The development of a full-scale system for concrete mix design optimization

The objectives of this research project are to develop a concrete mix optimization model that takes into account all of the known factors affecting concrete mix production, and customize the model to a specific concrete batch plant. Using historical data available from ready mix plants, an attempt was made to demonstrate the new mix design concept using as many variables as possible. Linear programming was selected to implement the optimization model using the LINDO optimization package. A prototype was developed and demonstrated a superior capabilities in meeting required specifications and recommending optimum mix ratios. It also showed that significant cost savings could be achieved by implementing the proposed model. Accordingly, a full-scale system was developed and tested. The developed system proved to have the capability to determine the optimum concrete mix design based on client's specifications, job requirements and available aggregate properties. It recommends the optimum combined aggregate gradation within code limits and predicts concrete properties. Furthermore, the system provides statistical information that can be used for quality control. Finally, as the actual concrete production data becomes available, it can be added to the database and used in future designs. This paper reviews the model development and implementation steps. It also summarizes the process of developing the full-scale system. Validation and reliability of the developed system are then discussed.