.png){kind=icon}
.png){kind=icon}
Return
Automated method for selecting optimal digital pump operating strategy
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2023.120509.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Commercially available hydraulic pumps exhibit lower performance at a wide range of operating conditions, resulting in a significant decrease in productivity and efficiencies as low as 30%. A new class of digital pump technology has been developed to overcome these limitations. This work builds upon a previously developed high-efficiency digital pump that utilizes high-speed on/off valves. There are four distinct strategies in which the digital pump technology can operate, and each of these strategies offers unique advantages that make it suitable for specific operating conditions. In this work, we introduce a novel approach that utilizes expert systems technology in combination with several automated machine-learning methodologies to determine the most efficient operating strategy. Different supervised machine-learning algorithms are investigated to predict the overall efficiency of each operating strategy. The best machine-learning model is then employed for selecting the most efficient operating strategy given any operating condition. Given the findings, the efficiencies predicted by the machine-learning model align well with the efficiencies measured through experiments. This innovative approach has resulted in substantial energy savings compared to the currently available pumps in the market.
Keywords:
Digital pump
Machine learning
Efficiency forecasting
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W
