Lean thinking focuses on value-added flow and the efficiency of the overall system. A part sitting in a pile of inventory is waste and the goal is to keep product flowing and add value as much as possible. The focus is on the overall system and synchronizing operations so they are aligned and producing at a steady pace.
Lean manufacturing is a manufacturing philosophy that shortens the time between the customer order and the product build/shipment by eliminating sources of waste (Appendix 1). Waste is anything that does not contribute to transforming a part to your customer’s needs. The results of the lean approach are illustrated in Figure 2 below. Lean manufacturing will take some waste out of the value-added activity shrinking it down as in the mass production approach, but more importantly, it reduces the pure non-valued added activities, which has the large impact on lead-time.
Traditional Results of Manufacturing Improvement
Amount of Time Eliminated
Lean Focus on Non-Value Adding Wastes
Large Amount of Time Eliminated
Time
= Value Added Time
= Non-Value Added Time (WASTE)
Figure 2: Traditional vs. Lean Approaches
Recognizing Waste
When our lean manufacturing process is in action, we should ask: what does the customer want from this process? This then defines market value and can be used to measure the lean outcomes [6]. Then it is worth asking what transformation steps are needed to turn materials entering the process into what the customer wants. Based on this we can observe a process and separate the value-added steps from the non-value added steps. As an example, this is demonstrated for a generic manual assembly operation on fan blade in turbine engine in Figure 3 taken from RR internal report [10]. There are many individual steps, but generally only a small number add value to the product. In this case only those steps highlighted in red add value. The point is to minimize the time spent on non-value added operations, for example, by positioning the material as close as possible to the point of assembly.
- Delivering components to the assembly line
- Walking to pick up the components
- Removing the package from the delivered parts
- Sorting the components before assembly
- Orientating the parts so they can fit the mainframe
- Picking up assembly tools and bolts for the components
- Walking 25 feet back to the mainframe on the assembly line
- Positioning the component of the mainframe of engine
- Walking to the power tool
- Reaching for power tool
- Pulling the power tool down to the component
- Placing the fixing bolts in the component
- Tightening the bolts to the engine mainframe with the power tool
- Return by walking 25 feet for the next component
- etc.
Red is value added, black is waste operations.
Figure 3: Waste in Manual Turbine Engine Assembly
A Model of Lean Manufacturing in Aerospace
The TPS in Figure 4 goals to illustrate the roots of concept -quality, cost, and delivery through shortening the production flow by eliminating waste. Traditional mass production focused primarily on cost reductions through individual efficiency gains within individual operations. It is learned from other studies [7] that in fact by focusing on quality – “doing it right the first time” - we could simultaneously reduce cost and improve quality. The focus of TPS is on total system costs by taking a value stream perspective.
Figure 4: Toyota Production System
The two main pillars of TPS are Just-In-Time and Built-in Quality, which are mutually reinforcing. Creating a JIT flow may result to increased quality. Without the inventory buffers of mass production, JIT systems will fail if there are frequent quality problems that interrupt the flow. At the centre of lean manufacturing are people who must bring the system to life by continually improving it. The TPS was translated to a lean turbine engine model shown in Figure 5. It includes all the elements of TPS, but shown within an assembly line with the engine as the centrepiece. The main model elements were based on the Rolls Royce business processes described elsewhere [8] and will be explained below.
Figure 5: Lean Manufacturing Process of Turbine Engine Production
3. Just-In-Time
Reengineering of the manufacturing process
The ideal for JIT is a one-piece flow. This means identifying families of parts that go through the same set of processes and dedicating a production line to that product family. All products assigned to the cell will go through those operations one piece at a time.
