Ken J. Nixon

Probabilistic interpretation of LDN confidence ellipses with reference to forensic applications

Data from lightning detection networks is often used for forensic purposes: to validate insurance claims or even to determine the cause of death. Stroke location and current estimates are subject to measurement error, and the dilution of precision is reported in terms of a median confidence ellipse and χ2 distribution. A method is presented to derive the probability density function from the reported statistics, and hence calculate cumulative probability densities. Two hypothetical cases are presented, with real lightning detection network data, to illustrate the obstacles constraining direct forensic application of the data.

Observations of lightning discharges on Brixton Tower

Observations of lightning strikes to Brixton Tower in South Africa are investigated in this paper. The tower is topographically situated in an ideal location to study lightning strikes, and has the benefit of a physical height of 250 m. Observations presented in this paper were made through photographic recordings of lightning events from November 2009 to May 2010. Lightning data from the Southern African Lightning Detection Network (SALDN) is used to match recorded lightning strokes to photographed events. The SALDN provides the associated stroke parameters for assessment of the events in terms of peak current values as well as rise and decay times. Particular focus is given to studying positive polarity events. This includes an investigation into positive strikes to the tower and an event of a downward positive flash that appears to initiate an upward flash from the tower.

Evaluation of the South African lightning detection network using photographed tall tower lightning events from 2009 – 2013

The performance characteristics of the South African Lightning Detection Network (SALDN) are evaluated using photographed lightning attachments to a tall structure over a four year period (2009-2013) after increased numbers (19 to 24) and repositioning of sensors in 2011. The structure is not instrumented and the evaluation only focuses on flash detection efficiency and location accuracy. During the four year period, 66 flashes to the tower were photographed and 185 negative SALDN strokes in 39 flashes were time-correlated with the photographs. It is seen that the flash detection efficiency increased from 50% to 76% after the upgrade with this being considered a low estimate given the majority of events are upward in nature. The median location accuracy is found to be 280 m and maximum location error is decreased after the upgrade.

Comparison between numerical and passive integration of flat plate electric field sensor output

It is required to determine the best configuration for sensors to be used to in an electric field study on lightning occurring in the Johannesburg region. The tests performed were to determine the best sensor configuration to be used and to validate sensor output against a known source. Using a pair of flat plates an impulse was applied to the flat plates resulting in a known electric field. The output of the flat plate electric field sensor, was then compared to the known input. It was shown that despite the resistive load being the simplest to implement, it resulted in problems that require more advanced acquisition tools. Whereas using the passive integration of a capacitor, resulted in a more accurate output, which does not require complex acquisition hardware.

Developing an approximation to the Heidler function - With an analytical transformation into the frequency domain

IEC 62305 utilises the Heidler function as the standardised lightning current waveshape because it mimics the properties of a real lightning stroke. There is no analytical solution to the integral of the Heidler function and this means that it is not possible to obtain an expression for the Heidler function in the frequency domain. There are several approximations that are used to overcome this shortcoming. This paper proposes a new approximation that is designed in the Laplace domain. Initial results show that the amplitude differs with that of the Heidler function by no more than 3.7%. Furthermore, the first derivative of the approximated current waveshape, dI=dt, differs with that of the Heidler function by 10.5%. This however can be explained by the steepness factor not being calibrated correctly for the same shape Heidler function. The approximation is created in the Laplace domain and therefore it is trivial to plot the frequency spectrum of the approximation. Frequency analysis shows that the approximation agrees with those of other researchers. It is concluded that the initial results are evident of a promising approximation to the Heidler function.

Can LDN data provide corroborating evidence for a Lightning event at a specific geographic location

In many cases of lightning attachment, the exact timing of the event is unknown (property damage, loss of life etc.) making comparison with LDN data difficult. A methodology for matching data from a Lightning Detection Network (LDN) with to a possible location of attachment using only the reported positions of strokes is proposed. A case study of lightning attachment to a tall structure (Brixton Tower, Johannesburg, South Africa) is used to evaluate of the methodology. Photographs of the lightning attachments are time-correlated with data from the South African Lightning Detection Network (SALDN) and these results are compared with the results of the methodology. It is found that the methodology returns 40% false positive results and further evidence is needed to distinguish the time-correlated matches from the false positive results. The median confidence ellipses provided for each stroke are also investigated but it appears that this does not provide the evidence needed to do this. It is concluded that LDN data can provide corroborating evidence of lightning attachment to a location but cannot be used in isolation.

Computer laboratory infrastructure in engineering education-a case study at Wits

Computer laboratories used in engineering education are investigated by considering the requirements, constraints and infrastructures involved. Three primary requirements are identified: to provide access and connectivity, to support the teaching and learning process and to provide assessment facilities. The use of fat client and thin-client infrastructures in engineering education is discussed using two computer laboratories at Wits University (South Africa) as a case study.

Volume Segmentation in a Stratified Vertical EWH Tank Using Steady-State Element Cycles for Energy Balance

An appropriate volume segmentation method for a multi-nodal energy model of a vertically oriented stratified electric water heater tank is determined through the use of energy balance during steady-state cycles under different thermal insulation schemes. A multi-nodal approach is chosen due to the transient limitations of single-node or bi-nodal methods during draw events. This paper only focusses on the steady-state condition to prepare for a multi-nodal model for draw events. This energy balance methodology will be extended to the tank in dynamic state.

A preliminary assessment of lightning activity around a South African mining town, eMalahleni

This paper presents the results of a preliminary study of lightning activity around eMalahleni, a South African town which is surrounded by an estimated 160 mines and 13 coal power plants. This area also falls within one of most polluted regions of South Africa with recorded particulate matter (PMio) concentrations of about 80 μgm-3 between 2009 and 2012. Cloud to ground (CG) lightning flash data from the South African Lightning Detection Network (SALDN) for 2013 was used to calculate the spatial distribution of overall flash density, positive flash density and average multiplicity in the region within a 250 km radius of eMalahleni. A total of 1127186 CG flashes were recorded in the area in 2013. The highest average flash densities in excess of 12 flashes/km2/year were observed over the regions that are part of the Drakensberg escarpment. Flash densities ranging between 6 and 12 flashes/km2/ year were also observed near some of the mines that are not on the Drakensberg escarpment. Positive flash densities less than 1 flash/km2/year were observed near mines and power stations compared to flash densities greater than 1 flash/km2/year observed elsewhere. This might be an indicator that both elevation and mining activities are affecting CG lightning activity in this region. However, more data analysis needs to be conducted before conclusions can be drawn on the effects of mining and elevation on lightning activity in the region.

A method for determining LDN confidence ellipse information for ground-strike points

This paper describes a method for determining confidence information for LDN ground-strike points. Ground-strike points can be determined using clustering algorithms and in the process, the strike point location is refined. However, the confidence ellipse information is lost when this is done. By treating stroke detections as multiple observations of the same event, Gaussian merging techniques can be used to develop a confidence ellipse for LDN ground-strike points.