35 0.3 0.8 0.9 7h 0.012 0.01 0.008 0.006 0.004 0.002 0 0.5 0.6 0.7 0.8 0.9 0.7 0.65 0.6 e 0.55 0.5 0.45 0.4 0.35 0.3 c) p3 =0.25 d) p3=0.20 Fig.6. Non-dimensional irreversibility i1 and effectiveness e as a function of p2 and the parametric curves 12 for a) p =0.50, b) p =0.33, c)p =0.25 d) p =0.20 for a counter-flow heat exchanger o.i h 0.08 1\ 0.06 0.04 0.02 0 0.5 0.6 0.7 0.8 0.9 p2 a) p =0.50 ^ 0.06 0.04 0.02 0 0.12 0.1 h_ 0.08 11 0.06 0.04 0.02 0 ^=1.11 0.5 0.6 0.7 0.8 0.9 b) 71=033 0.5 0.6 0.7 0.8 cK=0.25 0.9 0.12 0.1 0.08 0.06 0.04 0.02 0 0.5 0.6 0.7 0.8 0.9 d) ?r=0.20 Fig. 7. Relative non-dimensional irreversibility i1/q1 as a function of p2 and the parametric curves pT for a) p =0.50, b) p =0.33, c) p =0.25, d) p =0.20 for a counter-flow heat exchanger ii h e 1.1 1.2 0.6 0.7 1.1 1.2 h ii H E 1.1 1.2 1.1 1.2 p 0.12 0.12 0.1 L 0.08 h 1.1 1.2 1.1 1.2 k=1.07 7TT=1.14 Kj=1.17 1 1.1 1.2 1.1 1.2 162 Galovič A. - Živič M. - Can A. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 158-164 The above diagrams clearly show the relationship between the dimensionless values of i1 and the dimensionless parameters p2, p3 and pT. For each diagram it is possible to directly read the values of i1 and e for every case. It is obvious that the given values of p2 and p3 do not have a significant influence on i1, but they have an influence on e. The pT ratio has the greatest influence on i1, but it has no influence on e. This can be quantitatively seen from Figures 4a and 4d, where the values p3 = 0.5 and 0.2 and parametric values pT =1.11; 1.14 and 1.17 have the same value of p2 = 0.76 and 1.09 respectively. For p3 = 0.5, all operation points have the same value of e, i.e., e = 0.455, with the exergy destruction increasing from 0.0028 to 0.0078. On the other hand, for p3 = 0.2 and the same parametric values of pT and p2 of 1.09, the value of e is equal to 0.61. The associated exergy destructions are 0.0040, 0.0067 and 0.0076 respectively. It is obvious that parallel- and counter-heat-flow heat exchangers have very small values of the ratio i1 and i1/q1 for the measured operation points. It is not possible to make a comparison with respect to exergy destruction and heat-transfer effectiveness of the researched cases, because they did not have 4 CONCLUSION The presented analytical relationship between the dimensionless exergy destruction and the heat-transfer effectiveness of a heat exchanger seems to be rather convenient, because it relates the dimensionless parameters (p3; p 2 and pT as additional parameter) relevant to the operation of a heat exchanger and the exergy destruction of a parallel and a counter-flow heat exchanger. It is possible to include the values of the measured operation points into the presented mathematical model and simultaneously calculate the heat-transfer effectiveness and the exergy destruction of the considered heat exchangers. It can be concluded that the exergy destruction for each case of both investigated heat exchangers is small. The main reason for such a conclusion is the fact that the operation values of pT close to 1. Furthermore, by introducing additional exergy destruction due to the pressure drop in the model, it is possible to develop an analytical model for the minimization of heat exchangers’ exergy destruction. the same non-dimensional p variables. 5 NOMENCLATURE A0 C Ex i1 k Q& q1 S& T T overall heat-transfer area, m2 heat-capacity rate of the stream, W/K exergy, W exergy destruction, W non-dimensional exergy destruction overall heat-transfer coefficient, W/(m2K) heat-transfer rate, W non-dimensional heat-flow rate entropy generation rate, W/K thermodynamic (absolute) temperature, K ambient temperature, K Greek Letters J pT p2 p3 Celsius temperature, oC ratio of inlet absolute temperatures, K number of heat-transfer units ratio of heat-capacity rates e, p heat-exchanger effectiveness Subscripts 1 weaker stream 2 stronger stream destr destruction gen generated Superscripts ‘ inlet ‘’ outlet 6 REFERENCES [1] Bejan, A., (1977) Second law analysis in heat transfer, Energy, 5, (1977), pp. 721-732 . [2] Rant, Z. (1956) Exergie, ein neues Wort für technische Arbeitsfähigkeit. Forschung Ing. Wesens 22(1956), pp. 36-37. [3] Rant, Z. (1964) Thermodynamische Bewertung der Verluste bei technischen Energieumwandlungen, BWK 16(1964), pp. 453-457. Energijska in eksergijska analiza - Energy and exergy analysis 163 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 158-164 [4] Can, A., Buyruk, E., Eryener, D., (2002) Exergoeconomic analysis of condenser type heat exchangers, Exergy- an International Journal 2(2002) , pp.113-118. [5] Galovič, A., Virag Z., Mudrinič, S., (2003), Non-dimensional entropy analysis of condenser and/or evaporator type heat exchangers, Transactions of Famena, 27(2003), pp. 1-9. [6] Gregorig, R. (1973) Wärmeaustausch und Wärmeaustauscher, Verlag Sauerländer, Frankfurt am Main. [7] Galovič, A., M. Živič, M. Andrassy, (2003) Entropijska analiza sotočnih prenosnikov toplote, Strojniški verfwl49 (2), pp.l00-110. [8] Bejan, A., (1988) Advance engineering thermodynamics, John Willey and Sons, New York. [9] Bejan, A. (1996) Entropy generation minimizaton, CRC Press, New York. [10] Bošnjakovič, F., K. F. Knoche, (1997) Technische Thermodynamik, Teil II, Steinkopf Verlag, Darmstadt. [11] Çengel, Y,Boles, M., A., (1994) Thermodynamics: an engineering approach, McGraw-Hill Book Company, New York Authors’ Addresses: Prof. Dr. Antun Galovič University of Zagreb Faculty of Mechanical Engineering and Naval Architecture Ivana Lučiča 5 10000 Zagreb, Croatia antun.galovic@fsb.hr Dr. Marija Živič J.J. Strossmayer University of Osijek Mechanical Engineering Faculty 35000 Slavonski Brod, Croatia marija.zivic@sfsb.hr Prof. Dr. Ahmet Can Trakya University Faculty of Mechanical Engineering and Architecture 22030 Edirne, Turkey Prejeto: Sprejeto: Odprto za diskusijo: 1 leto 7.2.2006 25.10.2006 Received: Accepted: Open for discussion: 1 year 164 Galovič A. - Živič M. - Can A. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 165-172 UDK - UDC [620.179.13:620.193]:519.876.5 Kratki znanstveni prispevek - Short scientific paper (1.03) Numerični pristop toplotnih neporušnih preiskav skritih okvar A Numerical Approach to Hidden Defects in Thermal Non-Destructive Testing Srečko Švaič - Ivanka Boras - Mladen Andrassy (University of Zagreb, Croatia) Uporaba infrardeče (IR) termografije kot neškodljiva preizkusna metoda za odkrivanje napak pod površino ter tudi za določevanje jakosti korozije veliko obeta. Razen omejitev, ki so posledica IR kamere same in toplotnih lastnosti preizkušanega materiala, nam da IR termografija v povezavi z ustrezno numerično metodo sprejemljive rezultate. Numerično simuliranje prenosa toplote nam omogoči ločeno analizo primernih parametrov, ki določijo razširjanje toplote v materialu, kot so: jakost in trajanje toplotnega vzbujanja, lastnosti materiala in začetni pogoji, prav tako pa tudi časovno porazdelitev nekaterih parametrov. Primerjava numerične simulacije in termografskih meritev, opisanih v [5], je dala zelo dobro ujemanje rezultatov. Iz numerične analize je jasno razvidna pomembnost določitve trenutka, ko kontrast doseže svojo največjo vrednost. Analiza nam prav tako pokaže, da lahko relativno izgubo materiala in premer napake določimo z največjo točnostjo v trenutku, ko kontrast doseže svojo največjo vrednost. © 2007 Strojniški vestnik. Vse pravice pridržane. (Ključne besede: neporušno preizkušanje, termografija, metode nadzornih prostornin, numerično simuliranje) The application of infrared (IR) thermography for detecting defects under the surface as well as for the estimation of the corrosion intensity seems to have good prospects as a non-destructive testing method. Besides the limitations which are the result of the IR camera itself and the thermal properties of the material detected, IR thermography produces acceptable results when combined with an appropriate numerical method. A numerical simulation of heat transport makes possible a separate analysis of the relevant parameters that characterize heat dissipation in the material, like the intensity and duration of the heat stimulation, the properties of the material and the starting conditions, as well as the time distribution oj certain parameters. The comparison