Winning Formula for Cosworth's scheduling success

With experience spanning four decades, Cosworth Racing is a world leader in the design, development and manufacture of ultra-high performance engines for racing, rally and road car application. Cosworth Racing provides engine technology and design solutions for Formula 1, CART, World Rally car and motorbike racing, as well as racing engine technology transfer to the Premier Automotive Group of Ford Motor Company.

Founded in 1958 by Mike Costin and Keith Duckworth, the company has undergone several incarnations with Cosworth Racing now wholly owned by Ford since fourth quarter 1998. With 650 UK employees committed to continually driving forward design and production techniques for mission critical applications, Cosworth Racing turned to Preactor International for a specialist scheduling solution.

Cosworth Racing is intimately involved with its customers at every level ? from concept through design, manufacture and assembly, to ongoing service and development to remain competitive. In order to best meet the stringent demands of its customers, the company comprises of four individual projects that focus exclusively on its own area of expertise. These teams are Formula 1, American Racing Design (ARD), World Rally Car (WRC) and Commercial.

From a planning and scheduling perspective each project team is treated as a separate internal customer. While providing the benefit of allowing close collaboration of general ideas and solutions, this does present a number of potential difficulties ? namely the effective allocation of capacity!

Darren Dowding is Production Scheduling Manager and has 15 years experience within Cosworth. He explains some of the key planning and scheduling issues that the company faces. "The nature of motorsport is that things break or require continuous development. Consequently we work in a rapidly changing environment with short lead times where there can be no margin of error. It is simply unacceptable to supply parts late. The final customer date, race day, is immoveable." He continues, "The situation is further complicated by each internal team having its own core components in addition to an Ancillary Refurb and Prototype (ARP) division. While the former tends to take a 12-month view of general forecasting, the latter deals with the most immediate jobs ? including bespoke component developments. All of this has to be reacted to, planned for, and effectively scheduled."

Cosworth?s decision to use Preactor for its scheduling requirements followed an evolutionary route that effectively commenced in late 1997. This coincided with the decision to create manufacturing core teams specialising in producing particular components. Each core team is self-contained with respect to machines and manpower, with their own production engineers, schedulers and Team Leader. At this time Cosworth relied entirely on a ManMan mainframe Infinite Capacity System that was part of an MRP system dating back to 1971. When examining works orders, the system would show areas of overload but this then required a manual interrogation of each individual order to see where the overload had occurred. Any changes made to correct this also required a further manual investigation of all the orders to assess the implications of these changes. As Dowding says, "With between 2-3000 orders, this was a permanent job for several people. The only realistic objective we had was to make the orders fit our capacity ? optimising capacity was simply impossible. In addition to the time taken to investigate and affect any changes which wasn?t helped by the textual output format from ManMan, these could only be actioned as a batch the following day."

When the decision was taken to split into manufacturing core teams, tough new lead times were assigned to each team in order to improve efficiency. To aid this, Cosworth took its next step to automating its scheduling requirements by purchasing Microsoft Project. Dowding explains the rationale behind this decision. "We were originally using Excel spreadsheets with a rough 12-month forecast ? for example, we knew that in a particular month we may need 10 engine blocks but that was it. Further details about specifications would emerge nearer the time that they were required. From this overview, we then had to jump to the cumbersome information on the ManMan system. When we moved into teams, it became obvious that the team schedulers needed a more visual planning solution if they were to reach their targets. They simply could not do it otherwise."

To this end, Microsoft Project enabled Cosworth to make a significant leap forward. For the first time, schedulers could quickly see where production processes could be successfully overlapped. It also clearly showed the actual bottlenecks in the system as opposed to the perceived bottlenecks. As such, Cosworth gained a greater understanding of its true capacity. "Within a year, we had achieved our target lead times, in some cases down from 15 weeks to 6," enthuses Dowding.

Whilst the system was very visual and had delivered definite benefits, it was still very manual and very time consuming. For example, when a job had to be moved, the links between all the processes associated with that job had to be re-established manually. And while it had definitely brought a greater degree of education into the company as a whole, it still took a great deal of skill to interpret the answers provided. Dowding elaborates on a further area of concern, "Cosworth has always been committed to obtaining and retaining the very best workforce, which includes their innovative qualities. With this solution we were potentially alienating people from their decision making role in the job and we were concerned that they would end up more like clerks than thinkers."

