Feature Selective Validation of Automotive EMC Pre-compliance Tests

134 V. RUZEK, J. DRINOVSKY, J. CUPAK, FEATURE SELECTIVE VALIDATION OF AUTOMOTIVE EMC PRE-COMPLIANCE TESTS Feture Selective Vlidtion of Automotive EMC Pre-complince Tests Vclv RUZEK 1,2, Jiri DRINOVSKY 2, Jn CUPAK 2 1 EMC Dept., SKODA AUTO,.s., Vclv Klement 869, 293 01 Mld Boleslv, Czech Republic 2 Dept. of Rdio Electronics, Brno University of Technology, Technick 12, 612 00 Brno, Czech Republic vclv.ruzek@skod-uto.cz, drino@feec.vutbr.cz, xcupk02@stud.feec.vutbr.cz Submitted November 6, 2017 / Accepted December 5, 2017 Abstrct. The pre-complince tests of electromgnetic immunity re t present crucil issue for ll mjor utomotive mnufcturers. In current prctice, there re two bsic wys to implement these tests. The first one is bsed on significnt simplifiction of mesurement methods nd the subsequent trnsformtion of results for relevnt estimtion of certifiction mesurement results. The second is bsed on the ppliction of numericl methods nd the clcultion of the electric field intensity distribution in the cr body. The question, however, remins the degree of correltion between these methods, especilly when confronting the results of certifiction mesurement. This rticle provides n overview of the results obtined using both methods nd offers n innovtive view of the method of comprison using the FSV. Keywords Feture Selective Vlidtion, pre-complince testing, utomotive EMC, results processing 1. Introduction Numericl simultions hve become n integrl prt of seril production preprtion in ll sectors of the utomotive industry, electromgnetic immunity including. Numericl simultion llows to significntly shorten product development time nd reduce the costs of performed physicl tests. The open question is the degree of numericl methods vlidity with respect to rel test results. This leds to the decision whether it is possible to completely replce rel mesurements by numericl simultions, or use them s wy of certifiction [1]. The vst mjority of numericl methods currently used for numericl simultion of electromgnetic problems re bsed on Mxwell's equtions. Their overview cn be found in [2]. In the sme wy, high hopes re introduced [3] into the modifiction of the estblished mesurement procedures in order to simplify nd implement them with lowcost testing equipment. Such n pproch requires thorough knowledge of the method, s well s method for mtching the results obtined by mens of pre-certifiction mesurements to the certifiction results. An nlysis of vilble mesurement methods [4], [5] shows tht we hve the following primry test procedures: Tests bsed on the ntenn method (ALSE) [6], Current pulse injection (BCI) tests [7], [8], [9]. In both cses, we must consider wys to modify the method to simplify, reduce nd ccelerte tests. 1.1 Feture Selective Vlidtion Method Feture Selective Vlidtion Method (FSV) is modern, very effective nd objective method, prticulrly suitble for verifying the correltion of dt obtined using CEM (Computtionl Electro Mgnetism) methods nd vlidtion mesurements [10], [11]. The bsic motivtion for the ppliction of this method to technicl prctice is n effort to quntify: Repetbility of mesurement, Mesurement portbility, Effect of the opertor, Method vrition. The bsis of the FSV technique is to decompose the results into two components nd then recombine the results to obtin n overview of the overll consistency. The components used re the Amplitude Difference Mesure (ADM), which compres the mplitudes nd trends of individul dtsets (1), nd the Feture Difference Mesure (FDM) tht compres rpidly vrible chrcters (2) (s function of n independent vrible) of these sets. AMD nd FDM re then combined into the GDM (3) form. All ADM, FDM, nd GDM components re usble s pointby-point nlysis or s single globl mesurement [12]. For the ADM, FDM, nd GDM vlues, we cn use the following equtions (1), (2), (3). DOI: 10.13164/re.2018.0134 ELECTROMAGNETICS

