How SPC and AI Helped a Choc Ice Factory Reduce Giveaway and Save £800,000

From Choc Ices, to Chocolate Bars, to Chocolate Biscuits, and the ever increasing cost of chocolate, the need for accuracy in Chocolate enrobing was never more important than it is today.

So many products are produced with a layer of chocolate, sometimes containing nuts or other tasty ingredients. To the manufacturer, it is vitally important that the layer of chocolate is thick enough to guarantee a high quality and highly reproducible treat, whilst not being unnecessarily thick causing higher and possibly uncompetitive manufacturing costs. The optimum between taste and cost is usually achieved through trial and error during pilot production runs, prior to full scale production.

Once the optimum thickness has been determined, it must then be controlled during large scale manufacturing by statistical sampling. Sampling involves the taking of sample packs from the production line both before and after the chocolate is applied (known as chocolate enrobing).

Over-sampling costs money, due to wasted product, whilst under-sampling possibly leads to inadequate control, resulting in held or quarantined product.

Case Study

We installed a system in an ice cream manufacturing plant a few years ago to begin their transformation from paper recording to full digitisation. Apart from the obvious benefits of paper elimination, this transition was necessary to increase visibility and encourage real time actions to improve performance, sustain consistently high quality and reduce manufacturing costs.

When installing such a system we will always begin with one critical aspect of control rather than trying to improve everything at once. We chose one of the choc ice lines on which to run a weight control improvement programme. Concerns had already been raised due to the ever-increasing cost of chocolate. Though there was no concern about the quality of the finished product, the team wondered to what extent it would be possible to reduce the chocolate enrobing, whilst still producing a consistently high- quality product.

The Chocolate Reduction Process

We created a Statistical Process Control quality profile within the Harford system for the choc ice production on line one. For the first day, we simply monitored, to measure the weights of samples, both before and after chocolate enrobing. We found 5 samples taken every half hour throughout the day gave us quite consistent results. This statistical data gave the team the confidence they needed to try reducing the thickness of the chocolate coating. This was necessarily a gradual process so as to ensure the quality was not compromised during the systematic trials of reduced chocolate enrobing. Once the optimum was found, production line set-up changes were standardised for all shifts.

We have no wish to bore readers with the percentage reductions which led to optimisation, but we could work with any teams on any products to ensure similar outcomes. Whilst the products and factory locations may change, our system of SPC control delivers similar results. One reason Harford Statistical Process Control is so effective is that it also calculates the process capability in each case. This reduces the risk of overdoing it or over adjusting on too small an amount of data. Consequently, we can say with total confidence that we can optimise enrobing, whatever the product and whatever the extent of variation, without compromising the product quality.

SPC/AI

Our Statistical Process Control SPC, in conjunction with Process Capability, also invariably means fewer process adjustments and, therefore, less work for the operators. In fact, as the Harford Statistical Process Control weight/volume control programme, now over 30 years old, is one of the earliest examples of Artificial or Automated Intelligence. In modern parlance, it could more accurately be called Harford SPC/AI.

Dramatic Results

Back to the choc ice line, when we started working with the team to reduce the excess chocolate enrobing, the results were quite dramatic and within another week we had achieved the optimum…an acceptable level of chocolate take up with no loss of quality and fewer process adjustments.

The detail is irrelevant as it will be different for every factory and every product. Even another choc ice line would have different parameters sufficient to vary the process capability and control limits.

What’s important is to know how successful it was. The team, after our initial training and coaching, saved more than £800,000 in the value of chocolate no longer overpacked for the first year of operation, repaying their investment several times over.

But, we still had to add in the benefits from the use of paperless quality, label verification and OEE, plus the weight control benefits expected across a further 10 lines.

More about these other potential savings in future articles.

For more details contact Harford Control at sales@harfordcontrol.com or give us a call +44 (0)1225 764461.