Summary of “Controlling design variants”

Subtittle: Modular Product Platforms

Authors: Anna Ericsson, Gunnar Erixon

ISBN 087263514-7 1990  (Also G. Erixon, doctoral thesis 1998)

Preface:

the authors define ‘modules’ as items with well-defined interfaces, together forming a modular platform, from which a stream of derivative products can be efficiently developed, marketed and produced.

A design approach towards a modular product platform allows building a variety of final products, while controlling the internal complexity of diversity

Chapter 1 gives an historical perspective. We decided to skip that here.

Chapter 2 insists on incorporating design for manufacturing in the design phase. This is common practice today.

Chapter 3:

A modular design has two characteristics:

• The functional architecture is reflected in the physical architecture
• The minimizing of interactions between physical components

Modularization is the decomposition of a product into modules with specified interfaces, driven by company-specific strategy

An assembly is not always a module:  an assembly is often the result of assembly planning.  Assemblies are created because the product design does not permit the product to be assembled in a single continous flow. Having many sub-assemblies may indicate a poor product design.

A module is quite different from an assembly: a module is chosen for specific, corporate strategic reasons and the interfaces take easy assemblability into account

The drivers for modularization are

• In Design

Striving for product development efficiency, Carry over: isolate proven design solutions
Technology evolution: isolate design solutions that change
Planned product changes: roadmap items
to manage Variance in machine types
Different specifications: function elements with different spec
Styling differences: create diversity by e.g. other color

• In Production

Common units: repetitive production of the common units
Process and/or organization: technology specific modules (weld-assy)

• In  Quality

Separate testing: qualify the functionality of a module

• In Purchasing

Supplier availability: black box outsourcing of functional units, rather than parts

• In After sales

Service and maintenance: more (off –fab) repair of modules, Upgrading: as in PC

Recycling: ….

Chapter 4: How can a modular design be achieved?

Via “module function deployment” (MFD):

1. Derive customer demands, market related
   1. Use QFD?
2. Identify main functions that fulfill the demand
   1. functional decomposition, understand the function of parts, create functional independence
3. Which technical solution can be created as module?

Ranking of the possible design solutions per modularity driver, using the Module Indication matrix, or  MIM
(see figure, for better resolution see above)

• Some solution will score high on many modularity-drivers: make them a module
• A small number of scores might indicate that this part does not deserve a module of its own and is better integrated in another module

4. ‘Evaluate concept regarding interfaces/assembly module drivers’
   1. Here, a general rule is given for the number of modules that a product might consist of: the square root of the total number of assembly steps/piece-parts to be assembled
   2. Two ideal types of assembly are defined here:
      1. The hamburger type of stacking items one upon the other
      2. ‘base unit assembly’: assembling elements on the same carrier

An interface matrix is defined, which shows the character of the interface (either i or ii) and whether it is an geometrical fit and/or an energy/fluids/gas-transmitting fit.

5. Specify the modules, regarding design, cost, description of variants

Chapter 5 Example of analysis and synthesis a modular design in a vacuum cleaner

The five steps described in chapter 4 are followed through in the design of a vacuum cleaner.

In step 1 of the analysis of the vacuum cleaner, customer requirements are related to product properties (size, noise, weight, power, operating forces required), rather than to product functions.

In step 2 the analysis demonstrates how a design solution for one function leads to  next functional requirement, which leads to a next design solution, which…etc.

In step 3, it is analyzed which of the module-drivers works the hardest for each main part.

It is concluded that ‘reuse of an existing design’ (carryover) and the requirement for ‘common unit’ in manufacturing are applicable for many parts. This is a sign of a mature product with limited technical development.
Nine modules are proposed, suiting the analysis best.

Step 4 identifies this design as mainly being a ‘base unit assembly’

Step 5 is not worked out in the book

Chapter 6

Modular Function deployment in practice requires intensive effort, considerably more exhaustive than for a normal single new product. The intention is, though, that this will pay off in the development of consecutive models, since only new  features or planned design changes will be of concern in the coming development projects.

A case at Volvo Car Corp

Volvo introduced MFD in order to create commonality in the production system structure equipment in two plants (Gent and Gothenburg). The new platform concept was designed along with new products.  It resulted in more pre-assembly (outsource larger parts) and a shorter final assembly time.