RADIOENGINEERING, VOL. 27, NO. 1, APRIL 2018 135 ADM ( f ) exp, (1) FDM f 2FDM 1 f FDM 2 f, (2) f 2 FDM f 2 GDM ADM (3) where α, β, χ, δ re clcultion elements of the individul difference functions (DC nd Lo) nd FDM 1, FDM 2, re elements of the inverse trnsformtion of Hi nd Lo pssges. GDM is n overll scle complince of ADM nd FDM. A detiled derivtion of equtions cn be found in [11]. By computing the FSV, we obtin the reltionship between the numericl output nd its distribution in individul ctegories, with the number of vlidted dt points plced in the individul ctegories being preferbly expressed using the histogrm [13], [14]. In Tb. 1 we cn see n interpretive scle cpturing the quntittive nd qulittive spects of FSV vlues tht this method cn cquire. In this context, the gol of this pper is to present methodology for the vlidtion of numericl results nd experimentl mesurement tht includes the following. 1) Provide bsic rules for numericl modeling in the utomotive EMC re (model simplifiction nd division into functionl prts). 2) Describe n evlution methodology (E field intensity mesurements, choice of proper mesurement points). 3) Provide novel test setup for experimentl mesurement nd set rules for vehicle body dismntling. 4) Discuss the Feture Selective Vlidtion method ppliction in the cse of utomotive industry. 2. Pre-complince Immunity Test Bsed on Numericl Simultion As mentioned bove, there re two bsic wys of obtining pre-certifiction dt describing the electric field intensity distribution. The most promising is the ppliction of numericl simultions. In order to ssess their results, we firstly perform theoreticl nlysis of the resonnt behvior of the cr body, followed by prcticl clcultions. FSV vlue (quntittive) Less thn 0.1 Between 0.1 nd 0.2 Between 0.2 nd 0.4 Between 0.4 nd 0.8 Between 0.8 nd 1.6 Lrger thn 1.6 Tb. 1. The FSV interprettion mesure. FSV interprettion (qulittive) Excellent Very good Good Fir Poor Very poor 2.1 Estimted Field Distribution in Cr Chssis The own resonnt frequency, the resontor qulity, nd especilly the resonnt frequency chnge, cn be seen s result of severl filure types. From the point of view of this rticle, we will describe the chnge of resonnt frequency nd resontor qulity cused by the dielectric mteril plced inside the resontor. For the description of this sitution, we will dvntgeously use the Slter filure method [15]. The description is bsed on the idel sitution when the resontor is perfectly closed, its wlls hve n idel (infinite) conductivity nd there is perfectly homogeneous environment in the volume of the resontor. In such n environment, the distribution of the electromgnetic field cn be found by the nlyticl method. Assuming tht, we strt from homogeneous Mxwell equtions, we obtin wve equtions for the prtil components of the electromgnetic field E k E 0, (4) 2 2 H k H 0 (5) 2 2 where k is the wve number, E nd H the function denoting the distribution of the intct field for individul modes. These equtions cn hve n infinite number of solutions bsed on n infinite number of k vlues nd corresponding boundry conditions. For E E dv nd H H d V we get wveform, in this cse non-zero right side. The mening of the individul members on the right side of the eqution is the following: In nlogy with simple hrmonic movement, the externl forces represent n element tht chnges the resonnt frequency of the oscilltor. Thus we obtin modified electromgnetic field nd thus the resonnce frequency to the intct stte [15]. This is ffected by: ) Currents flowing through the resontor, b) Holes in the wlls of the resontor (imperfect resonnt cvity); c) The finl conductivity of the resontor wlls (cused by non-zero tngentil component). According to Slter's method we use the nlogy with resonnt circuit defined by RLC prmeters. For the resonnt frequency the following eqution cn be found 1 0 j 0 0 Q where 0 is the resonnt frequency of the LC circuit nd Q is the qulity fctor. For further considertion, we will consider tht the ngulr frequency 0 does not differ much from one of the resonnce frequencies. Another derivtion [15] gives us reltion (7) by which we cn esily express chnges in the resonnce frequency given by the bove effects: the currents flowing through the reson- (6)

