Xore's Boxray on-line elemental analyzers will already today let you have the performance you think you will need in a few years - but you actually need much sooner than that. Given the long service life of the analyzers it is good to keep this in mind, so you do not buy an analyzer today that will need to be replaced on only a few years.

There is a lot of talk in the mining industry about Big Data and Industry 4.0 and other buzzwords. What everyone seems to agree on is that there will be a lot of data and your process will be mapped and even controlled by algorithms.

What is not often discussed is how the models for processes are made. They are of course based on data, but just how much data do you need? Regardless of if your process is controlled manually or automatically the data that is the foundation for decisions needs to give an accurate image of the current situation, and also a sufficient history to show where things are going.

If you in the future plan to control the plant automatically to some degree, it is also easy to see that at that time you need a lot of data. What is not always obvious is that in order to make the models and algorithms needed to achieve this - you need the same amount of data, if not more, to be able to do this.

In other words, the data collection needs to start years ahead if you are to develop a well-functioning control system.

Process control is trying to answer some basic questions; where are we, where are we going and are we turning in any direction?

Another aspect is the history of the last 20 minutes or two hours, that you use to spot the trends in the plant and very importantly spot when the trends are changing. To do this you need a few samples to get an idea.

The problem you may run into with slowly collected data is that the oldest samples can be from a few hours ago, and that they are simply not relevant any longer. You do not just need a number of samples, you need all of the samples to be relevant.

The Boxray 24 and Boxray Compact analyzers have already today the fast analysis time you need to start building a database for your future needs, but you can draw benefits from that even if those plans are far away.

More data, and more accurate data, will let you make better decisions from day one no matter how you control your process.

How much data do you need? Simple sampling theory says that you need to gather data twice as fast as the rate of change you are recording. If the metal content in a certain place in your process changes every ten minutes you need to record that metal content every five minutes, and that is the absolute minimum.

In the plant there is easily 20 different locations to sample, and that is where you realize that the time available for analyzing a sample is a lot less than that. Not all samples require the same high sample rate, but time is still critical.

Another aspect is the history of the last 20 minutes or two hours, that you use to spot the trends in the plant and very importantly spot when the trends are changing. To do this you need a few samples to get an idea.

The problem you may run into with slowly collected data is that the oldest samples can be from a few hours ago, and that they are simply not relevant any longer. You do not just need a number of samples, you need all of the samples to be relevant.

The Boxray 24 and Boxray Compact analyzers have already today the fast analysis time you need to start building a database for your future needs, but you can draw benefits from that even if those plans are far away.

More data, and more accurate data, will let you make better decisions from day one no matter how you control your process.