Tekkem SlakerAI
This article addresses RDP Technologies in relation to the current AI revolution and how its premier product line has embraced and utilized it since 1998.
Our Tekkem SlakerAI is mechanically a fairly simple mixing tank, consisting of water valves, lime feed screws, a discharge and inlet valve, along with sensing elements - namely load cells and thermocouples.
In order to produce a system that could consistently produce batches at very high precision, regardless of the varying mechanical tolerances, changing atmosphere and water temperatures, varying plant water supply pressures, variations in the supplied lime and other items, we knew the controlling program would have to be created to adapt to everything, regardless of everything, including where it was installed.
The very first program was written with all of this in mind, and the end result ultimately had to include elements of what is currently referred to as AI (artificial intelligence). So, since 1998 we have been programming and over time improving our programs using this framework. The result is a product line that is unmatched in performance, reliability, and longevity.
A simplified walkthrough of how the Tekkem SlakerAI operates in terms of its learning ability follows:
The Slaker is a batch process, and within that process are several discrete stages to create the final full tank batch. Before, during, and after a batch is complete the program is continuously learning and adjusting parameters to create an extremely precise fully slaked slurry product that keeps lime in suspension and has a minimal grit content.
At all Times – Ambient temperature and plant water temperature is monitored to determine some proprietary parameters within the program.
Stage 1 – Initial plant water enters the Slaker through water valves mounted at the top of the tank. As water enters, temperature and weight are monitored to determine the exact target weight of water needed to slake the lime. This is continuously changing during the entire water addition based on the sensed elements discussed. In addition, the water valves must close at some point prior to target weight due to water pressure changes and response variation of the valve itself, as well as the location of the valves at the top of the tank. This is determined based on previous batch information and calculated during the entire stage.
Stage 2 – The amount of lime to be added is determined by the user setpoint concentration. However due to the varying individual nature of the screw and the current quality of lime being delivered to the system, as well as the lime inlet valve, proper shutoff of the lime screw prior to meeting its target weight is critical. This is determined based on previous lime feed information and calculated during the entire lime addition.
Stage 3 – The water and lime must mix for a precise time at a user setpoint temperature to create an excellent slurry at the exact setpoint. During this mixing the temperature is rising and reaches the desired temperature due to the precise additions of both elements. Slaking ends only after the program determines the slurry has reached the perfect slurry solution. Data is retained about the lime reactivity of this mixture and is used for calculations in the next batch.
Stage 4 – Water is added to the tank to create the final concentration of solution based on the user setpoint. The program uses the current live concentration to determine the final amount of water that must be added. As in stage 1, the water valves must close at some point prior to target weight due to water pressure changes and response variation of the valve itself, as well as the location of the valves at the top of the tank. This is determined based on previous batch information and calculated during the entire stage.
Stage 5 – Slaker contents are discharged into another tank. During discharge flow is monitored and the program determines when the tank no longer contains material that can be discharged. Any remaining scale or residual material is stored and the tank weight is programmatically re-zeroed to prepare for the next batch.
To summarize, RDP has created its intelligent control systems based upon what is currently known as AI, since 1998. We continue to develop our programs and stay far ahead of the curve, as our past and recent successes have clearly shown. There is the best, and then there is the rest!