136 V. RUZEK, J. DRINOVSKY, J. CUPAK, FEATURE SELECTIVE VALIDATION OF AUTOMOTIVE EMC PRE-COMPLIANCE TESTS tor, the holes in the resontor wlls nd the finl conductivity of the resontor. 1 d j ne H A S 2j Q 00 E EdV (7) 1 nheda 1 dv S J E. EE dv EE dv 0 0 0 Eqution (7) cn be considered s fundmentl expression of chnges in the properties of the resontor. The eqution cn be pplied now to the cvity resontor environment of the vehicle body. In this setup, we hve to tke into considertion the following conditions, which distinguish the idel resontor from the described rel one: ) The resontor is mde up of mteril with finite conductivity. b) In rel conditions, the wlls of the resontor re covered by lyer of dielectric mterils (coting mterils, fom-bsorbing mteril). c) The resonnt cvity, due to its irregulrity, llows the emergence of certin resonnt modes only in certin prts of it. d) The resontor hs lrge number of holes, significntly reducing its qulity. Fig. 1. The bsic vehicle body model. Color-coded body elements. Fig. 2. Supplementry prts of the numericl model geometry. 2.2 Field Distribution Obtined Using Numericl Simultions Bsed on the performed experiments, the MoM nlysis method ws chosen s the most pproprite [16]. The ReMesh tool [17] ws used for the model preprtion. For the solution of the eqution system, the 3D TriD field solver, which forms the core prt of the EMC Studio progrm pckge [18], ws used. The most importnt spect of the chosen pproch is the thorough preprtion of the numericl model, bsed on well-vilble CRASH models. Modifiction of such model consists of the following steps: reducing the extreme redundncy of the originl model dt, removing excess surfces of the reduced model, pproprite division of the model into functionl units, definition of model network size. Bsed on the bove described strtegy, the model is divided into two bsic prts: Welded nd mounted cr body model - Fig. 1, Other inserted prts (ggregte, front xle, dshbord reinforcement nd set structure) - Fig. 2. The experiment is performed by the exposure of the prepred model using verticlly polrized plne wve t the most criticl 30 to 220 MHz bnd. From prcticl experience we know tht exceeding the electric field strength Fig. 3. Loction of the electric field probe rry in the cr body. inside the body over 100 V/m limit cn cuse problems in the proper opertion of CAN bus communiction, sensors nd controls unit. For this reson, we will consider 100 V/m s the criticl limit. In order to mesure the electric field strength, n rry of virtul electric field probes ws plced in the body spce ccording to Fig. 3. The probe mtrix ws plced on the dshbord level bsed on theoreticl nlysis. Individul points were plced in the xes nd, ccording to their verticl position, were numbered 1 to 14 from the left side. The E1 E5 point groups then represent the steering wheel re, the E6 E9 re of the center console nd the E10 E14 pssenger irbg re. If the model is exposed by verticlly polrized plne wve (Fig. 4), we cn clerly identify the dominnt reso-

RADIOENGINEERING, VOL. 27, NO. 1, APRIL 2018 137 nnt frequency of 71 MHz in the group of mesuring points E1 E5, which mnifests t ll points in this group. It is the lowest self-contined resonnt mode of the interior of the bodywork with projection for ll mesuring xes. Intensity of the electric field detected especilly t points E2 E4 exceeds 500 V/m. Such high intensity is minly cused by the contributions of surfce currents flowing over the idelly conductive (PEC) elements of the net forming the steering wheel ssembly. However, given the enormous intensity of the electric field, we do not consider this result to be completely ccurte - we expect the rel intensity round 250 V/m. There is lso shift wy from the trend in E1. This point is significntly influenced by its proximity to the body edge nd thus the expected increse in the clculted field strength given by the computtionl method. The re of points E6 to E9 (Fig. 5) shows, unlike the previous group, more expected results. Agin, the dominnt resonnce frequency of 71 MHz is shown here, supplemented dditionlly by the frequency of 120 MHz, which