APPLICATION OF SEMI-ACTIVE SUSPENSION SYSTEM WITH FRICTION DAMPER AND ELECTROMAGNETIC DRIVE ON MOTORCYCLES

Transcription

1 APPLICATION OF SEMI-ACTIVE SUSPENSION SYSTEM WITH FRICTION DAMPER AND ELECTROMAGNETIC DRIVE ON MOTORCYCLES Aash Riazian & Soheyl Zakandy * Ayandegan Institute o Highe Education, Tonekabon, Ian * s.zakandi@aihe.ac.i ABSTRACT This pape investigates the eect o using semi-active suspension system with dampe iction and electomagnetic dive on motocycles. The advantages o this system can be pointed out to be the inceased passenge comot and saety. The cost o using this system is also aodable and can be oeed it as an option on motocycles. The model used in this pape was hal-vehicle model that it is a complete model o motocycles. Theeoe, the eect o ont and ea wheels is investigated simultaneously on suspension status and comot citeia o passenge. In addition, in ode to educe its sel-active vibation (chatteing) a balance contol logic appoach is used to emove the chatte. Keywods: motocycle, semi-active suspension system, dampe iction, electomagnetic dive; comot citeia o chatteing passenge comot. 1. INTRODUCTION Motocycle is one o the most useul vehicles in Ian that ae used by passenge o a long time. I the vehicle is lack o suitable suspension system, the vetebal column and uppe pats o bodies will be damaged. In almost all types o motocycles, suspension system is Inactive. Since the sping constant is invaiable, inactive suspension systems can have an optimal peomance in speciic situations and will not always be able to espond appopiately. Many Reseaches have aleady been done on active and semi-active suspension systems [1-6]. Most active suspension systems ae used in special cases such as militay and agicultual vehicles. These types o suspension systems equie a signiicant enegy souce o thei peomance and cuent souce powe supply (battey) o motocycles will be ailed to powe supply, and i it is necessay to add a battey, motocycle weight will be inceased signiicantly. Semi-active suspension system that was used by Kanopp and et al. [4] was outlined as the ist eseache in this ield. These systems ae against the costly, complex and enegy-consumed active suspension systems and have a wide application compaed to those systems. The eect o the application o semi-active suspension was studied with electohydaulic on the ea wheel o a motocycle theoetically and expeimentally by Segio et al. [5]. In this study, seveal dieent contol algoithms wee taken and thei eect was evaluated. The esults o this study indicate the eectiveness and the positive advantages o semi-active suspension systems. Cistiano et al. [6] installed o a semi-active suspension system on the ea wheel o a spot motocycle investigated the oad vibations o spot motocycles. Inteestingly, this study pesents a new contolle type using only one senso. The senso simply measues the ate o expansion o compession dampe based on data eceived om the senso and the ECU commands necessay ules to set the value o dampes coeicients. The algoithm has a positive impact on the motocycle suspension systems. Both ecent studies only have consideed the ea wheel suspension systems and the eect o both the ont and ea wheels to each othe has not been consideed yet. Dampes used in semi-active suspension systems have dieent types such as: electohydaulics, magnetics and iction luids (MR). The need o using iction dampes in semi-active suspension systems is the equipments o the contolle and contol algoithms [7]. Most o the hydaulic systems ae diven by the iction dampes. Since the hydaulic equipment can be costly and lead to incease o its weight, which is undesiable in motocycles, theeoe the iction dampes with hydaulic dive systems ae moe about vehicles and less about motocycles. Due to these disadvantages, the poposed electomagnetic dive is standing on a pole pemanent magnet and a coil is placed on the opposite side. By passing the electic cuent between the windings can be change the oce between them and the diving oce as the vetical oce on the shoes may be used. No study has been conducted to the possibility o using iction dampes contolled by magnetic stimulation. In this pape, the eect o a semi-active suspension system with iction dampe and an electomagnetic dive in motocycles has been investigated with the aim o impoving passenge comot. Hee, the hal-vehicle model is used, and the simultaneous eect o ont and ea wheels is consideed in the suspended status. 55

