MIS begotten gains

EWOS produces compound feed for rearing fish of the salmon family ­

principally Atlantic salmon and rainbow trout. The company’s factory at

Bathgate in Scotland employs 60 people producing some 60,000 tonnes per

annum of pelleted feed on three parallel, highly automated production lines.

Recently, the company made a substantial investment in a management

information system supplied by IEA Ltd of Cardiff, to enable its teams to

improve the performance of the business.

Interdependent production

In the continuous search for performance improvement, EWOS contacted PA

Consulting who undertook a review of the plant and its procedures. Part of

PA’s recommendations was to invest in a management information system (MIS).

Armed with the information from an MIS, management would be better able to

focus investment in new plant, plant improvements or in improving best

practice. PA also insisted that for a complex interdependent production

process such as that operated by EWOS, information collection and analysis

would need to be automated in order to provide the speed and accuracy

needed. However, automated data collection and sophisticated database

analysis is what normally makes MIS a very costly tool.

IEA’s solution differed from classical MIS in two critical ways. Firstly,

IEA was able to derive the vast majority of the data needed from existing

automation systems, and secondly, IEA used Windows NT Client/Server-based

packages for collection, analysis and presentation of information. EWOS

production processes are controlled by a number of islands of automation: a

BMS system for the stock control and intake; a Datastor system for batching

and the extruder; a Wenger system for the dryer; and an Intertech system for

the vacuum coater. The MIS has to interface with all four to collect the

required information ­ not a simple task. Some data, such as reasons for

machine down time, inherently require judgement and manual entry. Other

data, such as the results from quality control sampling processes, are

manual simply because the cost of mechanising the testing is prohibitive.

With a few exceptions then, data collection at EWOS is fully automatic. The

volumes of data are huge. Thousands of measurements are logged every minute.

To put it into perspective, the system generates 600MB of data per week.

At the other end of the system ­ management information reporting ­ the MIS

breaks the business process into three parameters: utilisation, yield and

efficiency. This can be applied to the entire factory, or to any section of

plant or process. The product of these, applied to the whole operation, is

the measure of how well the business is performing. A simple graphical

display of this one variable showing current actual compared to the average

for the last three months and the average for the last 12 months is a

powerful indicator of business performance. By adding the highest and lowest

short-term excursions to the graph, a simple graphical display of spread or

variability of the business performance can be provided.

Whilst a very useful tool for monitoring progress, these simple graphs

clearly do not provide enough information to analyse problems or inform the

planning of change. For these purposes, users employ a technique called

‘drilling down’. Simply put, drilling down means investigating the

successive levels of detailed measurements underlying a trend until the

cause can be identified. This sounds simple, but it is easy to become

overwhelmed with data while remaining starved of information. IEA’s system

uses a product from Oracle, called Express, to display information as 3D

stacked bar graphs (left), and allows users to display any parameter as a

function of any two others. These Ocubes’ are startlingly effective.

Of course, to make good use of the system, there must be some focus. The

sudden wealth of information can lead to starting too many competing

projects, resulting in effort being spread too thinly and little being

accomplished. Barbara Wood, production manager at Bathgate, emphasises this

point strongly: “Without MIS, the problem is how to measure processes,” she

explains. “With MIS, the critical management skill shifts to deciding what

to focus on. We decided to focus first on yield, and specifically on the

drying part of the process.”

It was well known to EWOS that more moisture ­ desirable in a product sold

by weight – had the effect of making the pellets more fragile and so more

likely to break up in the coating stage. The key question was: what is the

optimum level? The issue had defied analysis by manual methods. Using the

MIS, the impact of small changes in moisture content was analysed and the

value that produced the optimum result for the whole business was

determined. During this process, it became clear that achieving this optimum

value was so important to the business that it was worth investing in

improvements to the dryer control system to tighten up the already small

(2%) achieved spread of product moisture.

Together, these improvements have shown a significant return on investment.

During the same period, production of out-of-limits product that has to be

recycled was cut by 95%.