By Ron Beck, Marketing Strategy Director, Aspen Technology
The debate has ended, not implementing artificial intelligence (AI) is no longer an option. Every company should have an effective AI strategy, as the pace of innovation accelerates and presents increased opportunities to change businesses. Big data is the bedrock of AI, as the latter gains momentum via machine intelligence and insights by employing streams of business and operational data. Currently, companies are only utilizing a fraction of the data collected. This provides a huge opportunity for the implementation of a digital twin (virtual copy of the company’s assets and business information). A digital twin can unlock the potential of the massive data mountain in each company.
In 1984, “Neuromancer”, a scientific fiction novel, authored by William Gibson, captured the imagination of readers – as a prelude to the world of AI. Gibson envisioned the massive value and power that digital twins can bring and how they can change the world. Today’s technology has surpassed what writers in the past, can only dream of. More than three decades later, this vision materialized in the enterprise world. Virtual copies of physical locations provide an insightful way for companies to harness the true value of data, as AI helps humans access this massive world of multi-dimensional data. Turnkey solutions for digital twins lie in the power of AI and the interconnected industrial world to extract critical insights via data mining and leveraging domain expertise.
No longer dream time…
The future has arrived with advanced technology, available on demand. Now, the golden question is where to invest, as digital twins transform asset-intensive businesses, especially the ones from energy and chemical sectors. In today’s volatile, uncertain, complex and ambiguous (VUCA) marketplace, the deployment of digital twins can help companies achieve sustainability and operational excellence. Digital twinning technology is a business necessity that can greatly impact the industry. It provides a valuable model of the physical asset to help plants explore “what-if” scenarios safely and provide forecasting capabilities and advice on degradation, asset failure events and more. Digital twins also function as business models to optimize various business scenarios.
Based on models and real-time data, the digital twin is an evolving digital profile of the behavior of a physical object or process that optimizes business performance. This provides important insights into system performance, which in turn, leads to actions in the physical world. The digital twin takes advantage of asset data to stay updated and is increasingly made more intelligent by AI agents. First, the digital twin ensures that the process plant is modeled vigorously using engineering models, enhanced via AI techniques with embedded cost and risk models. Second, the operational digital twin ensures that plant operations are models and viewed virtually as planning, scheduling, control and utility models. Areas covered included planning and scheduling, demand models, distribution models, energy demand and supply, as well as control and optimization. Third, the operational integrity digital twin provides tactical and strategic decision guidance around prescriptive maintenance and real-time decision-making to maximize uptime, adjust production, minimize environmental impact and production losses, and prioritize safety. The digital twin also covers asset condition and sustainability.
Overall, companies need a future-proof digital reference architecture to structure the implementation of digital twins supporting collaboration and integration across business functions.
Powered by business value
Boston-based LNS Research conservatively estimated that more than 1,000 digital twin models may be required to model the operations of a typical refinery. The scale of digital twins will deliver value for the enterprise. For example, unit level models can deliver very high value returns for digital twins – involving process, asset condition, control and optimization online models. Energy and utility models, refinery and bulk chemical planning, specialty chemical scheduling, debottlenecking and de-risking, as well as emissions are very high-value opportunities for plants to adopt digital twin models. A new but important area, enterprise-level digital twins, enable rapid analysis of available enterprise profit opportunity options and effectively present actionable information at the executive level.
Examples of how digital twins can value-add to businesses:
- A Bolivian upstream company, YPFB Andina, has increased yield via an asset-wide digital twin model.
- A major US-based international refiner has adopted machine learning digital twins to improve refining uptime and margins, saving money in the process.
- Abu Dhabi National Oil Company (ADNOC) via Equinox, has constructed an asset-wide twin with a visualization dashboard to achieve better decision-making and sustainability milestones.
- Sustainability is a key business driver for Bharat Petroleum (BPCL). This resulted in the deployment of an integrated digital twin to optimize sulfur recovery – encompassing software, such as AspenTech’s self-diagnosing adaptive process control and engineering software models to improve contaminant removal, as well as visualization and key performance indicator (KPI) dashboard. Within six months, the implementation resulted in 90% reduction in sulfur emissions and economic value from recovered sulfur for sale.
- A service provider in Alberta oil fields, Process Ecology, mitigated costs in fine payments via online digital twin models, in the area of compliance and permitting.
- Via a digital twin, Momentive Specialty Chemicals optimized their supply chain daily, in the area of order fulfilment and working capital.
- Forbes has also reported that AspenTech’s customer, Perstorp, a leading global specialty chemicals manufacturer, has a model that contains 1.5 million variables and 500,000 constraints. Perstorp is reportedly evaluating supply/demand scenarios based on marginal contribution impact and plans to have the whole company in a unified model.
- A polymer producer implemented a multivariate analysis data analytics digital twin to manage a wider range of specialty chemical applications, where product quality is key and often problematic.
Companies are progressing with new, advanced technology, as it is necessary to be strategic and have a roadmap to get ahead. As companies invest in digital twins, it is critical to observe (at a high level) how this technology will help businesses overtake the competition. Beyond technology, companies should also take note of their organizational change and evolution. The organization’s ability to adopt, the level of enthusiasm and readiness, should be managed regularly, as business value creation is a key driver of technology.