Key Takeaways

• Collaborative Innovation: Drishya AI worked with IIM Bangalore to create a data-driven, AI-powered solution for alarm management in industrial systems, bringing in cutting-edge technology and academic rigor.

• Real Impact: The implementation reduced alarm floods by 40%, increased operator efficiency, and prevented costly downtime while ensuring seamless integration with existing systems.

• Global Recognition: This innovative solution was featured in the Harvard Business Review, demonstrating Drishya’s expertise in leveraging machine learning and deep learning to solve pressing industrial challenges effectively and sustainably.

• Scalable Solutions: Drishya’s expertise in AI and machine learning extends beyond alarm management, paving the way for smarter systems in the energy sector.

What happens when cutting-edge artificial intelligence (AI) meets academic rigor and real-world industrial challenges? The result is a transformative story of collaboration, innovation, and measurable impact, one that not only solves immediate problems but sets benchmarks for the future.

This is the story of Drishya AI Labs, a startup that is redefining the role of AI technologies in the energy sector. With support from the Indian Institute of Management Bangalore (IIMB) and recognition from the prestigious Harvard Business Review, Drishya has become a trailblazer in intelligent alarm management systems.

The Alarm Crisis: A Problem Demanding Innovation

On December 11, 2020, OSUM’s SAGD plant in Alberta faced an unprecedented operational challenge. Over 770 alarms were triggered within a single day, overwhelming plant operators. Many of these alarms were repetitive or irrelevant, creating confusion and delaying critical responses.

Operators struggled to manage this deluge of alarms. Each alert demanded attention, but the sheer volume of data overwhelmed the system’s design, leaving operators unable to prioritize tasks efficiently. This situation stressed the operators and risked operational delays, production losses, and potential equipment failures. The importance of predictive maintenance in preventing such crises cannot be overstated.

[Graph showing the alarm trend on December 11, 2020, with a sharp increase in alarms: 1 alarm from 6 AM to 7 AM, rising to 56 alarms from 11 AM to 12 PM, peaking at 190 alarms from 12 PM to 1 PM, and declining to 110 alarms from 1 PM to 2 PM]

The Distributed Control System (DCS), which was supposed to streamline processes, became a bottleneck. While the DCS handled automated tasks well, it failed to provide the intelligent alarm management required to navigate this crisis. The need for a smarter solution grounded in artificial intelligence and deep learning was evident.

This crisis also underscored a deeper issue: industrial systems are only as effective as the humans managing them. While automation reduces manual workload, it introduces complexity that often requires advanced tools and strategies, such as neural networks, to handle. Addressing these challenges requires a blend of technical innovation and human-centric design.

Another critical realization from this crisis was the importance of preemptive interventions. Instead of reacting to alarms after they occur, modern systems need the ability to predict potential disruptions and alert operators in advance. This is where machine learning and deep learning technologies shine, offering predictive insights that enable proactive measures and mitigate risks effectively.

This moment called for innovation—a solution that could handle alarm floods intelligently, reduce operator stress, and restore operational efficiency. Drishya AI stepped up to the challenge.

Alarm Management: A Growing Concern

The Need for Smarter Systems

Modern industrial operations rely on interconnected systems to monitor and optimize performance. These systems produce vast amounts of data, triggering alarms whenever anomalies occur. While this setup is meant to enhance safety and productivity, it also introduces challenges such as alarm overload. Digital transformation is essential to address these challenges effectively.

Alarm overload occurs when the volume of alerts surpasses an operator’s ability to respond effectively. In such cases, critical alerts can be lost in the noise, creating a dangerous environment where safety and efficiency are compromised. Artificial intelligence tools, such as neural networks and machine learning, offer the precision and scalability needed to manage these scenarios.

The Cost of Inaction

Failing to address alarm overload has significant implications. Operators may ignore or miss critical alerts, leading to delayed responses, increased safety risks, and operational inefficiencies. For OSUM, this could have resulted in production downtime and substantial financial losses, making it imperative to act quickly.

Alarm overload isn’t just a technical issue; it’s a challenge that directly affects human performance, operational outcomes, and even a company’s reputation in the market. Advanced technologies, including deep learning and specific tasks-oriented AI systems, have become indispensable for addressing such challenges effectively. Intelligent asset monitoring can also play a crucial role in mitigating these issues.

Furthermore, alarm overload can cause long-term issues, such as operator fatigue and desensitization. When workers are consistently overwhelmed by unnecessary alarms, they may begin to disregard alerts altogether, increasing the risk of missing truly critical warnings. Addressing this challenge is as much about safeguarding human performance as it is about improving system efficiency and processes.

Drishya AI: A Partner in Innovation

Drishya AI Labs, established in 2020, specializes in designing AI-driven solutions tailored to the unique needs of the energy sector and industrial automation. With expertise in machine learning, deep neural networks, and real-time data analysis, the company has quickly established itself as a leader in the field.

When OSUM sought a solution to its alarm management problem, Drishya responded with a comprehensive, scalable approach. The team’s focus was clear: to create a system that could filter, prioritize, and predict alarms, ensuring operators could focus on what mattered most.

Drishya’s commitment to delivering actionable insights and measurable results made it the ideal partner for OSUM’s operational challenges. The company ensured that its solution was not only efficient but also practical, integrating seamlessly with existing processes.

The company’s approach to innovation extends beyond technological solutions. By emphasizing close collaboration with clients, Drishya ensures that its AI models are not only effective but also user-friendly and practical for real-world application. This focus on usability is a key factor in its success.

IIM Bangalore: Adding Academic Depth to Management Education and Innovation

Drishya’s collaboration with the Indian Institute of Management Bangalore (IIMB) brought academic rigor to the project. IIMB’s Business Analytics and Intelligence program, under the guidance of Professor Dinesh Kumar, provided the analytical depth required to design and validate Drishya’s AI solution.

