164 / 2018-08-25 19:37:08

The Imbalance-Induced Resistive Heating Profile analysis for Low Voltage Transformers

Phase imbalance,transformer operation,energy loss,resistive heat profile

Phase imbalance in the UK and European low voltage (415, LV) distribution networks is a widely acknowledged issue. For these LV transformers (11 Kv/0.415 Kv), phase imbalance causes extra energy losses on the transformer windings. These energy losses induce the imbalance-induced resistive heat. Then, this heat will increase the operation temperature of transformers and further increase the operation energy losses (caused by temperature increase).
There exist some references that focus on the energy losses of transformers with the phase imbalance. Reference [1] calculate the yearly extra energy losses for transformers with the phase imbalance. Reference [2] using some examples to represent the energy losses of transformers with the phase imbalance. However, these references cannot give the details of representatives of the imbalance-induced resistive heat profiles. In industry, this profile can supply more useful information rather than the yearly energy losses.
The challenge for this research is that, in the UK there are too many LV transformers require to be analyzed, i.e. approximately 900,000. For analyzing all 900,000 LV transformers, the costs of computation and human resource are incredibly higher. Thus, for easily deriving the imbalance-induced resistive heat profile in industry, this paper using a statistical approach (Gaussian mixture models) to find the representative clusters of imbalance-induced resistive heat profiles, leverage only 800 LV transformers with one year secondary time-series phase current (magnitude) data. For using the Gaussian mixture model, it is because comparing with traditional clustering method (k-means and hierarchical), the clusters have unconstrained covariance structure. This characteristic is relatively suitable to find the representative clusters with the minor data situations. Comparing with analyzing 900,000 LV transformers directly, this approach is very economical saving. Furthermore, it can efficiently give the representative profiles for analyzing the imbalance-induced resistive heat. For realizing this approach, the 800 transformers are chosen from the business area of a UK DNO (Western Power Grid). They cover a good mixture of LV network’s geometry scenarios (urban, suburban, and rural) and LV network’s load scenarios (domestic, commercial, and industrial). For example, city center of Newport is chosen as an urban area with commercial customers; Lower Machen is selected as a rural area with domestic customers.
Through extensive case study, the representative imbalance-induced resistive heat profiles are given with probability density function. The case study also reveals that the imbalance-induced resistive heat can highly achieve approximately 30% of the no load resistive heat of LV transformer.
With this research, the imbalance-induced resistive heat profiles can be applied to calculate the LV transformer’s temperature rises, For the details of the temperature increase, other scenarios require to be considered: 1. ambient temperature of LV transformers; 2. ambient humidity of LV transformer; 3. heat dissipation method. The future work will consider these scenarios, and calculate the specific temperature increase and further the extra energy losses (caused by the temperature increase). Otherwise, this statistic approach is generic. For using this approach in other countries, it requires the 800 or more LV transformers operation data. For finding the accurate representatives of the imbalance-induced resistive heat profiles, this paper suggests the 800 or more LV transformers can be chosen in a more representative way as this research does.