2 Riazian & Zakandy Application o Semi-Active Suspension System 2. Fiction Dampe with Magnetic Dive Figue 1 shows a schematic design o a semi- active suspension system with iction dampe. A new type o electomagnetic dive was used in this suspension system. Vetical oce equied compessing the iction layes to the dampe moving pat and hence to poduce the ictional esistance oce against the motion is povided by a magnetic oce. To this end, a winding is consideed that becomes magnetic when the electical cuent is enteed. Figue 1: schematic design o a semi- active suspension system with iction dampe and electomagnetic dive. 3. Motocycle Suspension System Modeling The suspension system was used in this study has ou degees o eedom and its dynamic model is shown in Figue 2. Figue 2: The dynamic model o the suspension system o a motocycle Indices and ee to the ont and ea pats espectively. Theeoe, m and m ae unspaing mass o ont and ea ties espectively. k and k ee to the constant sping o the ont and ea suspension systems, espectively. Similaly, k 1, k 2, c 1, c 2 ae: the stiness coeicient and extinction coeicient o motocycle ont and ea ties, espectively. The amount o oce o the suspension system dampes ae vaiable and contollable and ae shown with u and u o the ont and ea systems, espectively. Cente o gavity G and its distance om two teminals o the motocycle ae shown with a and b. m and w ae the mass and momentum the motocycle aound the cente o gavity, espectively. Paamete θ epesents the angle o motocycle otation the axis pependicula and z epesents the height o the cente o gavity elative to the eeence level (the gound). In addition, the z, z, y and y ae the 56

3 Riazian & Zakandy Application o Semi-Active Suspension System height and the height o the post - the height o the suace o the eath o the ont and ea ties, espectively. y and y unctions with a phase dieence equal to each othe: y ( t) y ( t ) (1) in which ( a b) y (2) Vibations equations o eeed model is as ollows : m z F w F m z F mg a F b ( 1 1 2( z y ) c2( z y k z y ) c ( z y ) F m m z k ) F m g F k ( z a z ) u F k ( z b z ) u g (3) (4) (5) (6) (7) (8) Calculating the iction oce on iction dampes used in this study does not equied vey high pecision so the use o complex models o taction contol calculations is not cost eective. The use o advanced iction models o usually mico- motos that is consideed accuate at small displacements is impotant. So, the poposed model o the study is the simpliied model o Coulomb iction. 4. Citeia o passenge comot The most impotant citeia o judging the ca s passenges on the ide and comot level, the inluence o vibations o the vehicle acting in the way. Beaing level o the vibation equencies and intensities and, most impotant citeion o judging the condition o a vehicle [13]. Identiication o evaluation citeia and the eect o vibations on health, comot and pesons eel standads developed and ae also pesented including a standad ISO 2631 and BS 6841 can be cited as one o the ist Intenational Standad Oganization and second o Bitish Standads Institutions. These standads have been evaluated the eects o thee modes o vibation in sitting, standing, and eclining and the input equency ange o vibations that wee studied wee between 0.5 to 80 Hz. Level o exposue to vibations (eithe in tems o time and intensity o the vibations), the aithmetic oot mean squae ae measued; a ms 1 2 T aw ( t) dt 0 T 0.5 (9) Acceleation values o a w in the above equation have the weight coeicient accoding to Table 1 (om the standad ISO2631). Regading the vetical motion o the vehicle is consideed, we only oe vetical weight coeicients in the above table (w k ) ae satisied. This place is the measue place o the oot mean squae o the aithmetic mean o the measued acceleation, in act the location is the seat occupants on a motocycle, this point ae vay on dieent motocycles, so in the pesent study is consideed consistent along the vetical cente o gavity. Calculating the amount o a ms can be attibuted to the passenge vehicle in accodance with Table 2 (in accodance with the standads o BS6841 and ISO 2361). Classiication pesented in the tables and explanations o each categoy accoding to the aveage o people's eelings calculated and pesented. 57

4 Riazian & Zakandy Application o Semi-Active Suspension System Table 1: the weighted coeicients o equency in vetical vibation mode Fequency () Hz W k (Facto1000) Fequency () Hz W k (Facto1000) Table 2: Reaction o dieent people sitting inside the ca against dieent values o a ms accoding to BS 6841 and ISO2361 standads a ms (m/s 2 ) Reaction <0.315 Not uncomotable A little uncomotable Faily uncomotable Uncomotable Vey uncomotable >2.0 Extemely uncomotable The main citeion o the contol logic pesented in this study is based on educing the maximum value a ms o in othe wods, cause to incease passenges iding and comot. Balance logic The logic was oiginally developed by Sanka and Rakheja in 1985 and then by Stammes and Sieteanu it was developed in The pupose o the contol logic is the deceleation to the chassis and the vehicle body and the method is a unction o the amount o oce on the dampe sping oce must be opposite, which means that the sping oce is equal but opposite in diection. With this policy, the sum o oces acting on the vehicle chassis will be zeo. But we cannot always use this law and in some cases o the lowest amount. Fomulation o balance logic can be expessed as ollows : F k sin g( ) sping damping (10) k sping in the above equation is the sping stiness coeicient that i the ont suspension is consideed, the ea suspension is in ode i its value is equal to k, k. The elative displacement o chassis and ties is shown with z a z z b z o ont suspension o ea suspension (11) 58

