Optimized architecture manufacturing systems using witness models and simulation.

Minciu, Constantin ; Gandila, Sanda ; Anghel, Florina 等

1. INTRODUCTION

Witness contains a series of elements for discrete manufacturing (assembly marks) but in the mean time one can do a continuous systems modelling.

One can specify variables and attributes and the order of operations can be based on times, priorities or other criteria defined by the user. Simulated actions, from beginning and ending of simulated events, can be used for programming structures type like: For-Next, While-End, Go to-Label and use mathematics and logical operators.

It also has an animation module build when defining a model. A characteristic for Witness is the ability of stocking an unlimited number of variables and attributes.

Witness Optimizer module was projected in order to significantly reduce the period of time necessary for automatic search of the optimum solution.

The number of the modeled parameters that will be optimized is big and it includes: manufacturing times, operator's number, the vehicles speed, etc.

[FIGURE 1 OMITTED]

Witness Optimizer is not only looking for the optimum solution but it also makes a rapid analyze. By utilizing intelligence algorithms that measure and react to every assay, the best solution is usually found after trying at least 10% of all possible combinations. Witness Optimizer users can make a rigorous analyses for different production sceneries in a flexible system production. Below we give some of the most important advantages of the program:

* An intelligent help in production optimization;

* Fast confirmation of good solutions;

* Bigger amount of time for making more creative solutions;

* Confirmation that the solution found is solid, statistically speaking.

Witness Optimizer gives graphic or table results.

POWERPAQ is another module of simulation language Witness. It has functions that allow an automatic simulation by defining a simulation category of parameters which will be subjected to supplementary tests determining the optimum set of characteristics. Using more optimization techniques one can identify blockages and sensible point in the system, test options and develop more efficient alternatives.

2. MODELLING THE MATERIAL FLOW

When modeling a Witness discrete value material flow managing system we studied this system's behavior for a period of 240 hours functioning, with 3 shifts of 8 hours daily, one operator for each shift with monitoring role.

[FIGURE 2 OMITTED]

The development for a non-linear model process of a production system containing a tool store or not was made by giving information of medium value behavior based on standard operation curve.

Received commands are particularized by receiving moment and quantity. Both are modeled as random values with a given rate and a standard deviation. System's working performance is determined by planned capacity as an external entrance and MTBF (mean time between failure) as well as MTR (mean time to repair) taken from practical knowledge and observation of the production process Exit values are resolved command and unfinished production for the working system. After doing a calibration with a set of parameters the model behavior on a long period of time is similar to practical working systems behavior described thru operating curve. These parameters were determined thru simulation analyses (Bley, 1999). In practice, especially in lot production, real curve object oriented are different from operating curves based on medium values by 5-7 %. The model is composed from four partial models: the reference model is made of two partial models, information model and objects model. The information describing models is taken by method and activities models; these last two partial models form the reference model frame. The activities from the activities model are attributed to the individual phases. Here we can find data given by the simulations from CATIA V5R16 (Lee, 1999) and Deform (Patrascu, 2004).

Discrete material flow simulation from the studied system can be done in real time in Witness, in accelerated time (if we want to visualize in time behavior) or slowed down time (if we want to see a detailed actions that causes a critical moment during function). As earlier mentioned this is an accelerated simulation, lasts 240 hours, which aims for a medium time optimization of these flow's functioning. In figure 1 we can see the system's model after 64 hours of functioning. In figure 2 is given an example of a function report for c4 conveyor after working 161 hours. The meaning of the references from the functioning reports is as follows:

blocked = the system is blocked

broken down = = the system is interrupted

busy = the system is in function

busy/fill/empty = the system is in function / full / empty

empty = the system is empty

idle = the system is inactive

moving freely = the system is moving without constrains

setup = the system is being setup

wait repair labour = the system is waiting to be repaired

wait setup labour = the system is waiting to be setup

waiting parts = the system is waiting for parts

Based on the Witness reports and on calculus algorithms we determined that c4 conveyor is a material flow concentrator.

3. CONCLUSION

In this paper we illustrate a manufacturing architecture optimizing algorithm by a case study used in our research grants. The case study is basically a concentrated manufacturing system (Cotet & al., 2007) where we increase the productivity by eliminating the flow concentrator. As one can see in figure 3 we can do that if we double c4 conveyor capacity and we obtain a 20% increasing of the manufactured parts number for the studied period of time. In conclusion based on the data given by the process simulation through flow simulation we can increase the productivity with 20% for the studied system.

4. REFERENCES

Bley H. & Wuttke C.C.(1999), "Multiple Use of Simulation Models for Production Systems", Institute of Production Engineering / CAM, University of Saarland--Germany, 1999.

Cotet, C. E., Dragoi, G. & Carutasu, G. (2007). Material Flow & Process Synchronous Simulation In Concentrate Manufacturing Systems, Annals of DAAAM for 2007 & Proceedings of The 18th International DAAAM SYMPOSIUM, "Intelligent Manufacturing & Automation: Focus on Creativity, Responsibility and Ethics of Engineers", Katalinic, B. (Ed.), pp. 180-181, ISSN 1726-9679, ISBN 3-901509-58-5, Zadar, Croatia, October 2007, Publisher DAAAM International Vienna 2007.

Cotet, C. E., Dragoi, G. & Abaza, B., F. (2007). Multipolar

synchronous material flow & process simulation in diffused manufacturing systems, Annals of DAAAM for 2007 & Proceedings of The 18th International DAAAM SYMPOSIUM, "Intelligent Manufacturing & Automation: Focus on Creativity, Responsibility and Ethics of Engineers", Katalinic, B. (Ed.), pp. 179-180, ISSN 1726-9679, ISBN 3-901509-58-5, Zadar, Croatia, October 2007, Publisher DAAAM International Vienna 2007.

Lee, K. (1999). Principles of CAD/CAM/CAE Systems, Addison Wesley Longman, Inc., ISBN 0-201-38036-6, USA.

Patrascu, G. (2004). 3D Simulation of Turning Process using FEM Software, Proceedings of the International Conference on Manufacturing Systems ICMaS 2004, Constantin, I., Ghionea, A., Constantin, G. (Ed.), pp. 297-300, ISBN 973-27-1102-7, Bucharest, 2004 October, Editura Academiei Romane, Bucharest

Fig 3. Usage degree of the system.

Lathe with tool storage

Empty 11%
bloked 4%
busy 85%

c4 conveyor belt

Empty 2
bloked 24
busy 72

Note: Table made from pie chart.