18 August 2017
Blogs by author: Andy Rowland, Head of Customer Innovation: Energy, Resources and Manufacturing, BT.
If you were asked about the Internet of Things (IoT), what would come to mind? You would probably think about connected cars and homes, smart cities, and wearables. But would you think about the use of IoT in manufacturing or ‘heavy’ industry?
According to McKinsey, business-to-business (B2B) applications of IoT will probably create more value by 2025 than consumer uses, equating to nearly 70 per cent of the potential value. Yet General Electric found that only 3 per cent of industrial data is currently being used in a meaningful way. But as the price of sensors, storage, compute power and connectivity continue to fall, harnessing IoT data can drive dramatic efficiency gains for commercial organisations.
Data is fuelling the growth of Industry 4.0
Industry 4.0 is all about collecting industrial data from different sources into a data lake, and then running advanced machine learning and artificial intelligence (AI) to predict equipment failures and streamline production processes. Currently that data is kept in silos: enterprise resource planning systems, industrial control systems, and spreadsheets and documents used by engineers to keep track of supplies and people. But companies who are making efforts to make sense of all this data are gaining some really valuable insights.
As an example, one of our oil and gas customers was running very expensive electric submersible pumps, without any detailed predictive analytics. This meant that they would run until they failed. To avoid disruption, spare parts priced at $500,000 were kept nearby. But using Industry 4.0 analytics, they’re now able to compare pumps from 12 different manufacturers, work out which ones are the most business critical in terms of throughput and then predict likely failures. They can see which pumps are suffering from issues, e.g. excessive temperature or vibration, and give operators prescriptive guidance on how to extend their life. The solution initially monitors 200 pumps, and 197 parameters, using nearly 1000 analytics models. This equates to three million records a day, with data uploaded every 10 minutes. Following two man months of work, they’ve already seen a 20 per cent efficiency gain.
Securing your data and managing risk
When it comes to industrial data, security is paramount, as once connected, industrial control systems may become vulnerable to cyber-attack. In the past few years there has been an increasing number of cyber-attacks on the industrial control systems that underpin manufacturing plants and critical national infrastructure.
These days it’s impossible to keep the ‘bad guys’ out, and in fact they may already be within your organisation. This could be anything from a disgruntled employee to someone who clicks on the link from a phishing email. Security becomes even more of a concern if Industry 4.0 is used not only to guide employees on how to make efficiency gains, but for command and control (where the analytics decide on which processes need to be changed and then automatically execute against this). So good security housekeeping is critical, as is the use of advanced threat intelligence.
As Industry 4.0 analytics gets to the heart of company processes, data privacy also becomes a concern. How you optimise production, streamline maintenance, and use research to create value all involves highly sensitive data. Consequently you need to consider how you will process and transfer this data.
How technology can help
When you need to run analytics quickly and feed the results back into local processes, edge computing and storage can be the answer. You’ll also need private cloud compute for when you want to collate and analyse data from multiple sites. Between the two you will need a secure connection, with a good quality of service like MPLS, with suitable mechanisms to authenticate and protect your data. For authentication you might want to consider public key infrastructure (PKI) or Blockchain, and in the future quantum key distribution (QKD) to ensure data hasn’t been accessed on route.