Matlab Virtual Reality Simulations for optimizations and rapid prototyping of flexible lines systems

Transcription

1 Matlab Virtual Reality Simulations for otimizations and raid rototying of flexible lines systems VAMVU PETRE, BARBU CAMELIA, POP MARIA Deartment of Automation, Comuters, Electrical Engineering and Energetics University of Petrosani Str. Universitatii, Nr.0, Petrosani, Hunedoara ROMANIA Abstract: - In this aer a new aroach for simulation of Comuter Integrated Manufacturing Systems is considered and develoed in order to deliver raid rototying and to otimize the rocess. This aroach is based on the ossibilities of the Matlab to simulate comlex mathematical models and also to integrate simulation rocedures into design using the virtual reality environment. In this way we can fully simulate mathematical and visual comlex systems like Comuter Integrated Manufacturing solutions. This aroach will allow better and faster design of automation solutions. Key-Words: - robot mathematical model, CIM modeling, virtual simulation, comuter integrating manufacturing, flexible lines systems, raid rototying Introduction Due to increasing demand on fast delivery of comonents, technologies, manufacturing lants and facilities, new methods for rototying and develoment have to be investigated and imlemented in order to satisfy the time, recision, cost and rofitability of the new systems and roducts. One solution would be the introduction of a hase of visual simulation between the design and imlementation []. The roosed solution resulted from the design can be validated by a simulation in a visual environment, which will roduce a refined design solution where all the comonents will interact erfect []. This simulation can verify the mathematic models of different comonents and also offer an accurate visual reresentation of the interaction between all of the comonents of the roosed installation or lant. This method of a visual reresentation of different elements and the entire rocess in a real life like environment can roduce early modifications and changes in the structure of the installation to roduce otimal results []. Also this method can validate the individual elements choose for the imlementation from the oint of view of dimensional arameters, seed arameters, matching sizes and others. Problem Formulation Let s consider a CIM automation solution having robots, conveyers, a visual insection system and a machining unit used for flexible manufacturing. This automation solution is intended to roduce different roducts and have to comly with an increased degree of flexibility. For decreasing the costs of the overall solution, the idea is to use comonents with the best rice/erformance ratio. That means utilization of comonents from different roducers and the integration of them into a unified system. The comonents of the automation solution are a degree of mobility robot arm, a SCARA tye robot arm, a Cartesian robot, conveyers, rogrammable logic controllers, inverters, three hase squirrel cage AC motors and a video reconnaissance system. The degree of mobility robot arm will satisfy a dual role. The first task will be to rovide raw ieces for the machining unit from the magazine and also to retrieve the rocessed ieces from the machining unit and transferring them on the rocessed ieces conveyer belt []. The secondary role, and also otional, will be to feed the first conveyer belt with ieces destined for an otional assembly hase with the ieces rocessed by the machining unit. The SCARA robot will erform the task of transferring or/and assembly the ieces from the first conveyer belt to the machined ieces conveyer belt. The assembly oeration will take lace on a reinforced art of the second conveyer belt secially designed for that urose. The SCARA robot is connected to a visual reconnaissance system which erforms two functions. The first function is to determine the osition on X and Y axes and also the ISBN:

2 rotational vector of the iece on the conveyer belt so the robot arm can reach and retrieve the iece in the grier in a manner that will facilitate assembly oeration. The second function is to verify the dimensional arameters of the ieces so mismatching events can not aear in the assembly hase. The Cartesian robot will erform the function of alletizing the finalized ieces or assemblies, retrieving the rocessed ieces or finalized assemblies from the assembly station and disosing them to the receiving allet. The machining unit is custom made from three Fli X series axes coordinated by the motion controller through the secial Fli X series controller RDX. The cutting tool is a PAC mill attached to the vertical axis of the machining unit. The main roblem of this heterogeneous system is the comatibility between the different arts of the systems. To solve those roblems from the rototying stage we can create the virtual environment models of each and every art and then imlement those virtual models into a simulations environment. Choosing the solution to use the Matlab Virtual Environment we can simulate not only the hysical interactions between different arts but also the interactions between mathematical models that describes the function erformed by each art. The virtual model of the CIM installation can comrise the mathematical models for each of the individual comonents but is not mandatory. If the mathematical models of the comonents are included in the virtual model it is ossible to determine the minimum requirements for the comonents that have to be satisfied, in order to achieve the best erformance of the entire system [] cos sin( cos cos( sin( cos( cos( sin( ) d ) cos d Problem Solution The first ste in the simulation rocess is to describe the structure and the models of the elements which will be simulated. The following ste is to realize the virtual model for each of the elements and afterwards to integrate all the discreet virtual models into a unified virtual reality model. This final model will comrise all the elements of the CIM solution and will be used to validate the ositions of different equiments, to validate the workflow of the installation, to determine the seed requirements of different elements or even to study the solution for a new roject imlementation.. Simulation of the robotic comonents First was determined the kinematic diagram based on that the Hartenberg-Denavit method and the movement matrix. After the mathematical model was comleted the virtual rototye was constructed and the mathematical models were integrated in the simulation environment. First robot to be simulated with six degree of mobility is the articulated robotic arm. Based on the rocedures resented above the movement matrix was constructed. n T 0, = n n 0 d ) cos cos( ) o o o a a a () The elements of the matrix were comuted and the ositions comonents for X axis, Y axis and Z axis, 0, 0, 0, are resented bellow. d ) d ) cos( ) d ) cos d Starting from this data we build the virtual model of the robot which resents the same movement ossibilities and the same dimensions like the real version. The building of the virtual model is a for stes rocess. The first ste is to build the individual d ) sin( d ) d ) cos d d ) d d ) d ) () () () comonents which comrise the final object. The construction of the elements can be done in a secialized dedicated environment that can be found in the Matlab ackage and named v-realm or can be done using general urose D design software. ISBN:

