TY  - GEN
KW  - Roller Compaction 
KW  -  Expert System 
KW  -  Artificial Neuronal Net 
KW  -  Compression 
KW  -  X-ray computed micro-tomography
UR  - https://archiv.ub.uni-heidelberg.de/volltextserver/10111/
Y1  - 2009///
ID  - heidok10111
AV  - public
N2  - Roller compaction is a dry granulation process which is still not well understood since a large number of processing variables and inherent material attributes make the description of the mechanisms involved in the process extremely complicated. Neither the intermediary compact (i.e., the ribbon) nor the final product (i.e., the granule) is normally used as dosage form. Thus, the granule is usually processed for the production of tablets. When the substance needs to be dry-granulated prior to the tablet production, not only the initial composition of the formulation, but also the processing conditions (device parameters, intermediary steps, etc.), affect the workability of the granule and, consequently, the characteristics of the produced tablets. Thus, the development of a formulation and the optimization of the production process are necessary to meet the mandatory tablet requirements that warrant the pharmaco-therapeutic success of the drug. Expert Systems (ES) are tools that result from combining experience and computing. They can simplify the task of development of new formulations in an accurate, efficient and rapid way by processing, extracting and putting to anyone?s disposal data contained in their data base. This work pretends to join the effort of others on the characterisation of the compaction. The experiments described here supply key hints for the understanding of the process and the variability in the product attributes due to composition of the formulation and processing conditions. Finally, the information and the predictive models created during the experiments have been implemented in the development of an ES in order to demonstrate the benefits of these applications in formulation research. For that purpose, this work reports experiments that analyze the compact process visually and use a critical attribute of the ribbon, the solid fraction, to evidence the interactions between the powder and the rollers, and to describe the effect of the lubrication and other compaction conditions over the dynamics and the densification of the material between the rollers. Moreover, the suitability of a number of techniques for the characterisation of the ribbon solid fraction is analyzed and compared. Furthermore, after close examination of the ribbon quality and also of the granule and the tablets produced from different pharmaceutical substances it has been shown that these product characteristics are strongly affected by the initial composition and by the parameters and the conditions of the process. Moreover, predictive models based on computer general mathematic regressions and artificial neural networks, have been developed using gathered experimental data. The models identify satisfactorily the underlying relationships between the independent variables (composition and process parameters) and the product characteristics, and deliver prognoses of the properties of the intermediary (ribbons and granules) and the final products (tablets). Finally the models are integrated into an ES that has shown to be a user-friendly application that manages the information and retrieves the models to generate a report that includes predictions of the product attributes and helpful information to be born in mind during the production. ESs are extremely useful tools, especially for the pharmaceutical industry, as they ease the formulation development, reducing effort and expenditures, and increasing the consistency, the quality and the efficiency of the formulation tasks. In addition they ensure the permanency and the accessibility of the knowledge and assist the work of both experienced and novice formulators.
TI  - Roller Compaction of Pharmaceutical Ingredients : on the Undrestanding of the Compaction and the Use of Knowledge based Applications in the Formulation of Tablets
A1  - Miguélez-Morán, Andrés-Miguel
ER  -