of a numerical simulation and thermographic measurements presented in [5] shows a very good agreement of the results. The importance of determining the moment when the contrast reaches its maximum can be clearly seen from the numerical analysis. The analysis also shows that a relative material loss and the diameter of the defect can be estimated with the best accuracy at the moment when the current contrast reaches its maximum. © 2007 Journal of Mechanical Engineering. All rights reserved. (Keywords: non-destructive testing, thermography, control volume methods, numerical simulations) 0 INTRODUCTION The development of thermal non-destructive testing (TNDT) started with the introduction of the first infrared sensors. Basically, the method consists of the thermal treatment of a sample and of the measuring of its thermal response in the spatial and temporal domain, J=J(x,y,z,t). Experimentally, the method is based on recording the temperature dis- tribution of a reference sample surface over time, using one of the known methods of infrared (IR) radiation recording. The theoretical part is reduced to solving the problem of heat conduction through the sample in the 1D, 2D or 3D domain over time. The temperature field at the sample surface is the basis of all further analyses and conclusions on the possible presence of defects under the material surface and their parameters. The detection of a 165 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 165-172 defect expressed by the discontinuity of the temperature field at the sample surface depends on the changes of thermodynamic and physical properties of the sample. With regard to the temperature field, it is necessary to take into account all the additional influences important in forming the object temperature image (anisotropy of the basic sample material, variations in the surface state and the object’s thermal stimulation, noise from measuring equipment, etc.), and to distinguish between a real defect and the artefact. The paper presents the results of an examination of two models: one with defects underneath its surface made from a different material and the other with corrosion defects. The numerical simulation performed for both models is in fair accordance with measurement results. 1 THE BASICS OF THERMAL NON-DESTRUCTIVE TESTING Although the beginning of thermal non-destructive testing is attributed to the time when the first IR detectors were discovered in 1914, and almost all TNDT methods used today were developed from the 1960s until the end of the 1970s, these methods did not yield the results expected in comparison with other non-destructive testing (NDT) methods. A major change occurred when a thermodynamic approach to the problem of heat conduction and transfer was introduced into the measurement result analysis. Basically, the method may be reduced to a number of steps: thermal stimulation of the object, surface temperature recording (temperature response), data processing and decision making. The choice of the object thermal stimulation depends on the type of control to be performed (final testing, corrosion detection, improvement of visibility of impurities in the basic material, etc.). Rapid heating or cooling of the sample leads to the sudden appearance of hot or cold areas at the sample surface, and a high sensitivity of measurement may be achieved. Heating or cooling may be applied to the front or the rear side of the sample, depending on the purpose of the procedure. The front or the rear sample side refers to the surface to be inspected. Nowadays TNDT methods are divided into two basic groups: methods for preventive maintenance and methods developed for special needs of defect detection in the material [1]. The choice of a TNDT method depends on the type of defect expected underneath the material surface, the object accessibility from both sides and geometrical characteristics of the sample. The basic defect types in the object are as follows: surface or inside cracks, impurities in the material, surface roughness, variations of the coating thickness, poor clinging of the coating, surface delaminates and corrosion. The thermal properties of a solid may be regarded as functions of spatial coordinates X = A(x,y,z) and a = a(x,yz). Every interruption or steep change of these functions represents a potential defect (impurity) in the basic material. The analysis of the functions A(x,y,z) and a = a(x,y,z) represents the material analysis of the sample. The problems TNDT deals with include temperature functions depending on numerous parameters, such as spatial coordinates, time, dimensions of the sample and its thermal properties, defect dimensions and thermal properties, and heat-exchange parameters. All these parameters have to be taken into account when evaluating measurement results and defining the defect parameters, e.g., position, shape, dimensions and thermal properties. The basic parameter in TNDT analysis is a quantity defined as the contrast. There are several definitions of contrast [1]: a) Temperature contrast, defined as the difference of the temperatures of spots at the object representing a sound material and a material with impurities: AT(t) = Tnd(x,y,t)-Td(x,y,t) (1). b) Instantaneous (current) contrast, defined as the ratio of the temperature contrast and the sound material temperature: c = T„d (x,y,t)-Td(x,y,t) =1 Td (x, y, t) Tnd(x,y,t) T^(x,y,i) c) Normalized contrast defined as a difference of ratia: c = Td(x,y,t) Tnd(x,y,t) " T/(x,y,t) TJ"(x,y,t) ( ), where the quantities in the denominators represent the peak temperatures at the chosen object surface spots with and without defects. The selection of the contrast calculation mode during the analysis of the results also brings 166 Švaič S. - Boras I. - Andrassy M. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 165-172 along possible dependences of the chosen contrast type on the process parameters, the heat flux density in the first place, sample heating time, uniformity of the thermal stimulation, as well as the uniformity of the optical properties of the object surface, the distance between the spots of the object surface, the temperatures that are taken for the contrast calculation and other parameters. In order to achieve a simpler result analysis and to enable a comparison with other investigations, one should apply the contrast types that are less dependent on the process parameters and more dependent on the state of the material underneath the object surface as more appropriate. 2 NUMERICAL METHODS IN THERMAL NONDESTRUCTIVE TESTING A numerical simulation of heat transport through an object with defects makes it possible to evaluate the behaviour of various types of defects (position, geometry, defect material properties) with different initial and boundary conditions, without the noise that is normally present in experiments. A comparison of the results of the numerical analysis performed using the control volume method with the measurement results confirms the reliability of the numerical procedure. The procedure of the numerical analysis of heat transport starts with the three-dimensional non-steady heat-transport equation in rectangular coordinates: 33 ô [ „ 33 p C----=----- A----- H dt dx\ âx ô y\ ô y ô z\ ô z where: p - density, kg/m3 e - specific heat capacity, J/(kgK) X- thermal conductivity, W/(mK) 3 - temperature, °C x, y, z - spatial coordinates, m - heat source or sink, W/m3 Equation (5) is obtained by implicitly discretizing equation (4) [2]: a p xTp aE^E "*" aW **W "*" aN ^N "*" "T"(3rr XTg ~\~ Ü.J, XTj, ~T~ Ü.D XTn ~\~ U (5), with coefficients: Ay-Az X Ax- Az w Ax- Ay W7 a Ay-Az (sX Ax- Az Ax- Ay (sX v-Ax-Ay-Az + a°p-3p p-c- Ax- Ay ¦ Az At (7) (8) (9). (4), For each control volume the associated set of algebraic equations is to be solved. Boundary conditions are defined for all body surfaces according to the conditions in the experimental part. 3 DESCRIPTIONS OF MODELS AND MEASUREMENT PROCEDURE Investigations were carried out for two models: the first one simulated the existence of air in the steel plate and the second defects due to corrosion. Convection and radiation y = 0, y = h x = 0, x = b z = 0 z =S adiabatic boundary condition____________ adiabatic boundary condition____________ thermal simulation