Home Research
Leonardo De Novellis - Research Interests
His research is focused on the theoretical modelling of the CVT transmission and on the experimental investigation of the mechanical behaviour of the CVT Variator. The experimental activity is an important topic of his research since it serves to validate the results obtained with the theoretical model.
  • CVT transmission test rig
  • Chain link force measurement
  • CVT chain modelling

    CVT transmission test rig

    The test rig installed at the DIMeG is a torque-controlled rig, since it is possible to fix the load torque by means of an electromagnetic brake: a three-phase asynchronous 4 poles and 30 kW of power Siemens motor determines the rotation of the variator; the electromagnetic brake is used to control the resistant torque and is located on a countershaft connected to an output shaft through a synchronous belt transmission. An hydraulic circuit allows to regulate the oil pressure acting on the variator moveable half pulleys and therefore to change the gear ratio, and to lubricate belt and bearings as well. The rig is equipped with pressure, torque and speed sensors for the acquisition of main operative parameters. The moveable sheave axial position of the driven pulley is measured by means of a linear displacement sensor. In order to perform more accurate measurements, it has been developed a control strategy of the variator transmission ratio which is based on a Cascade approach: the ratio control is obtained by regulating the oil pressure on the driven side, for a fixed clamping force acting on the drive pulley. The control system has been modelled, designed and verified relatively to the system stability and global performance. The theoretical basis is the CMM model for shifting dynamics, proportional-integral controllers have been chosen in order to provide the system a feedback control loop.

    CVT test rig

    The CVT test rig

    Chain link force measurement

    We have developed an ad hoc measuring device constituted by a datalogger which is fixed at the chain and moves with it. The datalogger has been designed in such a way to be light and small, in order not to strongly influence the distribution of the chain forces. The electronic device records the strain data from a strain gauge and stores them in a flash memory card. The data can be easily downloaded through a serial connection after the test. In this way, we have been able to measure the tensile force acting on a single chain link in a wide range of working conditions. Our measurements have shown that a non-perfectly linearity (which has a clear theoretical explanation) exists between the clamping force and the link tension distribution. We also have found that the link tension distribution is less sensitive to the torque load, which mainly influences the local slip between the chain and the pulley and, hence, the time required by the link to cover the entire contact arc. We have also carried out a comparison between the experimental data and the theoretical predictions which have been obtained employing the CMM model and a FEM model of the chain: we have found a relatively good agreement which confirms the validity of the theoretical approach.

    The datalogger and the battery

    The datalogger and the battery

    The link tensile force distribution. Measured data are reported in blue color, theoretical prediction in red color. Data are shown for speed ratio τid = 1, primary clamping force SDR = 12kN and torque load TDN = 70Nm.

    CVT chain modelling

    The theoretical modelling of CVT variator is of great interest under several aspects, which go from designing and optimization of control strategies to the search for improvements in the mechanical design. Recently a very promising model, the so called CMM model (from the name of the authors: Carbone, Mangialardi, Mantriota), has been proposed to describe belt and chain CVT dynamics. The model points out the influence of the pulley deformation on the mechanical behaviour of the CVT variator; furthermore, the model has been refined by including a more accurate description of the pulley flexibility effect, thus achieving a better agreement with the experimental results, related to the evaluation of the clamping forces necessary to achieve the desired transmission ratio and to the variator traction capabilities. Besides to the continuous approach that has been followed with the CMM model, we have also developed a multibody model of a pin-jointed chain and carried out simulations of the system in some running conditions. Furthermore, we have compared the results of the aforementioned simulations with the outcomes of the CMM model for the same working points, in order to show similarities and/or differences in the evaluation of the main physical quantities that are involved in the mechanics of a CVT variator. Our research has shown a good agreement between the two considered models in terms of prediction of the variator performance, however we point out the large computational time required for the evaluation of one steady state working point in the case of the multibody model, if compared to the CMM model, which makes the discrete approach not immediately exploitable for practical applications.

    The tension force F calculated with the multibody model during one revolution at gear ratio τid = 1, load torque TDN = 50Nm and primary speed ωDR = 50rad/s

    The tension force F calculated with the CMM model during one revolution at gear ratio τid = 1, load torque TDN = 50Nm and primary speed ωDR = 50rad/s

Leonardo De Novellis


Room: +39 080 596 2817

Mobile: +39 320 149 7255

Fax: +39 080 596 2777

Email:  l.denovellis@poliba.it



Copyright © 2010 --- TriboLAB
DIMEG - Politecnico di Bari,  Viale Japigia, 182 - 70126 BARI (Italy)

Tel. +39 080 596 2746, Fax +39 080 596 2746/2777

Contact us at tribolab.poliba@gmail.com