At TAPPICon 2025 in Minneapolis, the Technical Association of the Pulp and Paper Industry TAPPI brought together industry professionals to discuss the latest advancements in data analytics and their applications in optimizing critical process phases in paper production, particularly during grade changes.
In paper industry, grade changes are a critical yet complex aspect of production. Frequent transitions between product types can lead to disruptions such as machine downtime, speed fluctuations, and increased off-spec production. These challenges not only affect throughput but also increase costs and complicate production planning. Traditionally, operators relied on their expertise to manage these transitions. Today, however, data-driven solutions are providing a more consistent and reliable way to navigate these critical phases.
Applying batch production principles into papermaking environment originated from recurring customer challenges. Many clients were looking for ways to standardize and improve consistency and efficiency during transitional process phases, such as product or recipe changes. These transitions involve repeatable process segments often around 30 minutes before and after a grade shift that closely resemble a batch in structure.
We observed together with our customers that different transitions often follow distinct patterns and should be treated individually. This opened the door to applying batch-like thinking: analyzing historical data to identify repeatable sequences, then using that insight to guide and optimize operations moving from manual intuition to data-driven standardization.
We realized that a grade change from, say, B to A is very different from A to C, and each should have its own model. Instead of relying on operator intuition, grade changes could be standardized through historical analysis and modeling—much like recipe-driven batch operations in other industries.
This realization made us introduce visual analytics tool we call the batchgram into papermaking context.
A batchgram displays how a process variable behaves across many similar cycles, for example, past grade changes from Grade A to Grade B. Visualized using historical data, it highlights optimal operating zones (green areas) and problematic ranges (red areas).
Unlike traditional trend graphs, a batchgram puts each live transition into context. Operators can see, in real-time, whether they are on the expected path or drifting toward risky territory. It also shows how much leeway exists in any given phase, helping users understand when precise control is needed and when there’s more flexibility.
It’s not just about a single golden batch - it’s about knowing the range of success and helping teams navigate that window with confidence.
An example of a batchgram visualization comparing multiple production cycles, with performance zones clearly indicated for process optimization.
While paper manufacturing is a continuous process, key moments like grade changes or machine restarts have a clear start and end point. This is where batch thinking shines.
Having worked with numerous paper mills, we've observed that transitions such as grade changes share many characteristics with batch production. By analyzing when process variables begin to shift ahead of a change, and how long it takes for stability to return, we can define a “batch window” – a time-bound segment that can be monitored and optimized.
Our observations have consistently shown significant variation in raw material use during these windows, even in otherwise successful transitions. Batchgram helps bring these inefficiencies to light, often in surprising ways. Deviations have been noted not only in fiber consumption, but also in parameters like pressure levels, indicating that multiple critical factors may drift during transitions. In one ongoing case, Batchgram supported the identification of a recurring spike in fiber use during specific transitions. By addressing this, the mill was able to reduce unnecessary raw material usage without compromising quality.
Batchgram real-world benefits
Applying batchgram models has already shown potential benefits in practical use:
As analytics and automation technologies mature, batchgram-style tools will evolve. More dynamic models that adapt to real-time conditions can help mills further refine their control strategies and workflows.
The data already exists. What we need is a better way to use it. This approach empowers operators and engineers with real-time insights that support smarter, more consistent production.
In an industry where every grade change counts, bringing a batch mindset to continuous production might just be one of the smartest moves a mill can make.
Up next in this blog series:
We’ll dive deeper into how real-time monitoring and predictive insights provide users with actionable insights that further reshape operator decision-making towards optimal conditions. Don't miss it!