of the object in a defined time interval, free convection and heat radiation after heating free convection and heat radiation, ambient temperature 3o, temperature of the object in ambient 3ob. Fig. 1. Boundary conditions Numerični pristop toplotnih neporusnih preiskav - A Numerical Approach to Hidden Defects 167 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 165-172 front surface 1,5 mm 0,9 mm 0,6 mm O o o 0,06 mm 0,15 mm 0,3 mm o o o rear surface 120____________^ -----------------------* Fig. 2. Model #1 geometry Fig. 3. Model #2 geometry 3.1 Model #1 glecting any changes in thermal properties of the material that may occur due to chemical reactions Figure 2 represents the model geometry. The involved in the corrosion process. model consists of two 146 x 155 mm steel plates. In The smooth front surface of the model was the base plate, which is 19 mm thick, cylindrical re- stimulated at the beginning of the simulation with a cesses 18 mm in diameter are milled, representing heat flux having a total energy of 78 J in a 5 ms time defects at various depths from the inspected sur- interval. The model was made of steel with known face. The 5 mm thick cover plate is tightly screwed thermal properties, i.e., thermal conductivity A = 32 to the base plate. The height of the cylinders milled W/(mK), thermal diffusivity a = 1.65-105 m2/s. It is into the base plate is in the range from 14 mm to 18 assumed that the thermal stimulation is uniform along mm, so their distance from the inspection surface is the sample surface. between 5 mm and 1 mm [3]. The properties of the used steel are as fol- 4 RESULTS OF THE NUMERICAL SIMULATION lows: thermal conductivity 2 = 35 W/(mK), thermal AND COMPARISON MEASUREMENTS diffusivity a = 9.96-106 m2 /s. The model is mounted into a wooden frame that is filled with thermal insu- The problem was solved numerically by us- lation, so that at the lateral sides of the model an ing the control-volume method. For solving the sys- adiabatic boundary condition may be assumed. The tem of algebraic equations, the Gauss-Seidel proce- thermal simulation of the model is performed using a dure was used. The total number of control volumes 500 W spotlight. The heat radiation of the spotlight was 15 960 for model #1, and 29 120 for model #2. is directed through an aluminium sheet channel to The control-volume mesh was adapted to the ob- the cover plate surface, and the temperature distri- served problem. The mesh was condensed in the bution is recorded on the opposite side of the model. areas where steeper temperature gradients were ex- For the numerical simulation a uniform thermal stimu- pected. lation of the object is assumed. The temperature dis- The time step was also adapted to the stabil- tribution at the object surface was measured using ity and accuracy requirements of the discretized the IR AGA 680 STANDARD camera. equation, so At = 0.04 s was adopted for model #1 and At = 0.001 s for model #2. The initial condition of 3.2. Model #2 the simulation assumed a model of uniform temperature. The model for the simulation of the corrosion defects was a steel plate with dimensions 120 x 4.1 Comparison of results 80 x 3 mm, as shown in Figure 3. The corroded areas are represented by six cylindrical recesses of 10 mm Model #1 in diameter. The depth of each defect represents a The investigations carried out on model #1 particular loss of material caused by corrosion. For were aimed to show the ability of the thermographic the numerical part of the investigation, corrosion is method to be used in the detection of defects under defined as a reduction of the material thickness, ne- the surface in objects made of materials with a good rear surface 168 Švaič S. - Boras I. - Andrassy M. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 165-172 28.21 28.01 27.81 27.61 27.41 27.20 26.99 26.78 26.57 0.02 0.04 0.06 0.08 0.10 0.12 0.14 COORDINATE X (m) 27.65 27.55 27.45 27.35 27.25 27.15 27.05 26.95 a) b) Fig. 4. Temperature distribution at the smooth surface of model #1 after 120 s: a) experimentally recorded thermogram, b) result obtained by numerical simulation thermal conductivity, and to show the possibility of detecting the defect geometry, i.e., its dimensions and position (depth) in the object. It was shown that the detection of defect shape was better in short observation times, more shallow defects are easier to recognize, the maximum instantaneous contrast is directly connected with the distance of the defect from the observed surface, regardless of its dimensions. Figure 4 shows a comparison of the experimental results and the numerical simulation. A relatively fair agreement between the results can be seen. In the thermogram on the left, in the upper right-hand corner there is an area of higher temperature, which can be explained by experimental noise rather than a measurement error. The contrast values are rather small, so in the recorded thermogram the deeper defects are barely noticeable. 0.06 0.05 0.04 0.03 0.02 0.01 0.07 0.06 0.05 0.04 0.03 0.02 0.01 AL/L = 50 % AL/L = 30 %A AL/L = 20 % 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 COORDINATE X (m) 0.02 0.01 0.07 0.0 0.0 0.04 0.0 0.02 0.01 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 COORDINAT x (m) 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 COORDINATE x (m) 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 COORDINATE X (m) Fig. 5. Numerically obtained temperature distributions on the front plate surface at 280 ms, 400 ms, 600 ms and 1000 ms time increments 0.12 0.1 0.04 0.02 0.07 0.0 0.0 0.04 0.03 Numerični pristop toplotnih neporusnih preiskav - A Numerical Approach to Hidden Defects 169 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 165-172 Model #2 The investigations carried out on model #2 were aimed to show the possibility of corrosion detection with thin metal plates. This paper presents the results of the numerical simulation carried out by the authors, and their comparison with the data obtained experimentally by Marinetti, Bison and Grinzato [4]. Temperature distributions at the plate surface for different time increments are shown in Figure 5. A comparison of numerical and experimental results, as described in the literature [4], using the temperature distribution along the white dotted line in the thermogram for the time increment 280 ms in Figure 5, displays a very good agreement. The corrosion degree estimation can be performed by using one of the inverse methods. The simplest algorithm relates the relative material loss and the instantaneous contrast in the following equation: AL L C(x,y,Ta) 1 + C(x,y,Tr) (10). The time interval during which the temperature response at the material surface is observed is most frequently expressed non-dimensionally, i.e., using the Fourier number: Fo L2 (11), where a is the thermal diffusivity of the material, (m2/s), ris the time interval, (s), and L is the thickness of the plate, (m). For large defects the heat transport can be reduced to a one-dimensional problem. The optimal inspection time is in the range Fo * 0.6 to 2.0, and Equation (10) is recommended. The characterization of smaller defects of complex shape is recommended in the interval Fo * 0.3 to 0.6 [5]. Due to the three-dimensional heat diffusion the defect detection is more difficult, so it is recommended to use the temperature derivative by time according to the following equation: M (x, y) T(x, y, T1 ) - T(x, y, T2 (12). In every case the shape of the corroded area is more precisely indicated in a shorter inspection time interval, but with a somewhat lower amplitude [5]. Figure 7 displays the temporal distribution of values according to Equation (10). In contrast to the data from [4], where the analysis was done for Fo = 0.68, the numerical simulation shows that the best results are obtained by inspection during the time of peak temperature contrast [3 ]. For all defects this was in the range Fo * 0.21 to 0.32. It is also important to select a reference point Tnd sufficiently far from the defect itself [6]. Table 1 Effective Estimation Estimation corrosion time of AL/L Fo corrosion AL/L Error % 0.02 0.05 0.1 0.21 0.242 0.264 0.2 0.292 0.175 0.3 0.31 0.264 0.5 0.32 0.455 0.0199 -0.5 0.0466 -6.8 0.0902 -9.8 -12.5 -12.0 -9.0 The estimation of corrosion for particular defects is shown in Table 1. The optimum time interval to estimate the defect contours (time interval during which the defect contours in the thermogram may be most accurately identified) is also the time interval for achieving the 0.90 0.80 0.70 0.60 0.50 0.00 0.02 0.10 0.12 0.04 0.06 0.08 COORDINATE x (m) Fig. 6. Temperature distribution on the dotted white line from Fig. 5 at time increment 0.28 s - numerical result 170 Švaič S. - Boras I. - Andrassy M. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 165-172 0.500 0.450 0.400 0.350 0.300 0.250 0.200 0.150 0.100 0.050 0.000 A A \ / \ \ / \ \ \ \ \ \ \ \ ** \ / \ / V - |> --__21 :=----= ---- ----------- ~ ----------- Timm i M M M M M M M M M M M M M M M M M M M M 1.50 Time (s) Fig. 7. Time dependent distributions according to Equation (10) peak DT, where the reading of the diameter of each particular defect is done at half of the amplitude. Figure 8 displays the line temperature distributions along the defects in the time instant Fo = 0.28 (t = 0.15 s). The effective diameter of all the defects was 10 mm. The measured defect diameters are given in Table 2. According to the literature data, material loss defects above some 20 % may be detected by ther-mography. The temperature time derivative according to equation (11) increases partially the defect visibility. In Figure 9, where the three-dimensional surface temperature distribution in the time instant Table 2 Effective corrosion DL/L Measured defect diameter mm 0.02 10.16 0.05 10.16 0.1 | 0.2 | 0.3 | 0.5 9.83 9.83 9.83 9.5 21.10a 21.001 20.90H 20.80H 20.70^ 20.60 | 20.50 — 0.04 0.06 0.08 Coordinate x (m) a) 20.64^ 20.62 20.60 20.58 20.56 20.60 — 20.58 — 0.00 0.02 0.04 0.06 0.08 Coordinate x (m) 0.10 0.12 b) Fig. 8. a) Line temperature distribution for cross-section a; b) Line temperature distribution for cross-section b 0.00 0.50 .00 2.00 2.50 3.00 0.00 0.02 0.0 0.12 Numerični pristop toplotnih neporusnih preiskav - A Numerical Approach to Hidden Defects ill Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 165-172 ^5^ ^sägg^S** Fig. 9. Temperature distribution for the time instant 0.1 s 0.1 s is shown, the smallest defect is not visible. On the other hand, the time derivative of temperature for t1= 0.05 s and t2 0.1 s in Figure 10 displays all the defects. 5 CONCLUSION The performed investigation shows that both the thermographic and the numerical methods may be successfully employed in thermal non-destructive testing. The numerical approach enables the simulation of the influence of particular parameters, thus enabling a more universal overview of the influencing values. With the simulation of a process for a Fig. 10. Temperature time derivative (0.05 to 0.1 s) model with a defined number and distribution of defects it was established that the mutual influence of defects is very important, which may be directly concluded from the experimental part of the investigation. The numerical simulation also indicates situations when the three-dimensional diffusion cannot be neglected. It was also shown that the selection of an optimum inspection time interval is essential for a high-quality evaluation of the results. It can be concluded that a high-quality approach to thermal nondestructive testing necessarily requires a close link between experimental and numerical analyses in order to detect and determine all the relevant defect parameters. 6 REFERENCES [1] D.P. Almond, P.M. Patel (1996) Photothermal science and technique, Chapman & Hall, London. [2] S.V. Patankar (1980) Numerical heat transfer and fluid flow, Hemisphere Publishing Corporation, Mc Grow-Hill Book Co, Washington. [3] Llhč-Boras, Š.Švaič (1997) Numerical simulation of the defect in specimen, 4th International Conference of the Slovenian Society for NDT, Ljubljana. [4] S.Marinetti, P.G.Bison, E.Grinzato (2002) 3D Heat flux effects in the experimental evaluation of corrosion by IR thermography, QIRT’02, Quantitative InfraRed Thermography 6, Dubrovnik, Croatia, 2002. [5] E.Grinzato, V.Vavilov (1998) Corrosion evaluation by thermal image processing and 3D modelling, Rev.Gen.Therm. Paris, France, pp. 669-679. [6] I. Boras, Š. Švaič (1998) Determination of the defect parameters in specimen by means of thermography and numerical methods, Proceeding of The International Society for Optical Engineering, San Antonio, Texas, USA, Vol. 3396, pp. 271-281. Authors’ address: Prof. Dr. Srečko Švaič Prof. Dr. Ivanka Boras Prof. Dr. Mladen Andrassy Faculty of Mechanical Eng. and Naval Architecture University of Zagreb Ivana Lučiča 5 10 000 Zagreb, Croatia ivanka.boras@fsb.hr Prejeto: Received: 15.3.2006 Sprejeto: Accepted: 25.10.2006 Odprto za diskusijo: 1 leto Open for discussion: 1 year 172 Švaič S. - Boras I. - Andrassy M. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 UDK - UDC 629.017.004.15:519.876.5 Kratki znanstveni prispevek - Short scientific paper (1.03) Metodologija načrtovanja zanesljivosti vozil med zasnovo A Methodology for the Design of Reliable Vehicles in the Concept Stage Predrag Popovič1 - Gradimir Ivanovič2 (1Institute "Vinca", Serbia; 2Faculty of Mechanical Enginireeng Belgrade, Serbia) Pri tako močni svetovni konkurenci je bistveno, da imajo motorna vozila veliko zanesljivost. Eden od načinov za doseganje tega cilja je metodologija načrtovanja zanesljivosti vozil. Stopnja načrtovanja je najbolj pomembna v dobi trajanja motornih vozil. Načrtovanje zanesljivosti v okviru faze načrtovanja je v neposredni povezavi z zanesljivostjo motornega vozila. Problem praktične metodologije za načrtovanje zanesljivosti motornih vozil je bil rešen z razvojem postopkov za ta namen. Osnova razvite metodologije predstavlja sodoben pristop k načrtovanju, ki zajema faze zasnove, predhodnega in glavnega načrtovanja z uporabo metode za načrtovanja zanesljivosti v teh fazah. Zaradi razpoložljivega prostora ta prispevek daje poudarek fazam zasnove načrtovanja zanesljivosti vozil. © 2007 Strojniški vestnik. Vse pravice pridržane. (Ključne besede: motoma vozila, zanesljivost vozil, načrtovanje zanesljivosti, numerično simuliranje) With world-wide competition so strong it is essential that motor vehicles have good reliability. One of the ways to reach this goal is the methodology for the design of reliable motor vehicles. The design stage is the most important in the life cycle of the motor vehicle. Reliability design, within the design stage, directly correlates with the reliability of the motor vehicle. The problem of the practical methodology for the reliability design of a motor vehicle was solved by developing a methodology for this purpose. The developed methodology represents a modern approach to design that comprises the phases of conceptual, preliminary and main design with the application of the method for reliability design within these phases. Due to the available space, this paper pays attention to the phase of the conceptual design of the vehicle's reliability. © 2007 Journal of Mechanical Engineering. All rights reserved. (Keywords: vehicles, reliability, reliability design, numerical simulations) 0 INTRODUCTION Reliability is a feature incorporated into a vehicle in the course of the design process, which is realised in the course of production by a high degree of technological discipline, and maintained in the exploitation by continual and stipulated maintenance and orderly usage. In designing reliability, it is necessary to predict or estimate the reliability of each motor-vehicle system element, as far as this is technically possible. The reliability is mainly determined on the basis of the ability of the given part or assembly or system to withstand the unforeseen overloading without catastrophic failure. The reliability of the vehicle elements (system, subsystem, assemblies, sub-assemblies, parts), especially of those critical in terms of reliability, is increasingly becoming the subject of special attention by vehicle designers and the automotive industry in general. The activities within the design process for motor vehicles and their components are often the result of quite opposing requirements, for example, low vehicle mass combined with high payload, high reliability and safety combined with maximum material savings, and small outline combined with maximum passenger