U radu se razmatra problematika modeliranja i eksperimentalne identifikacije karakteristika suhe dvostruke spojke s ciljem formiranja dinamičkog modela spojke te skupa eksperimentalnih procedura i postava za potrebe identifikacije i validacije modela. Razvoj više-fizikalnog modela suhe dvostruke spojke proveden je kao nastavak prethodnog istraživanja tokom kojeg su razvijeni i eksperimentalno validirani dinamički model pripadajućeg elektromehaničkog aktuatora te bazni model aksijalne dinamike spojke. Pritom je naglasak stavljen na proširenje postojećeg modela aksijalne dinamike spojke različitim efektima (npr. toplinsko širenje uslijed porasta temperature). Zajedno, ova dva modela (model aktuatora i model spojke) omogućavaju predviđanje karakteristike okretnog momenta suhe dvostruke spojke u funkciji pozicije motora aktuatora za široki raspon radnih parametara uključujući promjenu temperature i trošenje spojke. Model je strukturiran kao niz koncentriranih masa, opruga i poluga između kojih se odvija prijenos sila, topline te suho trenje. Struktura modela dobivena je razmatranjem strukture spojke dok su pojedini parametri modela eksperimentalno utvrđeni (uključujući mase, toplinske kapacitete, dimenzije i opružne karakteristike) primjenom razvijenih i nadograđenih eksperimentalnih postava. Pored mehanizma izgradnje normalne sile spojke u funkciji pozicije motora aktuatora, model ujedno uključuje toplinski model spojke koji omogućava predikciju temperatura komponenata spojke iz drugih operativnih parametara te eksperimentalno identificirane karakteristike tarnog materijala. Eksperimentalna identifikacija tarnog materijala uključuje dvije karakteristike ugrađene u model: (i) faktor trenja u funkciji triju operativnih parametara: temperature, brzine klizanja i normalne sile te (ii) trošenje tarnog materijala u funkciji disipirane energije i temperature. Dodatno, istražena je (iii) sklonost tarnog materijala podrhtavanju eksperimentalnom identifikacijom na malom uzorku i čitavoj spojci uz razmatranje korelacije između rezultata te (iv) utjecaj dubine trošenja tarnog materijala na faktor trenja i sklonost podrhtavanju. Predložen je novi model aksijalne dinamike s ciljem daljnjeg povećanja preciznosti predikcije okretnog momenta te pozicija pojedinih komponenti spojke tokom prijenosa momenta. Validacija novog modela pokazuje povećanu točnost u modeliranju pojedinih efekata u odnosu na postojeći model.
|Abstract (english)|| |
The thesis deals with modeling and experimental characterization of dry dual clutch characteristics with the goal of forming the clutch dynamical model as well as a set of experimental procedures and test rigs required for experimental identification and validation of the model. The development of a multi-physical model of a dry dual clutch is conducted as an extension of a previous research during which a dynamical model of a related electromechanical actuator and basic clutch axial dynamics model were developed and experimentally validated with the emphasis on expansion of the existent axial dynamics model by various effects (e.g. thermal expansion due to temperature increase). Together, these two models (actuator model and clutch model) enable the prediction of clutch torque characteristics as a function of actuator motor for a wide range of operating parameters including temperature change and clutch wear. Model is structured as a series of masses, springs, and levers in between which a transfer of forces and heat as well as dry friction occurs. Structure of the model is obtained by consideration of clutch structure while individual model parameters were experimentally identified (including masses, heat capacities, dimensions and spring characteristics) using the developed and upgraded test rigs. Torque transfer over dry dual clutch is a function of (i) normal force, (ii) coefficient of friction, (iii) clutch friction plate radius, and (iv) the number of friction surfaces. It should be noted that the first two parameters are dependent on various operating parameters, while the second two are constant for a given clutch. Normal force characteristics is influenced by various effects whose precise modeling is a necessary prerequisite for the development of the overall model which can be used as a base for the development of clutch control system. The effects which influence the normal force, i.e. the clutch axial dynamics include: (i) subsystem which generates the normal force including the effects of elasticity, friction and clearance of individual elements, (ii) effects of thermal dynamics on key components temperatures, (iii) effects of clutch thermal expansion effects on normal force development, and (iv) wear compensation mechanism influence. Coefficient of friction typically depends on three operating parameters: temperature, slip speed, and normal force. Additionally, coefficient of friction can also depend on friction material wear.The thesis is therefore separated into several chapters, where, in general, each chapter describes one of the effects which influence either the normal force or the coefficient of friction. Thus, the thesis consist of eleven chapters: (i) Introduction, (ii) System description, (iii) Test rigs, (iv) The existing actuator and clutch axial dynamics model, (v) Friction material characteristics, (vi) Clutch thermal dynamics, (vii) Clutch thermal expansion, (viii) Clutch wear, (ix) Clutch components spring characteristics, (x) New clutch axial dynamics model, and (xi) Conclusion. The Introduction chapter describes the basic principle of dual clutch based transmissions including the general advantages and disadvantages of such system as well the challenges related to the development of control system, specifically for system based on dry dual clutches. The System description chapter provides general description and the structure of transmissions based on dual clutches with the appropriate actuators as well as the specifics of dry dual clutches. The dry dual clutch which is investigated in this thesis is thoroughly described including the associated electromechanical actuator. The third chapter describes the Test rigs which were developed or significantly upgraded with the goal of supporting the modeling of dry dual clutch effort. The test rigs include (i) the clutch setup which includes the associated transmission, (ii) the CNC pin-on-disc type tribometer machine and the (iii) manually powered spindle based press. The structure of the Existing actuator and clutch axial dynamics model is presented in the fourth chapter. The existing axial dynamics model was developed during the effort to develop the model of the associated electromechanical actuator. Herein, the entire clutch/actuator