5 Riazian & Zakandy Application o Semi-Active Suspension System in ode to handle the numeical solution o equations 3 to 8, the ist time the tansomation matix and then compaed with the discete Eule method have been esolved. Figue 3 compaes the acceleation tansmitted to the vehicle chassis and body with Poughoban study [10]. In this study, the optimal contol o active suspension in hal vehicle model is applied. With given the similaity o this model with the vibating motocycle model used in this study. The bottom o Figue 3, the acceleation cuve in the ont and ea suspension models shows inactive suspension mode on Poughoban eseach [10]. The oughness unction o the oad that used in the study ((D'Amato et al, 2000) and (JS Lin, 1997) has also been as ollows : y 0.035(1 cos(8 t) 0.5 t 0.75 ( t) 0.025(1 cos(8 t) 3 t othewise (12) Figue 3: Acceleation vaiation o the motocycle, (a) and (b) ont body and ea body with semi-active suspension system and iction dampe, espectively; (c) and (d) ont body and ea body with passive suspension system and viscous dampe, espectively Compaing the ist ow o Figue 3 with second ows chats speciied that by using semi-active iction dampes the ate acceleations tansmitted to the ont and ea chassis signiicantly deceased ; o example, the maximum vetical acceleation o the vehicle in ont o a conventional viscous dampe moe than 4 m/s 2 ate the application o semi-active suspension iction, this numbe has deceased to about 2.5m/s 2.Although the use o semi-active iction dampe not only have beneit but also has some disadvantages, the most impotant othe undesiable phenomenon called chatte o sel-diving vibations. As can be seen in pats o the diagam, with the jeks huge 59

6 Riazian & Zakandy Application o Semi-Active Suspension System amount o momentum behind it is that they ae highe than usual. This phenomenon is called chatteing o sel diving vibations. The main eason o chatte is the natue o balance logic, because that it is a discontinue stategy. Accoding to the contol balance logic, i the amount o is less than zeo, the dampe oce will be equal to zeo othewise it will be equal to elastic oce o the sping, but with the opposite sign. Thus, by small changes in the speed and o envionmental conditions, cicumstances change and switch om zeo to non- zeo to the dampe oce o vice vesa is povided.it is natual that this phenomenon is hamul and contolled mechanical systems ae vitually unable to ollow commands with the intensity. Theeoe, it should be consideed a choice to its emove. Time Delay o Stimuli Function Any contol system has a time delay on its stimuli opeation. This delay could be due to actos such as the time o pocess computing, low mechanical equipment in pactice. It is o couse as the type o stimuli o this eseach is magnetic stimuli type, so its peomance much lowe than othe othe stimuli such as hydaulic stimulation. In the pevious example, i the time delay is consideed ideal and zeo. Figue 4 shows the ont and ea acceleation values ove the time, consideing the time delay o 30 ms. Compaing this igue with the uppe diagam in Figue 4 shows that the ate has declined shaply consecutive jeks and chatteing. Theeoe, taking into account the appopiate time delay can be eective in educing jeks but on the othe hand, excessive time delay inceases the peomance o semi-active suspension system will have a negative eect because it lags, only a potion o the elastic oce o the sping will neutalize and not all o it. Compaing the gaphs in Figue 4 with uppe gaph in Figue 3 clealy indicate that the volume o tansmitted acceleations is moe in case the system delay is consideed 30 ms. Figue 4: Acceleation vaiation o the motocycle, (a) ont body and (b) ea body with consideation o 30 sec time delay. Balance contol logic appoach to emove the chatte In aticles o emove chatte o this type o contolles, vaious solutions have been poposed. One o these methods is the use o Low-pass ilte on the output signals om the sensos to tack movement o the speed o vetical suspension system [8]. Since chatteing occu on the elatively highe equencies compaed to the nomal state occus, so this type o ilte can be used to detect the occuence o an chatte to pevent the passage o iational eedback One o the disadvantages o equency ilteing methods is that it equies a lot o inomation point to calculate the Fouie tansom o the intensity o the heavy computing load on the zeo point o each point within the ange o zeo. Anothe way to emove chatteing is the use o dead band on ecognized aound switch points and mode change; that means that instead o being zeo and o lead to thei multiple is zeo, it is consideed a bounday o each point since the width o the dead zone is consideed low, so movement in this aea is small. Theeoe, the elastic oce o the sping o handling in the aea, not too much. It is natual that the movements in the egion, the elastic oce o the sping will be balanced by the dampe. Also, in many cases, the height band o chatteing may be vaied that such cases can be eeed to suspension system in the study. Amplitude o displacement diagam jeks o speed will vay depending on oad conditions and suace oughness and conside the appopiate scope to dead band is not an easy task. Given the lage amount o this domain esulted in a complete loss o system peomance. Howeve, vey little consideation o it has no eect on chatteing emoval. Theeoe to emove chatteing and simultaneously 60