comfort. It is the vehicle designer who has the special responsibility to assess the effect of the technology he or she is converting, by way of his or her efforts and knowledge, into a complex product in accordance with numerous legal regulations. Covering the reliability requirements is inevitable, particularly when designing the motor vehicle’s critical parts and assemblies. Providing the 173 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 designed life with the requested reliability and safety, Maintaining the competing ability implies resolving taking into consideration the appropriate legal regu- numerous problems and issues of vehicle reliability lations, is one of the requirements that have to be and safety for the producer. fulfilled with modern motor vehicles. Since reliabil- The aims of motor-vehicle design are often ity and long life are the primary goals in motor ve- opposed to the specific dependability require-hicles design, this means that developing and ap- ments, i.e., to the requirements related to reliabil-plying reliability design methods and techniques ity, availability and maintainability. This is the represents a significant activity in the vehicle-de- reason for investing great efforts and consider-sign process ([2], [3], [8], [9], [14], [15], [17] and [18]). able means to meet these requirements. Integrating In the field of reliability, reliability-design the overall design goals with the reliability methods and techniques have been developed to requirements is conducted and mainly executed be applied in the phase of system design. However, in the beginning of the initial design phases by these methods have been developed on different establishing an appropriate reliability programme. bases, so it is now difficult to determine which one The framework of the motor-vehicle reliability is the most suitable for application in motor-vehicle programme should be closely related to applying design. Perceiving this problem, as well as the and mastering new technologies, as well as to posproblem of comprehensive vehicle design, it has been sible difficulties in the design process. In other concluded that it is possible to bind these methods words, it ought to ensure the data and procedures into a single whole, with certain improvements, within for solving of these difficulties. In order to have the methodology of vehicle-reliability design. In this an effective programme for providing vehicle respect, vehicle-reliability design methodology has reliability, it has to be planned and process-been developed ([20] and [21]), and the basis of the determined, i.e., with determined assignments, modern approach to technical system design, with activities for completing them, and methods and its phases of concept, preliminary and detail design, techniques for the efficient execution of the are presented in Figure 1. Due to the available space, activities. The basic tasks of a motor-vehicle this paper pays attention to the phase of conceptual reliability-enabling programme are as follows: design of the vehicle’s reliability. - determining the reliability requirements, - realisation of the reliability-design process, 1 RELIABILITY DESIGN - supervision and control of the required reliability. Vehicle users’ sensitivity to failures is Incorporating reliability requirements and increasing. It is very often of decisive importance building them into the structure, by applying reli- when deciding to purchase a particular vehicle. From ability-design methods and techniques, defines the the motor-vehicle user’s point of view, the most vehicle’s behaviour in terms of failure, and that is important issue is the one related to the vehicle the vehicle designer’s obligation and responsibility. functioning without failure, i.e., it is related to the The basic aim of vehicle-reliability design, i.e., of its length of operating time, during which planned and elements, is reflected in decreasing the failure-event indicated maintenance is provided. Failure data are probability and possible human and material losses. the starting point for the quantitative analyses of Reliability design has a key role in motor-vehicle the reliability, maintenance and safety of the vehicle development, since reliability analyses and param- elements. Bearing in mind the fact that the failure of eters are the entry data for analysing and realising a vehicle part or assembly is the basic concept for the other design tasks, for the purpose of providing any reliability analysis, vehicle designers have to high vehicle effectiveness and efficiency. pay attention to the following: E S I G N PRODUCTION MARKETING CONCEPTUAL PRELIMINARY DETAIL USAGE & SUPPORT * + + i k Fig. 1. Popovič P. - Ivanovič G. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 - decreasing and minimising the number of failures during operation, - providing necessary and sufficient warning before a failure occurs, - enabling the vehicle to continue operation at a lower level in the event of a failure, - decreasing the costs and shortening the time required for the repair or replacement of the failed part. The approach to motor-vehicle design that poses such and similar questions and finds the balance in a systematic manner between all the design requirements (functionality, reliability, maintenance, and logistic support) is the design approach and philosophy that ought to be accepted and applied, since it contributes to meeting the users’ requirements, legal regulations and to raising the quality of vehicle usage, i.e., attaining market competitiveness. The basis of vehicle-reliability design are the methods of reliability design and the experimental databases ([1] to [3], [13] to [16] and [20] to [22]). Reliability design of vehicles and their elements ought to cover the following: - the selection of parts and materials that have been standardised as much as possible; - review and evaluation of all the parts and materials, prior to adopting the design documents, covering the operating characteristics, operating and critical stresses, manufacturing allowances and other features of the parts, for the determined and required function; - applying only those constituent parts/assemblies that are capable of meeting the reliability goals, i.e., of meeting the requirements of the specification for reliability design. The process of systems-reliability design includes a series of proceedings and working methods that in their essence have the character of predicting certain states that should be achieved within the system development. The design methods are developed and perfected with the objective of discovering the critical states stated in the process of system design, i.e., identifying possible errors, omissions and shortcomings. The decisions and selections based on the state analysis and on the results obtained from applied reliability-design methods, to a large extent contribute to a qualitative, efficient and effective achieving of the process of system design. The numerous difficulties in this kind of work and in the application of reliability-design methods, as well as the inevitable costs, can all be significantly decreased by a systematic approach, good planning, organised preparation and computer support. The initial bases and necessary preconditions for a complete inclusion of the reliability analysis and design into the vehicle-design process are the following: - organised databases on failures, operating and critical loads and logistical parameters of motor vehicles and their elements; - a systemised number of reliability-design methods supported by computer programmes; - defined criteria for the selection of an optimum re- liability-design method; - the prior experience and knowledge of designers and the technological ability of motor-vehicle producers. In accordance with the above, a reliability-design methodology has been developed covering the phases of conceptual, preliminary and detailed reliability design, so as to enable the designed system to contain certain, previously set reliability indicators. The framework and reliability design approaches are given for each of the phases, containing the sequence of applying the reliability design methods, the reasons and purpose of their application and the manner of evaluating the achieved results. Due to the wide scope of these results, this paper shows the phase of conceptual reliability design, as a portion of the developed vehicle-reliability design methods. 