Modularization of a complete car is described as identifying ‘functional areas’, whereas the actual modules are ‘roof, side panel, doors, hub, rear hatch, shields, front floor, middle floor, rear floor, engine, gearbox, front suspension, rear axle, instrument panel, etc..

Again, these items, or groups of items, are checked against the module drivers in a MIM.

Engine, gearbox, doors, instrument panel and exhaust system score high on ‘should be modulized’.  Vice versa, strong reasons for modularization are found in process/organization of manufacturing, in purchasing, and in carryover of design.

Inner roof

This chapter describes how making the ‘inner roof’ into  a module gained valuable knowledge: the roof consisting of 40 parts, some 6 modules might be considered, since the optimal number of modules is the square root of the part count of the total product.

To build a roof assembly, six subassemblies were identified; other loose elements were better off being integrated in them. Two of these modules come in variants ( a roof base module with/without hatch, a sun shield in two variants

Experience gained in another Volvo-project showed that:

• Different competence groups in an organization give different weights to the module drivers: manufacturing wants modules for ease of assembly, in D&E carry-over is emphasized. Cross-functionality is required for balancing the module-definition-drivers
• Different parts have different scores on module drivers
• A holistic approach to modularization is required before a particular modular concept is adopted

There are yet other examples in chapter 6 which are not summarized here

Chapter 7 Modular Product Platform management

Highlights in bold letters

A companies’ strategy must be explicitly stated, understood by all and reflected in the heart of the company: its product. The product platform strategy should be derived from the company strategy. A platform development plan contains product features to be launched, cost reductions to be managed new technology to be introduced, etc.

The demand for frequent product introductions makes it impossible to develop a brand new product each time. Development resources are focused on one of the product functions. Design changes do not spread all over the machine. Primary development is discriminated from product development. In primary development, the technical evolution of new functions or products is determined, the technical solutions being unknown. In ‘product development’ the actual means are designed.

During the modularization process, areas in the product where it is strategically important to offer the customers variance are identified. Well thought-out interfaces will assure that the variations are concentrated to only these parts

The sales organization should understand what advantage the module concepts bring, because for Sales modularization brings limitations in customizing: the diversity spectrum is predefined.  If customers should demand a product variant that is outside the predefined structure it become a strategic question as to whether accept the order or not.

Product development process should be characterized by how things are done, not by who does them.

The concept phase is the most important part of a Product Generation Process since costs and product structure and fit to varying requirements are determined there.

Cross-functionality is crucial, QFD and MIM help to prevent departmental  preferences to overwhelm objective evaluations

Appendix

The appendix gives an overview of modularization and its effects on the design/manufacturing process. See the book for quantifications.

1. Interface complexity leads to a long lead time in development

Designing on modules can be done in parallel if the interface is simple.
Simple geometrical interfaces reduce assembly time.

2. Carry-over reduces development costs. The higher to complexity of a carry-over, the larger the reduction in development expense
3. The more black box outsourced modules, the less development capacity required. A modular design creates an opportunity to outsource development
4. A platform complexity and cost increases with the number of
   1. modules in each product variant
   2. module variants needed to build all product variants
   3. contact surfaces on the interfaces between modules

It is shown here that an ideal modular concept would have a complexity of 1,5  times the square root of the number of parts in the complete product. (page 116)

5. The total cost of running the assembly system (here called ‘system costs’) can be greatly reduced by outsourcing parts and modules. A company might focus on ‘architectural knowledge’: the ability to capture customer needs and translate them into the performance specifications of subsystems.
6. Modularity and lead time in assembly.
   If modules are assembled off-line, than functionally tested and than assembled into a product, the lead-time has an optimum if the number of modules is the square root of the total number of parts in the product. The value depends on the assembly time of parts to for the module, and of the assembly time of the modules to form the product.

The qualification of modules before applying them in the end-product results in less quality cost

7. Strategic variety is noticeable by the customer and is introduced in product design via marketing. Tactical variety is not obvious to the customer and serves manufacturing or design. Introducing diversity should underlay the following qualitative questions
   1. It is significant to the customer
   2. Will the customer be aware of it
   3. Where in the production chain will the diversity appear?
   4. Which variants are necessary?
   5. In which part of the product can variants be allowed