August 2018


Digital: Improve chemical production with the IoT

If chemical companies want to remain competitive and move forward in a changing world, they must rapidly adopt innovative technologies.

If chemical companies want to remain competitive and move forward in a changing world, they must rapidly adopt innovative technologies. Incorporating the Internet of Things (IoT) within these companies can provide important benefits. Combining the IoT with machine learning can move the chemical industry forward to work more efficiently and create better results.

How can the IoT improve chemical production? How can chemical companies use the IoT and machine learning with their processes?

Improve chemical production with the IoT

While many industries are embracing the IoT, it might not seem clear how it relates to the chemicals business. In his keynote presentation at the ARC Advisory Group Industry Forum, Andy Chatha, President and CEO of ARC, explained that the IoT can streamline many parts of industrial companies, including providing smart machines, offering better capacity for big data storage, and helping optimize systems and assets. The benefits of the IoT within this industry are far-reaching. They include better productivity, improved asset utilization and higher revenue.

Fostering innovation

Significant opportunities exist in research and development to create higher-value, higher-margin products at a faster pace, particularly in specialty and crop protection chemicals. Advanced analytics and machine learning enable high-throughput optimization of molecules, as well as simulation of lab tests and experiments for systematic optimization of formulations for performance and costs from test tube to tablet.

For example, BASF worked with Hewlett Packard to develop a super computer that can run simulations on and predict properties and performance of new industrial catalysts, crop protection products, materials and formulations. In addition, advanced analytics and machine learning can drive the allocation of best available resources to research projects in line with portfolio priorities. Screening of internal knowledge and patent data bases becomes possible to maximize the use of intellectual property and fill gaps therein. Machine learning can also help chemical manufacturers run simulations on sustainability and environmental impact across a product’s lifecycle.

Changing the game in plant operations

The IoT builds the foundation for machine learning in manufacturing and asset management. It can capture real-time data on an asset’s status and performance, process parameters, product quality, production costs, storage capacity and inventory (telemetry), inbound/outbound logistics, worker safety, pairing products with services, etc.

With advanced capabilities in capturing, storing, processing and analyzing data, a vast amount of plant, asset and operational data can be used in conjunction with advanced algorithms to simulate, predict and prescribe maintenance needs for assets. This benefit increases availability, optimizes uptime, improves operational performance and extends the asset’s lifetime. Digital asset networks start to emerge and bring original equipment manufacturers (OEMs), operators and service providers together on a single platform, with the goal to facilitate collaboration on common standards and increase the efficiency of operations and maintenance.

In this context, digital twins play a major role in managing asset performance and maintenance. Once plants and processes have been designed and engineered, digital twins can be used to train operators by simulating special plant and process conditions related to safety and/or performance—much like how flight simulators are used to train pilots. Digital asset twins can be used in maintenance to predict the impact of certain process parameters on asset performance, asset lifecycle and maintenance needs.

FIG. 1. Digital twins play a major role in managing asset performance.
FIG. 1. Digital twins play a major role in managing asset performance.

A 2016 industry report explains the concept of digital twins in such a way that organizations create value from information via the movement from physical to digital and then back to physical.1 Another industry report notes that a petrochemical company that used a digital twin model created a 20% improvement in product transitions.2 Even networks of digital twins (FIG. 1) are proposed to increase interoperability along the entire asset lifecycle, ultimately maximizing asset performance.3

Completely new opportunities for the chemical industry arise from distributed manufacturing/3D printing in terms of developing innovative feedstock and driving new revenue streams. While more than 3,000 materials are used in conventional component manufacturing, only about 30 are available for 3D printing. To put this into perspective, the market for chemical powder materials is predicted to be more than $630 MM/yr by 2020.

Worker safety can be enhanced by the addition of smart tags on wearables, which can help alert workers of exposure to dangerous substances (e.g., toxic gases), upcoming fatigues, as well as help locate employees and contractual workers in case of emergencies. Moreover, alerts could be triggered if employees work out of their designated or authorized working areas (e.g., connected worker).

Taking your supply chain to another level

Many untapped and potential IoT and machine learning technologies exist in the area of supply chain management. For example, using advanced analytics to increase forecast accuracy can lead to improvements along the entire sales and operations planning process and related key performance indicators.

Advanced analytics and machine learning can be used for mitigating risks of supply chain disruptions. For example, shipments can be automatically rerouted during natural disasters to meet on-time delivery goals and customer commitments at minimum costs.

An additional opportunity resides in optimizing the use of transportation assets and related costs. Transporting chemicals means considering special equipment and complex compliance requirements so that empty backhauls are the norm rather than the exception. This results in increased costs and suboptimal asset utilization. Machine learning can better leverage transportation assets and drive waste out of the logistics function.

Innovate by getting closer to your customer

Over the past several years, the chemical industry, as an asset-intensive industry, has been focusing its efforts on optimizing plant and asset operations. However, untapped potential exists to develop innovative, customer-centric business models and services. The following are examples of how chemical companies could benefit from leveraging the IoT and machine learning at the customer front end:

  • Utilize sensors and telemetry to implement vendor/supplier-managed inventory concepts and completely automate the replenishment process (“no” or “low touch” order to delivery)
  • Monitor customers’ manufacturing process parameters in real time via sensor technology, leveraging advanced algorithms to correlate process parameters with quality of (semi-) finished products, selling first-pass quality as a business outcome rather than selling products, and offering benchmark data as a service
  • Use advanced algorithms to better understand customer buying behavior/patterns and adjust product and service portfolios correspondingly, as well as to identify cross-selling opportunities to increase customer loyalty and share of wallet
  • Increase visibility into customer/market sentiment via capturing and processing unstructured data from social media, then respond with appropriate marketing campaigns and innovative service offerings.

Moving forward with the IoT

By combining the IoT with machine learning, chemical companies can move forward and gain positive business results. How do chemical companies use IoT technology? Industrial businesses already have built, or are just building, the foundations for incorporating the IoT and machine learning to become an intelligent enterprise.

In general, intelligent enterprises drive game-changing outcomes. They do more with less and empower employees through process automation. They deliver a best-in-class customer experience by proactively responding to customer expectations. They invent new business models and revenue streams. Intelligent enterprises differentiate with three key capabilities. They operate with:

  • Visibility—The ability to collect and connect data that was previously siloed and recognize unseen patterns
  • Focus—The ability to simulate the impact of potential options and direct scarce resources to the areas of maximum impact
  • Agility—The ability to respond faster to changes in the marketplace or the business and pivot business processes towards the right customer outcomes.

Overall, the IoT can act as a solution that helps the chemical industry keep up with changing times and better meet the needs of shareholders and customers. However, having clean and abundant data available to train algorithms and build high-quality models that predict high-quality results is pivotal to success. Another critical success factor is highly skilled data scientists. The lack of these types of professionals can be a severe constraint for rapid adoption of the IoT and machine learning in the chemical industry. HP


  1. Deloitte Insights, “Industry 4.0 and the chemicals industry,” Deloitte, June 2016.
  2. International Data Corp. (IDC), “The IoT imperative for energy and natural resources companies,” 2017.
  3. SAP, "SAP Point of View on the Network of Digital Twins," 2018

The Author

Related Articles

From the Archive



{{ error }}
{{ comment.comment.Name }} • {{ comment.timeAgo }}
{{ comment.comment.Text }}