model is validated against experimental data obtained from new clutch at room temperature. Previously, the actuator portion of the model was thoroughly validated, while the entire model was only validated for room temperature and static clutch conditions (without torque generation). Experimental characterization of three Friction material characteristics is given in chapter five: (i) the coefficient of friction, (ii) tendency towards shudder and (iii) wear. Experiments are conducted on a small friction material sample using a CNC pin-on-disc type tribometer machine. The coefficient of friction is characterized as a function of three operating parameters: (i) temperature, (ii) slip speed and (iii) normal force. The results of the experiments are used to calculate the coefficient of friction vs. slip speed gradient which is typically used as a measurement of tendency towards shudder. Friction material wear is modeled using the Archard’s law for which it can be shown that it is possible to reformulate the said law into a form which states that the worn volume depends on (i) temperature dependent parameter named wear rate and (ii) dissipated energy. Thus, experimental characterization of friction material wear was conducted by determining the wear rate values over a targeted span of temperatures. In-between the wear experiments, the coefficient of friction experiments were conducted. Thus the influence of wear depth on the coefficient of friction and tendency towards shudder was also investigated. Additionally, friction material tendency towards shudder was investigated on the entire clutch. The correlation of the tendency towards shudder results on both test rigs indicate that the results obtained on a small friction material sample can be used to predict the behavior of the clutch with respect to tendency towards shudder, although additional experiments are required to confirm/improve this capability. Clutch thermal dynamics modeling is presented in chapter six. The goal of this model is the prediction of clutch components temperatures from other operational parameters. It is necessary to know these temperatures due to their influence on thermal expansion and coefficient of friction, however, their measurement during regular clutch exploitation is technically problematic. Model is based on concentrated parameters, i.e. the clutch is described as a series of concentrated thermal masses with thermal fluxes appearing in-between them. Model validation has shown that the model error is below 10% of full clutch temperature range. Chapter seven presents the results of research related to Clutch thermal expansion of clutch components and its influence on the clutch axial dynamic. Initial modeling was conducted by a simplified analysis of individual components thermal expansion, i.e. initially the thermal expansion was only treated as a change of components position for two extreme cases: (i) fully open and (ii) fully locked clutch. Based on acquired results, the expansion of basic axial dynamic model was carried out by individual modeling of the decrease of clearance and the increase of force in the return spring. Model validation shows that the model describes the thermal expansion qualitatively well although significant possibilities for quantitative improvement exist, especially in the low torque region. Chapter eight presents the effort related to experimental characterization and modeling of Clutch wear related effects. Wear influences the clutch axial dynamic through two separate effects: (i) by decreasing the friction plate thickness due to loss of friction material and (ii) through the activation of the wear compensation mechanism which periodically compensates the effects of friction material wear (i.e. reduces the increased clearance). Initial experimental characterization was conducted by direct wear of the clutch using the clutch setup. The clutch was worn from approximately non-worn condition up to first mechanism activation. The results show that the torque characteristic falls as the friction material is worn (due to loss of normal force caused by the increase of clearance) and then is (approximately) restored after the wear compensation mechanism activation. In order to determine the influences of wear compensation mechanism activation, static characteristics were recorded at the initial mechanism position and at the position the mechanism should achieve after 10 and 20 activations (relates to fully worn clutch). Wear was emulated with machining, i.e. by mechanical removal of material in order to avoid lengthy clutch wear. The results show that the wear compensation mechanism activation influences the axial dynamics by lowering the torque characteristics. This can be explained with the increase of force in return springs due to change of press plate initial position. Clutch wear modeling was conducted through separate modeling of the two described effects. The friction plate loss of thickness was modeled as the increase of clearance where worn volume, thus also the reduction of friction plate thickness calculates from dissipated energy, temperature and wear rate. Wear compensation mechanism activation was modeled as a change in press plate initial position (i.e. reduction of clearance) which is activated each time the increase in clearance reaches targeted value. Model validation shows that the model describes the clutch wear qualitatively well although significant possibilities for quantitative improvement exist. The experimental characterization of the Clutch components spring characteristics is shown in chapter nine. The experiments were conducted using the manually powered spindle based test rig. Additionally, two components (one with soft and with steep characteristics) were also characterized using universal testing machine. The correlation showed that the press test rig is sufficiently precise for the intended application. Five different characteristics were obtained, namely the characteristics of (i) friction plate woven spring, (ii) the diaphragm spring, (iii) the return leaf type spring, (iv) the diaphragm spring lever support points, and (v) the flywheel bearing pack. The tenth chapter presents the new axial dynamics model which is developed with the goal of a further improvement of precision of torque prediction as well as the positions of individual clutch components. The validation of the new model show increased precision related to modeling of individual effects with respect to the existing model. Final, eleventh chapter presents the main Conclusions which can be made based on the results of the research presented herein including the summary of all results. Possible future research is shortly discussed.