A new automation solution has helped Stora Enso Kvarnsveden increase TMP production and reduce energy consumption and quality variations
July 2007
By Nigel Farrand
For a number of years the relative stagnation of paper prices together with ever increasing manufacturing costs have forced paper producers into a very tight corner. This is especially true for mills relying on thermo-mechanical pulping (TMP) for the majority of the paper furnish. Increasing energy prices are forcing concentrated efforts to improve the efficiency of the process without compromising quality. Control of TMP refining has always been a daunting task with many interacting variables and little real time information on which to base control assumptions. By using a combination of advanced control and optimization technologies, supported by pulp quality analyzers and in-line sensors, Metso Automation has developed the Advanced Quality Control Solution (AQC) for TMP production. At the heart of AQC are multi-variable model predictive controls (MPC) which have been implemented at a number of TMP mills. At Stora Enso's Kvarnsveden mill in Sweden, a trial on one refiner resulted in two separate projects extending AQC solutions to nine primary and two reject refiners.
Windows on quality
The Kvarnsveden mill already had a quality window concept using the so-called 'F1' and 'F2' factors to represent fiber length and fiber bonding characteristics derived from laboratory tests. The F1 and F2 targets take into account the fact that TMP furnish quality is much more than freeness alone and provide an easily understood indication of quality. For a number of years the mill operators were manually adjusting the process to stay within defined F1/F2 quality windows to produce the optimum papermaking furnish. However, as with any manual test, the information came too late for a more aggressive approach to control. On-line Metso PQM analyzers provided freeness and long fiber content information but long sampling intervals and over 20 minutes latency chest retention often resulted in less than ideal control decisions by the operators.
In 2003, as a part of activities to optimize production, Kvarnsveden accepted a trial of the Metso AQC on refiner line R16 in TMP plant 2. The objective of the trial was to decrease pulp quality variability while at the same time reducing the refiner's specific energy consumption by increasing production. "Energy was the driving force behind the decision to use AQC, we needed to fully utilize the capacity of the refiners," says TMP superintendent, Martin Jansson.
Positive results
Based on the positive results from the trial during 2004, all seven primary refiners (R13 to R19) were equipped with the new technology the following year. Martin Jansson, takes up the story, "Prior to having AQC, operators needed years of experience to run the refiners. The new system provides us with a cruise control; operators now have more time to watch the process while the automatic control maximizes the production speed within well-defined quality limits."
In 2006, the AQC system was further expanded to cover two primary refiners in TMP 1 and the two TMP 2 reject refiners. According to, assistant superintendent, TMP, Torbjörn Holten, "The TMP 1 conical disk refiners present more of a challenge having more parameters to control than the double disk units in plant 2. They are also more sensitive to control changes and very hard to run manually."
Quality is on target, production is up and energy savings have been realized. "Total energy usage in the mill is approximately 2.5 TWh/yr, of which the TMP plant accounts for almost half. AQC makes a significant contribution to more efficient energy usage in refining," says Jansson.
Reducing variability and increasing production
Unlike single loop controls with one measured variable and one controlled variable, AQC's multi-variable model predictive controls manipulate several process variables to achieve target values for several controlled variables. Using dynamic models of the refining process, control moves are calculated within preset limits to bring the refiner as close as possible to the desired operating point. AQC controls freeness and long fiber fraction, refiner and blow line consistency, motor load and production level. The AQC control achieves the targets of these control variables by manipulating refiner gaps, dilution flows and screw speed.
MPC requires reliable process measurements updated frequently enough to be able to reflect process dynamics. While the key measurements are available from the PQM, the update interval can be long. More frequent measurements are required for continuous control, which makes it necessary for the control to simulate, with software, some of the measurements. These "software sensors" for freeness, long fiber average and blowline consistency are updated every 5 seconds to provide the control with a faster response than otherwise possible. When measure variables are available, the software sensor model is corrected for model error. For instance, every time the PQM outputs a new measurement the model error is calculated, filtered in time, and applied as a correction to the quality estimates for each line.
Cruise control - on
The operator inputs the target ranges for all of the controlled variables; freeness, fiber length, consistencies and motorload. The control then manipulates screw speed, dilution flows and gaps within the preset limits to produce stable quality pulp. Within the set limits, as process conditions allow, the control will maximize production, thereby lowering specific energy consumption. The refiner line is run at optimum capacity at all times. With reduced quality variability and more stable production, AQC can actually increase the production capacity of a refiner line.
Both the mill and Metso Automation closely monitor the performance of AQC with automated reporting online performance evaluation tools. According to Jansson, "AQC has brought more awareness about the process and running conditions." Part of this is the daily performance report, automatically generated by the AQC system and emailed to participants in advance of the morning meetings. Containing a complete graphical record of mini-trends from the previous 24 hours it also summarizes the important F1 and F2 quality windows for all controlled refiners. "The F1 and F2 windows contain the quality parameters we need most. They provide fast confirmation that things are going well or alert us very quickly to provide some remedial action in the mill," says Holten. Preconfigured trends and control status displays are also accessible from the WebMon tool, very much in use by the engineering and development staff.
Service support 24/7
Stora Enso Kvarnsveden has a long-term service agreement with Metso Automation providing round-the-clock remote service in addition to quarterly service visits. "The remote service works," says Jansson, "but it's good to have someone here on a regular basis." Alarms generated by the system are e-mailed to a Metso call center, where an engineer will be alerted to log into the system and fix the problem. In many cases, especially at night, the mill will only be aware of the problem when an email report detailing the alarm resolution arrives.
