Our approach to simulation
We use our industry experience to break down and simplify design problems. Often, we find that a calculation carried out on a spreadsheet has more impact and value in a feasibility study than a complex simulation. At the other end of the scale, we have developed algorithms and written software to solve complex optimisation problems commonly encountered in pharmaceutical facility design.
Since it informs the sizing of primary process equipment, the mass balance is the most important simulation in many facility designs. We have vast experience in this area and utilize a range of applications, for example:
- SuperPro Designer: We use this flowsheet-based tool for detailed mass balance simulations. Often in combination with SchedulePro for subsequent detailed scheduling. The use of both packages reduces the need for redundant data entry and provides a streamlined workflow.
- BioSolve Process: This spreadsheet-based application allows rapid mass balance generation. It is more specifically geared towards biopharma processes. We use BioSolve Process to generate easy-to-follow, parameterised mass balances for front-end designs of biopharma facilities.
- Excel: During early-phase design, the best approach is often to develop a lean mass balance from first principles in a clear, relatable format. The use of complex software packages can often hinder the rapid decision-making and relatability needed, for example, during feasibility studies. In these instances, we often develop mass balances in Excel. This allow us to maximise client interaction.
A key element of pharmaceutical facility design, process scheduling is used to validate the sizing and count of process equipment, to generate workable solution preparation philosophies and to assist in sizing of utilities and estimation of personnel requirements.
- Graphical scheduling: During short-duration early-phase design, often there is a lack of granular data and the design team need to rapidly gain a high level insight into the process. Rather than carrying out a detailed simulation using a software package, it can be more instructive to take a graphical approach. Gantt charts or equipment occupancy charts generated in Excel or Visio can provide visual impact. They can also be easily modified in workshops.
- Finite capacity scheduling: We use SchedulePro for scheduling recipe-driven processes, for example, drug substance production. This approach works well for for clash resolution around solution preparation, CIP and utilities. The resultant production schedule is used to validate solution preparation philosophies and utility sizing.
- Discrete event simulation: Witness Horizon is utlized to carry out discrete event simulations (DES). DES is a powerful method for modelling stochastic (random) processes and its also well-suited to scaled-out processes. Accordingly, we use it to model logistics, personnel movement, and for processes such as autologous cell therapy. With Witness Horizon, our specialists can create dynamic 2D and 3D models to produce realistic process digital twins.
Data analysis and optimisation
We do not constrain ourselves by relying on industry-standard software and methods. Our Simulation Bureau has experience in mathematical optimisation, data analysis and data visualisation. We can tailor or develop solutions to novel design problems.
- Solution preparation optimisation: Given dozens or hundreds of solutions to prepare in a multi-product biopharmaceutical facility, with multiple processes running simultaneously at different cadences, how do you decide on the size and quantity of solution preparation vessels? We have developed software to solve this problem using mixed integer linear programming. This allows us to develop rapid, robust designs that can optimise for captial cost or for cost of goods.
- Data analysis and visualisation: Industry standard simulation software often produces poor-quality graphics and cannot output useful statistics. We can query the results of simulations carried out in proprietary software packages and perform our own post-processing to generate more useful outputs, illuminating and informing design decisions.