The use of automatic control strategies can help ensure consistent, high-quality, error-free electrode production – a critical factor in high-end lithium-ion battery production and providing the performance and reliability required of these ever-evolving and increasingly in-demand power sources.
This interview with Chuck Blanchette, Field Marketing Manager for Thermo Fisher Scientific, outlines several approaches to automatic control which are ideal for electrode manufacture.
By outlining the use of control loops, measurement systems and approaches to categorizing and managing the different variables and sources of errors prevalent in electrode production, Chuck provides practical insight into this achievable means of optimizing electrode production.
How are measurements of typical electrodes performed?
A typical line will have a bare substrate such as copper or aluminum, and we measure that bare substrate with a scanner frame to get the grams per square meter. A topcoat is then applied either in a stripe, patch or continuous pattern and when we measure it again while it is wet, we can determine the total weight of the coating and substrate.
The electrode will then go through a dryer, and we measure it again to get the total dry measurement of the coating and substrate. We are doing this to subtract the weight of bare substrate and provide a coating-only measurement for the wet and the dry. Afterward, the web is flipped over, and the process is repeated on the other side.
As the material passes under the gauge, the gauge scans back and forth while the material is moving across it, resulting in a zigzag measurement path.
If the scanner is scanning faster, you will get more coverage. The scanner’s direction across the web is called the ‘CD direction,’ and the direction of actual product flow is called the ‘MD direction.’
The main reason that people apply a gauging system to their production line is to allow for the visualization of product quality. They want to see how much the product weighs, how thick and how uniform it is across the CD direction of the material.
They do this through different displays, including 3D and 2D cross-section displays. A 3D display will include a cross-section of the CD of a scan, but you can also see upwards of a hundred scans in the MD direction. This allows you to see trends in the MD or the CD.
The 2D graph is purely a cross-sectional view of the CD profile with two main pieces of information: how thick the material is and how uniform it is. The scan’s average thickness is the average of all the points in the CD direction - this is the value that a battery manufacturer compares to the target thickness value.
The scan’s uniformity is typically called its ‘CD spread,’ which can be looked at as a peak to peak, highest minus lowest or a signal value or the standard deviation of the material across the scanned area. This number should ideally be as flat as possible.
There are two different sets of alarms for these values because you could have perfect material on target and have a terrible profile. You could also have a perfectly flat profile, but the target is out of spec.
The average thickness is controlled by MD control algorithms such as pump speed control, and the uniformity is maintained with CD control algorithms such as slot die gap control.
Image Credit:Shutterstock/asharkyu
How is a variation of the CD and MD values categorized and managed?
Variation generally characterized as CD variation addressed with the CD control algorithms or MD variation controlled by the MD control algorithms. MD variation is caused by pump surges or pump speed variations, line speed or roll speed variations, and process variations such as a rolls out of round , temperature fluctuations and such.
We have categorized this further into short-term machine direction variation and long-term machine direction. Long-term machine direction variation is the amount of variation from scan to scan to scan, and we can control this with MD control algorithms that control things like the pumps.
Short-term variation is not controllable by any automatic means, but we can measure and display this using the gauging system and then optimize the production line to see how much of an improvement you have made by doing so.
CD variation is a function of things like the opening in the slot die or the gap settings in a roll coater.
An excellent example of the types of actual control algorithms for these different variations is the role of MD control in the slot die process.
The pump speed that feeds the slot die and the amount of material that goes through the slot die is controlled by the MD control algorithm. Roll coating would involve controlling the applicator roll speed and blade position, while calendaring may also include controlling the material coming through the gap in the press line.
CD control for a slot die process will involve controlling the individual die bolt or individual gap opening thicknesses across the CD direction using an automatic profile control system. For CD control on coating and calendar lines, you would be controlling these via side-to-side gap control - the left gap versus the right gap.
There is another category that we call optimized control. This is not a direct control of thickness or uniformity, but it does improve the throughput of the production line. For example, after narrowing the distribution curve with CD control, target management control allows you to automatically shift the target downwards to save material while staying within spec.
We can also leverage feedforward control for coating and calendaring to predict how the line speed will affect the thickness. This can be modeled, meaning that we do not have to wait for a thickness error to go through the gauge before we can adjust the actual thickness.
Image Credit:Shutterstock/immersionimagery
Where does the nominal MD control fit into these processes, and how are errors corrected?
Nominal MD control involves taking the end of scan average value and comparing this to the target. This is done in the thickness controller loop, which is designed to determine any errors in thickness.
Typically, this loop controls pump speed, meaning that the thickness error will now be converted to a speed target speed error, and a speed target will be set as an RPM value for the speed control loop.
This RPM error would be fed into the speed control, which would regulate the pump’s speed. We can also have disturbances in speed or thickness be fed back into these loops to maintain the overall thickness.
The other consideration for these types of control loops is that there is usually a lag distance between where you are controlling the pump and where you are measuring it with the gauge. It will take some time for the control action to get down to the gauge before this can be measured and fed into another control action.