3 The second ste in building the virtual model is to set the movement hierarchy structure of the elements. Multile elements can be connected together in structure where the movement of the nth element is determined by the osition of the (n- )-th element in the kinematic chain and his own movement arameter. The third ste in creating the virtual model is to create the scene of the modelled assembly. The scene has to be also constructed in the v-realm environment, and in order to assure the accuracy of the simulation, all the elements have to be laced to secific coordinates of the scene referenced to world system coordinates of the v-realm environment. The result of this third ste is a virtual rototye, which has the same arameters and movement ossibilities like the real one but lacks the interface connections with Matlab Simulink and also lacks the mathematical model that interact with Matlab environment. The forth ste in the simulation rocess is to link the Matlab environment to the simulation scene. The link is done through the visual sink element find in the Matlab Virtual Reality ToolBox. This element connects the Matlab environment and the virtual scene created by the v-realm software. Also this element selects the arameters of the scene that will undertake modifications from the Matlab variables. The variables used to control the virtual model can have redefined value or can be real time editable. Another otional ste can be added to the simulation rocess. This ste imlies utilizing a scrit to run the simulation fully automated in order to observe the entire rocess of a comlex installation during a comlete functional cycle. The scrit can be constructed by using a multidimensional Matlab variable which contained the movement values for all the arameters to secific moments in time. Other imortant comonent of the CIM system is the SCARA Yamaha robot. This tye of robot is often utilize in the ick and lace tye of alications. The advantage of the SCARA tye robots is due to their unique arrangement of their joints which allow comlex movements and very high seeds. Having the information for the SCARA robot we build the mathematical model and then the virtual reality model in the Matlab Simulink using the vrealm tool for virtual reality scene. The simulation scene once constructed is linked to the Matlab Simulink environment with VR-Sink tool, where the arameters of the scene which will be modified are chosen. Adding the corresonding rotational and translation movement vectors to the model will comlete the structural art of the design. Also, to the model have to be added elements that will ermit user to control the model and to simulate manual control. The comlete simulation scene for the SCARA robot is resented in Fig.. The same rocess is used to build and integrate the virtual models of the Cartesian robot and Fig. SCARA robot virtual reality real time simulation machining unit into the virtual simulation environment of Matlab Simulink. ISBN:

4 . System Simulation After all the comonents of the CIM lant are individual simulated and the virtual reality model for each is constructed, the next ste imlies the connection rocess between all those comonents. First the D ositions for each comonent have to be established. The coordinates have to be secified for each comonent of the virtual reality model. Due to hierarchical system that was used reviously the coordinates have to be secified when the virtual reality scene is constructed just for the first element in the hierarchical chain. The coordinates used are absolute coordinates of the virtual world. These coordinates lay a role in setting the virtual reality scene by they do not influence the simulation rocess which is done using simulation vectors for each of the elements. The virtual reality simulation scene was constructed using the Matlab Simulink tool v-realm. The simulation scene is intended to be used in dual mode. The first method is to manually maniulate the elements using altering arameter devices that Matlab Simulink offers. This method give the user the ossibility to modify the arameters so can verify the behavior of each comonent alone decouled from the system. This method will also rovide the ossibility for the user to study different solution for movement of different arts or comonents of the CIM installation. The second way that the scene was intended to be used is to run following a scrit. This second method alied to the virtual reality model offer the ossibility to study very comlex Fig. Matlab Simulink to level simulation design oerations that haened on the workflow of the system. With the ossibility to visually exerience the comlex workflow of the manufacturing system the designer can imrove the design of different arts of the lant before the system is build. Other benefit that occurs from this second method of virtual reality interaction is that the timing of different comonents and rocesses can be verified, corrected and otimized before the construction hase. ISBN:

