Using HLA for factory simulation

Starting with a short analysis of the current situation in the field of factory simulation and an overview of current tendencies in the manufacturing area, the article introduces the work carried out in order to integrate HLA and existing simulation tools. It presents the simulation tool SLX and the visualization tool Skopeo, which were both utilized to perform a prototype federation of a manufacturing plant. The article is concluded by looking at further chances for HLA to also become a standard in the field of factory simulation. 1 Current situation in factory simulation Simulation has become an important tool for planning the complex process structure of a company's business. Whereas during the last few years simulation was mainly focused on a few key activities of a company’s overall business it recently became more important to consequently involve simulation in all parts of a factory’s life cycle. In the design stage, simulation can be used to determine capacities of certain production facilities as well as to estimate efficiency and throughput of the future factory. The same simulation model might be used to evaluate different control strategies in subsequent stages. Also, the simulation model can support the training of personnel before the factory is even finished. During the last few decades a variety of simulation tools has been developed. Most of them include model elements specialized in a certain area of application. Each simulation system usually has its own paradigm of how to map the real world's objects into the simulation model. Each product comes with its own version of a model editor and has its own way of presenting the simulation results ranging from simple text-based output to sophisticated three-dimensional graphical scenes. Nevertheless, the great amount of time spent on acquiring input data, programming the simulation model, and the considerable effort of adequately visualizing the simulation results have been often criticized in the past. They conflict with current tendencies in the manufacturing industry. It is therefore necessary to research new concepts of factory simulation. 2 Tendencies within manufacturing The shortening of a product's life cycle, the rising variety of customized products, as well as increased competition in the market force companies to transfer new technologies into production more rapidly in order to gain advantages over their competitors. This requires simulation models not only to be created cheaper and with less time but also to be easier to adapt. These improvements can be accomplished by increasing the reusability of simulation models, or at least parts of it. Often plants of a certain manufacturing branch consists of similar components. For example, there are several car manufacturing plants in Germany but only very few manufacturer of paint shop equipment. Therefore, the same hardware might even be installed in factories that produce cars of different brands. Suppliers of such equipment could take advantage of that fact and could already offer simulation models for their components. These models could be used by the customers to be included in their overall factory simulation model. That opens up a new market for the supplier and is also an advantage for the customer since it decreases development time for simulations. Another tendency in the manufacturing industry is the globalization of the market. Complex products are very seldom manufactured by only one company. Usually components are produced by different companies and assembled at a different location. Capacities of warehouses are reduced and often parts are delivered "just-intime" as needed, which increases the dependency between companies. Therefore, it will become more important in the future for a simulation study to also include the simulation models of the company’s suppliers. This raises the need for interoperability of simulation models to solve logistical problems. Interoperability is also important because it increases flexibility of the simulation model. In case of an alteration of co-operation between companies, it is easier to adapt the simulation model to the new circumstances. If all simulation models adhere to the same standard, the component that simulates the old partner can be replaced by a new one. Interoperability is especially important because different industry partners use different simulation tools according to which tool is most feasible for their purpose. Reusability of their models can only be achieved by setting up a standardized interface through which interoperability is granted. Due to the increased graphical capabilities of today’s computers, there is also a new field of application for computer simulation: user training. Although many complex devices are computer-controlled, some of them still need to be supervised by humans. Incorrect operation can often cause substantial damage. For this reason, simulation tools are already in use, for instance, to qualify railroad engineers to operate Germany’s high speed trains or for training drivers to operate harvesting machines. These tools, however, have the restriction that the teacher and trainee have to physically be at the same location. Also, collaborative work is not simulated adequately. Here, distributed simulation will become more important. A very powerful application of distributed simulation is also the training of personnel for coordinating control stations within a factory. For training purposes, the trainee might use a one-by-one replica of the control panel whereas the real machines might be transparently replaced with a simulation model running on a different computer. Once the trainee is experienced enough, he/she will not actually control a simulation model but instead the real machine. This provides a high level of reality within the simulation. To conclude the section above, it can be said that the current problems that simulation faces can be improved by consequently applying the concepts of reusability, interoperability, and distributed simulation to factory simulation. Hence, HLA has a promising basis to become an accepted standard within the field of factory simulation, too. 3 Towards an HLA based factory simulation As mentioned in the first section, most factory simulation developers prefer to use specialized simulation tools rather than programming languages like C++. By being convinced to move to C++, developers could use the advantages of HLA immediately, but it would also imply the loss of an often highly specialized simulation and modeling environment. In most cases, this is not acceptable, therefore, propagation of HLA within the field of factory simulation will only be successful, if there are appropriate simulation tools available. Currently, this is not the case. For this reason, research work focused on opportunities for the integration of classical stand-alone simulation tools into the High Level Architecture. In general, adapting a simulation tool to be HLA-compliant will require modifications of the source code as well as inside knowledge of the software. In particular, it is necessary to have access to the internal event list and detailed knowledge of the internal representation of data types within the simulation tool. For this reason, it can only be accomplished through a close co-operation with the software provider. The simulation tool SLX has been selected as a platform to gain first experiences with HLA-based factory simulation. Reasons for this decision were the availability of the required inside information and the availability of a library for factory simulation. As a first step, SLX needed to be integrated into HLA. This was accomplished by Steffen Strassburger [1]. The results of his work will be described in chapter 4. Another part of the research focused on adapting a visualization tool. For this purpose the Java-based tool Skopeo developed by K.C. Ritter [2] has been chosen. The major advantage of this tool is the availability of its source code since it was developed at the University of Magdeburg. A description of HLA-Skopeo is given in chapter 5.