Navigating uncertain times with Digital Manufacturing

The COVID-19 pandemic has led small and medium-sized businesses in India to join industry giants in accelerating the use of technology and analytics to transform their manufacturing processes. Rightly so, “future-ready” organizations are 1.7 times more efficient and 2.8 times as profitable compared to companies at lower maturity levels, as revealed by recent Accenture research involving 1,100 C-suite executives from 11 countries.

Accelerated by the pandemic, there are multiple factors driving the digital enhancement of the manufacturing sector. The waves of the pandemic are leading to significant demand variation. For manufacturing companies, this means:

• Focus on quality – During periods of low demand, customers’ propensity to switch between products and suppliers increases. That means manufacturers need to pay closer attention to their product quality.
• Focus on cost competitiveness – Low-demand scenarios typically lead to a hyper-competitive market environment, prompting a relentless drive from manufacturers to squeeze out costs to remain profitable.
• Adjust to changing production scenarios – Constrained by demand, manufacturers need to run plants efficiently to cope with varying production scenarios (sometimes as low as 20% capacity utilization).

In the wake of the pandemic, businesses are also witnessing a scarcity of skilled workers. Nearly all manufacturing plants were running with limited on-site presence, during the peak of both the COVID-19 waves, due to employees and/or their families being exposed to the virus. In addition, the experienced workforce in India is ageing and will eventually retire, taking with them precious knowledge and skills. According to a recent Centre for Monitoring Indian Economy (CMIE) report, the share of workforce older than 40 is rising steadily.

To address these challenges, companies are utilizing digital technologies to improve operational efficiency of their manufacturing processes to drive better safety, throughput, quality, yield and cost structures. Some of the broad trends include:

Information technology (IT) and operational technology (OT) integration

To holistically process data to enable intelligent decision making, manufacturers are streaming data from OT sources (Distribution Control System, Programmable Logic Controllers, Supervisory Control and Data Acquisition) and IT Sources (Enterprise Resource Planning systems like SAP) into one database. These databases are typically stored in a cloud environment, making intelligent computing and decision-making possible remotely. A leading oil and gas company recently developed an integrated command and control platform that integrates data from across its countrywide network, including fuel retail outlets, fuel trucks, oil installations and depots, LPG bottling plants and additional industrial and commercial locations. The platform uses machine learning (ML) techniques with the data to trigger automated alerts and actions, in case of equipment failures or hazardous situations.

Advanced analytics-led performance improvement

Machine learning-based models can help leverage historical data to run a manufacturing plant with the optimal efficiency. These help reproduce historical best-fit actions taken by experienced control room operators.

• Soft sensorization – Physical sensors for measuring quality are often costly and unreliable; hence not installed for critical processes. Offline lab measurements take long time, making it difficult to control or improve the quality of products. Soft sensors, developed through advanced algorithms, can accurately predict quality in real time or ahead of time. The algorithms typically leverage historical data related to raw material quality, relevant process parameters and offline product quality to build a virtual sensor. Manufacturing companies in India are developing soft sensors for most of the critical quality parameters where physical sensors are not available. For example, a soft sensor for paper stiffness and tear strength can help indicate in real time the parameter changes required to maintain paper quality.
• Digital Twins – These are virtual replicas of physical processes based on machine learning models with multiple inputs such as raw material parameters, process parameters and manipulated variables and multiple outcome variables such as yield, throughput, and quality. The digital twin helps adjust to changing production scenarios through recommendation of the right settings of the manipulated variables to optimize the outcome variables. It is further used for “what if” simulation. For example, a digital twin for a coke oven can suggest coking time and regenerator temperature in real time to maintain coke qualities and suggest the leanest coal blend to charge to the ovens.
• Video and image analytics – Analytics of images and video (sometimes in the form of big data) are key to solving various manufacturing problems. Advanced artificial intelligence (AI) or ML models are used to process these kinds of data and help in pattern recognition, ultimately delivering benefits ranging from productivity enhancement (such as crop production for sugar, pulp and paper manufacturing), enhancement in product quality assessment as well as tracking of workforce productivity, safety and social distancing. The tubes division of a leading steel manufacturer has developed a real-time video analytics-based solution to provide alerts if workers are violating safety norms e.g. moving without safety personal protective equipment (PPE) or treading to potentially unsafe zones.
• Predictive asset maintenance – As the waves and demand scenarios normalize, it is important that plants quickly ramp up throughput and therefore, they cannot afford unplanned downtime during this period. AI/ML-based algorithms that utilize equipment and process data to provide early warning signals for equipment anomalies are increasingly being adopted to maintain plant uptime. For example, paper manufacturers can use plug-and-play sensors and machine learning techniques to predict failures in rotating equipment like paper rolls and reduce unplanned downtime.

Remote Operations Center

To reduce dependence on onsite presence and localized decision-making, companies in multiple industries, including those in the utilities, oil and gas, metals and mining sectors, are considering setting up remote operations centers. Today’s remote operations centers are based on connected real-time asset management, predictive analytics, and process optimization. Remote operations centers typically help in growing revenue, achieving operational excellence, and improving capacity utilization.

The above transformation-led activities are now supported by technological advancement in 4G, LTE and 5G networks as well as edge computing. India is witnessing the beginning of a new era for digital manufacturing and it will open a world of new opportunities.

Disclaimer: This content is provided for general information purposes and is not intended to be used in place of consultation with our professional advisors. The article tilted “Navigating uncertain times with Digital Manufacturing” is the property of Accenture and its affiliates and Accenture be the holder of the copyright or any intellectual property over the article. No part of this article may be reproduced in any manner without the written permission of Accenture. Opinions expressed herein are subject to change without notice.