in the cse of the previous group of points s resonnt ws reflected only mrginlly. The distribution of electric field strength in the re of points E10 to E14 (Fig. 6) is significntly influenced by two bsic phenomen: the presence of controls on the dshbord reinforcement nd the distribution of surfce currents on the pssenger door including the proximity of the door to the rest of the bodywork (high current density per segment). The E12 nd E13 points from the hlf of the monitored bnd clerly record the decrese in intensity cused by the wekening of the externl bond. Point E14 is then ffected by currents flowing long the body of the vehicle s well s point E1 nd exhibits completely different behvior thn the points locted more inside the body. Due to the similrity of the results in individul groups, for future work we cn consider only representtive points usge in the given group (steering wheel, vehicle center, pssenger irbg). Fig. 5. The field strength obtined with the TriD t points E6 to E9. Fig. 6. The field strength obtined with the TriD t points E10 to E14. 3. Experimentl Mesurement To verify the ccurcy of numericl simultions, experimentl mesurements of the electric field strength were performed under the sme conditions s for numericl simultion. Fig. 4. The field strength obtined with the TriD t points E1 to E5. 3.1 Mesurement Configurtion For purposes of mesurement, the body of the Škod Octvi ws dismntled ccording to the sme rules s for the construction of the simultion model. The test cr thus consists of the sme prts tht hve been numericlly modeled for simultion purposes. The test ws crried out by exposing the body of the vehicle to the field in the 30 MHz to 220 MHz frequency rnge with verticl nd horizontl polriztion. The intensity of the test field ws chosen ccording to the sme clibrtion curve s for numericl simultion with nominl intensity E = 30 V/m. For the mesurement, semi-nechoic chmber ws used.

138 V. RUZEK, J. DRINOVSKY, J. CUPAK, FEATURE SELECTIVE VALIDATION OF AUTOMOTIVE EMC PRE-COMPLIANCE TESTS Fig. 7. Skod Octvi cr body fter dismntle process. Fig. 9. The electric field strength mesured in the bodywork t points E1 E5. Fig. 8. Illustrtive figure of the mesurement. 3.2 Obtined Results The methodology of mesuring the intensity of the electric field ws gin bsed on the loction of the individul mesuring points in the mtrix with the designtion E1 E14, grouped gin into 3 res. The mesurement results obtined in xis 1 cn be seen in Fig. 9 to Fig. 11. An overll view of the frequency response of the electric field indictes tht for most of the tested frequencies, the intensity mesured inside the vehicle is lower thn the intensity of the clibrtion field. This mens tht the vehicle body forms n obstcle to the penetrtion of the wve into its interior spce. Only loclized resonnt frequencies re the exception for this condition. In ll groups of mesurement points, dominnt resonnce t 78 MHz is evident, especilly when the body is exposed by verticlly polrized wve. The highest electric field strength (310 V/m mesured t point E2) ws chieved t this frequency in the steering wheel re, where individul mesuring points re locted in the direct vicinity of the conductive prts of the steering column reinforcement nd the intensity of the electric field ws gretly ffected by currents flowing long their surfce. 4. Results Comprison Using FSV In order to express the degree of relibility nd suitbility of the proposed pre-certifiction simultion, con- Fig. 10. The electric field strength mesured in the bodywork t points E6 E9. Fig. 11. The electric field strength mesured in the bodywork t points E10 E14. sistent comprison of the results obtined through individul pproches is necessry. For better understnding of the difference interprettion before using the FVS method, the direct comprison of the numericl simultions results nd the mesured dt is shown in Fig. 12 to 14. The full

