COMPUTER INTEGRATED TECHNOLOGIES AND KNOWLEDGE MANAGEMENT:CONCEPTS

CONCEPTS

Process Chains

A process chain is a set of rules or steps carried out in a specific order in order to carry out the completion of a defined process. Processes can be executed either in sequence or in parallel. A typical example of a product development process chain is portrayed in Figure 19.

A process chain is characterized as:

• A process, divided into subtasks, in which contributions to the creation of virtual products take place

• A series of systems in which computer support for product-related subtasks is guaranteed

• A series of systems for organized support of the complete proces

• Mechanisms for the adequate exchange of data between systems Computer support of process chains is related, on one hand, to the processing and management of product data and, on the other hand, to the organization of the supported process and the handling of associated data.

The processing and management of product data entail all tasks that concern the generation, storage, transfer, and function-related editing and provision of data (Reinwald 1995).

General classification of process chains suggests a determination of the depth of integration of the product data handling. Classification can be divided into:

• Coupled / linked process chains

• Integrated process chains

• Process chains with common data management

The exchange of data within a coupled process chain takes place directly between the supporting systems. Thereby, information transferred between two respective systems is aided with the help of system-specific interfaces. This realization of information transfer quickly reaches useful limits as the complexity of the process chains increases. This is because the fact that the necessary links grow disproportionately and the information from the various systems is not always transferable (Anderl 1993).

Characteristic of integrated process chains is a common database in the form of an integrated product model that contains a uniform information structure. The information structure must be able to present in a model all relevant product data in the process chain and to portray the model with the use of a homogenous mechanism. A corresponding basis for the realization of such information models is offered by the Standard for the Exchange of Product Model Data (ISO 10303—STEP) (ISO 1994).

Basic to a process chain with common product data management based on an EDM system is the compromise between coupled and integrated process chains. With the use of common data man- agement systems, organizational deficiencies are avoided. Thus, the EDM system offers the possibility of arranging system-specific data in comprehensive product structures. The relationship between documents such as CAD models and the respective working drawing can be represented with cor- responding references. A classification system for parts and assemblies aids in locating the required product information (Jablonski 1995).

On the other hand, with joint product data management, conversion problems for system-specific data occur in process chains based on an EDM system. These problems exist in a similar form in coupled process chains. This makes it impossible to combine individual documents into an integrated data model. References within various systems, related to objects within a document rather than the document itself, cannot be made.

Besides requiring support for the processing of partial tasks and the necessary data transfer, process chains also require organized, sequential support. This includes planning tasks as well as process control and monitoring. Process planning, scheduling, and resource planning are fundamental planning requirements.

The subject of process planning involves segmenting the complete process into individual activ- ities and defining the sequence structure by determining the order of correlation.

In the use of process plans, the following points are differentiated:

• One-time, detailed process planning (all processes are executed similarly)

• Case-by-case planning of individual processes

• One-time, rough planning of a generic process, followed by deeper detailing and specification of individual processes.

The first basic approach is typical for the majority of commercial work flow management systems. Because these systems are intended for routine tasks, their application is unproblematic only when they concern the support of completely determinable processes. Product development processes are characterized by the fact that their achievements are not precisely predictable in advance. For that reason they are treated as nondeterministic processes.

The second approach requires an independent sequence for each process path. This corresponds to the generation of a network plan using methods such as CPM (critical path method), MPM (metra potential method), and PERT (program evaluation and review technique). These techniques, as part of the project management, are employable for activities such as proper scheduling (Burghardt 1988).

The third approach involves the strategic use of rough supplementary outlines of generic processes with those of detailed, individual processes. A requirement for successful implementation of the strategy is the ability of the supporting system to map process structures hierarchically.

For scheduling, a closed sequence plan is assumed to exist. The activities and correlations must be determined beforehand. Therefore, the minimum task of the scheduling part of the process is the calculation and determination of time limits, critical paths, and buffer times. The CPM, MPM, and PERT methods are implemented for this purpose, and the incorporation of these operations leads to the calculation of complete process and buffer times using both forward and backward calculations.

The duration of a procedure depends significantly on the use of resources required for fulfil- ment of the process. In this respect, interaction exists between the scheduling and capacity planning (Burghardt 1988).

Capacity planning for the product-development process involves proper quantity and time allo- cation of coworkers, application system capacities, and other partial tasks of an individual process (Golm 1996). Therefore, the goal of capacity planning is to ensure the on-schedule processing of the entire process and the uniform utilization of resources.

