Time to level up our defensive measures in Operational environment using the MITRE ATT&CK framework. It underscores the growing cyber threat to critical infrastructures and details how the framework aids defenders in identifying, understanding, responding to, and mitigating attacks that impact the energy vertical. We’re in a time that highlights the necessity of adapting advanced threat intelligence and tailored cybersecurity strategies to the unique needs of OT environments, thereby strengthening our resilience against common threats as well as mitigating the more sophisticated adversary.

We are a company with purpose. At Universal Robots we want to create a world where people work with robots, not like robots. Our mission is simple - Automation for anyone. Anywhere. We continuously redefine automation. We’re a moon-shot technology company; we inspire our employees and global network of partners to think differently and innovate ground-breaking technologies that move people, industries and the world forward. We are committed to every customer's success. We’re trusted advisors, dedicated to building long lasting partnerships with our customers. By getting to know their individual needs and challenges we’re able to focus our vast network of partners, innovative technologies and leading expertise to provide solutions that deliver real value and on-going success. We ensure the process of buying, deploying and using our automation solutions is as frictionless as possible. Our success lies in our customers’ success. We collaborate, it's in our DNA. Collaboration is our greatest strength. We listen, share ideas and work closely with our colleagues, our partners and our customers to find the best solutions to their challenges. Our desire to share expertise and continually improve has fueled our commitment to building the most passionate and innovative automation community and technology platform in the world. A powerful ecosystem of people and technologies working closely together to find a better way.

Machine learning methods have the potential to ‘learn’ complex physicochemical processes from large, properly prepared data sets. These models can be used to leverage historical data to drive better process decisions. Custom machine learning models of the coagulation process were developed and deployed at a drinking water treatment plant serving over 100,000 people in western Canada. This presentation will discuss the project, including model development, piloting, and integration with the plants operational system, and the results of the first full year of operation.

Model preparation included using various data selections, processing techniques, and model architectures. Coagulation is a complex process impacted by several physical, chemical, and operational conditions, and modelling required significant domain knowledge. The models underwent a desktop assessment using process data and yielded promising results, including a 24% reduction in coagulant chemicals needed to meet treatment objectives. Models were subsequently tested full scale at the water treatment plant in May – June 2022. During the testing period the models provided coagulant and polymer dose recommendations during both stable and rapidly changing raw water conditions. During stable conditions the models helped operational staff to reduce the coagulant chemical used while exceeding treatment objectives. In rapidly changing, poor raw water conditions the model provided immediate adjustment recommendations that helped maintain excellent treated water quality. In February of 2023 the models were integrated with the plant’s supervisory control and data acquisition system using edge computing hardware with no internet connectivity to avoid increasing the cyber attack surface of the water treatment plant.

Results of the first full year of implementation have included a 21% reduction in coagulant chemicals used, equating to approximately $260,000 in chemical savings and 230 wet tonnes less chemical sludge being produced and sent to the landfill. More important than the cost savings, the models have helped operational staff adapt proactively to rapidly changing raw water conditions and protect public health through consistently providing clean drinking water, regardless of the raw water conditions. The water treatment plant has also won awards in western Canada for producing the best testing water, all while using model recommendations.

Due to the critical nature of drinking water treatment, significant risk mitigation strategies were used to deliver this project. A review of these strategies will be shared, as well as work that is being undertaken to assess the transferability of the technology to other facilities.

First introduced in 2012, PlantESP Loop Performance Monitoring is now the market-leading and most innovative control loop performance monitoring (CLPM) solution. PlantESP taps into a facility's existing data historian and it applies a targeted library of key performance indices (KPIs) to uncover a range of common issues that undermine production throughput and efficiency. Join Alessia Testani for a demonstration of Control Station’s PlantESP Loop Performance Monitoring solution and learn how recent innovations are making plant-wide process optimization easy.

Asset-intensive industries like oil & gas, mining, energy and utilities use complex equipment such as rotary, compressors, haul trucks, and turbines in their day-to-day operation. Any unplanned downtime or major unforeseen equipment failure negatively impacts production, which affects the organization’s financial performance. However, increasing instrumentation (“smart technology”) of equipment and infrastructure and wireless communications are enabling organizations to acquire volumes of asset performance data and become proactive in monitoring the condition of these assets.

Furthermore, analytics are enabling organizations to develop sophisticated models of asset performance, predict component and equipment failure and assess the health of in-service equipment. Driven by predictive analytics you can now detect even minor anomalies and failure patterns to determine the assets and operational processes that are at the greatest risk of problems or failure. Advances in analytic algorithms enable organizations to identify signs of possible failure well in advance of previous methods. What-if analysis allows an organization to investigate potential scenarios to determine the most appropriate (economic, efficient, safe) means of responding to pending equipment failure. Automated decision management can then recommend the best action to take in anticipation of equipment problems.

This session will explain how the capabilities of Predictive Maintenance and Quality solution are being used by oil & gas, mining, energy and utility organizations worldwide to integrate relevant equipment data, including real-time, build models that predict maintenance needs, monitor asset performance, provide timely alerts, and recommended appropriate actions. These integrated capabilities allow these organizations to deploy limited resources more cost effectively, maximize equipment uptime and enhance quality and supply chain processes.

Use-case examples:
Predict the failure of a monitored asset in order to fix it and avoid costly downtime
Identify the root causes of asset failure to take corrective actions
Minimize product quality and reliability issues to meet customer delivery schedules
Others...