RADIOENGINEERING, VOL. 27, NO. 1, APRIL 2018 139 line shows the results of the pre-certifiction mesurements, the dshed line then shows the dt obtined using numericl simultions. In ll sets of dt we consider the vlues obtined in xis 1 nd points E4, E7, E12, exposed by verticlly nd horizontlly polrized wve. The points bove were chosen s reference in the given re. Point E4 represents the electric field intensity distribution in the vicinity of the steering column nd the steering wheel, the min resonnt element of the vehicle interior in the frequency rnge being investigted. Point E7 llows the description of the intensity of the electric field in the re of the infotinment system disply nd represents essentilly the center of the vehicle in its longitudinl plne. The lst point E12 repre- Fig. 14. Comprison of vehicle cr body mesurement nd simultion for points E10 E14. sents the vlue of the electric field in the right-hnd interior. Ech of these points is dditionlly locted from the electrodynmic point of view in different prt of the dshbord. The point E4 is surrounded by metl reinforcements (dshbord, steering wheel nd steering column) on both sides, the point E7 is locted ner the chrome frme of the infotinment disply nd the point E12 is locted on the plin surfce of the dshbord. Fig. 12. Comprison of vehicle cr body mesurement nd simultion for points E1 E5. 4.1 Comprison of Dt Sets The verge score of ll the points compred is shown in Tb. 2. The vlues below the limit of 0.5 re mrked. The course of the individul FSV prmeters for point E7 is then shown in Fig. 15 to Fig. 17. We cn notice tht very good degree of complince of both proposed pre-certifiction methods hs been chieved, especilly in the cse of verticl polriztion of the test field intensity. It is obvious tht the proposed numericl simultion chieves good results with pre-certifiction mesurements, especilly in the ADM AV prmeter. The verge vlue cross ll mesured points nd polriztion ADM AV = 0.38. This result ws chieved in prticulr by setting the prmeters of the vehicle simultion model. In the model, the dominnt resonnce fetures of the interior trim (dshbord reinforcement nd set structure) re used with the correct choice of sub-mesh locking on the internl stiffeners nd body edges. POINT/ PARAM. E4 V E4 H E7 V E7 H E12 V E12 H TOTAL ADM AV 0.41 0.51 0.37 0.31 0.37 0.31 0.38 FDM AV 0.49 0.48 0.49 0.46 0.49 0.46 0.48 Fig. 13. Comprison of vehicle cr body mesurement nd simultion for points E6 E9. GDM AV 0.70 0.74 0.67 0.58 0.68 0.58 0.66 Tb. 2. Overview of FSV scores t mesured points.

140 V. RUZEK, J. DRINOVSKY, J. CUPAK, FEATURE SELECTIVE VALIDATION OF AUTOMOTIVE EMC PRE-COMPLIANCE TESTS From the ADMi wveforms, it is obvious tht the gretest differences between mplitude mgnitude obtined by numericl simultion nd mesurement re in the position of very nrrow frequency mxim (bsolute difference E = 320 V/m t point E4, verticl polriztion). On the contrry, these vlues re not very high, but we generlly see higher intensity vlue obtined by numericl simultion. This is due to the modeling of the body s perfect electric conductor compred to rel-body mesurements mde up of lossy mteril. It would therefore be possible to consider mteril differentition of body prts mde up of different types of steels in order to increse the ccurcy of the clcultion. It is lso evident tht good results of numericl simultion were given in verticl polriztion, when contributions of the frequency mxim with respect to the shpe of the cvity re given in prticulr by their own resonnt modes of the body. Frequency mtching FDM AV chieves good results with its verge vlue over ll mesured points nd polriztion FDM AV = 0.48. Differences between polriztions re not consistent. In the cse of verticl polriztion, the shpes of the dominnt frequency mxim re very high, but their positions re not locted identiclly in both wys. The dominnt resonnt pek in the 70 MHz bnd is loclized to 72 MHz using numericl methods, while mesured t 78 MHz. We know tht this mximum is cused by the resonnce of prts of the steering column nd the steering wheel stiffness nd is prticulrly evident in point E4. As mjor cuse of this frequency difference, we suspect the limited possibility of discretiztion of the smll components constituting this ssembly, which in the numericl model ccording to the rules ws modeled by single piece, which cuses the shift of their resonnt frequency when exiting different lengths of conductive elements. Conversely, t 120 MHz, given in prticulr by the resonnce of the center console reinforcement, both methods re in very good greement. This prt of the cr is mde from elements whose size cn be esily modeled nd therefore the resonnt lengths for both methods re not different. The totl GDM AV complince level does not rech the required limit t ny of the points. The min reson is tht the numericl method used nd the setting of the simultion model contin conscious shortcomings given by the pre-certifiction test. It is cler tht perfect model (especilly the size of the discretiztion network elements nd very detiled mteril description of individul body prts) could be mde to increse the level of complince to GDM AV level of 0.5. Fig. 15. Frequency dependence of ADMi for point E7 (verticl). Fig. 16. Frequency dependence of FDMi for point E7 (verticl). Fig. 17. Frequency dependence of GDMi for point E7 (verticl). 4.2 Proposed Techniques for Improving FSV Complince A key spect for chieving high degree of complince of pre-certifiction tests nd numericl simultions presented with FSV is the correctness of prepring individul dtsets for the input method. It is therefore necessry to tke the following mesures in these res. 4.2.1 Numericl Simultion 1) Obtining pproprite source dt nd bsic modeling. The most commonly used source dt for EMC simultions re crsh models in the utomotive industry. The bsic problems tht hve to be solved in obtining numericl model hve been identified. Reducing the extreme redundncy of the originl model dt. Removing excess surfces of the reduced model. Approprite division of the model into functionl units. Definition of model network size. Error correction nd control of the resulting model. 2) Preprtion of the numericl model. To ensure dequte qulity, we set out the bsic rules for discretiztion of the model. Respect of phse-dependent distributions of currents nd chrges long the model structure. The mximum segment size is 1/10 of the shortest wvelength.

RADIOENGINEERING, VOL. 27, NO. 1, APRIL 2018 141 Using loclly higher density of discretiztion close to the model nd cvity edges. Respecting the 1:5 boundry rtio of ech tringulr segment. Incresing locl mesh density (sub-mesh) in the vicinity of simulted conductors due to high current nd chrge grdients. Determintion of the mximum rtio between the length nd dimeter of the wire segment (including the thickness of the shething) to 3. Ensuring correct connection of tringulr segments: segments must not overlp, intersect, nd must be connected through common endpoints. 3) Selection of the test field source including its position. We recommend using verticlly nd horizontlly polrized plne wve in position from the front of the model. We lso recommend simultion from the side of the model (with the sme prmeters) by horizontl polriztion to excite other custom resonnt modes of the body cvity. 4) Methods of monitoring nd mesuring simultion results. It is recommended to monitor the model using electric field intensity distributions, virtul field electric field mtrices, current mp lyouts on the bodywork nd virtul voltge nd current probe on the hrness. 4.2.2 Simplified Mesurement Method 1) Vehicle selection. To verify numericl simultions, we idelly chose the bre body of cr equipped with prts of the cr tht hve mjor influence on the distribution of the electric field strength in the bodywork. Typiclly, it is the dshbord reinforcement, including the steering wheel nd sets (to solve the problem inside the vehicle). 2) Dismntling the cr for the test (relevnt only when verifying the numericl simultion). A key spect is the mintennce of high-qulity conductive connection of ll used prts of the cr. 3) Reliztion of the pre-certifiction workplce. Due to the wy we perform the rdited electromgnetic field test, we need shielded workplce. Regrding the high finncil demnds for the purchse or rentl of shielded chmbers, we recommend the use of shielded tents. 