Figure 6 gives an example of batch processing versus one-piece flow. In the batch processing case some rectangular titanium shapes for a fan blade are cut to create a hollow mainframe of the blade, along with some stiffeners, if any. This is done in large batches which are moved as large batches to be cut into more specific curved shapes. These parts must be sorted before they are cut into the actual shapes needed. This batch cutting leads to a large pile of inventory which must be moved to another buffer and then sorted through to be sub-assembled, and finally the subassemblies are moved and sorted through to get the parts needed to construct the actual blade. Notice how much non-value added work there is on this process all of the moving and storing and sorting is pure waste.
The alternative ideal from a lean manufacturing point of view is a pure one-piece flow that is shown in the bottom of Figure 6. In this case you would cut just the material you need, pass it on do the final cutting, pass it on, do the subassembly, pass it on, and build up the blade. While it may not be feasible to make one and move one, the smaller the batch size the better from a lean manufacturing point of view, within feasible limits.
Figure 6: Batch process vs. One-Piece Flow in Fan Blade Assembly Process
The lean manufacturing is even speed-up by new innovations introduced in the process. Hollow fan blades (Fig. 7) currently enclose a strong, stiff metallic structure to maintain the cross-sectional profile of the blade when subjected to the large forces in flight. Introducing nanotechnology-based process the stiffeners in the blade core are replaced by synthetic foam with nanoscale fillers to stiffen hollow fan on large civil aircraft engines. In the cavity-fill fan concept [11], the light-weigh core replaces this metal structure simultaneously acting as both strengtheners and vibrations reducing elements. The next generation blades are composed of a titanium root and composite internal/external manifolds with nanomaterials reinforcement.
Figure 7: Innovations Reduce Process Time, Performance and Minimise the “Waste”
Reengineering of the production space
Figure 8 provides a large scale picture view of the turbine engine assembly process [10]. Traditionally, the process was organized by functions. For example all the fan blades are processed in one shop, whether curved or flat, and the engine mainframe is processed in a separate shop. Large batches of plates and profiles are processed and then pushed into storage. They are then pushed into subassembly where they need to be sorted as it is shown in Figure 6. All parts must go through the same welding shop which often becomes a bottleneck for process.
The bottom part of Figure 8 shows a typical arrangement in the modern aerospace. In this case the process is organized by “product line.” Product line does not mean separate engines, but rather similar part families. In this case flat blocks go through one set of processes and a separate set of processes are reserved for curved blocks. For example all the flat plates are cut in process lanes, as are straight profiles, and then small batches are brought to the flat block line for assembly. Figure 8 shows a process that actually segregated the control and welding shop into two shops, one for flat blocks and one for curved blocks. The flat and curved components are then outfitted in separate areas and finally assembled together in the engine mainframe. Notice the convoluted paths materials take in the functional batch process and how clean and smooth the flow is in the product-flow process.
Figure 8: Functional-Batch vs. lean Product-Flow Process
Figure 9 shows a layout of a traditional functional flow compared to that of a lean product-oriented process; there is a real paradox. It seems that if we segregate operations by product family and duplicates some resources, e.g., the welding booths, it would take more space. The main reason for this is the dramatic reduction in inventory when our process changes to a lean flow-inventory, the space needed to store it and move it, often takes as much space as our value adding process. Since lean flow reduces inventory it saves on the inventory carrying costs.
Figure 9: Lean Product-Flow Free-Up Space
Figure 10 suggests there are many other savings, perhaps more important than inventory carrying costs. One of the most important benefits is improved quality. The quality benefit comes because of the shorter feedback loops when what a worker cuts this morning is actually assembled into a block this afternoon, instead of weeks from now. With large batches of inventory many quality problems are hidden and only become visible when our down steam customers (e.g., the block assemblers) try to use the material and it does not fit. Productivity also improves as a result of reducing all the non-value added time spent handling and handling again materials. Productivity is also increase since identifying problems and solving them in real time takes less labour hours than finding and fixing problems that have accumulated over weeks. One of the largest benefits of continuous flow is shrunk lead times which allows you to quote shorter lead times to your customer and also to increase the utilization of assembly line therefore generating more revenue in the same period of time.