Tasks of Process Control and Monitoring

The task of process control and monitoring is to guarantee the fulfilment of the process based on the sequence, schedule, and capacity planning determinants. The following individual tasks must then be resolved:

• Activation of processable actions

• Process monitoring of contents

• Process monitoring of time

Integrated Modeling

Modeling of virtual products includes all phases of the product life cycle, from product planning to product disposal. The aim is complete integration of all the development processes for efficient product development.

The following system characteristics are strived for through the integration of design and pro- duction planning systems:

• Increased productivity during product development

• Acceleration of the product development process

• Improvement of product and technical documentation quality

An increase in the productivity of product development phases through the integration of design and manufacturing plans is based on preventing the loss of information. A loss of information often occurs between coupled and unrelated systems because of integration difficulties. With a lack of proper integration, detection and editing of data take place numerous times for the same information content. Therefore, using product models derived from information within both the design and man- ufacturing data, it is possible to integrate specific work functions. The models then serve as a basis for all system functions and support in a feature approach-type form.

The feature approach is based on the idea that product formation boundaries or limits are explicitly adhered to during design and production planning stages. This ensures that the end result of the product envisioned by the designer is not lost or overlooked during the modeling phase. To realize the form and definition incorporated by the designer in a product, it is essential that an integrated design and production planning feature exist.

Accelerating the product development process is achievable through the minimization of unnec- essary data retrieval and editing. Efforts to parallel the features of concurrent engineering tasks require integrated systems of design and production planning.

Higher product and documentation quality is achievable with the use of integrated systems. Through integration, all functions of a product model are made available to the user. Any errors that may occur during data transmission between separate systems are avoided.

Because of the high responsibility for cost control during product development, estimates for design decisions are extremely necessary. Cost estimates reveal the implications of design decisions and help in avoiding costly design mistakes. For effective estimation of costs to support the design, it is necessary to be able to access production planning data and functions efficiently. Integrated systems provide the prerequisites for the feedback of information from the production planners to the designers, thus promoting qualitatively better products. The following forms of feedback are possible:

• Abstract production planning experience can be made available to the designer in the form of rules and guidelines. A constant adaptation or adjustment of production planning know-how has to be guaranteed.

• Design problems or necessary modifications discovered during the production planning phases can be directly represented in an integrated model.

• Necessary modifications can be carried out in both production planning and design environ- ments.

Methodical Orientation

The virtualization of product development is a process for the acquisition of information in which, at the end of the product development, all necessary information generated is made available. As- suming that humans are at the heart of information gathering and that the human decision making process drives product development, the virtual product-creation methods must support the decision maker throughout the product creation process.

Product-creation processes are influenced by a variety of variables. The form the developed prod- uct takes is determined by company atmosphere, market conditions, and the designer’s own decisions. Also influential are the type of product, materials, technology, complexity, number of model varia- tions, material costs, and the expected product quantity and batch sizes.

Product development is not oriented toward the creation of just any product, but rather a product that meets the demands and desires of the consumer while fulfilling the market goals of the company. The necessary mechanisms must provide a correlation between the abstract company goals and the goals of the decisions made within the product development process. For example, to ensure the achievement of cost-minimization objectives, product developers can use mechanisms for early cost estimation and selection, providing a basis for the support of the most cost-effective solution alter- natives (Hartung and Elpet 1986). To act upon markets characterized by individual consumer demands and constant preference changes, it is necessary to be able to supply a variety of products within relatively short development times (Rathnow 1993; Eversheim 1989). This means that product struc- turing must take into account the prevention of unnecessary product variation and that parallel exe- cution of concurrent engineering is of great importance.

Methods that are intended to support the product developer must be oriented toward not only the contents of the problem but also the designer’s way of thinking. Therefore, a compromise must be made between the methodical problem solving strategy and the designer’s creative thought process.

The product-development methods can be separated into process-oriented and product-oriented categories. The main concern in process-oriented product development is the design. The objective is the indirect improvement of the design or, more precisely, a more efficient design process. Product- oriented methods concern the product itself and the direct improvement of the product.

Various process-oriented methods are discussed below. Fundamental to VDI 2221 guidelines is the structuring of the product-development process into partial processes or phases. This structuring takes place independently of the developed product, the company, market conditions, and the decision maker. The demand for a general strategy results in individual steps being described at very abstract levels. Consideration of product, company, and market-related influences is incorporated into the design throughout all processes. The impact of individual influences on methods can only be clarified with examples. The worth of a new method lies in the general systemizing of the product development process. This method is not suitable for immediate derivation of concrete product development pro- cesses. Rather, it describes an ideal, flexible product-development process. The structure and content of the development process are then molded by the product, company, or market-related influences (VDI-Gesellschaft Entwicklung Konstruktion 1993).