5 Because the manufacturing rocess is described like a chain of events that roduce at the end a roduct, we can say that the manufacturing rocess is as strong as his weakest link That means that if we want to verify if a manufacturing rocess can run at a certain seed, due to this method, the weakest link, slowest link, can be identified and relaced so the entire manufacturing system will roduce better and faster from the first day. In order to run the virtual reality simulation scene in an automated scrit we have to create a time driving comonent that will comrise the status of all the arameters of the lant at different moments in time. That comonent will rovide for the virtual reality simulation scene the events that roduce the reaction of the environment at certain moments in time. The virtual reality model will react by modifying his arameters, calculating the mathematical exressions, and animating the scene accordingly. This comonent will be a multidimensional Matlab worksace variable that will load the values from a linked Matlab file. The multidimensional variable will comrise on the first column the time vector. For each osition of the time vector a value it set or calculated for the same osition in the following column vectors reresenting the arameters that will be modified in the simulation rocess. Fig. Comlete CIM system autonomous model simulation vs. real installation The simulation model links all the arameters of the virtual reality simulation scene with the multidimensional variable. The routing element is disatching vectors values to each selected comonent. Each vector from the multidimensional variable is referenced to his unique simulation arameter. The values from the multidimensional variable are transmitted by the routing element to the to level control blocks as can be seen in Fig.. Each of those blocks has inside a model which ermits switching from manual to automated simulation mode for each of the comonent. These blocks deliver the control signals accordingly to the models inside to the Matlab Simulink linking tool which offer the connection between Matlab environment and the visual scene created with vrealm tool. Also these blocks rocess the inut value and if that values exceed the maximum allowed numbers the values are limited to maximal or minimal values. This is done in order to revent wrong values to erform actions on the virtual reality simulation scene. In order to successfully animate the simulation model we have created a multidimensional Matlab worksace variable. This variable is linked to a.mat extension file which comrises all the data. After the multidimensional variable is created and saved, it will be necessary to be loaded into the worksace of the Matlab environment using the aroriate tools. The simulation window in running mode for the entire CIM system is deicted below in Fig.. The simulation is running comletely autonomous based on the data rovided by the multidimensional variable. Conclusion The Matlab virtual reality simulation used for the CIM installations can rovide multile advantages from the design stage to the secondary revision or refitting. ISBN:

6 A main advantage is reresented by the ability to test how different comonents match in terms of dimensional and seed arameters before the comlete design rocess is finished and before the comonents are bought. Another advantage rovided by the virtual reality simulation using Matlab environment is the fact that mathematical models are not mandatory to develo the simulation scene, making the utilizations of the alication much user friendly. Using this system to model and simulate the function and the functioning of the CIM lant we can make verifications in terms of dimensional arameters, seed arameters, work flow, visual confirmation, before the building rocess had started. This rocess of virtual reality simulation for a CIM lant can make the installation run at otimal arameters from the first time when the installation starts. The future develoment imlies using feedback information from the real installation and to monitor the functioning of the lant using the virtual reality model. Further develoment is to combine the virtual reality model, Matlab environment and custom translator software rograms in order to simulate and afterwards build the rograms for different comonents in an automated fashion. For further research the Matlab virtual reality model of the CIM is connected with the D-Sace simulation and rocessing board in order to otimize the functionality and efficiency of the installation. References: [] Fawaz K., Merzouki R., Ould-Bouamama B., Model based real time monitoring for collision detection of an industrial robot, Elsevier Mechatronics journal, Vol. 9, 009,.9-70 [] Bi Z.M, Wang L., Advances in D data acquisition and rocessing for industrial alications, Robotics and Comuter- Integrated Manufacturing, Vol., 00,.0- [] Tomiyama T., Gu P., Jin Y., Lutters D., Kind Ch., Kimura F., Design methodologies: Industrial and educational alications, CIRP Annals Manufacturing Technology, Vol 8, 009, - [] Stark R., Israel J.H., Wöhler T., Towards hybrid modeling environments Merging deskto- CAD and virtual reality-technologies, CIRP Annals Manufacturing Technology, Vol 9, 00, 79-8 [] Ha S., Kim L., Park S., Jun C., Rho H., Virtual rototying enhanced by a hatic interface, CIRP Annals Manufacturing Technology, Vol 8, 009, -8 [] Leba M., Po E., New Distributions Proerties and Alications in Digital Control, Robotics and Automation, Proceedings of the 7 th WSEAS International Conference on Signal Processing, Robotics and Automation (ISPRA 08), University of Cambridge, 008,. -8 ISBN: