Seismic evaluation and qualification of transformer bushings

Abstract

Earthquakes are a significant threat to the integrity of the electric transmission system; prolonged, widespread electric outages can cause significant hardships and economic losses to the whole country, and degrade public safety. Much of the electric system is in highly active seismic regions around the United States, including some of the large urban areas. Consequently, both the probability and consequence of earthquake damage to electric service reliability, infrastructure security and cost control are extraordinarily high. Post-earthquake functioning of utility systems is viewed by emergency responders and society in general as an absolutely vital need for rapid response, recovery and preservation of public safety. Furthermore, building an electric system that is more resistant to seismic motion damage will reduce the consequences and costs of electric service disruptions caused by earthquakes. Previous research studies have resulted in significant knowledge of electric system seismic behavior and have led to substantial improvements in key areas; they have also identified remaining vulnerabilities in the electric system, and several areas of high-value seismic research that can lead to a more reliable, robust and resilient electric system. One of the least understood seismic behaviors in the electric system is the interaction between the bushing assembly and the transformer housing in large substation transformers. This study investigates the seismic response of the combined transformer-bushing interaction to enable future design, analysis and physical seismic qualification of transformer bushings. Current evaluations and qualification procedures based on IEEE 693 2005, or earlier, protocol consider both the bushing and the transformer in isolation from each other, but fail to explicitly recognize the influence of bushing installation on, and interaction with, the transformer in actual in-service conditions. Such interactions include the flexibility of the transformer tank's walls and roof. These features may lead to acceleration amplifications from the base of transformers to the mounting flange of transformer bushings (seismic loads) beyond those currently covered in the standard, and to modes of vibration that are unforeseen, resulting in failures that cannot be addressed by existing standards or engineering guidelines. This study focuses on determining the dynamic characteristics and behavior of transformer bushings. Extensive testing was performed in the laboratory for this purpose. The bushings were essentially tested on a fixed base to measure the stiffness, frequencies and damping of the bushings itself and then installed on a mock-up transformer (rigid frame with flexible top plate) and the bushings were further tested for the dynamic properties. The change in dynamic characteristics because of plate flexibility on the frame was an important factor ignored in previous studies. The change in dynamic properties was also verified by numerical model based on mechanics and structural computations. Bushings are mounted at the top of transformer plate, which are flexible in construction. During earthquakes this plate vibrates in the vertical direction as well as in rotation, causing a significant change in the input excitation. The rigid body accelerations and displacements caused by the flexibility of plate were quantified and a simplified model was developed to capture the flexibility of the plate in the vertical direction as well as rotation using discrete springs in corresponding directions. In previous studies, it was noted that the transformer body is flexible in horizontal direction and hence some amplification of acceleration was anticipated during earthquake excitations. Actual transformers are expensive and too massive to be mounted on a shake table for qualification purposes. Therefore it was recommended to use a rigid frame mounted on the shake table and tested to twice the input motion to qualify the bushings. A factor of two was used to take into account the flexibility of the transformer tank. Previous studies also indicated that this factor of two was not adequate because the actual acceleration amplifications at the bushing frequency were higher than two. For this reason, it is mandatory to measure the actual demand on the bushing and compare it to the capacity of the bushing at critical locations. This study shows that the adequacy of transformer bushings should be determined by comparing expected demands measured in laboratory settings to their strength capacity measured or calculated at critical locations. The findings in this dissertation are supported by extensive laboratory study of various types of bushings and by computational models developed to improve understanding.