DS Smith uses predictive simulation to unlock unusually complex production insight

DS Smith is a leading international provider of sustainable packaging solutions, offering plastic-free packaging, integrated recycling services and sustainably sourced containerboard and speciality papers.

The challenge - Understanding production flows to support planning

DS Smith needed a clear view of production flows at its Featherstone plant to determine whether it would need to further invest in new equipment or modify existing processes to meet evolving customer requirements.

However, production and internal logistics processes were complex and difficult to capture accurately in simulation, particularly when it came to deckling, the process where cardboard is cut from sheets on a corrugator.

DS Smith Operational Excellence Manager Andrew Gray explained: “250 human years of experience have gone into our processes, which we needed to capture for simulation to deliver value. The challenge was capturing this accurately, so we could have confidence in the simulation outputs, build trust among colleagues and maximise its potential as a tool.

The company turned to Haskoning and our Witness predictive simulation software for help.

The solution - Modelling scheduling, downstream processes and material movement

During the Haskoning discovery process, it became clear the project could be broken down into two distinct challenges:

• A simulation modelling exercise of downstream processes and material movement
• The complex task of accurately replicating the company’s scheduling processes, including the deckling algorithm used in corrugation

Simulating corrugator scheduling logic for the first time

The deckling algorithm was a level of complexity our Haskoning simulation experts hadn’t come across in 40 years of building simulation models. Effectively capturing it was critical to validating the model and gaining trust in simulation as a tool.

Haskoning Simulation Consultant Joel Wilson explained: “We had to simulate the deckling algorithm as closely as possible and then bolt on all the ways the planning department tweak it. The tricky element was modelling the human interaction, such as planners moving jobs, splitting runs and responding to storage constraints. We had to reflect reality without making the model too complicated for users.”

Building trust with robust validation

Given the complexity of the deckling algorithm, model validation was crucial for building confidence in simulation outputs. Over a month-long refining process, we validated and tweaked against past production data.

“A major lesson was that you can’t rush the process,” said Andrew. “The fact we invested in closing the data gaps and nailing the validation has meant we’ve been able to take the model to the wider business with confidence.”

Demonstrating flexibility beyond expectations

The resulting model involves hundreds of variables and thousands of lines of code. The experience has, according to Andrew, been “absolutely superb,” with the Haskoning team always available to work through issues and keep everything on track.

“What surprised us was how much we could actually change once it was working – everything from speeds to job complexity and scheduling decisions,” said Andrew. “It’s very flexible, and small parameter tweaks can make a huge difference. But once you trust the model, you can finally see what’s possible.”

The impact: Clear, actionable insights to drive informed scheduling and investment decisions

Planning for new business and investments

From testing new customer orders to validating operational choices and informing capital investments, the simulation model has become a trusted decision-making tool at DS Smith.

One early use case involved a potential new customer, with the team wanting to test their ability to deliver the specific requirements, as well as the effect on broader production. The simulation has also validated scheduling strategies and been used to evaluate a multi-million-pound capex decision. As Andrew explained: “We’re no longer relying on opinion and experience. Instead, we have real data: we can run scenarios, identify trade-offs and get to an answer faster, rather than learning and realising bottle-necks in real-time which impacts on the factory efficiency.”

Growing confidence in the power of simulation

The success of the project has increased confidence in simulation across DS Smith. Discussions about modelling additional plants and processes are ongoing. And Andrew has presented the simulation at different levels in the business, often sparking what he describes as “lightbulb moments”.

For Haskoning, this project demonstrates that tackling the most challenging elements is often where simulation delivers the greatest value. As Joel reflected: “We’re proud to have succeeded with such a complex project. It shows our determination to solve problems that matter to our customers.”