2 THE PHASE OF CONCEPTUAL RELIABILITY DESIGN Realisation of the process of system reliability design begins with developing a logical model covering all the available levels of the system design. This implies an organised and systematic approach to using design methods and techniques related to reliability, so as to ensure that the system being designed contains certain reliability indicators that have been stipulated in advance. The final reliability-design objective is installing the stipulated reliability into the system structure ([13] to [16], [20] and [21]). Identifying and defining the essence of the problem and the selection of the most acceptable solution, i.e., determining a set of quantitative and qualitative reliability requirements, represent the basic goals of reliability design within the phase of the conceptual designing. The correct outlining of Metodologija načrtovanja zanesljivosti vozil - A Methodology for the Design of Reliable Vehicles 175 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 the problem, i.e., of the questions, and the timely obtaining of the correct answers at the beginning of the reliability-design process is of special importance for the further realisation and for the success of reliability-design process. Therefore, it is necessary to answer the following questions: - What is the requested, specified level of reliability? - Why has that level of reliability been requested and specified? - How do we realise the reliability requirements and their elements? - What are the reasons for not being able to attain the requested reliability? - Is the additional reliability improvement necessary? Reliability design in the concept-design phase is primarily oriented towards defining the reliability specification and selecting the most acceptable solution from the point of view of a reliability-requirements meeting, which means that the reliability of vehicles and their elements is analysed. The process of vehicle designing is started by translating the users’ requirements and needs into the specification for designing, i.e., into the design assignment within of the creation of the predesign. The concept-design phase also defines the design goals from the point of view of meeting of the standards and regulations an of acceptable limit values, such as the following: the maximum torque burdening the transmission, the maximum operating speeds in individual transmission degrees (the tow diagram), the maximum speed and acceleration, the transmission ratio in the main transmission, and the maximum allowed level of noise. Determining the reliability requirements for each requested vehicle function was conducted on the basis of the identified stipulated functions and the desired reliability level for the covered mileage ([20] to [22]). The biggest challenge for a reliability designer, and thus for the whole motor-vehicle designers’ team, is the initial design activities, when it is possible to perform changes and modifications, practically without any significant material investments and loss of time. Therefore, the battle to successfully incorporate the reliability into the vehicle structure is either gained or lost on the issues of the accuracy, the comprehensiveness and the timeliness of the responses to the previously defined questions, as well as on the issue of forming of a precise and complete reliability specification. When aiming at providing a qualitative development of the concept design phase the literature was reviewed and research was conducted in the field of motor-vehicle reliability ([1], [5] to [8], [11] and [12]). The basis of vehicle-reliability design methodology are the phases of concept, preliminary and detailed design and the following methods: quality function deployment (QFD), reliability specification , reliability block diagrams (RBDs), reliability functions (RFs), reliability prediction and modelling (RPM), reliability allocation methods (RAM), failure mode, effects and criticality analysis (FMECA), fault tree analysis (FTA), stress-strength interference (SSI), reliability qualification testing (RQT) and design review evaluation (DRE). In defining the reliability requirements at the level of the vehicle-reliability concept solution and pre-design, the first seven methods were incorporated. In this respect, the reliability algorithm was defined in the conceptual phase, as shown in Figure 2. 2.1 The Algorithm of the Reliability Design Conceptual Phase The designing of a motor vehicle is shown in the form of a flow diagram, describing in a graphic manner the activities and sequences of the characteristic activities and of the activities in the realisation of the design processes with feedbacks. The flow diagram, based on the following of the system procedure, which is also an algorithmic one, provides a complete understanding of the reliability design problem, so that no major portion is overlooked or omitted. The structure of the system-reliability design proceeding of the developed methodology, as stated earlier, is based on three basic phases: the concept, the preliminary and the detailed, while Figure 2 shows the conceptual phase. These phases often overlap and are executed alternately, with a returning to previous phases through a feedback mechanism. The mainstay of the developed reliability-design methodology is establishing the starting basis for the reliability design and to improve the designers’ activities, rendering the improved proceedings of operation for a logical process of considering and deciding, using the methods and techniques of reliability design. The concept reliability design covers the vehicle system level, where by applying the reliability design methods the systems are considered, i.e., options are evaluated and a decision is made regarding 176 Popovič P. - Ivanovič G. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 the most suitable solution, according to the established criteria devised on the basis of the vehicle reliability-design specification. In the phase of the vehicle-reliability concept design, the established sequence of applying reliability-design methods is based on their accuracy and complexity. Namely, the reliability methods that are relatively simple are applied first, while the attained results have a lower confidence level, i.e., they are of an orientational character. With the small step forward, a complex method (FMECA), with a higher confidence level, was introduced into the designing process, through which the initial results are corrected in an analytical manner, and substantially more accurate reliability parameters are defined ([11], [16] to [18] and [23]). In the phase of concept reliability design, a reliability block diagram (RBD) is formed on the basis of knowing the vehicle system structure, the system functioning manner, i.e., the effect of system and subsystem failure on the vehicle’s operation. Considering the structure of the vehicle and its systems, from the point of view of functioning, and that of the requested functions, defining is enabled of the satisfactory and unsatisfactory vehicle/ system operations, as well as the manner in which to achieve it. The system, which is divided into functional blocks (systems and subsystems), represents the appropriate connections in the reliability block diagram, which is defined on the basis of the goal function and on the functional block failure. These in turn bring about a vehicle/system failure. RBD represents the basis for defining the vehicle-reliability function. Upon defining the RBD and the reliability function, a further design proceeding is realised in accordance with steps 7 to 15. 