Navigance Whitepaper Optimizing with Intelligence

Navigance Whitepaper Optimizing with Intelligence


Optimizing with Intelligence

Using data-driven analysis and machine learning to boost chemical process efficiency. A white paper

Technologies such as machine learning, a type of artificial intelligence (AI), are reshaping how business works across the world. Manufacturing – and chemical production specifically – are no exception.
Companies using these technologies effectively to analyze data gathered in their plants can enjoy increased transparency of their process performance and ways to improve efficiency, often using just their existing assets. So how and where do you start doing the same?

Developing and supporting AI with in-house resources can be cost and time intensive. Without deep knowledge or prior experience, it may also be difficult to know the right way to start, to deliver the right outcomes for you.

Chemical facilities are characterized by highly complex, interlinked processes. This means the control systems currently in place may not be enough to allow optimization. And factoring all relevant variables into established plant modelling techniques is impractical.
Machine learning offers an effective answer. Its algorithms can learn to identify complex, non-linear patterns and relationships in plant data, enabling adaptable, plant-specific models that provide a truer picture and the foundation for process optimization. From there, you can quickly and continuously adjust control settings to changing conditions, demands, and factors, making efficiency and profitability gains.

The right partner can help you digitalize quickly, effectively and much more cost efficiently. They can guide you through a successful implementation and provide ongoing support to make the project a success.

Navigance combines machine learning that unlocks hidden potential in your plant data with years of process expertise to help you optimize within weeks, around the clock, and without tying up your in-house resource and skills. As it’s important to take a structured approach to digitalization, this paper is designed to help you get started. It looks at how technologies such as machine learning are shaping your sector’s future, and the strategies its players should consider in order to stay ahead. Plus, it shows how the data already building up in plants – most likely in your plant – is a potential untapped asset. A platform for introducing process optimization and a widely interconnected operation.

Artificial intelligence – a somewhat nebulous and misrepresented term – is a core element of the Industry 4.0 revolution. It focuses on interconnectivity, automation, and harnessing cyber-physical systems, real-time data, and the Internet of Things to make the smart factory possible.

It’s technology that learns from experience and adjusts to new inputs to perform tasks such as visual perception, speech recognition, decision-making and language translation, which previously needed human intelligence.
The idea is that, by taking these tasks on for us, it leaves us free to concentrate on the areas where we can add most value. And it’s expected to change the way we manufacture products and process materials forever.

Forecasts suggest its value to the manufacturing industry as a whole – estimated at USD 1 billion in 2018 – will reach USD 17.2 billion by 2025. That’s a compound annual growth rate (CAGR) of 49.5% in just seven years.

Connecting the dots
The Internet of Things (IoT) describes interrelated computing devices, machines, objects – and even people – that can transfer data over a network with no need for a human to interact with another human or computer. Day to day, we increasingly experience their effects.
In an industry context, the Industrial Internet of Things (IIoT) links sensors, instruments and other devices in plants to digital systems and industrial applications that monitor, collect, exchange and analyze data.
What today’s intelligent technologies offer is to bridge the gap between current forms of automation, which follow set guidance or programming, and more advanced forms that learn and adapt for themselves. And very powerful technologies are now available within the IoT, IIoT and cloud services.
In terms of economic output, Accenture and Frontier Economics estimate AI could boost gross value added (GVA)* by almost USD 14 trillion by 2035 – an increase of nearly 45% compared with today. Over the next decade, technologies like machine learning will drive innovation in the manufacturing sector. What opportunities could adopting new technologies early bring you to outperform your peers?

Joining the fourth industrial revolution means building on established IT infrastructure and tapping into the great stores of diverse data building up right across operations, as we’ll see.

Artificial Intelligence in Manufacturing market, by region (USD billion). Source: MarketsandMarkets



The major goal for chemical manufacturers is finding ways to enhance the reliability and efficiency of their production processes, while constantly improving their safety and sustain-ability standards. Given the already high CAPEX intensity in the sector, it means they must often do that within the confines of the assets they already have.

