BLOG

What is data-driven plant performance optimization?

Factory owner, Updates, Blog, lean manufacturing, oee

What is data-driven plant performance optimization?

5 Mar, 2019

This article seeks to clarify the benefits of data-driven plant optimization using Overall Equipment Effectiveness (OEE) , explain the data cycle as it applies to plant optimization and the key areas of industry that can be improved through implementation.

The 21st Century has seen major changes in the way industrial processes and machinery are analyzed and optimized. Technological advances have heralded a new age of innovation, often referred to as the 4th Industrial Revolution or Industry 4.0. Wireless interconnectivity of machinery and devices (Internet of Things or IoT) has allowed for more efficient collection and monitoring of data. Modern industrial processes can now be scrutinized, analyzed and adapted more than ever. To stay profitable, companies are under more and more pressure to reach tough industrial KPIs.

But what does this mean in practice for businesses that implement advanced data systems and optimization processes?

 

What is OEE?

OEE stands for Overall Equipment Effectiveness (OEE). It is a lean manufacturing term, originally introduced by the Japanese. Before moving on to the practicalities and benefits of data-driven plant performance optimization, it’s important to grasp the concept of OEE and understand its impact on modern lean manufacturing and industrial processes.

OEE is a measure of how effectively industrial equipment and plant performs. This is done by calculating a percentage, with 100% representing perfectly optimised plant performance. It is measured in three areas – quality, performance and availability. A process that scores 100% in all three areas would show good quality output, as fast as possible, with zero machine downtime.

OEE can also be linked to the losses caused by less than 100% optimization, making it a powerful way to reduce costs and improve profits. OEE involves collecting a lot of data from all plant and machinery operations, analyzing it thoroughly and gaining insights to improve performance.

 

The main benefits of OEE and data-driven plant performance optimization

The obvious benefit is that plant and asset managers will have greater insight into the operation of plant and machinery, allowing them to make better-informed decisions. They can quickly pinpoint plant that is underperforming and addresses the problems quickly. Companies that introduce sophisticated Human-Machine Interfaces (HMI) will gain a serious edge over those that don’t, as the enhanced control system will increase productivity dramatically.

Increased automation is also a major benefit. For example, a spare part that frequently needs replacing can be monitored closely for wear by a sensor. The system can be set up so that when the part is at a predetermined level of wear, a replacement is automatically ordered, therefore eliminating machine downtime while waiting for delivery of parts and avoiding plant failure.

Another major benefit is that modern optimization systems are capable of automatically adjusting production volume according to demand or manufacturing material prices. Previously, this would have involved lengthy research and calculation.

In fact, all areas of industry will benefit from the increased use of data in terms of increased efficiency and productivity. When combined with machine learning algorithms, data output can be used to allow industrial systems to effectively streamline themselves. Targeted aggregation and in-depth analysis and simulation of data can be fully automated for cost savings and productivity increases.

Data has always existed, but now it’s possible to access it via high-performance gateways, which are further supported with the introduction of OPC UA over TSN.

 

The four stages of data-driven plant optimization

A typical data cycle for plant optimization is as follows:

  1. Record and monitor data – Real-time recording of data such as operation speed, wear and tear, motor usage, energy usage, etc. is carried out using IoT sensors.
  2. Upload of data – All data is wirelessly uploaded to a cloud storage system often referred to as the industrial cloud.
  3. Analysis of data – The data is automatically fed into simulation software or a virtual plant model and data insights generated.
  4. Reports and action – Reports are automatically produced and can be sent remotely to the relevant people. In some cases, automatic action will be taken such as ordering replacement parts.

 

The key areas of data-driven plant performance optimization

The following areas can be optimized through the efficient use of data collection and analysis across various sites:

Energy Usage

In today’s environmentally conscious world, businesses are under pressure to use more sustainable practices and reduce energy consumption. The reduction of energy usage will also result in considerable cost savings over the long term.

Plant performance data systems can be implemented that monitor and target energy usage. For instance, the current being used by an electric motor can be carefully analyzed during operation, meaning that the electrical input can be reduced during periods of low usage.

Process Efficiency

Using data analytics to improve process efficiency will help companies to maintain a competitive edge. As industrial processes become more complex and technical, using targeted data will help to keep them manageable and streamlined.

For instance, manufacturing companies may use multiple machines and plant operations throughout product fabrication. Data analytics can identify any stages that are slowing production down, and alternative workflows or tweaks to existing ones can be implemented to speed things up.

Control loops are a major area for potential improvement in most plant processes. Research has shown that around 50% of control loops require fine-tuning; 25% are ineffective and 25% show declining performance over time. Approximately 33% of control loops are still manually controlled and 25% use commissioning parameters which are no longer applicable. Often, a control or systems engineer will singlehandedly manage dozens or even hundreds of control loops, making it impossible to achieve optimal performance. Data systems can help to alleviate all these control loop problems.

An effective process data system will:

  • Analyze control states and provide reports and KPI calculations
  • Provide hierarchical plant detail and overviews
  • Generate data-driven insights and optimization suggestions
  • Highlight maintenance issues or automatically take preventative actions
  • Create control loop reports and in some cases suggest improvements

 

Plant and Machinery Usage

As mentioned earlier, plant and machinery can be fitted with multiple wireless sensors to measure wear and tear. This means early intervention can be made before the machinery fails, creating a pro-active approach to maintenance, rather than a reactive one.

Well-implemented plant data analysis can minimize machine downtime and eliminate complete failure or serious damage.

Cybersecurity

Another key area that can be easily overlooked but is of critical importance in today’s world is cybersecurity. Experts warn that industrial cyber attacks are becoming more common and are likely to increase in the coming years.

Cyber attacks can take the form of stealing intellectual property or sabotage of operations. With global terrorism on the increase, large-scale industrial attacks designed to cause chaos will become more common. With effective data analysis and in-built security systems, any operational anomalies will be quickly flagged up, negating the risk of serious damage.

 

The future of data-driven plant performance optimization

As data collection and analysis systems and simulation software become more intelligent, the scope for predictive performance and automated responses will improve data-driven plant performance.

For instance, before integrating a new piece of machinery into a complex system, existing data can be coupled with advanced simulation modelling software and specification data from the manufacturer of the machinery to predict the effect it will have on production. The software can then suggest changes to the workflow or process without the need for an in-depth engineering study.

We are entering an exciting new phase in plant and machinery optimization, with data analytics and smart systems at the forefront of delivering efficiency and cost savings. Businesses that embrace these new technologies and methodologies will stay ahead and maintain a competitive edge in this fast-changing industrial landscape.