AI Control Techniques for Stabilization and Energy-Saving of Pneumatic Conveying Systems

For a typical pneumatic conveying system, various combinations of mass flow rate of solids (𝑚_𝑠), mass flow rate of air (𝑚_𝑓) and pressure drop (Δ𝑃) can be considered for the conveying of a given bulk material and pipeline. However, there are only a few combinations located in the vicinity of the optimal condition that can convey the bulk material at the design or desired capacity while consuming the lowest amount of energy.

Traditional control systems exhibit several limitations, particularly in their lack of adaptability to changing operating conditions. When operating conditions change (e.g. due to varying bulk material properties), the pneumatic conveying systems will continue running with the same parameter inputs as the initial design or operating condition. Moreover, the pre-designed operating parameters are typically conservative with a certain safety margin to prevent unstable flow or pipeline blockages. Consequently, most pneumatic conveying systems operate under sub-optimal conditions, leading to excessive energy consumption. Therefore, it is imperative to develop “smart” operation/control strategies and technologies capable of addressing these issues, which can lead to improved performance, reliability and significant energy savings.

This study proposes an intelligent approach that employs both Artificial Neural Network (ANN) and fuzzy control techniques to achieve stable and energy-saving operation of pneumatic conveying systems, especially when subjected to external variations. ANN is used initially to automatically develop accurate Pneumatic Conveying Characteristics (PCC) and determine the optimal operating conditions for the system. A fuzzy controller is then developed and utilised to maintain real-time control, ensuring the system operates around the optimal condition in response to external disturbances.