Several of the mills built during the period have since closed. The reasons for their failure include poor design, poor water management, high energy consumption and high raw material cost due to the increased demand for waste paper. Our recycle mill in Northampton Pa. was one of those ill-fated plants. In 2001, the mill was purchased by Belkorp Industries Inc. and renamed Newstech PA, LP.
Newstech PA, is a deinking pulp mill with a design capacity of 420 air-dry metric tons/day. Our engineering group has been tasked with studying the redesign and feasibility of reopening and operating the plant.
At an early stage of the process study, it was determined that some form of process simulation software should be employed. When conducting a process study, it is important not to overlook any aspect of the operation. It is, therefore, very important to go through every control loop. Several different types of software were initially used to obtain a mill-wide mass balance, but always with the assumption of steady-state conditions.
VIRTUALLY THE REAL THING
Ful, dynamic process simulation means never having to say you're sorry for poor design decisions after the plant starts up.
To replicate what really goes on in the mill, a full dynamic process simulation iwas called for, something that would show tank levels raising or lowering, valves reacting to meet certain setpoints and so on. Of the simulation tools available, we found Aurel Systems’ CADSIM Plus (CP) dynamic process simulator fit the bill.
The software has been used extensively during the study but in actual application the platform has proved to be more than a just simulator—it has been the tool used for the redesign of the entire plant. The simulator has allowed Newstech to build the process flow, input all the specifications and then use them to determine the projected quality, yield and energy consumption of the process on a mill-wide basis.
At the early stage of the study, we worked with a consultant who employed dated steady-state process simulation software. When it came time to compare numbers, the consultant argued that certain water loops were short on water when the CP balance indicated otherwise.
This was puzzling until we realized that their balance model the simulation was based on had no valves, pumps, or tanks. All the consultant’s software possessed was a series of blocks which they had designated as separators and other key system elements.
In a CP simulation, when a water or pulp loop requires more or less water, valves open or close to satisfy that need, similar to what actually happens in an operating mill. This prompts tank levels to rise or fall as needed during the simulation. In this case, it’s easy to observe if the tanks are sized too small. Not having this capability produces more work and may possibly cause an error on the part of the process engineer.
Moreover, looking at a CP balance is much like looking at a process flow diagram, because the software utilizes a CAD drawing front end that allows units and streams to be drawn in any form–from block diagrams to P&C quality drawings.
A Process Snapshot
After getting the balance to converge, the controllers tuned, and everything else running smoothly, a snapshot of the process can easily be captured. Once that is done, converting to AutoCAD is one button click away. We found that with the old steady-state software, we would still have to do a process control study, chest tanks size agreement, and finally, spend hours redrawing the flowsheet in AutoCAD.
Our process begins with two Valmet batch pulpers that discharge into a dump chest. As part of the redesign, we needed to know whether or not the existing dilution chest for the pulpers was too small for the given tonnage that we wanted to push through the mill.
With CP we were able to simulate the batch sequence by inputting all of the parameters, such as batch times, batch consistency, discharge consistency, and discharge flow rate. After determining the parameters, we were able to see how quickly the pulper dilution chest would be depleted of water.
In the original design, the pulper dilution chest was on level control. The simulation model allowed us to trend the surge caused by the pulper on the white water system. The simulation helped us realize that a constant flow with the overflow routed back to the dilution chest was the optimum solution. This eliminated the surge in the first loop white water chest that occurred every time the pulpers were being diluted for discharge.
Engineering was also able to increase the size of the simulated dilution chest until the group felt comfortable with the level during discharge. However, since the pulper dilution chest was already present in the process, the group concluded that a second unit would have to be installed.
An Easier Calculation
Most of the water in the front loop comes from pulper dilution, because at that point, the stock is diluted from 8-95.5% moisture. It then became apparent that the most efficient way to cool down the first loop would be to lower the pulper dilution chest temperature enough to account for the heat added in that loop from motors and from the excess in the previous loop. The mass and energy balance made that calculation very easy since the total BTU value is calculated by the software.