Someone to watch over machines

Early adopters achieve multimillion dollar savings soon after deploying latest sensor-and-AI-driven asset health monitoring and diagnostics solution.

It’s a marriage of trusted wisdom and new tech smarts, pairing decades of experience with cutting-edge algorithms, to deliver a new way of monitoring machines in heavy industrial operations. 

Since Baker Hughes announced its alliance with Augury in October 2021, there’s been much work to combine the startup’s tech with Baker Hughes' own asset-management solutions

The result is a solution called Machine Health, an end-to-end machine monitoring and diagnostic solution that combines Internet of Things (IoT) hardware, software, and services enabled through purpose-built Artificial Intelligence (AI). It is designed to deliver prescriptive insights and help customers reduce downtime. With full-stack sensing, outcome-focused AI, and customer success, it doesn’t just flag a machine issue before it leads to downtime or failure - it diagnoses the problem and prescribes a solution. Machine Health is part of Cordant, a new integrated suite of solutions for asset performance management and process optimization.  You can find out more about it here.

The time to value from when customers begin monitoring the health of their machinery with the sensor-and-AI driven solution is typically 2-4 weeks, with return on investment (ROI) coming in at 300% in the first six months.  This is much faster than general-purpose AI where results like this can take years to achieve.

Baker Hughes, with its Bently Nevada product line, has been at the forefront of condition monitoring and protection for over 60 years. “As digital transformation became increasingly more critical, we knew that AI would be a key component to deliver the next wave of outcomes. You could say our experts know as much about machines as doctors know about the human body. We’re combining that deep knowledge with Augury’s strength in the AI tech space for machine health,” explains Sequoia Murray, Bently Nevada Global Customer Success Leader, Baker Hughes. “We’re connecting AI and our historical knowledge to bring this solution to heavy industries.”

Launching the INVISTA pilot

In September 2022, the rubber hit the road when the team deployed the first commercial implementation of Machine Health at INVISTA’s Kingston nylon manufacturing plant in Ontario, Canada.

‘’We fully realize that we have to embrace technology to become more efficient and more sustainable. And that's what we're trying to do in our plants,” says Bonorden.

Murray adds, “INVISTA is a long-time partner of Baker Hughes and already uses our System 1 software and Ranger Pro wireless sensors.’’

“They wanted to further decrease maintenance costs and reduce unplanned downtime,” she says. The expectation was that more value could be unlocked by adding prescriptive analytics to the existing hardware and software infrastructure.

The pilot was deployed across 50+ machines – including pumps, agitators, extruders, blowers, fans and conveyors.  Results flowed fast. “We started getting data and generating value immediately after the installation.” Two weeks after deployment, Machine Health made its first “machine save” when the machine health correctly flagged an anomaly, which pinpointed the root cause, and prescribed the mitigating solution. It saved INVISTA US$110,000.

The power of a purpose-built algorithm

Augury cut its machine monitoring teeth mainly in the food and beverage manufacturing industry. The alliance with Baker Hughes broadens its industry expertise. “We’re in the heavy industrial space, so in general we will see larger value generated,” explains Carlos Gomez, who leads the Baker Hughes Augury Strategic Alliance.

Machine Health’s AI algorithms run on decades of high-quality data collected from comparable industrial machines and set-ups around the globe. “Our AI algorithms have been built on more than 300 million run time hours across more than 100,000 machines” says Murray. Leveraging this data reduces time to value, time to ROI, and time to deploy which differentiates it from more traditional AI that has historically over-promised and underdelivered.

“We have purpose-built sensors operating with a purpose-built platform, which means we can plug and play very quickly,” says Murray. While 2-4 weeks to the first ‘save’ is the baseline, Murray says time to insight, or the first machine ‘flagging’ can be as little as 1-2 days.

$4.2+ million saved in under three months

Since deployment, the INVISTA Kingston pilot has delivered more than US$4 million in savings, with Machine Health flagging everything from process improvements to avoiding costly catastrophic failures.

The plant’s reliability program is managed by a lot of technology and one single person, INVISTA Reliability Millwright, Taylor Leeson. With Machine Health added to the existing system, Leeson now leads a team of AI sentinels, constantly communicating with him from across the plant and onto his desktop dashboard.

“The algorithms give me the opportunity to see what needs my attention the most,” says Leeson. “I love the anomaly detections; they make my life easier.”

The INVISTA management team sees how Machine Health is letting Leeson apply his human expertise better, too. “It keeps Taylor from coming in every morning and looking at the whole ocean,” says Bonorden. “It’s like having a big brother… somebody tapping you on the shoulder to let you know, ‘Hey, you need to go fix this’.”

Adds Murray: “It gives our customers the ability to refocus their day and eliminate a lot of manual work, particularly when you’re in a situation of managing a high asset count with just a few – or in this case one – personnel taking care of those assets.”

Machine Health has been described as a magnet that helps customers find the needle in the haystack, actively combating alarm fatigue. “The needle being the problem you’re looking for in your haystack of equipment,” Murray says. Leeson sees it as even more powerful. “It’s much more than just a magnet – it's like a tractor beam.”

Building on the data for future success – and savings

This is where people power comes in: Baker Hughes vibration analysts are on hand 24/7 via the dashboard.

“People are very much a part of Machine Health,” says Murray. “We have the AI component analyzing data 24/7 to automatically flag events teamed with an experienced human vibration specialist. The customer communicates back and forth with the vibration analyst via the chat function in the platform about the prescriptive alerts.” Additionally, each customer is paired with a success manager to ensure adoption and maximize value realization from the software.

Data-based diagnoses that need further investigation are explored together. “We’ll present the data that supports a diagnosis and ask the folks on the ground to do further checks if needed. We’ll use the additional data to zero in on the analysis. It’s very much about teamwork and INVISTA has been a great teammate”, she commented.

The intuitive dashboards keep a running log of all changes and improvements to the machines. A simple click shows events, alarms and their causes. The data from each event, including the chat, additional supporting data – even video snippets and pictures – are saved on the System 1 Machine Health platform, building even bigger brains for the future.

Unsurprisingly given the success of the pilot, INVISTA and Baker Hughes team have now extended the Kingston program into 2023. “They’re helping us perfect this solution,” concluded Murray.

Leeson sums up how machine health has changed his job by saying, “You guys are driving me to be better, showing me the value in the deep dive of condition monitoring. It’s not just me – we are more of a team than you think.”

Meanwhile, the AI’s smarts keep multiplying and never need a break.