One challenge to tackle is process instabilities, which can have a negative effect on both product quality and yield. Another is systems failures, which cause loss of production. Both translate directly into revenue loss. Meanwhile, inefficient use of resources not only means unre-alized potential, it also inhibits sustainable working, when the responsibility and scrutiny to do so have never been greater.
Another common challenge is how to reliably measure all prop-erties relevant to controlling and optimizing the manufacturing of a product. Understanding the causes of process inefficiencies after they’ve happened is hugely complex and time consuming. Predicting when and where they’ll happen is a whole new level. Attempting to do so can be problematic and resource heavy. It means defining, measuring, analyzing, modelling and imple-menting specific actions, but still leaves a level of uncertainty. So, it’s no surprise manufacturers are turning to industrial AI solutions to help spot, anticipate and help stamp out process inefficiencies.

Embracing digitalization
For years, the chemical industry has looked for ways to improve operational performance through automation and, more recently, digitalization.

At a production level, the potential benefits range from increasing output and revenues to reducing costs, raw materials, and energy use. Beyond the operational, it also gives manufac-turers scope to accelerate innovation, optimize throughout the value chain and keep pace with market and customer demands, ensuring a roadmap to future success.

With digitalization, the tools, services, and solutions needed to manage complex operations and achieve new levels of reliability and profitability have become more readily accessible. For instance, powerful machine learning solutions are now available as IoT and cloud services.

They’re not a silver bullet in themselves. First, producers must pinpoint their core business challenges, then explore the digital solutions likely to fit best, for the greatest return over time.

The data dilemma
These days, data is everywhere in production plants but rarely being used to its full potential. That’s because most companies are data rich but information poor.

Often, data sources from multiple unit operations aren’t system-atically linked and analyzed. There’s also the sheer complexity of chemical manufacturing and its countless interconnected factors and overlaid effects. It all makes the numbers hard to penetrate and gain meaningful insights from.

Linking process control data with quality control and yield data is a positive first step, helping to spot yield losses and the root causes of quality issues. However, prescriptive recommenda-tions – intelligently analyzing and recommending actions to optimize production based on predictions of future trends – demand an extra layer.

Now, with the capabilities industrial AI and machine learning offer, chemical manufacturers can unleash the potential hiding in their process data and continuously adapt how they work for maximum safety, reliability, efficiency and profitability.

What is machine learning?
Machine learning is one of the building blocks and a specific subset of AI. It enables computers to spot patterns and rela-tionships in both their experiences and available data, and so learn from them.

Using its statistical methods and adaptive algorithms, computers can find previously unknown solutions and make decisions based on real data, with minimal human interven-tion, rather than follow explicit programming instructions.

Why you can’t afford to ignore AI.
Companies that have introduced industrial AI in the chemical sector are already seeing big benefits.
A good example is one world-leading raw materials manufacturer introduced two real-time optimizers into its phenol production process as part of first digital transformation project.
Using machine leaning and predictive models to recommend process improvements, it grew output of its phenol intermediate product by 2.5% and production by over 5,500 tons a year3. That translates into a value gain of USD 5.5 million.**
Just one of many illustrations of the potential adopting machine learning early can unleash.

Make the paradigm shift and stay ahead.
McKinsey believes AI is nothing short of a paradigm shift. A move away from hard-coded, expensive, inflexible, first-princi-ple-based solutions to adaptive self-learning solutions that use vast data and machine learning to smart effect.
The technology offers the potential to introduce clarity and predictability, reducing errors and uncertainty while boosting productivity. It does away with labor-intensive, laboratory testing to weed out inefficiencies. Instead, a robust, real-time online view helps optimize plant production and control.
Companies that grasp the opportunity early will be more able to anticipate and adapt to changing conditions, meet customer demand, accelerate product innovation and optimize across the value chain. And, ultimately, generate higher margins and outperform their peers.
In contrast, the risks and costs of not adapting can be high – especially for those with volatile margins and capital-market pressures.
If you have substantial assets, have already been gathering machine and process data or recognize you need to, now’s the time to take the next step. To read, interpret and act on it. To improve the efficiency of your existing plants. To secure and improve your bottom line. And to outpace your rivals.

You can start now.
Digitalizing is not so much an option as a strategic necessity. It’s a way to access and analyze data like never before and maintain your competitive edge. The good news is the building blocks for the change are already there.
These large data stores are a possible treasure trove; one that might readily be used to create algorithms to improve everything from throughput to energy consumption to profit. But where do you turn for the expertise and capacity needed to catch up quickly?

To make or to buy?
New technology investments usually come with a choice over whether to make or to buy what you need. With technology such as machine learning, there’s less of a choice – at least for now.
Putting it to work for you demands a range of skills: from how to build its algorithms to how to collect and integrate the data needed to train them, and ultimately oversee that training.
This graphic shows the mix of skills a business might typically need to develop AI for itself. A mix that requires a large invest-ment of people, time, and money.


