The new QESTMonitor module uses data from QESTLab and QESTMix to automatically assess key parameters, raise alerts if results fall outside predefined or statistical limits, and make recommendations for mix adjustments. The sensitivity of the system to each key parameter can be adjusted according to the desired safety and risk profile by grade and mix family— allowing, for example, higher safety margins for high strength and project mixes. Most alerts are based on values generated by a sophisticated evolutionary prediction model, which notifies technical supervisors of issues as soon as early age results are available.

QESTMonitor comprises of the following elements:

• Setup of mixes and other data;
  
• Assessment parameters;
  
• Recommendations and alerts; and
  
• In-depth analysis of data;

In the Setup the user defines mix families (related mixes from one or multiple plants that will be assessed together), assessment period and frequency (daily, weekly, monthly, none), the trigger parameters that define the user customisable rules for recommendations and alerts, as well as the prediction methods used.

Concrete strength, density, yield, combined and individual material gradings, batch quantities and testing frequencies are then automatically assessed over night at the defined intervals. For each parameter the system computes average values, determines and eliminates outliers, re-computes average values and compares them with the trigger values defined in the setup. Thereby variability and average values will be checked. Depending on the outcomes, alerts will be raised or recommendations for mix adjustments made.

It is important that the need for adjustments is flagged early. This is why multiple regression models are used to determine predicted concrete strength values. The earliest predictions can be made based on actual batch data; refined values become available as soon as density and early age strength results are available. The system is able to adjust the focus, for example mix families which are frequently tested may use prediction models bases on just the family’s data, while those that are infrequently tested may include values from a larger prediction group.

For users who like verification of the alerts and recommendations, comprehensive statistical reports and graphs are available for in-depth analysis. All outliers are identified and comprehensive statistical and raw data are stored for each assessment. Old assessment data may be purged.

Once a business is confident that the trigger parameters are set appropriately they can rely on the recommendations and alerts without checking the underlying data.

For more information on QESTMonitor, please contact Spectra QEST sales.