It was this, combined with a desire to bring Finite Capacity Scheduling (FCS) functionality into the equation that lead Cosworth to investigate a FCS solution. Like many companies, the search began at the Computers In Manufacturing (CIM) show and Dowding recalls visiting a significant number of vendors at CIM 1999. "At the time we had to decide between an ERP solution with FCS functionality that closely matched our requirement, or buying a standalone FCS system. However, 60% of all ERP vendors I spoke to all used Preactor for their scheduling functionality. The other 40% didn?t appear to understand how FCS worked at all. Even the other standalone FCS vendors just looked like younger versions of Preactor."

A conversation with Stan Jonik of Preactor Reseller SFJ Systems essentially finalised the decision. ?It was immediately apparent that Stan was very experienced with the world of engineering and manufacturing as well as software. This was essential for us because we knew that we would only get best value from a vendor who genuinely understood how our business ran and what we needed to achieve. It confirmed to me what I had quickly suspected", continues Dowding, "that Preactor was the original and the best." Six months later, a team from Cosworth attended a Preactor User day at the company's headquarters in Chippenham where all their remaining questions were answered by demonstrating how the Preactor software could be configured to meet Cosworth's needs.

In June 2000, a decision was made to purchase a P300 system that would be run in a standalone context but alongside the existing system. It was decided to install this in the most challenging section because if it could successfully handle the scheduling requirements here, it could handle them anywhere. The implementation consequently took longer than if it had been in a simpler environment, Dowding estimates eight months as opposed two. The configuration required close collaboration between Stan Jonik and the relevant scheduler from Cosworth and involved reviewing and adapting many existing business processes. During this process, it was decided to upgrade from the P300 to the APS system due to the complexity of the processes involved.

In early 2001, a new scheduler was appointed and a decision was taken to train him solely on Preactor. Within a month, this section was running on Preactor alone and confirmed to Cosworth that this was the right way forward. Two more APS systems were purchased in addition to two P300 solutions and the entire implementation was completed on budget, and on time. This included some sophisticated customisation written by SFJ Systems to meet Cosworth's particular requirements to view flow and overlap of manufacturing operations, to get to even shorter lead-times. Stan Jonik describes what this entailed. "The real challenge was to create a generic system for each of the six installations, one that would meet everybody's requirements but which could be moulded into a unique tool for each scheduler. In this way support is kept to a minimum and modifications and improvements can be rapidly spread to all applications".

Dowding recalls the surprising place that Preactor's benefits first became apparent. "We were having a core team meeting in our ARP section which has to handle the most immediate changes and requirements. Before Preactor, the meetings would be full of 'What if' questions that were answered largely by gut feel or intuition which would or wouldn't be borne out over the weeks. This time, the scheduler simply took the 'what if' scenario to Preactor and returned to the meeting a few minutes later with a hard copy of what actually would be the consequences. In that meeting, we saw not only the transition from guestimates to definites, we also saw the beginning of a very short process by which all the teams came to trust completely the information from Preactor."

This in itself has enabled Cosworth to achieve its aim of keeping its workforce involved. By generating quick, reliable, accurate information that is easily interpreted, the workforce was given back a degree of self-determinism that had been impossible with the ManMan system. "One of the best things about using Preactor is that it keeps proving itself over and over again, which creates and maintains confidence from the shopfloor to the customer", comments Dowding. Other benefits include greater visibility, greater flexibility, as the impact of any changes can be seen and actioned immediately and significantly more time to resolve real manufacturing issues. As Dowding says, "Preactor has taken us from a planning situation which would have been impossible several years ago to one where we can alter the entire schedule in a matter of minutes."

As to the future, Cosworth has a series of further developments that it plans to implement. Pre-eminent amongst will be the introduction of Dynamic Materials Control, a feature available within Preactor APS that will allow automatic pegging of individual processes on individual components within a work order. Other anticipated developments include a consolidation of data sources with Preactor acquiring increasing responsibility for generating all relevant data, and a dynamic link to either the existing MRP system or a replacement. Finally, Cosworth also plan to fully automate the scheduling system by integrating it with its Shop Floor Data Collection (SFDC) system.

When it comes to a conclusion, Dowding is emphatic, "When people ask me what has Preactor achieved, I delight in telling them a great deal, but that it still has so much more to offer due to its incredible flexibility. I believe our existing system has at least two years further refinement which will only continue to increase its worth to us as a company."