7 Riazian & Zakandy Application o Semi-Active Suspension System appopiate solutions, we need to update it accoding to the speed and path conditions which incease the complexity and volume calculations. Anothe method to emove the chatte phenomenon has been consideed in aticles and it is a combination o uzzy logic with vaiable stuctue contol [9]. Fuzzy vaiables in the pesent system can be consideed displacement and elative velocity o the suspension system. Using this appoach can lead to a apid pocess o switching and changing light conditions and theeby be educed jeks dastically. The method that consideed hee, without loss o accuacy, compaed to othe methods o computing has lowe numbes, and thus moe speed. Fist enteing to chatteing mode was diagnosed hee, then the amount o eedback inomation can be appoximated by a unction and this unction is actually the best unction o the cuve o degee3 passing a total o 20 ealie points. As chatteing occus on balance logic in the hoizontal axis coodinate system, so the condition diagnosed o this phenomenon, o example, o ont wheel has been consideed as ollows: Subscipt i indicate the system status in cuent time (t). So the pupose o (i-1) and (i-2) ae the system state at time (t-dt) and (t-2dt). Fist condition in equation 13, indicating that the ate o change o speed at the instant t i at time t i-1 changes in the tems o dieent sign. The second line o equation 13 indicates the second condition o the existence o the jump and chatte. This condition is expessed that chatte phenomena in the system, due to the natue o balance contol logic must occu nea the hoizontal axis coodinate system. So shaps in boundaies aound the hoizontal axis o the gaph in the coodinate system, not subject to chatteing phenomenon, may be indicate the exteme points. The pupose o this section is to educe time delay and also chatteing Contol. In the pevious example, the value o the time delay was 30 ms is educed to 20 ms in this example and also chatteing contol is peomed as descibed above. The ist ow o the diagam in Figue 5 shows the vaiation o acceleation with espect to time the ont and ea o the model. As mentioned peviously, we judge about vehicle status by investigate the acceleation tansmitted to the chassis and the passenge. By compaing the gaphs in Figue 5, the elie o some amount o acceleation is detected. Howeve, caeul judgment in this case depends calculated a w using equation 9. Table 3 shows the oot mean squae o the momentum aithmetic in the last two examples with an equivalent amount in an inactive suspension. Table 3: Root Mean Aithmetic Squae o Acceleation in two ecent examples and compaed with inactive suspension Method : Inactive suspension Balance contol with 30 Balance contol with Paamete ms time delay emoval chatteing steategy and 20 ms time delay a w a w a w a w, a w and a w in table 3 ae the squae oot o the aithmetic mean o the acceleation at the gavity cente, ont and ea o the vehicle. A ew points om the esults pesented in this table ae emovable: a w, a w and a w values in semi-active suspension mode is less than inactive mode that this epesents, the positive eect o contol system on semi- active suspension. By compaing to table 2 it is ecognized that all o obtained values is less than a minimum value o a w to povide vehicle (0.315). So the iding position and comot o passenges in both examples is not only uncomotable but also in the atio o 3 to 6 times is a the distance o upset the bounday. The a w value (eithe in the cente o gavity o ont and ea) in the last example is lowe than its pevious example. Although in this case the delay was educed om 30 to 20 ms, but using chatteing contol stategy, the oot mean aithmetic squae o the acceleation is educed 10 to 20 pecent. The bottom ow o Figue 5, the Poughoban esults [10] is also given o compaison. Although the pesent study used semi-active suspension and active suspension in the Poughoban eseach [10] but elatively little dieence between the acceleations values and also even lowe in some cases. Fo example, the maximum negative acceleation at the ea o the vehicle declined but thee was an incease in its volume which geneally it can be consideed a simila point o both o them. 61