4) Choice of test method. We recommend using substitution method to provide the most pproprite ccess to the test nd clibrtion of the necessry performnce. 5) Choice of pre-certifiction test equipment. We need to crry out the pre-certifiction test with devices operting on the sme principle s the certifiction exm. However, the choice of chep nd ffordble equipment with much lower power nd frequency rnges is non-stndrd. We recommend the use of USB low-cost RF genertors nd BiLog ntenns. 6) Anlysis of electric field strength in different models. Apprent differences were observed when performing the test on the bodywork nd in n ordinry vehicle. Observing the pek frequencies lredy found on the bodywork, it hs been shown tht the bndwidth of these frequency peks is substntilly greter for production cr. This behvior results from the distinct qulity of the resonnt cvity, which is significntly higher in the cse of the bodywork, s only the conductive prts of the vehicle re locted inside the vehicle. On the contrry, the cr's spce in the production is equipped with number of plstic, fbric nd glss components tht significntly reduce the qulity of the cvity. 5. Conclusion The rticle provides look t the vilble pre-certifiction methods of electromgnetic field tests in the utomotive industry nd compres their informtion cpbility with the FSV method. From the presented findings, the FSV method cn be pplied without difficulty to the utomotive environment, however under the conditions defined in Sec. 4.2.1 nd 4.2.2, only relevnt input dt settings (or the results of the pre-certifiction procedures) cn yield relevnt results. The FSV method provided n objective picture of the consistency of both proposed methods. It follows tht both proposed methods cpture the frequency position of the mximum in the body quite well. In the cse of compring the mplitude mtch, worse results hve been chieved due to the necessry simplifiction of the description of body prts nd the neglect of elements tht cuse the locl mxim to be ttenuted in the rel body. It turns out tht the key element in numericl modeling is thorough model preprtion nd good network qulity tht uniquely determines the ccurcy of surfce current flow. Without their proper knowledge, it is not possible to correctly identify the distribution of electromgnetic fields in the bodywork using the MoM method to obtin the relevnt output of the numericl simultion. In the cse of pre-certifiction mesurements using the modified method, the clibrtion of the test signl performnce plys key role. The min problem is the frequency gin of the trnsmitting ntenn, which must be covered by sufficient mplifier power reserve. The recommended pre-certifiction methods of mesurement re the chepest implementtion of tests not only from the point of view of cquisition nd opertion of the testing technique, but lso of the time used for their preprtion nd implementtion. Acknowledgments Reserch described in this pper ws finnced by the Czech Ministry of Eduction within the frmework of the

142 V. RUZEK, J. DRINOVSKY, J. CUPAK, FEATURE SELECTIVE VALIDATION OF AUTOMOTIVE EMC PRE-COMPLIANCE TESTS Ntionl Sustinbility Progrm under grnt LO1401. For reserch, infrstructure of the SIX Center ws used. References [1] RUZEK, V., DRINOVSKY, J. Aspects of EMS precomplince testing. In Proceedings of 9th Interntionl Conference Vscký Cáb 2011. Semily, 2011, p. 117 121. ISBN 978-80-214-4319-8 [2] SADIKU, M.N.O. Numericl Techniques in Electromgnetics. 2nd ed. Boc Rton: CRC Press, 2000. ISBN 08-493-1395-3 [3] DRINOVSKY, J., SVACINA, J., RUZEK, V., ZACHAR, J. Electromgnetic comptibility in utomotive industry. Elektrorevue Internet Journl. 