Integrated Manufacturing

Real-time shop-floor data collection, scheduling and performance analysis offer dramatic improvements in productivity

Most UK manufacturers are having a tough time, faced as they are with increasing competition from developing economies in eastern Europe and the Far East, as well as the pressures of global sourcing. Engineering companies are addressing these challenges in a number of ways, including the use of advanced machining tools that are capable of one-hit machining and lights out operation, thereby minimising the labour cost advantage enjoyed by many overseas competitors. That said, efficient productive machines do not necessarily equate to efficient production; many managers will stress that it is door-to-door times that matter – not floor-to-floor.

One-hit manufacturing certainly minimises the number of operations that an individual part will be subjected to, but most machine shops that are involved in diverse small-batch production with some 15 or more machines – regardless of their type and capacity – will find scheduling and maximising overall efficiency a headache. Indeed, David Trowell, managing director of Seiki Systems Ltd, Brighton (Tel: 01273 680411) – a long-established company offering a modular approach to the total IT solution for manufacturing – says that it is almost impossible for these manufacturers to achieve an efficient operation without an efficient integrated manufacturing system.

“I wonder how many production managers can state with accuracy: the real-time status of their work-in-progress; the capacity that they have available and the location of any bottlenecks; which jobs are on time and which are late; and the consequence of any delayed operations, including subsequent ripple effects. If a specific operation is delayed, all subsequent operations on that machined will be delayed; and if that part requires subsequent transfer to other machines, the operations on those other machines will also be late, leaving what could be significant holes in the production schedule. If you don’t know how to accommodate this ripple effect, then I fail to see how you can effectively run your business. In the current competitive business climate, it is imperative that UK manufacturers use every means available to optimise their efficiency.

“If you can’t readily see what capacity is available, how can you utilise it? It could be that you will have 50hr of spare milling capacity in seven days time, but you won’t be able to sell it if you don’t know that it’s there. Conversely, what a loss of credibility – and perhaps money through penalty clauses – of you make delivery promises on capacity that is not available.”

Real-time benefits

Mr Trowell says that, to achieve high levels of visibility and control, manufacturers need access to real-time systems and to real-time information on many factors relating to machine operation status. “Paper-based  systems can collate data and even present data graphically, but this will be some time after the event and, in all likelihood, will not be accurate.”

The type of data that establishes the real-time status of work-in-progress, available capacity, which jobs are late and the consequences of delayed operation is derived from what Mr Trowell calls shop-floor data collection (SFDC) at the Component Level. This is one of our four key areas of SFDC which, with the appropriate software systems, can automatically present and archive – in real time – all relevant manufacturing data at the touch of a button.

Another consideration is those areas of manufacturing where Key Process Information (KPI) is an essential part of product acceptance – such as in-cycle probe data and fixture torque settings that have to be recorded for every component in a batch. That said appropriate software can make further use of this data. “For high-level products such as those found in aerospace, you need to know what the associated data was for a specific batch, record this data and distribute it as necessary. With in-cycle probing and un-manned operation, we can measure successive parts to establish any tolerance drift and distribute this data to an SPC system. This might predict that parts will be out of tolerance after the seventh component. With predictive analysis software, we can even forecast at what time this will happen and automatically initiate appropriate action. All relevant data is recorded automatically and without any paperwork. Indeed, in many cases if you can’t supply KPI data, you will not get the job.”

Mr Trowell’s third key area of SFDC is quality level data, of which a good example is NC programs, together with related manufacturing documents such as solid-model drawings and sub-sets of component drawings. The main areas of interest are: access control (who accessed the program); batch details; and details of any changes to the program (such as changes to the feed rate to eliminate tool chatter), because – in many cases – not recording this can be a serious contravention of quality standards. Edit approval is also included.

“All this data forms a history log for the audit trails demanded by specific quality standards, and there are many things that can be done to ensure that the standards are complied with to the letter. For example, we can prevent an operator getting his next NC program until he has returned the current one that he has just edited. Systems can log the time, date and batch details, and even provide a summary of the program changes made; and when the program is returned, the system flags up to the programming department that the program needs verification and approval before it can be used again – an action that is also logged.”

