Expanding upon the science and art of the possible when it comes to the convergence of software, networks, applications and clouds, AspenTech, one of the world’s leading suppliers of asset optimization software solutions, is taking safety seriously and pushing its team to solve the most complex process manufacturing challenges.
While the core of AspenTech business has been built around helping large industrial companies with massive capital investment in sophisticated design and process management, the value of those investments increases through extending the useful lives of the mechanical equipment running the process. In such asset health management, the Aspen Mtell product not only improves the bottom line, but the safety and lives of their customers’ employees.
Their aspenOne software is embedded into the operations of many of the world’s largest oil & gas, chemical, engineering & construction, pharmaceutical, food, beverage, and consumer packaged goods companies, and is used to improve the design and operations of workflow and activities in complex industrial environments.
With over 2,100 customers around the world, the company is in a unique position to bring innovation into industrial settings and, in the case of safer plant design, the company captures data from sensors and maintenance systems to enhance its advanced process safety software.
“We’ve been working for decades, stepping out of the box to find better ways to solve for our customers’ ultimate goal – zero accidents, zero environmental incidents, and zero breakdowns,” said Mike Brooks, Sr. Director Asset Performance Management at Aspen Technology. “We’re serving companies in heavy industries that are asset-intensive and, in some cases, dangerous. Our founder Larry Evans made safety a focus from the beginning.”
Evans, in the 1970s, led a consortium while still an MIT professor of chemical engineering around the opportunity to apply computer automation to other engineering disciplines – civil, mechanical, electrical, and chemical.
“As software and technology continued to evolve, we made an upshift and began developing an Autonomous AgentTM methodology where software can handle more and more work – then we added machine learning,” Brooks continued. “We realized that with sufficient warning, minor problems could be fixed before they became major problems. From there, we began to see indications that lead to signatures across diverse sensors that we could apply across many applications. For example, our software was able to predict, based on a specific signature, when an electric submersible pump leaked, and used that data to let the operator know they were going to have a problem in eight weeks.”
Brooks looks at their solutions supporting “prescriptive” rather than “predictive” maintenance solutions – a solution that gives an active set of recommendations to fix problems, often with 90% accuracy of analysis and reporting with 30 days lead time and more.
This can save lives and money and contributes to continual improvement of processes, given that the learnings from across large groups of assets help refine the analysis further with increasingly powerful AI and machine learning software.
“We helped a rail system virtually eliminate catastrophic failures of locomotives,” Brooks said, across 4,250 of them. By gathering data via lube oil samples, we were able to develop 21 different data points for analysis – from the mix of metals – carbon – calcium – water – and were able to predict up to 120 days in advance when a locomotive should roll into the shed for maintenance. This leads to better planning, more uptime, and more safety and security for the employees working the trains.”
Brooks calls AspenTech’s evolving IIoT offering “just the start of the journey. We’re increasingly able to collect more data given the availability of less expensive sensors and can more easily leverage their software on mobile assets, not just fixed assets.
“For our customers, the value of avoiding breakdowns is massive, financially – within Manufacturing 1.4 trillion dollars a year lost when equipment stops working,” Brooks said. “The human cost is also huge – we are keeping employees safer but also more productive.”
AspenTech is going beyond physical assets to digital assets, checking the health of software-based systems as part of the physical equipment – important as more and more new equipment is coming online and could OEM the AspenTech application. “We’ve seen this in large refinery deployments, where we’re not stopping at physical sensors, but are aligning what we measure with digital systems as well. In this case, we’ve been able to drive the earliest recognition of hydrocarbons going into the atmosphere as an issue in the process, thirty days before the issue can be spotted by human observation. A month in advance, with 90% accuracy, our customer is now able to plan while also responding to impending emergencies before they become real emergencies.”
Improving safety goes beyond the obvious “right thing to do” for employees. Time is money, and when care is taken to not only ensure the machines but also the machine operators, there’s more general uptime, people stay safe, and the customers increase profitability.
