 Hello, my name is Lisa Mears and I'm a PhD student at the Technical University of Denmark. I would like to introduce our paper which is published in Biotechnology and Bioengineering. This work describes the application of a mechanistic model for online monitoring of 550-litre fermentation processes which are operated at Novozymes. The process of interest is a filamentous fungus strain which is run in the fermentation pilot plant at Novozymes. This is the Novozymes Fermentation Pilot Plant Hall which is located just outside of Copenhagen in Denmark. These are 550-litre stainless steel fermenters. A goal here is to optimise fermentation processes which are applicable to the production scale. A key challenge of fermentation process development is the lack of available sensors for key process parameters, for example the biomass concentration, the product concentration and the substrate concentration. There's therefore an interest in modelling and monitoring tools to try to predict these parameters online in real time. If we can achieve this goal there's an opportunity for improved process monitoring and advanced control strategy development. The central part of this work is a dynamic mechanistic process model which predicts among other things the dissolved oxygen concentration, the product concentration, the biomass concentration and the mass in the system which is changing over time in these FedBatch processes. Since we are applying the model online as a monitoring tool, we have available process measurements which we can use in order to estimate parameters. We use a stoichiometric balance by the measured oxygen uptake rate, the carbon evolution rate and the ammonia addition in order to solve for unknown rates, specifically the product formation rate and the biomass formation rate. If we look at the model calibration we can see that there's a good fit to the experimental data. In particular if we look at the biomass concentration and the product concentration we see there's a very good fit to the data. This shows the full model output for one batch but we also see that for eleven batches which we have available there is a good fit for both the product concentration and also the biomass concentration. We have shown that the model has been successfully calibrated using a historical dataset of eleven batches. The purpose is to have an online monitoring tool and therefore it's important to validate the model online with new data. The model was then implemented with fourteen new batches online in this pilot hall. We use MATLAB to solve the model and also we use MATLAB timers in order to schedule the data collection from the OPC server and the model solving and parameter estimation. The data is read from the OPC server every thirty seconds and the model is solved every five minutes. This means that a user has an updated prediction of the biomass concentration and the product concentration every five minutes. This is very valuable as the alternative is manual sample analysis. The results of the model validation show that we have a good fit to fourteen new batches of data which were solved online. The results for the product concentration are shown where we have a root mean summer squared error of sixteen point six percent and a Janus coefficient of one point five. This shows that the relative errors in the calibration set and the validation set are of a similar magnitude. We therefore feel we have a well calibrated model and a good fit to the experimental data. In this work we describe the importance of having a robust mass model and this also includes a consideration for the evaporation rates in the system. This is since the air conditions are changing over time, they're not constant over the year. We show in the paper the different temperature profiles and relative humidity profiles that were expected in Denmark over one year. This gives the following results for the evaporation rates we would expect in the system. You can see that there's a large change in the amounts of evaporation expected in terms of kilograms per week. Thank you for your interest in this work and for watching this video abstract. If you have any comments or any questions please feel free to get in touch with the information provided. Thank you.