2.2 The Method of Quality Function Allocation in the Reliability-Design Process In accordance with the algorithm, and for the purpose of translating users’ needs and requirements (step 1, Figure 2) into quantitative and qualitative indicators, the method of quality function deployment (QFD) is applied, i.e., the procedure for identifying all the factors affecting the design requirements with the aim of meeting users’ requirements and the necessary methods and responsibilities for their control. Quality function deployment goes beyond the domain of reliability, since among the users’ requirements there are also the users’ wishes, but this is a useful and systematic way of pointing out the necessary activities of designing and control, aimed at enabling reliability. The quality deployment function is applied for the purpose of translating of users’ requirements into product characteristics, and in that way quality deployment is translated into the user-oriented quality function. The final objective is including the user’s “voice” into the process of product development and design, in this case vehicle development and design, regardless of whether it is a brand new or a modified product, with the aim of realising users’ requirements with respect to quality. 2.3 Reliability Specification The model that takes into account users’ wishes, requests and needs for the purpose of attaining users’ satisfaction, interpreted through the “quality houses”, enables an equilibrium between the operating characteristics of the requested reliability levels and the maintenance needed, i.e., the degree of availability of a freight vehicle. In the course of the reliability-specification forming, compromises are quite possible and exchanges between reliability parameters and maintenance parameters for the purpose of accomplishing the requested availability degree, with minimising of the construction costs. The reliability specification represents a starting point for reliability analyses and estimations, i.e., for the reliability design, and it is the constituent part of the documents related to the stipulated vehicle characteristics and performances. Defining the reliability specification of the vehicle, i.e., of its systems, is based on: - the reliability data available from the manufacturer, distributor and user of the vehicle, - requirements and needs of the user, as the main sources of information, - many years of following up of the vehicle in operational conditions ([19] to [21]). The specification (steps 2 & 3, Figure 2) implies defining the reliability level for the designed mileage and estimated number of failures. For instance, it can be adopted that there are no more than 10% of failures at 300,000 km of covered route. That implies that the number of vehicle systems failures will not be greater than 10%. In that respect, the reliability specification of the vehicle, as well as of its elements, where this level of reliability is expected, is as follows: Metodologija načrtovanja zanesljivosti vozil - A Methodology for the Design of Reliable Vehicles 177 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 MARKET - USERS * Manufacturers " Standards and * Scientific Publications Needs Analysis and System Users' Requirements Defining | 1 | List of Needs and Requirements Analyses and Reviews: Users' Requirements Users' Claim System maintenance Market Trent SYSTEM RELIABILITY CONCEPTUAL DESIGN J Translating of Users' Needs and Requirements Into Design Requirements | 21 Specification for Vehicle Design 13 I Specif, for System Design QFD Method Defining of Internal and Outside Influences on the System Design 4 System Reliability Specification Forming of the System Reliability Basic Block | 51 System Reliability Block Diagram^ | 61 System Reliability Function Qualitative and Quantitative System Reliability Requirements^, System Reliability prediction System Reliability Allocation System FMECA method 7 %:* Parameter \** y** Parameters 11 RPN: Parameters 9 ^mî Parameter io f,i Factor 12 13 Ç Factor L ~Hl4| h>i Parameter" T f»™. Factor Checking Rfi) ADS =Y,K (19). 11 In the case that the condition given in Equation (18) is satisfied, further designing is continued in accordance with steps 16 and 17, Fig. 2, within the phase of preliminary reliability design. Otherwise, in accordance with step 17, Fig. 2, it is necessary to take corrective measures according to the given algorithm. 3 EXAMPLE where in the above expressions: - fi, the factor of failure intensity mean value, - lDi, the designed failure intensity value. On the basis of the lDi design values, the reliability design value is determined in the phase of the concept reliability design, in the form: RDi (t) = e- (14). According to expressions (3) and (14), the designed vehicle reliability-function value is: RDS(t) = e-^'=Y\RDi(t) = e -<1>« (15). Using a logarithm, the following is obtained: (16) Due to the wide scope and the complexity of vehicle-reliability designing, this paper gives an example of motor-vehicle systems reliability design, i.e., of the mechanical power transmission system (hereinafter the PTS) of a freight vehicle within the phase of the concept reliability design. However, all that is presented for this particular system can be applied to other freight-vehicle systems as well, and to the vehicle as a whole. Structurally, this system comprises the following: clutch, gearbox, universal joint and rear axle. The PTS’s reliability depends on the reliability of its constituent parts, while the designing and the selection of parts have to be in accordance with the defined stipulated functions of those parts. The stipulated functions are those entered by designing into an item (system, subsystem, or part), 1 f RPM 2 180 Popovič P. - Ivanovič G. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 jointly comprising its overall operational ability. The functions, the PTS represents a complex system. quality of the reliability analysis greatly depends on Bearing in mind the fact that in the functional sense the ability of the designer to identify all the stipulated the PTS is a sequential link, it can be stated that the functions of an item, i.e., to classify the stipulated transmission system reliability design process is a functions, achieving a definition of the priority. very complex and demanding task that requires a According to the number of constituent timely establishment of a reliability programme. subsystems, assemblies and parts, their mutual Taking into consideration the great limitations relations, as well as the great number of stipulated regarding reducing the number of constituent parts/ Ratia: S^^ © HIGH yS >«. O MEDIUM yS >v A SMALL X ^V i MAX 1 MIN OGOAL Î O t A t A O t t 0 t î t t î A O T O O \ DESIGN ^REQUIREMENTS USERS' >v REQUIREMENT^ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 SIGNIFICANCE Power Transmission with min. Loss 5 © © O O O O A O A A O A A A A A O O O Efficient Resistance Overcoming in Vehicle Operation 5 © O © O A A o O A Gradial Connecting and Disconnecting from Engine 5 O © © Safe Start and Steady Operation 4 A © © © O O O O O A Protection from Dynamical Over Loads 4 O O O © O © A A A O Low Level of Noise and Vibrations 4 O A © © © © © © © O A O A O A O Driver’s Comfort 3 A O o O A A Long Life 3 o O A o O © O © O © O O A © © O © © Maintainability - Low Costs 3 A A O o A A A A O © A High Reliability Level 5 A © © © o O © A © © A O O O O © © O © © Small Size and Mass 2 A A A A © © O © ORGANISATIONAL DIFFICULTIES SET GOALS VALUE Fig. 3. Metodologija načrtovanja zanesljivosti vozil - A Methodology for the Design of Reliable Vehicles 181 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 assemblies, accomplishing the stipulated system reliability can be obtained by applying reliability design methods and techniques. In accordance with the algorithm in Figure 2, the text below shows the proceeding of reliability design in the phase of concept reliability design for the PTS. Step 1, Figure 2: Using the QFD method, the vehicle user’s needs, requirements and leanings were identified in connection with the PTS, obtained on the basis of processing and analysing the filled-in questionnaires and interviews with a large number of freight-vehicle users. Numerous initial data expressed through the freight-vehicle users’ requests and wants, were previously carefully “cleaned” to remove any impreciseness and fuzziness. After that, research of the relevant and measurable system characteristics, i.e., the design stipulations, was conducted, as well as the estimations of interdependence intensities, thus determining the design stipulations that have a major influence on the overall freight-vehicle users’ requirements. The