Situation Before Implementation
Every day, a single staff member logged in to multiple remote monitoring sites and transcribed generation figures into Excel. Including the first-pass assessment of anomalies, this took roughly 160 hours per month, and checks fell behind during busy periods. In some months, delayed detection of anomalies actually led to greater losses in electricity sales revenue, so operational burden and lost opportunity were occurring at the same time. Furthermore, when managing the validity periods of connection study responses overlapped with responding to additional utility queries, it became difficult to prioritize between monitoring and application work.
What We Did
We redesigned the monitoring flow into three stages—"Data Collection," "Anomaly Detection," and "Notification"—and set up an overnight batch to aggregate the previous day's results. Operations were changed so that a daily report is distributed before the start of the working day each morning. Anomaly detection now uses two criteria—warning information on the system screens and the deviation rate from the plant group average—and the prioritization of on-site inspections has been standardized. In addition, "Sites Requiring Review," "Sites in Submission Preparation," and "Completed Sites" can now be viewed on the same dashboard in the weekly review, eliminating the fragmentation of project management.
Results
Monthly workload fell from roughly 160 hours to roughly 10 hours. Manual work can now focus on "reviewing distributed reports" and "deciding on responses," eliminating dependence on specific individuals while maintaining monitoring quality. The time from anomaly detection to initial response has been shortened, which also helps minimize the risk of interrupted electricity sales. As a result, operations can continue at the same quality even when staff change, and the reproducibility of the O&M organization as a whole has improved.
