The Effect of Automation on Efficiency and Productivity: Evidence from a Hydro Power Plant
How to cite
Aksoy, Cevat Giray; Choudhury, Prithwiraj. 2024. "The Effect of Automation on Efficiency and Productivity: Evidence from a Hydro Power Plant." AEA RCT Registry. AEARCTR-0014406. https://doi.org/10.1257/rct.14406-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
Automation is rapidly transforming industrial operations across various sectors, promising increased efficiency, reduced human error, and improved productivity. However, the actual impact of automation on these outcomes remains empirically understudied in many contexts. This research aims to fill this gap by investigating the effects of implementing an advanced automation system in a critical infrastructure setting: hydroelectric power plants. Our study leverages the implementation of a new Supervisory Control and Data Acquisition (SCADA) system across multiple hydroelectric plants, offering a unique opportunity to conduct a large-scale, real-time experiment on the impacts of automation. While automation is expected to improve various aspects of plant operations, its actual effects on key performance metrics remain to be empirically tested. By employing a randomized controlled trial (RCT), this study will compare the performance of industrial operations under full automation, semi-automation, and manual operation modes. The research is conducted in partnership with a major energy company, focusing on their hydroelectric plants as a representative case of complex industrial operations. The results will provide insights into how automation affects productivity, efficiency, and overall operational performance in settings where precision, reliability, and real-time responsiveness are crucial.
Intervention
Our study leverages the implementation of a new Supervisory Control and Data Acquisition (SCADA) system across multiple hydroelectric plants, offering a unique opportunity to conduct a large-scale, real-time experiment on the impacts of automation.
Experimental design
Randomization and Treatment:
The experiment will take place at a power plant over a 30-day period, with each day divided into shifts. A cross-randomized schedule will be used to randomly assign the operational mode to each shift, reducing potential biases.
Treatment 1 (Full Automation): The automation system will have complete control over operations, including process management, load control, and system optimizations.
Treatment 2 (Semi-Automation): The automation system will assist human operators, who will retain some control over decision-making and operations.
Control (Manual Operation): Operations will be conducted entirely by human operators without the assistance of the automated system.
Design details
Cited by
No citation data available for this registration yet.