TY - BOOK

T1 - A Methodology for Dynamic Belt Simulation

AU - Fimbinger, Eric

N1 - no embargo

PY - 2021

Y1 - 2021

N2 - This thesis presents a methodology that enables the modelling and simulation of dynamically interacting belt models in Discrete Element Method (DEM) simulations. In conventional DEM simulation setups, belt models are typically modelled as rigid surfaces, which are further applied with a contact model that induces a movement into bulk material particles that contact with these belt-representing surfaces. Accordingly, such rigid belt models are not able to depict dynamic interactions - neither with bulk material particles conveyed on the belt nor with system components that further interact with the belt, such as idlers and pulleys. Particularly for the numerical simulation of conveyor systems comprising belts that significantly influence system characteristics due to dynamic belt behaviour, however, the consideration of belts as dynamically interacting objects is required. Such systems are, for example, sandwich or pipe conveyors. Furthermore, such dynamically interacting belt models are also required for simulations in which the analysis of particular effects are of interest, such as belt deformation/deflection effects. The belt simulation methodology developed and presented in this thesis explicitly addressed the simulation of such complex systems where dynamic belt behaviour is inevitable. This methodology is generally based on using a bonded-particle belt model (BP belt), which is furthermore initialised with a specific geometrical shape, more specifically relating to belt initialisation in almost-final state. These two areas - the general setup of a BP belt and its initialisation in almost-final state - form the major parts of the methodology. The general setup of a BP belt is defined fundamentally to show a single layer of rectangularly arranged/bonded cuboidal particles. The bondings that connect those belt particles are further defined by an enhanced bonding model, which is explicitly extended to enable the representation of belt-typical flexibility characteristics. The essential method with which such a BP belt is initialised is introduced as belt initialisation in almost-final state. This method comprises the computation of a BP belt with a specific complex shape approximating an assembled belt within a specific belt system. Therefore, an algorithm was developed that enables the conversion of such a given belt geometry, provided as a CAD model, into a corresponding BP belt. This developed conversion algorithm was also implemented into a software tool (BeltConverter), allowing convenient use via a GUI. As a further enhancing feature, initial belt velocity can be applied to the converted BP belt, thus allowing the initialisation of an already running BP belt. Especially noteworthy in terms of using this initialisation principle is the significantly reduced pre-simulation effort required for assembling such a belt model, which is generally reduced to imperceptible levels. Illustrations of applying the presented methodology on several different exemplary industry-relating applications highlight the various benefits of the methodology, such as in terms of computational efforts required, and ultimately reveal the methodology's favourable suitability for DEM simulations comprising dynamically interacting belt models.

AB - This thesis presents a methodology that enables the modelling and simulation of dynamically interacting belt models in Discrete Element Method (DEM) simulations. In conventional DEM simulation setups, belt models are typically modelled as rigid surfaces, which are further applied with a contact model that induces a movement into bulk material particles that contact with these belt-representing surfaces. Accordingly, such rigid belt models are not able to depict dynamic interactions - neither with bulk material particles conveyed on the belt nor with system components that further interact with the belt, such as idlers and pulleys. Particularly for the numerical simulation of conveyor systems comprising belts that significantly influence system characteristics due to dynamic belt behaviour, however, the consideration of belts as dynamically interacting objects is required. Such systems are, for example, sandwich or pipe conveyors. Furthermore, such dynamically interacting belt models are also required for simulations in which the analysis of particular effects are of interest, such as belt deformation/deflection effects. The belt simulation methodology developed and presented in this thesis explicitly addressed the simulation of such complex systems where dynamic belt behaviour is inevitable. This methodology is generally based on using a bonded-particle belt model (BP belt), which is furthermore initialised with a specific geometrical shape, more specifically relating to belt initialisation in almost-final state. These two areas - the general setup of a BP belt and its initialisation in almost-final state - form the major parts of the methodology. The general setup of a BP belt is defined fundamentally to show a single layer of rectangularly arranged/bonded cuboidal particles. The bondings that connect those belt particles are further defined by an enhanced bonding model, which is explicitly extended to enable the representation of belt-typical flexibility characteristics. The essential method with which such a BP belt is initialised is introduced as belt initialisation in almost-final state. This method comprises the computation of a BP belt with a specific complex shape approximating an assembled belt within a specific belt system. Therefore, an algorithm was developed that enables the conversion of such a given belt geometry, provided as a CAD model, into a corresponding BP belt. This developed conversion algorithm was also implemented into a software tool (BeltConverter), allowing convenient use via a GUI. As a further enhancing feature, initial belt velocity can be applied to the converted BP belt, thus allowing the initialisation of an already running BP belt. Especially noteworthy in terms of using this initialisation principle is the significantly reduced pre-simulation effort required for assembling such a belt model, which is generally reduced to imperceptible levels. Illustrations of applying the presented methodology on several different exemplary industry-relating applications highlight the various benefits of the methodology, such as in terms of computational efforts required, and ultimately reveal the methodology's favourable suitability for DEM simulations comprising dynamically interacting belt models.

KW - belt modelling

KW - belt simulation

KW - belt initialisation

KW - Discrete Element Method

KW - DEM

KW - bonded particle

KW - almost-final state

KW - dynamic belt model

KW - deformable model

KW - flexible model

KW - virtual prototyping

KW - belt system development

KW - 3D simulation

KW - Gurtmodellierung

KW - Gurtsimulation

KW - Gurtinitialisierung

KW - Diskrete Elemente Methode

KW - DEM

KW - bonded-particle

KW - einbauzustandsnahe Form

KW - dynamisches Gurtmodell

KW - verformbares Modell

KW - flexibles Modell

KW - virtueller Prototyp

KW - Gurtsystementwicklung

KW - 3D-Simulation

U2 - 10.34901/MUL.PUB.2021.3

DO - 10.34901/MUL.PUB.2021.3

M3 - Doctoral Thesis

ER -