A dynamic model of tablet film coating processes for control system design
Aqueous film coating is a common and important step in the production of most pharmaceutical tablets. Controlling this process is beneficial to maintain stable and optimal operation and the use of a model capable of describing its main physical phenomena is fundamental for control system design. This paper presents the development of a dynamic macro scale model of a pan coater based on energy and mass transfer equations. Model complexity is kept low to simplify parameterization and allow for real-time applications, such as the development of system observers to estimate critical unmeasured variables. Pilot plant data from seven batches, ran according to designed experiments, are used for model calibration and validation. The final model adequately captures the global trajectories of the main process variables; mean differences between experimental and simulated values are below 2°C for the outlet gas and tablet bed temperatures, and 2% for the outlet gas relative humidity.
Introduction
The practice of applying an external coat to pharmaceutical tablets has its roots in the 19th century (Cole et al., 1995) and is performed for multiple reasons, ranging from aesthetical to increasingly functional (Cole et al., 1995; Seo et al., 2020). This process is normally executed in batches (Ketterhagen et al., 2017) and the main equipment used for tablet film coating is the pan coater (Merkus et al., 2018), which essentially consists of a horizontally rotating perforated metal cylinder equipped with baffles to enhance stirring (Fig. 1). Heated air (preferably dry) enters the drum at the side opposite to the tablet bed, and is drawn through it by a fan, and finally exhausted from the equipment (Cole et al., 1995). A pumping system connected to spray guns transfers the coating solution to the tablet bed. If a pneumatic spray gun is used, compressed air atomizes the coating solution droplets and changes the pattern of the spray zone (Porter et al., 2009). The film thickness required to guarantee product coverage is expressed as a target weight gain established during the drug’s development stage (García-Muñoz and Gierer, 2010).
Batch coating follows basic operational steps or phases: coating solution preparation (Merkus et al., 2018), tablet loading, pre-warming or heat up (to minimize water penetration into the tablets) (Cole et al. 1995), coating or spraying, cooling or drying (Merkus et al., 2018), and unloading. During production, the end of the spray cycle is usually determined after a given batch duration is reached or by the application of a pre-determined amount of coating solution (Porter et al., 2009). Batch processing time varies depending on batch size and target weight gain but rests in the order of a few hours (Aulton and Taylor, 2013).
At present, tablet pan coating processes are normally only equipped with feedback loops for regulatory control of the main process variables (e.g., inlet or outlet air temperature, inlet air flow rate, coating solution flow rate, and pan speed) at a fixed set point (Porter et al., 2009). According to Yu et al. (2014), this is considered as the least advanced (albeit mostly commonly adopted) method of supervisory strategies applied in the pharmaceutical industry, consisting of maintaining materials and process variables as constant as possible (i.e. a fixed recipe approach). As a result, disturbances can lead to variable product quality and opportunities are lost in terms of reducing batch time and increasing process efficiency.
Process modeling and simulation are key to control system design and evaluation. However, due to the complexity of the coating process, it is impractical to model the entire process in detail; the choice of model type, scale and complexity depends on its finality. For process control, dynamic models, capable of representing the systems transient responses, are required (Roffel and Betlem, 2006). Also, it is acknowledged that macro-scale thermodynamic models are useful for predicting key process parameters known to impact film coating quality (Ebey, 1987; Prpich et al., 2010; Pandey et al., 2014b; Kestur et al., 2014).
While multiple thermodynamic models of tablet coating processes have been proposed in the literature (Ebey, 1987; Ende and Berchielli, 2005; Cha et al., 2019), with brief overviews given in (Ketterhagen et al., 2017; Prpich et al., 2010; Agrawal and Pandey, 2015), most consider steady-state operation only. One exception is the work by Page et al. (2006a) and Page et al. (2006b) in which a dynamic model specific to a Bohle Lab-Coater was developed. However, it is not clear how this model would perform if applied to coaters with different configurations. More specifically, the added complexity of separating the drum into two zones, which requires the estimation of a tablet exchange rate between them, may not be necessary for control system design and assessment. Additionally, no heat loss is considered nor is the predicted humidity data explicitly validated.
Garcia et al. (Garcia Muñoz et al., 2012) proposed a more comprehensive dynamic model capable of both tablet temperature and water content prediction. However, at the time it was presented, the model was composed of several hundred equations and variables, requiring substantial parameter estimation efforts.
Within this context, the objective of the current work is to develop a simple dynamic process model, capable of adequately describing the heat and mass transfer and thermodynamic conditions of tablet coating processes to support control system design and validation. Additionally, this model also allows for the estimation, though the use of a system observer, of the tablet bed conditions, which are normally unmeasured but important variables.
The intent of keeping model complexity low is to minimize both parameter estimation and computing efforts to allow for real-time applications. Models that run in parallel with the process open the door for the implementation of control and optimization tools in real-time (Chen et al., 2020), which is typically considered ideal. ‘Digital twin’ is the modern term used to designate the virtual representation of a physical process and real-time implementation capabilities are considered a key feature of this framework (Chen et al., 2020).
Following the definition of the model objectives and detail level, section 2 describes the main steps involved in its development. In sequence, the experimental data, collected as described in section 3, is used to estimate the parameters of the equation system (as shown in section 4). Finally, the performance of the model is evaluated in section 5.
Excipients mentioned in this article beside others: Hydroxypropylmethylcellulose
Read more on a dynamic model of tablet film coating processes
Cecilia Pereira Rodrigues, Carl Duchesne, Éric Poulin, Pierre Philippe Lapointe Garant, A dynamic model of tablet film coating processes for control system design, Computers & Chemical Engineering, 2023, 108251, ISSN 0098-1354,
https://doi.org/10.1016/j.compchemeng.2023.108251.