Safety design is not new, but previously it has been manual, based on using spreadsheets vs. software leaving more room for human error. To reduce and replace manual workflow, including in the data transfer process between tools, engineers have the tools to make more effective use of their time.
Front-end engineering design (FEED) tools are advancing and perform ongoing analysis to improve process safety and reliability. For example, these tools enable engineers to more easily conduct safety studies include pressure safety valve checks, depressurization, dynamic views of startup and shut down, and response to emergencies (for example compressor surges which could be catastrophic).
By using simulation data from integrated engineering application environments, which AspenTech specializes in, equipment which can cost millions of dollars and be operated and maintained by dozens of employees, can be analyzed using calculations that help predict potential vulnerabilities.
Today, process safety engineering is moving more and more into the real-time world, where after being engineered, tested and deployed, sensor data can be collected and leveraged for everything from automated actions based on AI and machine learning, to notifications and alerts to those humans responsible for ensuring oil rigs, factories, chemical plants, energy grids and other complex mechanical systems are operating safely and efficiently.
“There are things humans can’t see; while our systems look at 50 dimensions or more, for example, humans in the same context see an average of three,” Brooks said. “Machines can capture miniscule changes we don’t see, and they become the means to create signatures feeding more automation and oversight.”
While Brooks is exceptionally proud of the progress being made and the company’s vision, he credits the experts in the field for ensuring a real difference is made using the platform. “We are careful to bring together our data scientists with the people operating equipment, facilities and systems because they understand the reality of their work more intensely than we do. Machine learning is less than 10% of the solution. While we provide an automated machine learning kit, to sample, validate, cleanse the data, look for patterns and such, this information is of no use without humans leveraging it.”
The company uses notifications and alerts triggered automatically when the signatures that have been set up reach certain thresholds. “We’re integrating this into process, workflow, human and machine behaviors. If we see something is not normal, we’re able to report in detail to the right people at the right time. Our ‘Failure Agents’ are noticing changes and aberrations faster and more accurately every day. We also are improving through adjusting, learning from the data – and our customers’ teams.”
Process safety, while not a “revenue source” for the manufacturers of equipment, and the consumers of that equipment (large industrial enterprises) does bring tremendous “risk management” value to the business.
There’s a high risk associated with avoiding doing everything possible to improve safety, including reputational and legal, so finding strategies to align the use of “digital” strategies for industrial equipment and plant safety aligned with operational improvements creates a strong case for investment and advancement.
“We operate in near real-time now and that’s good enough for now,” Brooks said, with polling every 4-5 minutes being more than enough. “Sensors at scale give us more data to work with, and with improvement in network technology and reasonable costs, we see future opportunities to automate more in near real-time.”
Brooks is excited about the opportunity to share signatures and knowledge of systems across machines, and across industry groups. “We’re doing more and more to incorporate information on the types of machines, vendors, models, dates manufactured, service schedules, and more – this is helping us create standards we can apply across other companies using the same machines, and even across other machines.”
Regulations from organizations like OSHA in the US (Occupation Safety and Health Administration) are a good thing, Brooks says, as they protect human lives and limbs. “Technology is having a significant impact on how we keep people safe and businesses running profitability. What we do is becoming part of the fabric of quality manufacturing, including appealing greatly to insurance companies and others who understand the power of ‘risk management’ in the world of Industry 4.0.”
Software and the IIoT has tremendous potential going forward, for both legacy machines, which can be retrofit with sensors and connected to systems, to new equipment being designed for a future of more automation and quality refinement.
Precise and accurate answers reduce mistakes and improve decision making, thus automating the safety process even as more and more mechanical automation and digitization is being developed.
In short, more visibility? Better decisions.
More automation? Better outcomes.
Greater accuracy saves time and money.
But more to the point - achieving greater safety saves lives.
Edited by
Ken Briodagh