Another example of process-oriented methods is the concurrent engineering method, which covers all topics from product development to equipment and production planning. The primary objective of simultaneous engineering is the reconfiguration of the development process, with the intention of reducing development time while improving product quality. As opposed to the traditional approach, in which a series of steps is followed and feedback occurs through long loops, the development tasks within concurrent engineering allow for the parallel execution of many tasks. Therefore, development times are substantially reduced and the old system of following predetermined steps is avoided. This concept relies heavily on the exchange of information between the various departments. This should extend beyond company boundaries to include the equipment manufacturer in order to tie production equipment planning into the product development process. The main focus in the implementation of these methods, besides the use of computer aided technology, is the reorganization of the company structure. In the foreground are measures to stimulate the exchange of information, such as the formation of interdisciplinary teams (Eversheim et al. 1995; Krause et al. 1993; Bullinger and Warschat 1996).

The design-to-cost method is a product-development approach that bases design decisions on cost-sensitive criteria. Here, evaluation standards are not only production costs but also the costs incurred throughout the life of the product. This method is particularly applicable for complex prod- ucts with relatively long life cycles. The method is broken down into target costing, cost-oriented design, and cost control. Within target costing, a goal for final costs is broken down for individual product components. The final cost is determined from the results of a market analysis or from a clear depiction of customer demands. Relative cost and manufacturing cost data can be used to aid the development of design alternatives that fall within the early cost-oriented design process. With the aid of cost estimate models, the life cycle costs of each alternative are assessed and the most cost effective variant selected. In the area of cost control, the design costs are compared to and contrasted with the original end cost objective. If the cost objective is not met, deviation analyses are necessary to determine required alternatives.

Product-oriented techniques are characterized by product formation, task formulation, and the objectives pursued. Product-oriented tasks vary considerably, according to the multitude of products and tasks. A few examples of these techniques and technique groupings are presented later in this chapter.

Design rules and guidelines specify how a product is designed according to previously determined objectives. Well-known design rules and guidelines exist particularly for production, assembly, er- gonomic, and logistics-oriented processes as well as resource-efficient and recycling-oriented product design (Pawellek and Schulte 1987; Krause 1996; Pahl and Beitz 1993; Kriwet 1995).

Simulation technology has become important in the field of product development. The primary objective in the implementation of simulation technology is the early acquisition of information on product characteristics before the product even exists. The knowledge gained aids in the assessment of the respective development results. Costly and time-consuming mistakes during development phases can be recognized early and avoided. Through an iterative strategy, it is also possible to optimize systematically particular product characteristics or qualities (Frepoli and Botta 1996; Scho¨n- bach 1996).

One requirement for the application of simulation techniques is the creation of a model, which allows for the investigation and breakdown of real product tasks. With the use of computers, it is possible to design complex models that allow answers to relevant design questions to be found. These models were previously realizable only through the tedious production of prototypes. The application of simulation technology is especially helpful when numerous product variations must be considered, a situation not economically feasible with prototypes.

In the future, systems must be able to provide all relevant information in the user-desired form and make additional mechanisms available that reveal the consequences of a decision made during product development. The provision of information must include not only product specific data but also comprehensive information. Data covering all topics from design guidelines and rules to design, solution, and measurement catalogs to company individual knowhow is necessary. Solutions devel- oped for similar products need to be accessible not only in general form but in a form that takes product-related tasks and context into account.

Rapid Prototyping

Due to the ever-increasing demand for shorter product development cycles, a new technology known as rapid prototyping (RP) has emerged. RP is the organizational and technical connection of all processes in order to construct a physical prototype. The time required from the date the order is placed to when the prototype is completed can be substantially reduced with the application of RP technology (Figure 20).

RP is made up of generative manufacturing processes, known as RP processes, and conventional NC processes as well as follow-up technologies. In contrast to conventional working processes, RP processes such as stereo lithography, selective laser sintering, fused deposition modeling, and lami- nated object manufacturing enable the production of models and examples without the use of forming tools or molds. RP processes are also known as free-form manufacturing or layer manufacturing.

Prototypes can be divided into design, function, and technical categories. To support the devel- opment process, design prototypes are prepared in which proportion and ergonomic models are

incorporated. Design prototypes serve for the verification of haptic, esthetic, and dimensional re- quirements as well as the constructive conception and layout of the product. RP-produced examples are made from polycarbonates, polyamides, or wood-like materials and are especially useful for visualization of the desired form and surface qualities.

For function verification and optimization, a functional product example is required. The appli- cation of production series materials is not always necessary for this purpose. Functional prototypes should, however, display similar material strength characteristics. On the other hand, technical pro- totypes should be produced using the respective production series materials and, whenever possible, intended production line equipment. The latter serve for purposes such as customer acceptance checks and the verification of manufacturing processes.