If something happens before the control action reaches the gauge, we have a compensation feature called ‘transport lag compensation,’ which cues up all the control actions. This feature understands what is coming up and continuously looks at further errors before automatically modifying control actions accordingly.
Essentially, it will anticipate and correct any other errors before seeing the last control action. Some people refer to this as a ‘Smith Predictor.’
In the case of roll gap control or calendaring control, this is a combination of MD and CD control because it is done using the same actuators.
Typically, actuators increase and decrease signals that control the gap at the left and right positions of the roll set, interacting with each other for MD and CD control. However, the control loops should be decoupled from each other.
The error in the CD (profile) would be categorized as a flatness or tilt error, whereas the MD (thickness) error would be classified as an offset error. These control loops calculate the offset error separately from the flatness or tilt error.
There is also a third error for the center zone - the crown error. Where production lines can control the center zone, they can use that error to do so.
These three separate loops are designed to work independently to create the control moves to be output. Before the control moves are output, they are summed together so that only one set of signals is sent to the PLC as an increase or decrease digital signal for the left and right sides of the roles.
What are some of the most common approaches to optimizing these processes?
One commonly employed optimizing control is target management control. Target management control takes advantage of the quality gains from CD control to optimize material usage.
A typical process would see a lot of the material on target, some of this will be approaching the lowest spec limits, and some will be approaching the upper spec list.
As you run a CD control, the recorded bell curve will get much taller and narrower because more data points will be on target, and there will be fewer data points near the extremes of the lower and upper limits.
Once you do this and the base of the bell curve is narrower, you can shift this bell curve. What target management control does is automatically shift this bell curve downwards towards the low spec one without going under the lowest spec.
This means that the product is still well within spec, but the target is shifted. If you compare the shifted target to the original target, that is where the material savings can be made.
Slot die control is the most common CD control for electrode lines. Most lines use slot dies and are currently managed manually, but there is a trend towards automatic profile control with slot dies.
Automatic slot dies can be adjusted manually. Their die bolts are modified and controlled by feedback from the measurement profile. The main reason for this is to maintain the flatness and average thickness of the coating.
This closed-loop measurement system uses measurement data calculated based on the high-resolution data from the gauge. The gauge outputs high-resolution data, and this data is re-segmented into zones where each zone is mapped to a specific die bolt.
The die bolt zone data is used to feedback changes in power levels to each die bolt. An automatic slot die is physically similar to a manual slot die except that the automatic die has a fixed die side with a fixed lip and a flexible die side with a flexible lip connected to the automatic die bolt assembly.
When the operator adjusts this with a wrench, it will close the gap in that area of the die under that die bolt.
The feature that makes this automatic is that each die bolt assembly also has a thermal expansion block in the die bolt assembly itself. By adding a cartridge heater to the die bolt assembly, you can control the thermal expansion of that bolt block. Our control loop controls the duty cycle of the cartridge heater. If we heat it, the thermal block expands and closes the lip at that location.
There are other actuator types, for example, thermocouple feedback, where the feedback for the die bolt adjustment is temperature. In this case, the slot die has a thermocouple for each die bolt, and the die bolt temperatures are regulated using a temperature controller.
The gauge sends temperature set points to the temperature controller based on the error in the measurement instead of sending power levels.
Another alternative setup involves robotic mechanical die control. One variation of this approach requires servo motors on each die bolt to turn the die bolt as if an operator was turning it with a wrench.
Another setup sees a robotic arm move to the chosen bolt automatically and turn this. In this case, the APC algorithm will send the die bolt changing arm to the proper position via the communication link before telling it how much to turn the die bolt.
What are some of the benefits of an automatic control system?
Using an automatic control system allows battery manufacturers to get the most value out of their investment in the system. Key benefits include optimizing the manufacturing process beyond what a manual operator could achieve, ensuring faster, more consistent control actions, increasing the uniformity and the output of your product and maintaining tighter tolerances.
Other tangible benefits include material savings through reduced scrap, reduced material usage and increased production via reduc
ed startup and change over time. Automatic control systems are also vital to increasing machine efficiency and product quality.
About Chuck Blanchette
Chuck Blanchette is the Field Marketing Manager for the Flat Sheet Gauging Business, Web Applications and is responsible for the external support of the product portfolio and training worldwide. He has worked for Thermo Fisher Scientific for over 30 years in various development engineering, product management and marketing positions and has a wealth of experience in gauging, web applications, manufacturing processes and instrument and sensor design.
This information has been sourced, reviewed and adapted from materials provided by Thermo Fisher Scientific – Solutions for Industrial and Safety Applications.
For more information on this source, please visit Thermo Fisher Scientific – Solutions for Industrial and Safety Applications.
Disclaimer: The views expressed here are those of the interviewee and do not necessarily represent the views of AZoM.com Limited (T/A) AZoNetwork, the owner and operator of this website. This disclaimer forms part of the Terms and Conditions of use of this website.