It is important to understand that, as mainly a software company, our strategy is to hold most of our technology as trade secrets rather than applying for patents. In our experience patenting software and algorithms is quite difficult (examiners and patent offices lean against granting). Plus once an application is published it remains public forever, even if the patent is not granted. So we are very careful with such applications. Cloud based software is a bit different to that of a typical hardware company where you ship the complete product to customers - who then holds the rights to disassemble and re-create (understand) the design. To protect yourself against this design leakage you really need a patent (or incredible complexity). But with cloud software the customer never really gets access to the algorithms (the internals) - just the outputs. So it can be maintained safely as a trade secret.
At the moment we collect vibration and acoustic information to inform the customer of changes in vibration state over time. Vibration growth may indicate escalating rates of transformer degradation or excessive load conditions. The absolute level is also important as high levels of vibration and acoustics can impact nearby residents and create customer and public relations problems. Finally the over-aching strategy with the multi-parameter sensor is to correlate between sensors to create higher probability status assessments. We do this presently for example with temperature and load, and DGA and electromagnetic PD measurements. We are in the process of expanding the load analysis to include vibration and acoustics levels.
A more advanced aim is the use of vibration to help locate and characterise actual winding faults. Over the past few year the University of Queensland been conducting research in this area. From time to time I provide advice on these projects (in my role as an Adjunct Professor). One technique in development is of particular interest to DTS. I encouraged the application of a provisional patent early this year so detailed discussions can begin. The technique is suitable for integration into our sensor and I am activity considering this enhancement in our roadmap.