If you can’t spare them, how can you cost-effectively set up and manage an advanced ecosystem? One that helps you work smarter, maximize value, deliver on your most pressing KPIs and, ultimately, secure your competitive edge?
The answer for most manufacturers and producers is to join forces with external experts in machine learning and data analytics.
It’s also worth making sure the partner you choose has extensive knowledge in your domain or is, at least, already working with several companies in the same industry. That’s because algorithms need to be carefully selected and combined to consider chemical process specifics. And those trained with more data – covering more states and situations a plant or asset could encounter – typically deliver better-quality outcomes.

Most business know they need to digitalize but lack a strategy for how. The gap between ambition and execution can be wide because they’re often unsure where to start converting poten-tial savings into actual ones.
Even plants with large numbers of connected sensors and systems, gathering huge volumes of data, will find it of little use without defined plans for how to store, harness and monetize it effectively.
With Navigance, your starting point is easy.
At the heart of Navigance are state-of-the-art artificial intelli-gence and data analytics technologies to unlock the potential in your operations.

They analyze all relevant historical and real-time process data from across your plant, spotting patterns and recommending settings to help meet your objectives through a single online dashboard. That’s backed by insights and advice from Navigance’s own process and catalyst experts. They work as your partner to help optimize performance within weeks – then around the clock.

We’ll be there for every step of your journey. Starting with feasibility checks to determine your digital readiness and fast, pain-free deployment. Right through to ongoing analysis, recommendations and expert advice to help you continuously optimize, boost process efficiency and your profitability.

Advanced technology meets world-class expertise
Navigance GmbH is an independent subsidiary of specialty chemicals company Clariant.
Our team brings years of experience spanning data science, software development, UX design, engineering, process, and catalyst expertise plus deep-rooted industry knowledge. We’re passionate about all these areas and bring them together to create value for you.
The result is Navigance. State-of-the-art software and exceptional service. Helping you unlock hidden potential in your plant and optimize with intelligence.

Digitalization is here and it’s accelerating. Systems, tools, processes and devices are increasingly interconnected, improving both how we live and how we work.
Manufacturers that embrace the new technologies making it all possible will be the ones that stay relevant and ahead of the competition.
The key challenge is how best to use the Industrial Internet of Things to improve reliability and efficiency, reduce costs, strengthen security and compliance, and keep innovating.
Many plants already gather large amounts of data on their production floors. However, few have the systems to sufficiently extract, compile and filter data from across unit operations, then analyze it to inform faster, better decision-making.
This rich gold mine of data is crying out for the support of intelligent industrial technology such as machine learning. Its algorithms can be trained to interpret and learn from process data. They can pinpoint the levers to pull to increase process efficiency and hit the operational KPIs you’re targeting.

Navigance translates these insights into prescriptive recommendations that help optimize control of process variables. It also flags signs of potential future issues early, before they impact production and quality, costing time and money.

Such technologies need human experience and expertise to train and make the most of them but investing in these in-house can be hard for businesses to justify.
So Navigance provides everything you need as a service. Years of process, catalyst and industry experience, the latest tech-nology plus real-time advice to help you continuously optimize your operations and maximize your bottom line.
We’ll take the leg work out of your first steps into digitalization and tailor a solution to suit your plant, its setup and the critical KPIs we identify together.

Navigance is ready to go. Are you?

[Footnotes and references]
* Gross Value Added: a measure of output value for goods and services produced in a certain sector.
Think of it as that sector’s contribution to economic growth, similar to a country’s GDP.
** Calculation based on a market price of phenol of ~$1000 recorded in January 2020.
1 Artificial Intelligence in Manufacturing Market by Offering (Hardware, Software, and Services), Technology (Machine Learning, Computer Vision, Context-Aware Computing, and NLP), Application, Industry, and Geography – Global Forecast to 2025 by
2 How AI boosts industry profits and innovation by Mark Purdy and Paul Daugherty – Accenture research (2017)
3 Cepsa Press release: “Cepsa optimizes its chemical processes with artificial intelligence-based technology” intelligence%E2%80%93based-technology
4 AI in production: A game changer for manufacturers with heavy assets – McKinsey & Company white paper (2019)



Share this on:

[user_registration_form id=”41351″]