Figure 10: Benefits of Creating Flow [3]
4. Built-In Quality
The aerospace industry traditionally recognises the importance of quality in terms of reliability and safety for passengers. Built-in quality is much more effective and less costly than inspecting and repairing in quality, especially in more complex and precise nanomaterials products. Accuracy and reliability control refers to a body of statistical and problem solving tools that can help do the job right the first time at design and then at manufacturing as the lean process. With very low levels of inventory there is no buffer to cover ourselves in case there is a quality problem. Problems in operation A will quickly cause a shut down of operation B.
This problem is multiplied when there are a whole series of operations. The problem of serial unreliability is illustrated in figure 11. As this figure shows, even four fairly reliable operations individually (85% to 90% reliable) can lead to low overall system reliability (62%).
Figure 11: Serial Low Reliability Leads to Uncontrolled Failure
5. Lean Quality Logistics Supply Chain
Although new assembly manufacturing of the turbine engine is designed according to the principles of lean manufacturing, it is also the responsibility to improve the supply chain. Especially with the start of the Trent 900/1000 series engines, as well as the coming nanotechnology-based processes, a high number of newly delivered components should be organized and controlled [9] (Appendix 2). The lean manufacturing can only succeed if the supply chain is organized in the same lean way, oriented at the flow of only the right materials. Suppliers’ deliveries need to be synchronized with the new demands of the assembly lines. In this part, the Kraljic-matrix (Appendix 3) is used to identify an approach to quality and planning assuring efforts, and when to delegate to or collaborate with suppliers to improve cooperation of Rolls Royce with its suppliers regarding supply chain delivery efforts.
Organisation and control of the outsourced phases with the supplier is possible by collaboration based on knowledge and information exchange, i.e. materials flow, information and control flows (figure 12). Materials flow is a physical supply of an engine part. Information part includes different data sources: drawings, filed history etc. Control flow provides a movement data of a part to predict forecasts and storage levels. As the flows can be independent of each other they might identify quality problems on an early stage. In case of a damage part during the assembly line, the control flow can show the first possible replacement part upstream and activate the recovery process.
Figure 12: The Delivery Process Can Be Divided Into Three Flows
Perhaps the most intensive way of supplier collaboration is integration of design phases, and planning phase (figure 13). When suppliers are already involved at early stage of defining product characteristics, customized products will have a shorter reaction time to a customer order. Additionally, working together on the design and process characteristics supersedes the quality inspection at delivery of the parts. When a planning interface is organized as well, like sharing forecasts, production proceedings and storage levels, a JIT delivery is possible. Combined with the customer order specifications, delivery in a kit will be a form of supplier collaboration.
Figure 13: Integration of Part & Process Design and Planning Supersede Actions at the Integrator’s Site, and Causes a Faster and Better Response to a Customer Order
Analysis of lean supply scheme has to determine, which phases should be outsourced, and how to collaborate with suppliers. The scheme is dependent on engine part and supplier characteristics (size, volume etc.). A supplier of a large and critical part (such as fan blade) will need more control over all cycles of design and production. The ‘Kraljic-matrix’ (Appendix 3) may help to understand the lean supply chain, combining profit impact and supply risk (figure 14). The matrix identifies four types of products: non-critical, leverage, bottleneck and strategic.
For non-critical parts, the firm must perform all three delivery phases, because of the low risk and turnover of the suppliers. Suppliers of leverage products will deliver JIT due to the high financial volume. An additional commission effort will happen in the DC. Strategic suppliers will deliver JIT. The same holds for bottleneck part suppliers. The high customization level tends to a kit-delivery, so quality agreement and JIT delivery would be suitable.
