Data Explorer
Select a recorded period to investigate Queensland grid frequency or its deviation from 50 Hz. Both views use AEST and support selections of up to 24 hours.
Time Series
When did the frequency change, and how? Plot original observations or clock-aligned means, zoom into disturbances, and download the selected measurements.
Open Time Series →Histogram
Which frequencies occurred most often? Group original approximately one-second observations into frequency bins, showing counts or percentages. Download the bin totals or original records.
Open Histogram →Switch between views with the same selected period and metric. Histograms use original observations, not the averaged points on a time-series graph. Missing measurements do not contribute to bin counts.
Original observations and clock-aligned means
Original observations preserve each recorded measurement and its timestamp. Each measurement covers approximately one second.
Clock-aligned means average readings whose timestamps fall within fixed 1-second, 10-second or 1-minute clock intervals. A reading at 12:00:00.900 belongs to the interval from 12:00:00 up to, but not including, 12:00:01. One reading retains its value; multiple readings are averaged; intervals without readings remain empty.
Time Series defaults to 1-second clock-aligned means. Longer intervals smooth variation and may hide brief peaks. These means group observations by timestamp; they do not reconstruct measurements over exact clock boundaries. For rate-of-change analysis such as RoCoF, the actual elapsed time and any averaging must be considered.
Frequency histograms use original observations; RoCoF histograms use eligible signed one-second rate estimates calculated from them. Histograms default to Frequency, 2 mHz bins and Measurement count. Shared links retain their specified settings.
CSV downloads support further analysis in Excel and other tools. For comparisons across multiple days, download separate periods and combine them offline.
RoCoF — 1-second interval
Both Explorer views offer signed rate of change of frequency in mHz/s, calculated as [f(t) − f(t − 1 s)] / 1 s at recorded endpoints. The earlier value is linearly interpolated from original readings, using the same method as the summary maxima. No clock-aligned averaging is applied. Invalid or discontinuous windows are excluded, and both endpoints must fall within the selected period. The original-record download retains the underlying frequency readings. This one-second estimate does not resolve subsecond peak RoCoF.