In a digital world, why is value still hard to find?

Organisations are gathering more data than ever before — but turning that raw information into value can be a major challenge.

At all levels of an industrial system, from the technician in the field to managers in the boardroom, there are a number of barriers that make it difficult to take advantage of the collected data, particularly in the fast-paced world of operational (OT) systems.

Issue 1: Scattered and siloed information

The information businesses have tends to be scattered — not just physically (having to go to several different places on-site to access information), but also in the various role-specific systems and platforms in use.

It’s important to note that this is not always a bad thing. Forcing people to use an unsuitable enterprise-level tool causes inefficiency and staff resentment — the staff deserve the right tools to do their jobs. But if you want to get a full understanding of the system or its parts, it can be quite painful to access all of these systems individually.

Consolidation is the process of gathering all of the various sources of information so they can be accessed in one, consistent place. This includes:

• Live sensor values
• Historical recordings
• Event databases and logs
• Documentation, manuals and training videos
• Downtime reporting and scheduling
• Maintenance information
• Staff and roster information
• Quality and inspection feedback
• Online databases (market rates, weather etc.)
   

Figure 1: Scattered and siloed information. For a larger image, click here.

A single point of data by itself isn’t very useful, but when there is a variety of different types of information from many sources (including non-sensor information), a wide range of analytics can be performed — from simple calculations to building complex machine-learning models. From a basic trend to power-efficiency analytics that take complex system states into account.

Issue 2: Complex identities

Some systems are well structured, so it’s easy to find the needed information. Unfortunately, many others — particularly on sites that have been around for some time — have a huge number of data points with names that aren’t readable, even to experts.

It’s not uncommon to spend hours looking for the point of data needed, only to find that the one you thought was right was taken out of service several years ago.

Even the names of hardware can vary — a single part might have a different name for the electricians than it does to maintenance and process people.

This messiness makes it extremely hard for people to engage with the data collected, even for the most technical and long-serving staff. For those who are new to the system or less technical (such as analysts), it’s impossible without needing the constant support of the engineering team.

It’s good practice when consolidating data to ensure that it’s human readable. Instead of having to know that the main tank temperature is PLC4.AIC1.42_AI, an analyst can go to the ‘Main Tank’, an operator to ‘TK_04’ and maintenance people to ‘TK4’ and they can see the temperature — along with all of the other details captured about the tank.

Issue 3: Data ownership and replication

Your data is exactly that — your data. It’s frustrating to come across vendor lock-in, where data is locked away or difficult to access. Or sometimes the opposite is true, and it is necessary to send the data up to the cloud and hand proprietary information to third parties, just to do relatively simple tasks that could easily be done on site.

Having data replicated in multiple locations causes its own unique issues. Which values should be believed when there are disagreements? How to deal with forecasts, which are based in the future and will change over time? If depending on feedback from the cloud, what happens to cloud data when the internet connection is down?

The ideal solution is that the consolidation system should deliver and not replicate the information. Instead of being yet another data store, a consolidation system should leave important data where it already is, so information is always up to date and reliable.

Issue 4: The OT/IT barrier

The technology of the operational (OT) network is very different to that of your IT and enterprise networks.

Outside the usual firewalls that keep them apart, the biggest difference between the two is the timescales they operate at. Management decisions happen in days or weeks, but process decisions need to happen in seconds. In the OT network, data has to be brought in and shared fast enough to help those quick decisions

That difference in timescale is why so many IT technologies fail in the OT network.

When consolidating information, the ability to gather data from a variety of different network layers and technologies is needed, from direct access to industrial hardware to cloud APIs. You also need to work with not only different time zones, but the radically different timing of enterprise and automation systems.

Issue 5: Change brings chaos

Change is unavoidable. Over time, new sensors will be added, equipment retired and even pieces of the data backend replaced.

Unfortunately, many custom data-driven solutions end up being quite fragile to change — they fail when data points are renamed or sources of data are moved.

But custom solutions have huge potential to help find value in your specific systems, workflows and operators. If the fragility wasn’t an issue, it would be possible to create unique value to uncover genuine commercial advantage.

A beauty of having consolidated, human-readable data is that it can help abstract the data, creating a ‘buffer’ between the sources of information and the data-driven values it can give (such as reports, analytics, AI and models).

The benefits of abstraction include:

• Data-driven solutions (alerts, reports etc.) can automatically include new and remove retired equipment, without any human intervention.
• Data backends can be changed or even completely replaced without disrupting any data-driven solutions.
• A single solution can be deployed across a number of similar facilities (such as multiple sites) without needing to make individual changes.

Issue 6: Lack of openness

It is of course very important that this data be open. There are a huge number of open-source and publicly available tools to help with data analytics, and they’ve never been so easy to use.

Easy and consistent access to data through an open, consolidated solution means you can innovative solutions such as machine learning in your process systems.

Bringing it all together

Consolidation is the first of the three key steps in turning data into business value. After this comes analytics (where value is added to the data) and delivery (where it is delivered into the hands of the people and systems that need it).

Fortunately, software systems like ARDI (from Australian company Optrix) are now available to help address the issues talked about, informing and empowering staff and driving quality, production, waste and energy improvements based on real-world data.

While many are still cloud-based and management-focused, there are new solutions, including ARDI, that are designed for low-level, high-speed process data. They run on site, are scalable for even small applications, and are focused on delivering real outcomes to people at all levels of the organisation.

Top image credit: iStock.com/NanoStockk