8 Riazian & Zakandy Application o Semi-Active Suspension System Figue 5: Acceleation vaiation o the motocycle, (a) and (c) ont body, (b) and (d) ea body, ate eliminating o chatteing. Implementation o balance contol logic on a motocycle Motocycle which consideed hee, is a spot moto poduced by Italy Ducati Co. and the ollowing chaacteistics (units ae in the SI system): Like the pevious examples, the moto speed is consideed 40 kilometes pe hou and oad inequality unction no dieent. t should be mentioned that the opeating conditions in this example is also the balance contol logic with emoval chatteing stategy but with delay 10 ms. Figue 6 shows gaph o ont and ea motocycles acceleations compaed with the conventional inactive suspension. In ode to bette intepet the above chats, a w values ae calculated and pesented in table 4 o ont and ea motocycle suspension, both semi- active and inactive. As can be seen om this table, the a w and a w values o semi-active suspension have been educed signiicantly as compaed to the inactive suspension, which shows signiicant impovement in the condition o passenge comot in the motocycle. Table 4 : the aithmetic mean squae oot o the acceleation in the case o motocycle Method : Inactive suspension Semi active suspension Paamete a w a w

9 Riazian & Zakandy Application o Semi-Active Suspension System Figue 6: Acceleation vaiation o the motocycle, (a) and (b) semi-active suspension system, (c) and (d) active suspension system. 5. CONCLUSION In the semi-active suspension systems cannot be oced by dampe in all conditions othewise it exacebates body vibation and acceleation incease, and hence the oce applied to the body o the motocycle is being led to educed passenge comot. Due to the discontinuity in the oce o the semi-active dampe is occuing, the occuence chatteing o successive jeks would be inevitable. Chatteing is one o the weaknesses o this type o suspension and obvious that we should have a stategy to educe the intensity o the pogam. In this study the stating point o chatteing phenomenon is obseved quickly and then by choosing a smooth cuve (with espect to the oveall shape o the cuve duing 20 pevious stages), an appopiate amount to be consideed o dampe oce. This pocess will be continued until the exit cuve om hoizontal axis o the gaph. The diagam esults o vibations o this study ae compaable to active suspension esults and maximum vibation amplitude was appoximately equal to it. The esults o semi- active suspension on active Suspension paticula in passenge comot paamete wee supeio. Accoding to the above cases, and given the simplicity and deceasing the design cost and installation o this type o suspension (especially the iction dampe with magnetic stimulation) can say that o some vehicles such as motocycles will be have a elative advantage compaed to active suspension. 63

10 Riazian & Zakandy Application o Semi-Active Suspension System 6. REFERENCES [1]. Fedespiel-Labosse GM (1954) Contibution a l etude et au peectionement de la suspension des vehicules. J de la SIA, FISITA, pp [2]. Hedick JK, Womely DN (1975) Active suspension o gound suppot vehicles - a state-othe-at eview. ASME-AMD 15:2 40 [3]. Colla DA, Aboul Nou AMA (1988) Theoetical compaisons o vaious active suspension systems in tems o peomance and powe equiements. Poc IMechE Advanced Suspensions Coneence, London, UK, pp 1 9 [4]. Kanopp DC, Cosby MJ, Hawood RA (1974) Vibation contol using semi-active oce geneatos. J Eng Ind 97: [5]. Segio M. Savaesi, Cistiano Spelta, Andea Moneta, Filippo Tosi, Luca Fabbi, Loenzo Nado (2008) Semi-Active Contol Stategies o High-Peomance Motocycle. Poceedings o the 17th Wold Congess; The Intenational Fedeation o Automatic Contol: 4689: [6]. Cistiano Spelta, SegioM.Savaesi, LucaFabbi (2010) Expeimentalanaly analysis o a motocycle semiactive ea suspension. Contol EngineeingPactice 18; [7]. Emanuele Guglielmino, Kevin A. Edge (2004) A contolled iction dampe o vehicle applications. Contol Engineeing Pactice 12; 4: [8]. Ming-Lei Tseng, Min-Shin Chen (2010) Chatteing eduction o sliding mode contol by low-pass ilteing the contol signal. Asian Jounal o Contol 12; 3: [9]. Guglielmino E, Stammes CW, Stancioiu D, Sieteanu T, Ghigliazza R (2005) Hybid vaiable stuctueuzzy contol o a magnetoheological dampe o a seat suspension. Int J Vehicle Auton Syst 3; 1: [10]. Nase Pooghoban, Design o optimum contolle o automobile suspension system, Maste thesis, Tabiz Univesity, (2009). [11]. B.M. Eickho (1993) Passenge comot citeia, Bitish ail eseach, pape