These delivery strategies may cause controlling efforts. The high customization level in the airplane industry locates most parts on the right hand side of the matrix. Especially these strategic and bottle neck parts cause a high level of collaboration or supplier development, to secure the supply risk. The problem is that quality cost money. These non-recur ring costs will not always be price-efficient for bottleneck suppliers that should be avoided. This is possible through, for example, system integration, see figure 14. Sourcing more parts to a single supplier lets it move towards the strategic quadrant, where supplier development will be effective. Figure 15 shows the strategies for all quadrants.
Figure 14: Moving Suppliers Out of the Bottleneck Quadrant via System Integration
Figure 15: Supplier Strategies per Quadrant
Conclusion
This report presented lean manufacturing as a philosophy, a way of thinking and a system of production, not a set of individual tools. Moreover, lean manufacturing requires an enterprise-level view of the value stream from raw materials to the finished engine delivered to the customer. The case study shows that the use of lean practices helped the firms:
- improve product design and quality,
- reduce assembly tooling and time,
- motivate staff to make improvements,
- eliminate non-value-added processing steps, reducing production labor hours,
- reduce shop cycle times,
- reduce work in process levels,
- shift from discontinuous to continuous (and even pull) production flow,
- reduce production floor space.
Due to new aircraft engines programs such as the Trent 11xx and Trent 9xx family, an increasing number of parts will have to go through the quality and logistics organization of the firm. The lean manufacturing also demands an accurate response of the supply chain in quality and timing. The increased supply risk needs to be assured trough more intensive, but selective supplier collaboration and development of the Rolls Royce supplier strategy. Recommendations on the level of JIT and kit deliveries require further study, since these definitions can also be categorized. For example, JIT can vary from an hour up till several days before assembly of the engine start. A sub-assembly (for example, fan blade) can be seen as a kit, but all these parts delivered, unassembled in a box, can also be considered as a kit.
The studies in this report have shown that the lean operations can result to reducing waste, speeding delivery, and meeting changing customer requirements. RR has the lean operations and thinks to remain competitive in the business. However, to conclude I wish to post a question: if you work in a hi-tech manufacturing, is your firm changing quickly to make improvements in its operations? If your employer is part of the civil aircraft industry, your competitors are, and your customer expects that you will, too!
Rolls Royce Centre located in Sheffield has confirmed to use several aspects of this report to improve the design process and quality assurance logistics in the fan blade materials (Appendix 4).
References
-
Rolls Royce Aerospace Market Outlook, 2006.
- T. Ohno, Toyota Production System, Productivity Press, Cambridge, MA, 1988.
- Y. Monden, Toyota production system - an integrated approach to just-in-time, Third ed. Engineering & Management Press, Norcross, Georgia, 1998.
- W.M. Feld, Lean manufacturing: tools, techniques, and how to use them. The St. Lucie Press, London, 2000.
- M. Holweg, “The genealogy of lean production”, Journal of Operations Management xxx (2006) xxx, in press.
- R. Shah, P.T. Ward, “Defining and developing measures of lean production”, Journal of Operations Management, xxx (2007) xxx, in press.
- F.A. Abdulmalek, J. Rajgopal, “Analyzing the benefits of lean manufacturing and value stream mapping via simulation: A process sector case study”, Int. J. Production Economics, 107 (2007) 223–236.
- Y. Yusuf, A. Gunasekaran, M. S. Abthorpe, “Enterprise information systems project implementation: A case study of ERP in Rolls-Royce”, Int. J. Production Economics 87 (2004) 251–266.
- Videotape, The Road to World Class Manufacturing, Rolls Royce Visual Media Services, 2005.
- “Developing of damping materials for visco-filled blisk concept”, Internal Rolls Royce UTC report, University of Sheffield 2001.
- “Polymeric damping filler for hollow fan blades – material development for the RC104B blisk”, Internal Rolls Royce UTC report, University of Sheffield 2003-2004.
Appendix 1
The Seven Wastes in Manufacturing are the following:
1. Over production – Producing more material than is needed before it is needed is the fundamental waste in lean manufacturing. Material stops flowing.