2012, vol. 14, no. 3, p.1 8. ISSN: 1213-1539 [4] PIGNARI, S., CANAVERO, F.G. Theoreticl ssessment of bulk current injection versus rdition. IEEE Trnsctions on Electromgnetic Comptibility, 1996, vol. 38, no. 3, p. 469 477. DOI: 10.1109/15.536077. ISSN 00189375 [5] ADAMS, J. W., CRUZ, J., MELQUIST, D. Comprison mesurements of currents induced by rdition nd injection. IEEE Trnsctions on Electromgnetic Comptibility, 1992, vol. 34, no. 3, p. 360 362. DOI: 10.1109/15.155856 [6] ISO 11452-2: Rod Vehicles Component Test Methods for Electricl Disturbnces from Nrrowbnd Rdited Electromgnetic Energy Prt 2: Absorber-lined Shielded Enclosure. Ed. 2. Genev, Switzerlnd: ISO copyright office, 2004 [7] ISO 11451-4: Rod Vehicles Vehicle Test Methods for Electricl Disturbnces from Nrrowbnd Rdited Electromgnetic Energy Prt 4: Bulk Current Injection (BCI). Ed. 3. Genev, Switzerlnd: ISO copyright office, 2013. [8] ISO 11452-4: Rod Vehicles Component Test Methods for Electricl Disturbnces from Nrrowbnd Rdited Electromgnetic Energy Prt 4: Hrness Excittion Methods. Ed. 4. Genev, Switzerlnd: ISO copyright office, 2011. [9] HILL, D. A. Currents induced on multiconductor trnsmission lines by rdition nd injection. IEEE Trnsctions on Electromgnetic Comptibility, 1992, vol. 34, no. 4, p. 445 450. DOI: 10.1109/15.179277 [10] COLEBY, D. E., DUFFY, A. P. A visul interprettion rting scle for vlidtion of numericl models. COMPEL: Interntionl Journl for Computtion nd Mthemtics in Electricl nd Electronic Engineering, 2005, vol. 24, no. 4, p. 1078 1092. DOI: 10.1108/03321640510615472 [11] DUFFY, A. P., MARTIN, A. J. M., ORLANDI, A., et l. Feture Selective Vlidtion (FSV) for vlidtion of computtionl electromgnetics (CEM). Prt I - The FSV method. IEEE Trnsctions on Electromgnetic Comptibility, 2006, vol. 48, no. 3, p. 449 459. DOI: 10.1109/TEMC.2006.879358 [12] ORLANDI, A., DUFFY, A. P., ARCHAMBEAULT, B., et l. Feture Selective Vlidtion (FSV) for vlidtion of computtionl electromgnetics (CEM). Prt II - Assessment of FSV performnce. IEEE Trnsctions on Electromgnetic Comptibility, 2006, vol. 48, no. 3, p. 460 467. DOI: 10.1109/TEMC.2006.879360 [13] JINJUJ, B., ZHANG, G., WANG, L., DUFFY, A. Credibility evlution of uncertinty nlysis results of EMC simultion. In 2014 3rd Asi-Pcific Conference on Antenns nd Propgtion (APCAP). DOI: 10.1109/APCAP.2014.6992803 [14] JAUREGUI, R., POUS, M., SILVA, F. Use of reference limits in the Feture Selective Vlidtion (FSV) method. In 2014 Interntionl Symposium on Electromgnetic Comptibility (EMC Europe). DOI: 10.1109/EMCEurope.2014.6931054 [15] TESAR, J. Gs Permittivity Mesurement by Resontor Method. Diplom Thesis. Msryk University, Brno, 2010. (In Czech) [16] HARRINGTON, R. F. Field Computtion by Moment Methods. Pisctwy (USA): IEEE Press, 1993. ISBN 978-0-7803-1014-8 [17] EMCOS, GEORGIA. Hrness Studio User s Mnul. 162 pges. [Online] Cited 2017-04-25. Avilble t: www.emcos.com [18] EMCOS, GEORGIA. EMC Studio: Computer Simultion Softwre. 130 pges. [Online] Cited 2017-07-28. Avilble t: https://www.emcos.com/?products=emc-studio About the Authors... Václv RŮŽEK ws born in Tábor, Czech Republic, in 1985. He received his M.Sc. from Electrotechnicl Mnufcturing nd Mngement in 2009. His reserch interests include pre-complince electromgnetic comptibility issues nd numericl simultions. He is n EMC specilist in ŠKODA AUTO.s., responsible for pre-development issues, numericl simultions nd he leds region Chin from the EMC point of view. Jiří DŘÍNOVSKÝ ws born in Litomyšl, Czech Republic, in 1979. He received the M.Sc. nd Ph.D. degrees in Electronics nd Communiction from the Brno University of Technology, Brno, Czech Republic, in 2003 nd 2007, respectively. His Ph.D. thesis ws wrded by Emil Škod Awrd in 2007. Since 2006 he hs been n ssistnt professor in Electronics nd Communiction t the Dept. of Rdio Electronics, Brno University of Technology. His reserch ctivities include selected topics of EMC, EMI mesurements, nd EMS testing. He is lso interested in specilized problems of rdiofrequency nd microwve mesurements. Jn CUPÁK ws born in Olomouc, in 1986. He received his Mster s degree from the Technicl University of Brno in 2011. He joined OZM in 2014. He begn his creer t the compny of Chromservis, where he worked in reserch nd development of dvnced lbortory equipment designed for CBRN defense. He is experienced in HW design, EMC nd the development of new electronic components. Also, he hs been responsible for the development of specil electronic devices for the mesurement of physicl properties.