Jack Welch once said that the 1980s would be a ‘white-knuckle’ decade of intensifying industrial competition – and that the 1990s would be tougher still.
In many ways, however, the 1990s were just the start of a massive reshaping of the global economy that will continue for the next 10 to 20 years. A recent McKinsey research report shows that the `topple rate' at which companies lose their leadership positions, has doubled in the 20 years to the mid-1990s. The driving force behind this process of ever intensifying competition is primarily communications technology, which is giving rise to globalisation and economic liberalisation in those pockets of the globe that just a decade ago were still fairly isolated. In this environment of ever increasing competition, manufacturing incumbents are seeing smaller, low cost and more agile competitors gaining increasing market share. Large corporations are finding that their decision-making cycles are just too slow for a world requiring realtime remedies. The answer is to adapt or die, to become as nimble and adept to change as the new rivals that are using technology to their advantage.
Within the continuous manufacturing environment, the smooth running of the production process is of paramount importance. Yet feedback on plant performance often occurs only on a weekly basis. By the time downtime, slow running and quality losses are recorded; valuable, irrecoverable production time has been lost.
Solution chosen by consumer goods companies
To address this need, Pragma developed On Key Performance Manager, a scalable solution that allows clients optimal benefits that suits their particular budget constraints. The solution can be implemented at three distinct levels of complexity, each providing a compelling cost/benefit advantage. Performance Manager has been running for a number of years at some of the largest consumer goods companies in South Africa including Parmalat, Clover, Nampak and KWV.
An effective response - the 80/20 principle
Before we discuss specifics, let us look at a few general principles of performance measurement. Firstly it should be noted that the objective of performance measurement within the continuous manufacturing industry is to calculate OEE (overall equipment effectiveness) and OPP (overall plant performance). OEE is a simple multiplication of three measures - availability, production rate and quality rate, each expressed as a percentage in design capacity. OPP adds a fourth term to the multiplication - the amount of time actually worked as a percentage of time allocated as operational time. This might be influenced by things such as strikes, or lunch breaks that run too long. OEE and OPP can be measured against any constant, ie, production line, operator, product or shift.
Since performance measurement can produce a flood of information, Pragma contends that the objective should be to focus on the 20% downtime reasons that caused 80% of the downtime. It is an established fact that many phenomena in life follow the 80/20 principle. 20% of church-goers will give 80% of the offering and, 20% of the population will control 80% of the money and 20% of all people will cause 80% of all car accidents. This is called the Pareto principle.
Manual capturing
The first level of complexity involves the manual capturing of performance information, the only requirement being workstations where data can be captured. At this level performance information is restricted to downtime duration, downtime reasons and quality losses due to product rejection. While speed losses are not measurable under this system, OEEs can still be calculated since production speed can be deducted from the operational time and production quantity measures. This figure is however an average for the production period and does not represent realtime data.
This system does, however, have its limitations. The most serious impediment being that downtime data is not always captured, especially for minor stops. Over a period of time these minor stoppages can add up and cause serious errors in the recorded data. There is also the danger that data is not accurately captured, either on the job cards or when the information is fed into the system. Speed loss reasons remain unknown while quality loss reasons might also be unknown, in effect limiting analysis to downtime reasons.
PLC I/O sensor capturing
The next level of performance measurement allows for the automatic capture of downtime reasons. Most manufacturing concerns already have PLCs that provide information from sensors on the machines. These can be read into the On Key system without additional capital expenditure for the client, providing 100% accurate downtime profiles. This allows for exact machine utilisation information. If shift periods are now logged onto the system the performance during specific shifts can be automatically calculated.
Actual realtime running speeds per product can now be compared to design speeds, but variances do not have reasons attached to them. Quality rate information obtainable from PLCs can now be fed directly into Statistical Process Control, ERP and quality systems, bypassing the need for manual (and error prone) transferring. In addition to the PLC readings required for OEE calculations any other measured variables such as temperature, pressure and vibration can also be captured into the system and shared with other systems.
The system now provides realtime information on availability, production rate and quality rate, but is still subject to data errors in that not all downtime reasons are discernable from PLCs and that speed loss reasons are unknown.
HMI capturing
The third and most advanced stage of performance measurement requires the rollout of HMIs (human machine interfaces) on the plant floor. HMIs allow for an additional layer of information to be obtained. This data collection process begins with the operator that has to log on in order to operate the machine, allowing performance to be evaluated on an individual basis. Whenever a machine stops for longer than a predetermined time the operator is required to insert a downtime reason. This now also applies to reasons not available through PLCs. Speed loss reasons are also captured in this way, once speed losses exceed a specific limit.
In order to ensure accuracy when data is being entered operators do not choose from a list of codes, but have to select the actual reason. Their options are also restricted to options pertinent to the particular operator and machine.
The OEE information is now as accurate as can be expected, broken down into specific reasons per efficiency loss type and up to date virtually to the second. Now that the data is accurate the On Key software allows the user a multitude of statistical tools for drawing comparisons and optimising production.
Actual production can be measured against the production plan, which is directly downloadable from the ERP. Production rate per product can be optimised and actual speeds per product per machine can be compared. The production schedule is also available on the HMI, allowing the operator to measure himself and make required preparations for upcoming production runs. Overall accuracy is improved through the elimination of paper-based production schedules. Order fulfilment information as well as all other production information is directly sent to the ERP. Measurements that might need to be done manually on the production floor are captured into the SQL database. The end result is one database containing all process related data.
A tight ship with minimised losses
The end result is a paperless production environment requiring no post manufacture administration. Production information is directly available to management, greatly improving its decision-making capability through accurate and effective yet simple graphs and reports.
For more information contact Louis Volschenk, Pragma, 021 943 3900, [email protected], www.pragma.co.za
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version