Machine performance

The fourth area of SFDC relates to machine performance: it is overall equipment effectiveness (OEE), which – with the appropriate software – can be calculated within seconds of a job being completed. This is in marked contrast to those companies employing people whose sole job is the manual gathering of data with which to calculate OEE figures – a laborious task that often takes days or even weeks to complete. “An increasing number of companies are stipulating that these figures must be available with each batch of work, because they give an indication of the supplier’s efficiency; if the figures are low, they may indicate that too much is being paid for the parts. Within the past 12 months, most major manufacturing companies have been pushing their suppliers to supply this data.

“Senior managers no longer need to wade through reams of data relating to cycle times, stoppage times and the like; for a whole plant, there can now be a single sheet of numbers that gives percentage figures for OEEs. A world-wide manufacturing company can look at cumulative OEE figures for all its sites and immediately pin-point plants that are under-performing. At a local level, senior managers can look at data for groups of machines – even individual machines – to see where attention needs to be focused. This type of performance data can also be used to justify the purchase of new equipment or machinery, and to highlight machines that are not being fully utilised, or machines that are suffering a particular reliability problem.

“If you don’t have access to this detailed breakdown data, introducing efficiency improvements becomes extremely difficult and somewhat haphazard. You have to know exactly what is dragging efficiency down, and you will not get the type of data needed if you are relying on manual systems and on operators telling you what problems they have been encountering. You will only find out what happened last week, not details of trends and the reasons for these.”

Real Time Scheduling

ERP (enterprise resource planning) is the system that drives the entire order management process, undertaking activities that include estimating and quotations, order management and integrated stock booking, process planning, stock management, purchasing administration, manufacturing requirements management, cost calculations and comparisons, invoicing and delivery management, plus management reports and statistics. However, the information that is important to the production manager is: what parts to make, in what order and by when.

A real-time scheduler will interrogate the MRP data, and it will establish the set-up time, the production time and the number of parts for each job, then calculate the total production time. The system combines this data for all machines, equipment and related processes for the forecast production, and graphically depicts all the data – including part numbers, individual operations, due dates and forecast finish dates – on a calendar that incorporates shift patterns.

At a glance, users can identify spare capacity, bottlenecks and jobs that will run late; they can also assess – in seconds – the effect on the finish date of running specific machines through lunch breaks, extended shifts or additional shifts. A real-time scheduler also allows the user to assess the implications of changing the scheduling of jobs, together with the multiple consequences for subsequent operations on other machines. Moreover, once the user validates a job in the schedule, the system confirms that all the relevant manufacturing data – such as verified NC programs – is available. If it is not, the system will send an intranet message to the CAD/CAM department, saying that a specific program is needed by a specific time.

“With these systems, the user not only sees the manufacturing status in real time, thanks to the integration with real-time SFDC, but can also make predictive decisions and solve problems that could occur in a few days time. For example, if a machine goes down or into alarm condition, the user is not only immediately aware of this, but also able to see the full implications of this condition on the schedule and assess various options. Furthermore, for every second that a particular machine is down, the scheduler is pushing back all the subsequent jobs on that machine; and of there are subsequent operations on that job on other machines, they too are being pushed back in real time. I just don’t know how a busy manufacturer can run a reasonable-size machine shop efficiently without real time facilities.”

The real-time products highlighted here are a natural extension to existing systems. The development path of many companies will have seen the separate installation of various automation systems, including ERP, CAD/CAM, DNC and even paper-based monitoring. These systems are essentially separate ‘islands of automation’ and, once companies have got this far, efforts to improve performance tend to run out of steam. However, the availability of real-time systems is changing the situation. “Users can now buy modules that plug into this ‘jigsaw’. These modules do not replace anything, but they bring the likes of ERP and CAD/CAM together at the point of manufacture with shopfloor information delivery systems and real-time data collection.”