design parameters target values were also defined, as well as the guidelines according to which designers can influence the users’ requirements. The ”quality house” for the PTS, obtained by the QFD method, is shown in Figure 3. In accordance with steps 2 & 3, Fig. 2, the specification for the PTS of VT10%=300,000 km was adopted, according to what is stated in defining expression (1). In accordance with this value, the stipulated level of PTS reliability was stipulated (step 4, Fig. 2), of: tf =A(300,000)= 0.9 (20) for the designed time t, i.e., for the covered route of 300,000 km. Step 5, Fig. 2: Based on the analysis of the functional-technological connection of the PTS elements, it was concluded that the PTS fails if there is a failure of the clutch or gearbox, or if there is a failure of the universal joint or drive gear. Therefore, it is a case of independent event failures. On the basis of this, the RBD represents a sequential link of 1-clutch, 2-gearbox, 3-universal joint and 4-drive gear, as shown in Figure 4.b. For example, Figure 4.a. shows the RBD for the vehicle. In defining the RBD for the vehicle, it was accepted that the failure of any system (from S1, S2 to SPTS - power transmission system, up to Sn), causes vehicle failure as well. In accordance with step 6, Fig. 2, and on the basis of the specified reliability, expression (20), the defined RBD, Figure 4.b. and the adopted exponential system-failure distribution, the “mechanical power transmission reliability function” was defined in the form: Rs(t) = YlR,(t) = e^ Z4 0.9 (21). On the basis of this value, for the designed number of kilometres of t = 300,000 km, the failure intensity of the PTS is: A =- nM) =- ln0.9 = 0.3512.10 -;b-1 22 t 300000 (2) where: - R/t), As, the set value of failure reliability and intensity of PTS, - R/t), A., R1(t),A1, R2(t),A2, R/t),X3, R/t), A4, of the failure reliability and the intensity of the subsystem (), i.e., of the clutch, gearbox, universal joint and drive axle, respectively, - t=300,000 the freight vehicle designed number of covered kilometres. After defining the reliability function, and on the basis of the available literature and reliability research performed on vehicles in operation, predicting and allocating failure intensity was performed. In that respect, the failure intensity values of i* according to Expression (5) (step 7, Fig. 2) Sn a) Vehicle 1 Clutch 2 Gearbox 3 Universal joint 4 Rear axle b) Power Transmission Fig. 4. 4 182 Popovič P. - Ivanovič G. Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 were adopted and given in Table 1. This table also gives the failure-intensity values allocated by applying the method of equal distribution A** according to expression (6) and A*** according to expression (7) (step 8, Fig. 2). On the basis of A*, A", A"*, the mean values XMi were determined, in accordance with expression (8) (step 9, Fig. 2), and given in Table 1. Using the FMECA method, a failure analysis was made and the values were set for the failure criticality degree assessment for the RPNi according to expression (10) and given in Table 2, (step 11 Fig. 2). On the basis of the RPNi and expression (11), failure criticality factors were determined and given in Table 1, (step 12, Fig. 2). Table 1 According to steps 10 & 13, Fig. 2, i.e., to Expressions (9) & (12), the values of factors fM, i fi were determined and given in Table 1. On the basis of AM- and these factors, according to Expression (13), i the failure-intensity designed values were determined for ADi i of the PTS subsystems and given in Table 1. According to step 15, checking the set values for AS and RS(t), according to Expressions (20) & (22) with respect to the designed values, ADS, according to Expression (16), Table 1, and RDS(t) according to Expression (15), and consistent with Equations (18) & (19), is a positive one, i.e., we find that: 4 Y,ÄDi =0.3474-10 [km] !=1 (23) A= 0.3512 -106>;L Failure rate Ax10-6[km-1] Factor Failure rate Ax10-6[km-1] Subsystem (i) Predicting of reliability l*i Allocation reliability Mean value l Mi l Di Equal distribution l** i AGREE l*** i fu /rpm fi 1-Clutch 0.04 0.0878 0.017 0.0483 0.1390 0.0210 0.1745 0.06064 2-Gearbox 0.09 0.0878 0.13 0.1026 0.2952 0.2750 0.2851 0.09907 3-Universal joint 0.05 0.0878 0.034 0.0573 0.1617 0.2050 0.1834 0.06371 4-Rear axle 0.16 0.0878 0.17 0.1392 0.4001 0.3100 0.3550 0.12338 Checking 0.3474 1.0000 1.0000 1.0000 A,DS=0.3474 Table 2 RELATED UNIT FMECA FORM FAILURE COMPONENT KIND EFFECT CAUSE Power flow interruption Clutch Loss of function Power transmissi on Gearbox Universal joint Rear axle Power flow interruption Power flow interruption ........................]...................... Power flow .interruption Loss of function Loss of function Loss of function Main spring failure Spindle output failure (stuck, worn out) Telescope failure (stuck, worn out) .................................................................... Flat gear failure ......................................... CONTROL MEASURES Measuring of rigidity Replacement 2 Testing of operating loads Replacement ESTIMATION OF CRITICAL DEGREE BASED ON Probability of Failure - PF NEGLIGIBLE LOW MEAN 4-6 Failure Demerit Value - FDV Probability of Failure Remedy -PFR NEGLIGIBLE 2-3 LOW 1__ 2-3 CRITICAL HIGH MEAN 4-6 MEAN ACTUAL STATUS PFi 3 10 8 240 FDVi 10 ............................................................................................................................................................... PFRi 10 180 RPNi 243 160 ESTIMATE OF CRITICAL DEGREE: Risk Priority Number – RPN RPN = PF x FDV x PFR 1 2-3 LOW 4-6 Estimated value of RPN < 50 MEAN 50 -100 HIGH 7-8 HIGH 7-8 LOW 7-8 HIGH 100 - 200 CRITICAL 9-10 CRITICAL 9-10 NEGLIGIBLE 9-10 CRITICAL >200 Metodologija načrtovanja zanesljivosti vozil - A Methodology for the Design of Reliable Vehicles 183 9 2 8 Strojniški vestnik - Journal of Mechanical Engineering 53(2007)3, 173-185 Rs (t) = e* = 0.9 < RDS (t) = l\Ru (t) (3), f i fa f = where N shows the number of balls, fr is the shaft’s rotating speed, a denotes the contact angle of the balls and races, fb, fi, fo, and express, respectively, the frequencies of the defective ball, the inner race, and the outer race, and d and D are the balls and pitch diameters, as illustrated in Fig. 1. Vibration signals measured on the machine’s surface are normally embedded in background noise, and therefore, high-precision techniques should be established for detecting and/or diagnosing machine failures. Consistent with other findings in the literature, to resolve the frequency content of a signal using the short-time Fourier transform, a sufficient data record is required. It is well known, however, that when the latter technique is applied to a large number of samples, the time localization is lost. This becomes more significant for non-steady signals when the detection of transients and movements (containing drift, trends, abrupt changes, etc.) is required. The ineffectiveness of this algorithm in such cases may lead to poor results and wrong conclusions. The use of wavelet packets has recently been considered for analyzing bearing defects, using coif4 wavelets [4]. The capability of this technique in concentrating on a desired portion of the frequency content of a signal in the time-frequency domain has received increasing attention in different applications. In this paper, a Meyer-packet-wavelets algorithm is proposed for bearing-fault detection. Accordingly, an effective technique is designed based on the relevant WP tree to reduce the number of computations. The outline of the paper is as follows. In Section 2, a brief review of wavelet packets and the Meyer filter is presented. In Section 3, the proposed wavelet-packet algorithm is described for bearing fault detection and is evaluated using the simulated and real-time data. In Section 4, a method is considered for reducing the required computations, and the conclusions are presented in Section 5. 1 WAVELET PACKETS Assume that the quadrature mirror lowpass and highpass filters of an orthogonal wavelet are, respectively, given by h(n) and g(n). The wavelet-packet coefficients are then defined by subsampling the convolutions of djp(n) with h(-2n) and g(-2n) as [5]: d2j^(ri) = d^(n)*h(-2n) (4), where 0