Real-Time Production Analytics
With the internet of things gaining traction and more and more businesses getting more data driven in their decisions, it has given rise to the industrial internet of things. The output of which results in the ability to perform real-time analysis on key problem areas and solve issues as they arise.
The practical deployment of industrial internet of things has been a bit of a dark art and not achievable for many, as business have legacy technology and equipment that is not quite geared for modern internet driven analytics, cost and technical resource barriers also inhibit businesses from leveraging the power of real-time inter-connected analysis.
Modern developments in PLC to TCP/IP interfaces have allowed older technology and existing investments in PLC technology to be brought into the modern ethernet age by allowing real-time data to be easily populated and streamed into databases for analysis. Allowing all businesses to start exploring the realm of the industrial internet of things.
The business benefits of leveraging the industrial internet of things is the ability to provide a feedback loop that directly impacts changes within the production environment in order to reduce variance, equipment downtime and measure productivity changes. An example of which is measuring the volume variance of packing equipment within the tile adhesive environment which has a direct impact on ISO quality requirements and ensuring customer satisfaction through strict specifications.
So how do we measure production variance in real-time and what are some of the outputs we can expect. The first decisions to be made are over how many samples do we want the variance to be measured and in what intervals do we spread our specifications.
Example: We have one machine that can pack one bag of 20kgs every 5-9 seconds and we would like to measure how many bags fall within a certain specification range ie 19.0 kg – 19.2 kg or 21.00kg – 21.20 kg over the last 30 bags and if indeed we have a normal distribution around our required specification of 20kgs per bag.
Building a data model within OQLIS would looks like this.
SELECT A.Row, A.BagWeight, A.BagWeight AS "Bag Weight Min", A.BagWeight AS "Bag Weight Max", A.BagWeight AS "bag Weight AVG", A.DateTime, A.BagCountSequence, A.PackerHead, A.LowerBound, A.UpperBound, A.Closest50gram AS "Bag Weight Closest 50 grams" FROM (SELECT ROW_NUMBER() OVER (PARTITION BY PackerHead ORDER BY DateTime DESC) AS Row, PHT.BagWeight, PHT.DateTime, PHT.BagCountSequence, PHT.PackerHead, 19.8 AS LowerBound, 20.4 AS UpperBound, CASE WHEN BagWeight <= 19.8 THEN -1 WHEN BagWeight >= 20.4 THEN -1 ELSE 0 END AS InBounds, ROUND(PHT.BagWeight / 5, 2) * 5 AS Closest50gram FROM PackerHeads PHT) AS A WHERE Row <= 30 AND A.BagWeight > 0
Exploring this data model through OQLIS and plotting the data for each machine with upper and lower bounds will give us an instant visual indication of where the sample results are trending.
Adding a chart for each machine to the dashboard will allow to view the detail of each machines performance over a given sample size.
Which will allow real-time access to statistics all within the production environment allowing a constant feedback loop on key data and allowing fast resolution of any issues that may creep up.
If you want to enable real-time analytics across your data contact firstname.lastname@example.org.