Umetrics Suite Blog

How Pulp and Paper Companies Are Reducing Production Costs Using Multivariate Data Analysis

April 23, 2020

While many other industries have implemented multivariate data analysis software for process optimization and control, it is still not very common in the pulp and paper industry. However, multivariate data analysis has a very promising potential for both cost reductions and quality improvements in pulp and paper mills. No capital investments are needed, the implementation can be done remotely, and the software typically requires no permits.

A Sartorius analysis of pulp and paper mills that use multivariate data analysis software for process optimization shows that companies can save millions of dollars per year, as evidenced by successes in the pharmaceutical, food and beverage, and other industries. Below are some industry examples of how data analytics has already saved money for pulp and paper companies.

Umetrics paper and pulp industry cost savingsMultivariate data analysis can help paper and pulp manufacturers save money, reduce waste and improve processes—without expensive investments in capital equipment.

Like in other industries, pulp and paper companies need to maintain consistency and quality in production while at the same time keep a high production rate. The pursuit of higher productivity puts a lot of stress on the equipment. Interruptions can be very costly, and a high variability is bad for quality. Thus, any measures to keep a consistent production – or even improve it – without exhausting or damaging the hardware can imply big savings.

Whereas capital investments sometimes may be necessary to improve the production process, these investments generally come at a high cost, take time to implement and often require regulatory permits. In addition, new personnel or extensive training might be required.

Cost Savings and Risk Reduction with Multivariate Data Analysis

Multivariate data analysis software is another means to improve the production process that has not yet been explored by many pulp and paper companies. Those that have implemented process analytics, however,  have realized huge gains in terms of cost savings and quality improvements. The use of analytics software shows that existing operations, and existing hardware, can be optimized to a higher degree. The software is usually far less expensive compared to capital investments, it does not typically require permits, it is quick to implement, and the implementation can even be done remotely without having to slow down or interrupt operations during implementation.

Process optimization and control by use of multivariate data analysis means that pulp and paper mills can significantly reduce risks in their production. They can also realize savings from faster startups or quicker grade changes with real-time process monitoring. Furthermore, reduced process variability provides consistent quality and reduces spoil. Other optimization opportunities are in the improvement of steam and power generation, including recovery boilers, resulting in additional savings on fuel, as well as keeping emissions under control.

Quick Return on Investment for Pulp and Paper Companies

An analysis by Sartorius of pulp and paper mills that have implemented SIMCA and SIMCA®-online for process optimization and control shows that the return on investment on the software cost is at least six times, more likely nine times or more, per year. In short, companies can save millions of dollars in just a few months with the software. Below are a few examples of process improvements with multivariate data analysis in the pulp and paper industry:

In one company the goal was to improve recovery boiler efficiency by washing as much of the solid as possible in the recovery boiler to produce more steam for the plant. The plant manager had calculated that a 0.25% change in solids would decrease steam loss worth over $100,000 a year.

By improving the process with SIMCA analysis software, the company achieved a 0.5% reduction of solids and estimated reductions over the next year to be 1%-2%.

[Watch the recorded webinar to learn more]

Another example of process optimization is a paper specialty company that implemented SIMCA and SIMCA-online for real-time assessment of product quality.

By assessing quality in real-time, and thus reducing the dependence on lab tests, the company was able to reduce start up time by up to 50% with the corresponding reduction of spoil. This has led to savings of nearly €300,000 on one machine.

Yet another example of process optimization is a global paper manufacturing company that implemented SIMCA-online to monitor its production process in real-time and to predict paper quality output. The result was a reduction of operational costs and a more efficient use of resources, while the company at the same was able to maintain paper quality levels. 

[Read paper manufacturing case study]

Multivariate Data Analysis Software—How It Works

With multivariate data analysis software, a model of the production process at optimal performance is created, based on multiple variables that have a causal effect. Historical data that characterize your product or process, such as yield, output and quality parameters, composition, pressure, temperature, flow rate and other physical units, are used to build the model.

The data can come from both online and offline sources, such as data from process control systems as well as data from lab analyses. There are no limits to the amount or types of data that can be used. The system can take hundreds of thousands of data points, or even millions in some cases, to provide confidence in the model.

The trajectory of the process at optimal performance is displayed visually as a center line on a screen with upper and lower limits of confidence that function as warning and alarm levels if the process is going out of control. Instead of monitoring a large number of single variables on multiple screens, the operator can now monitor one multivariate trend on one screen, in real-time, and see how the process is moving along. Alarms notify the operator if the process is deviating and which variables that need to be changed.

blog10-pulpandpaper

Predictive Control for Process Optimization

The benefit of multivariate data analysis is not only better process monitoring, but also the different levels of control it provides. One of the main advantages is so-called predictive control, meaning that the software will predict how the process is developing and in which direction it is going. This implies that it is possible to change the variables early on, if the process is predicted to start deviating. This can be done automatically in the form of a closed loop, or it can be done manually, where the operator can change the indicated variables. In other words, predictive control is a very valuable tool to ensure production success.

Fast and Easy Implementation

For those not used to multivariate data analysis, implementation and model building might seem like a daunting task. However, it takes only a few hours to implement the software and a few days to build the models. The implementation can be done remotely outside of the manufacturing plant. Sartorius provides training when SIMCA-online is installed and data scientists at Sartorius guide customers through the model building and general use of the software. Sartorius can also build the models for the customer, if desired.

Book a Free Demo

Want to Know More?

In a previously recorded webinar, Bob Davis, Key Account Manager at Sartorius, provides a more in-depth explanation of multivariate data analysis and how it can be used to save costs and improve quality in the pulp and paper industry.

Watch previously recorded webinar part I

Don't miss your chance to register for part II of this webinar series on May 14 that will focus on using multivariate data analytics software to reduce chemical consumption and emissions for pulp and paper mills.

Part II of this webinar series is offered in two sessions on May 14:

1. Thu, May 14, 2020 2:00 PM - 3:00 PM EDT

Register for webinar session I

2. Thu, May 14, 2020 9:00 AM - 10:00 AM EDT

Register for webinar session 2

 

 

Topics: Multivariate Data Analysis, Manufacturing Quality Control, Manufacturing Processes

Stefan Langner

Written by Stefan Langner

Stefan Langner is Market Manager for Process Industries (Food & Beverage and Chemical) markets at Sartorius Stedim Data Analytics.

Leave a comment

Search the Blog

    Subscribe to the Blog

    View the:

    Data Analytics Glossary of Terms

    List of Webinars

    Get a free trial