Prototypes can be utilized for individual parts, product prototypes, and tool prototypes. Normally, product prototypes consist of various individual prototypes and therefore require some assembly. This results in higher levels of dimension and shape precision being required.

The geometric complexity represents an essential criterion for the selection of the suitable pro- totype production process. If, for example, the prototype is rotationally symmetric, the conventional NC turning process is sufficient. In this case, the presently available RP processes provide no savings potential. However, the majority of industry-applied prototypes contain complex geometrical elements such as free-form surfaces and cut-outs. The production of these elements belongs to some of the most demanding tasks in the area of prototype production and is, because of the high amount of manual work involved, one of the most time-consuming and cost-intensive procedures. RP processes, however, are in no way subjected to geometrical restrictions, so the build time and costs of producing complex geometrical figures are greatly reduced (Ko¨nig et al. 1994).

Another criterion for process selection is the required component measurements and quality char- acteristics such as shape, dimensional accuracy, and surface quality.

3.4.1. Systemization of Rapid Prototyping Processes

With the implementation of CAD / CAM technology, it is possible to produce prototypes directly based on a virtual model. The generation of the geometry using RP processes takes place quickly without the requirement of molds and machine tools. The main feature of the process is the formation of the workpiece. Rather than the conventional manufacturing process of a clamped workpiece and material removal techniques, RP processes entail the layering of a fluid or powder in phases to form a solid shape.

Using CAD models, the surfaces of the components are fragmented into small triangles through a triangulation process. The fragments are then transformed into the de facto RP standard format known as STL (stereo lithography format). The STL format describes the component geometry as a closed surface composed of triangles with the specification of a directional vector. Meanwhile, most CAD systems now provide formatting interfaces as part of the standard software. Another feature of this process is that CAD-generated NC code describing the component geometry allows for a slice process in which layers of the object may be cut away in desired intervals or depths.

Starting with the basic contour derived from the slicing process, the workpiece is subsequently built up in layers during the actual forming process. Differences exist between the RP processes in process principles and execution.

RP processes can be classified by either the state of the raw material or the method of prototype formation. The raw materials for RP processes are in either fluid, powder, or solid states (Figure 21). As far as the method of prototype creation, the component forming can either be processed into direct, 3D objects or undergo a continuous process of layers built upon one another (Figure 22).

Digital Mock-up

Today the verification and validation of new products and assemblies relies mainly on physical mock- ups. The increasing number of variants and the need for higher product and design quality require a concurrent product validation of several design variants that are based on digital mock-ups. A digital mock-up (DMU) can be defined as ‘‘a computer-internal model for spatial and functional analysis of the structure of a product model, its assembly and parts respectively’’ (Krause et al. 1999).

The primary goal of DMU is to ensure the ability to assemble a product at each state of its development and simultaneously to achieve a reduction in the number of physical prototypes. Oriented to the product development cycle, tools for DMU provide methods and functionality for design and analysis of product components and their function. Modeling methods are divided into several cat- egories: space management, multiple views, configuration management of product variants and ver- sions, management of relations between components, and also incomplete models. Simulations are targeted to analyze the assembly and disassembly of a product as well as the investigation and verification of ergonomic and functional requirements. Constituent parts of a DMU tool are distin- guished into either components or applications. Components are the foundation of applications and consist of modules for data organization, visualization, simulation models, and DMU / PMU corre- lations. Main applications are collision checks, assembly, disassembly, and simulation of ergonomics, functionality, and usage aspects. Because a complex product like an airplane can consist of more than 1 million parts, handling a large number of parts and their attributes efficiently is necessary. This requires that DMU methods are optimized data structures and algorithms.

Besides basic functions such as the generation of new assemblies, the modification of existing assembles and the storage of assemblies as product structures with related parts in heterogeneous databases, advanced modeling methods are required in order to ensure process-oriented modeling on the basis of DMU. The main methods in this context are (BRITE-EURAM 1997):

• Organization of spaces: Allocating and keeping open functional and process-oriented spaces in DMU applications. The major problem is the consideration of concurrent requirements con- cerning the spaces.

• Organization of views: A view is an extract of the entire structure in a suitable presentation dedicated to particular phases and applications.

• Handling of incomplete models: Incomplete and inconsistent configurations during development process are allowable in order to fulfill the user’s requirements regarding flexibility. Examples are symbolically and geometrically combined models with variable validity.

• Configuration management of product structure: Management of versions, variants, and multi- use.

Besides these methods in a distributed, cooperative environment, consequent safety management has to be taken into account. Therefore, a role- and process-related access mechanism must be imple- mented that allows the administrator to define restrictions of modeling related to roles. The application of such technologies enables a company to manage outsourced development services.