2. Producing defective products – Defective products impede flow and lead to wasteful handling, time, and effort.
3. Inventories – Material sits taking up space, costing money, and potentially being damaged. Problems are not visible.
4. Motion – Any motion that does not add value to the product is waste.
5. Processing – Extra processing, not essential to value-added from the customer point of view, is waste.
6. Transportation – Moving material does not enhance the value of product to customer.
7. Waiting – Material waiting is not material flowing through value-added operations.
Fast Changeovers and Levelled, Mixed-Model Production
When we are thinking in a batch processing mode we would like to achieve economies of scale for each individual piece of equipment. So the logical solution is to build large batches of product A before changing over to product B. The result is large batches, and as we have discussed the system inefficiencies associated with large batch production.
In lean manufacturing, we want to keep batch sizes down and build what the customer (external or internal) wants. In a true one piece flow, we could build in the actual production sequence of customer orders (e.g., A,A, A,B, A,B,B, A,A,B etc.). The problem with building to an actual production sequence is that it is irregular and causes you to build parts irregularly. In the example earlier, if each letter represented a batch of product A or B you would need the parts for three batches of A in a row. Thus, you would have to have enough parts on hand for three batches of A. And it is even possible you will get orders for nine As in a row and need enough parts for this large number. To smooth out parts usage and send a more level set of orders to upstream operations it is often better to level the schedule. So instead of building to the actual customer demand you would notice you are making six As for every three Bs. You could then create a level production sequence of: AABAABAAB. This is called levelled, mixed-model production because you are mixing up production but also levelling the customer demand to a predictable sequence which spreads out the different product types.
There are a variety of ways of levelling the flow of work in engine assembly as follows:
- Using Temporary Employees
- Cross Training Employees
- Careful planning
- Standardized Times for Processes
- Standardized Designs
- Balancing Processes across the assembly
- Tact Time Planning.
Appendix 2
Supply chain and means of delivery
The supply chain of turbine engine consists of many component suppliers and one final integrator – Rolls Royce, who mainly assembles the parts into the final product and delivers it to the customer, the aircraft. The integrator is responsible for the lean design, production and assembly of all parts and can decide where nanotechnology suppliers produce, and, if they are responsible for the design (figure 1). This principle is called ‘build to print’ and ‘design and build’ respectively.
Order and delivery process from suppliers requires three classical delivery phases, which are primarily done by distribution centre (DC). These phases are quality inspection, storage and commission effort. For the delivery phases, the final integrator once again has opportunity to outsource them to the suppliers (figure 2). A lean manufacturing approach gives agreement lets the supplier deliver a new dimension to the delivery straight into the storage.
Figure 1: Sourcing Possibilities for the Final Integrator
Figure 2: Sourcing Possibilities for Final Integrator in the Delivery Process
Appendix 3
Kraljic-matrix
The profit impact is estimated as a turnover that supplier delivers to the final integrator. The supply risk derives from the operational supply risk, rather than from market factors such as the number of possible suppliers:
- operational supply risk,
- originality of a part or customization level,
- failure probability, control and make up possibilities,
- quality level parts,
- transparency of supply chain,
- replacement lead time,
- importance to non-stop assembly line.
Figure 3: The Kraljic-Matrix
The combination of the operational aspects may not be estimated as a single value. The originality or customization level is applied as the dimension of the supply risk and thereon allocates the means of delivery to the quadrants. These characteristics will then request a certain level of failure probability, control and make up possibilities that form the supplier strategy.
The uniqueness or customization level of a part is dependent on how often it is installed. The customization level can be expressed in the way a part is standard or not. The failure probability can be seen as the quality level of a part, in other words: the chance for it to break down or not to fulfil the task it was designed for. The probability of failure means how this possible failure would then be discovered. The control states how fast such a failure can be solved, for example by using a replacement part. It is difficult to estimate these values; however, what can be stated are the influences of these aspects on the supply risk, so tools can be developed to decrease these risks.