Using AI to Achieve Operational Excellence

Artificial Intelligence (Al) and Machine Learning (Ml) offer definite advantages in enhancing operational excellence of cement plants.
Dr SB Hegde, Professor, Jain University and Visiting Professor, Pennsylvania State University, USA, writes about the expanding and accelerating use of AI in the cement sector in a bid to reduce operating expenses while increasing yield, enhancing quality and lowering emissions.

The three main factors driving cement producers’ adoption of Artificial Intelligence (AI) are as follows:
computing power connected devices algorithms

In its daily operations, the cement industry faces a variety of difficulties related to profitability, cost control, quality versus throughput, emissions and environmental sustainability. Cement manufacturers can overcome these challenges thanks to the many benefits that AI offers. The game-changing technology that many cement producers have been waiting for is the ability to perform sophisticated data analytics and intelligent optimisation supported by AI. Artificial Intelligence is like a formula that achieves goals in new situations. The formula adapts to change rather than remain a static algorithm.

Cement manufacturers can achieve key performance indicators for operational excellence, connected workers, connected processes, and sustainability. Here are a few typical applications of AI in the manufacture of cement.

The following are the value drivers in a cement plant where AI and Machine Learning (ML) will be of great help and they are as follows:SustainabilityProcess performanceAsset performanceConnected workforceOperational excellence.

Security

ML, a subset of AI, is the principle that a machine can learn without human intervention, developing its own algorithm to improve the performance of a specific task. ML can only solve problems formulated for it. Not every optimisation method it learns from data makes sense in real life or delivers tangible benefits.

Deep Learning is a more sophisticated version of ML used to perform more complex tasks or to produce data needed for decision-making. It uses multi-layered neural networks for a more powerful way to filter and process information. Neural Networks is a set of algorithms loosely modelled on the way the human brain processes information.Ai and Sustainability
AI will be essential in achieving environmental sustainability goals, not just in terms of reducing emissions but also in terms of energy management and optimisation. As a result, operating costs and profit margins will immediately improve, and new business models for high-tech, low-CO2 cements will be possible. Cement plants are constantly working to stay within the daily SO2 emission limits and the hydrate consumption that goes along with them. They have a lot of process limitations to balance. Due to complex dynamics and the variability of feed and fuel sources, manual operators using PID control tend to keep ‘safe distances’ from process constraints, which reduces plant profitability.

Continuous Process Improvement

For plants to operate more profitably, traditional advanced process control (APC) solutions successfully address processes like clinker-to-cement ratio reduction, fuel switching and thermal efficiency. They include:
Increase feed by over 3 tph while reducing specific energy by 20 kcal/kgDeliver overall productivity increases of 4 per cent with better and more consistent cement quality.

Tying analytics and APC together will enable re-modelling and tuning in an automated way and optimising additional variables. Many technology suppliers are also working on utilising data collected through cement information management systems to address challenges that have not yet been tackled such as cement quality prediction.

Traditionally, cement strength is assessed after 28 days, which is obviously too late to make adjustments. As a result, plants frequently ‘over deliver’ on product specifications. On the day of sampling, technology providers are using ML and data-driven soft sensors to forecast 28-day strength, enabling quick process adjustments.

Setting new Blaines targets each day is required for this. Additionally, it means that cement plants will be able to sell more products with the proper specifications.

Asset Performance Management

Utilising AI for asset performance management (APM) is a significant improvement in how the maintenance and reliability team works with other departments. Depending on shifting production objectives, AI makes sure assets are available at the time and performance level required by the operations. Because complex systems interact in unexpected ways and are constantly changing, it is challenging to predict how assets will react and respond to different factors (like age or operating condition). Problems are frequently invisible to the human eye. AI/ML models can be continuously trained with pertinent datasets in order to provide precise target parameter predictions in close to real-time and to avoid failures. These datasets demonstrate in-depth knowledge of asset behaviour as well as cement processes.

AI-enabled APM is the most economical

method for extending the life of older and newer assets and determining the best time for scheduled maintenance turnarounds (one of the biggest costs in cement plants). Cement plants may be able to operate more efficiently and with remote management thanks to predictive asset models. Operating a cement plant with three shifts of just three people would be possible at the time of writing, which is during the global COVID-19 pandemic. The remote teams working from their homes would have complete access to data that would inform them of the condition of all the assets in the plant if they used an APM solution powered by predictive asset models.

Connected Workforce as a Change Catalyst

By analysing how operators interact with control systems and how quickly they react to alarms, AI will increase workforce productivity. AI is able to learn which priority alarms call for quicker responses. The visibility of these alarms will then be improved by filtering and rationalising them to enhance performance.

The use of mobile technologies, smart glasses and human-centric control rooms will increase the industry’s appeal to the next generation of engineers. By enabling more people to access low-code/no-code solutions, enabling them to capture ‘tribal knowledge’ on a common platform, innovate, and produce better results, technology can aid cement manufacturers in the development of their employees’ AI capabilities.

Secret to Operational Excellence is Visibility

Companies can optimise production and find the best operating points to increase margins by transferring knowledge and process methodology from higher performing to lower performing facilities.

Future modular and prefabricated construction will require less cement, so AI will be crucial in restructuring operations to maintain profitability as cement demand declines. AI helps with better supply chain management planning by analysing previous procurement methods.

Enabling autonomous operation
Without a doubt, the digital revolution in the cement industry is driven by more and better data, which is collected directly from connected machinery, processes, soft sensor models and other systems.

The degree to which equipment, processes, plant operators and corporate management are connected by digital and automation technologies is unprecedented. The concept of an autonomous plant will eventually become a reality thanks to advanced data analytics and artificial intelligence.

Prospects for the Future

1. AI will collect data from a wider range of sources, such as sophisticated sensors, instruments, historians and databases. Even though data from 20 years ago may not be available, there will be more data available so that AI algorithms can advance.
2. As AI implementation becomes simpler, more cement producers will use AI to meet sustainability goals. Since manual setpoint adjustments alone are not sufficient to achieve those goals. The use of AI will help the cement industry become more reputable and reduce its carbon footprint.
3. Based on outside disturbances, unit area models will continuously retrain themselves. With increased confidence as a result of autonomous operations, cement companies will feel more at ease with having fewer human operators in a plant.
4. AI will help them become autonomous themselves. Core areas like operational excellence, process performance and asset performance will continue to provide value.

ABOUT THE AUTHOR
Dr SB Hegde is currently a visiting professor at Pennsylvania State University in the United States of America and a professor at Jain University in Karnataka, India. He had held ‘Leadership Positions’ in significant and top cement businesses both in India and abroad. He has more than 154 research papers that have been published in both national and international journals. Dr Hegde was also the recipient of the Global Visionary Award.