The team at IIMB worked closely with Drishya to analyze historical alarm data from OSUM’s plant, ensuring the AI models were both accurate and aligned with industry standards. This collaboration demonstrated the power of combining academic insights with industrial expertise to solve complex problems effectively.

By leveraging IIMB’s extensive knowledge of data science and management practices, Drishya was able to develop a solution that was both innovative and practical, setting a new benchmark for intelligent systems.

This partnership also highlighted the importance of academic involvement in real-world problems. By connecting theoretical frameworks with industrial applications, institutions like IIMB can drive innovations that benefit both academia and industry.

Solving OSUM’s Alarm Overload

Drishya adopted a structured, three-step approach to tackle OSUM’s alarm crisis:

1. Understanding the Problem with Data Exploration, Classification, and Machine Learning

The first step involved a thorough data analytics and analysis of three years’ worth of alarm data from OSUM’s SAGD plant. This analysis revealed critical insights into the types and frequencies of alarms, enabling the team to classify them as:

• Chattering Alarms: Alarms triggered multiple times in quick succession, creating unnecessary noise and distractions.

• Nuisance Alarms: Alerts that were irrelevant to immediate operations, further contributing to overload.

• Stale Alarms: Persistent alerts indicating unresolved issues, which required long-term interventions.

This classification allowed Drishya to focus on reducing noise and improving alarm relevance, paving the way for a more efficient system.

2. Developing Artificial Intelligence-Powered Predictive Models

Using machine learning algorithms, deep neural networks, and predictive analytics, Drishya developed predictive models capable of identifying patterns in alarm triggers. These models provided operators with proactive insights, enabling them to address potential issues before they escalated.

The predictive capabilities of the system ensured that operators could focus their attention on critical alarms, improving overall efficiency and response times.

3. Real-Time Alarm Management

The final step involved integrating the AI solution with the plant’s Distributed Control System (DCS). This integration allowed the system to filter and prioritize alarms in real time, ensuring operators only received the most relevant notifications.

By reducing cognitive load and streamlining workflows, this real-time system significantly enhanced operator performance and decision-making capabilities.

The system’s real-time nature also meant that it could adapt to evolving conditions within the plant, ensuring continued reliability even as operational dynamics changed.

Results: A Smarter, More Efficient System

Immediate Improvements

The implementation of Drishya’s solution led to measurable improvements, including:

• 40% Reduction in Alarm Floods: The system eliminated redundant and nuisance alarms, allowing operators to focus on critical issues.

• Enhanced Operator Performance: Streamlined notifications and predictive insights improved response times and decision-making.

• Avoidance of Downtime: Proactive alerts minimized disruptions, ensuring consistent production and operational stability.

Long Term Impact

Beyond these immediate benefits, the solution provided a scalable framework for alarm management, which could be applied to other industrial systems. This long-term impact underscored the value of Drishya’s AI-driven approach.

Expanding Applications - Beyond Alarm Management

Drishya’s expertise isn’t limited to alarm management. The company is exploring applications in digital transformation, intelligent asset monitoring, production optimization, and real-time decision-making, helping industries unlock the full potential of AI systems.

As global industries embrace AI, Drishya’s solutions are setting new standards for efficiency, safety, and sustainability, ensuring companies remain competitive in an evolving landscape.

Drishya is also exploring partnerships with other industries where alarm overload is a critical challenge, demonstrating its ability to adapt its solutions to diverse operational environments.

Lessons Learned

• The Role of Collaboration: This project demonstrated the value of partnerships between startups, academic institutions, and industrial leaders. Collaboration fosters innovation, combining diverse perspectives and expertise to solve complex challenges.

• The Power of Data: Accurate data and advanced analytics are critical for developing effective AI systems. By leveraging historical data, Drishya delivered a solution that not only addressed immediate issues but also provided actionable insights for future improvements.

Why Choose Drishya AI?

Recognized by Harvard Business Review, Drishya AI is a leader in applying AI to solve industry-specific challenges with a focus on:

• Tailored Solutions: Drishya designs customized AI models to address specific industrial challenges, ensuring maximum impact.

• Proven Expertise: With a track record of success, including recognition in Harvard Business Review, Drishya stands out as a leader in industrial AI solutions.

• Collaborative Approach: By partnering with institutions like IIM Bangalore, Drishya combines theoretical insights with practical applications, delivering results that matter.

The Future of Drishya AI in Oil and Gas

• Scaling Innovations: Drishya aims to expand its AI capabilities across the energy sector, from predictive maintenance to real-time optimization of operations.

• Transforming the Industry: As the energy industry evolves, Drishya’s solutions are helping companies navigate the transition to smarter, more efficient systems, enhancing operational efficiency.

Recognition in Harvard Business Review

The success of Drishya’s alarm management system was featured in the Harvard Business Review. The coverage highlighted the project as an exemplary case of leveraging AI to solve industry-specific challenges. You can read it here.

The HBR spotlight emphasized how Drishya applied machine learning, deep neural networks, and predictive analytics to address alarm overload. It also highlighted the company’s focus on measurable outcomes, operator efficiency, and system reliability—highlighting their ability to align AI with specific tasks and measurable outcomes. This case demonstrates the difference academic and industry partnerships can make in driving innovation.

Conclusion

Drishya AI’s journey from tackling alarm floods to earning recognition in Harvard Business Review is a testament to the transformative power of AI. By combining academic rigor with industry expertise, Drishya delivered a solution that resolved OSUM’s immediate challenges and set a new standard for alarm management in the oil and gas sector.

As industries embrace smarter technologies, Drishya AI remains at the forefront, empowering companies to achieve greater efficiency, reliability, and safety.