Adverse space weather conditions have been shown to be directly responsible for faults within power networks at high latitudes. A number of studies have also shown space weather to impact power networks at lower latitudes, although most of these studies show increases in GIC activity within networks not directly related to hardware faults. This study examines a GIC event that occurred in New Zealand's South Island power network on 6th November 2001. A transformer failure that occurred during this day is shown to be associated with a change in the solar wind dynamic pressure of nearly 20 nPa. Measurements of GICs recorded on the neutral lines of transformers across the Transpower network during this event show good correlation with a GIC‐index, a proxy for the geoelectric field that drives GIC. Comparison of this event with GIC activity observed in the Transpower network during large space weather storms such as the “2003 Halloween storm,” suggests that solar wind shocks and associated geomagnetic sudden impulse (SI) events may be as hazardous to middle latitude power networks as GIC activity occurring during the main phase of large storms. Further, this study suggests that the latitudinal dependence of the impacts of SI events on power systems differs from that observed during large main phase storms. This study also highlights the importance of operating procedures for large space weather events, even at middle latitude locations.

In the mining industry, primary digging units such as wheel loaders are critical components due to their position at the start of the process chain. Consequently, the cost of unexpected downtime is high: this motivates efforts to provide an early warning of faults using remote diagnostics. Machines are equipped with sensors that measure machine health. Some sensors are highly correlated and a model based on machine learning techniques can leverage such relationships across sensors to detect within-group abnormalities. Autoencoders are auto-associative artificial neural networks which are trained to compress and rebuild the original input with minimal loss. The information is stored in the lower dimensional hidden layers as an internal coding. This is susceptible to a phenomenon called spillover, where the error in a single input can propagate through the network, corrupting the coding and biasing the entire reconstructed data. A denoising autoencoder is a more robust variation on the traditional autoencoder, trained to remove noise and build an error-free reconstruction. We created a denoising autoencoder to utilize the noise removal on corrupted inputs, and rebuild from working inputs. While this technique is novel to this problem it remained susceptible to spillover. We show our findings and discuss future anomaly detection techniques in correlated data.