Lifelines 404 - Improvements to Modeling Substation Equipment

These modal models can be integrated with standard finite elements to model structural response modifications due to mass or stiffness changes resulting from support structures. The linear analytic studies will recommend methods for modeling equipment and major components. The significance of the experimentally-derived, modal model serves at least four purposes. First, the analytical models for representative substation equipment should be relatively simple compared to those that would have been derived from standard finite element techniques. Secondly, the effects of boundary conditions can be easily assessed by posing ‘what if’ questions to the ME’Scope Structural Dynamics Modification (SDM) software. Thirdly, the lumped parameter model of the substation equipment should be easily integrated with more elaborate finite element models of the support structure. Lastly, the modal models for the response of the substation equipment due to low- and moderate-level shake table excitation should provide some guidance regarding the appropriateness of assuming a linear analytical model.

Figure 1. Bushing Vibration Tests at Bonneville Power

Larger View

The nonlinear analytical studies have been devoted primarily to single-degree-of-freedom (SDOF) models of bushings exhibiting the hysteric load-deformation behavior defined in earlier bushing tests conducted at UCB’s Richmond Field Station. The purpose of these studies has been to quantify the excitation/response levels at which a linear model can no longer adequately predict the dynamic behavior of a bushing. The basic approach has been to calculate the dynamic time history response of the nonlinear SDOF system and then to calculate the system’s frequency response function (FRF) under different levels of excitation and/or different levels of stiffness non-linearity. Although the FRF is only theoretically defined for a linear system, the graphical changes in the FRF for various nonlinear excitation/response levels have been tracked. One study has noted that for a given level of excitation, the FRF’s of the nonlinear SDOF tend to be characterized by a SDOF system with a lower fundamental frequency with increased damping compared to its linear SDOF system counterpart. Preliminary FRF studies of a nonlinear two-degree-of-freedom system exhibit similar conclusions as those of the nonlinear SDOF system.

Figure 2. Bushing #1

Larger View

Define substation equipment to be tested in UCI laboratory

Define substations to conduct in-situ tests

Conduct modal analysis tests of substation equipment – provide results

Conduct shake table tests of substation equipment – provide preliminary results

Conduct in-situ tests at substation – provide preliminary results

Conduct in-situ tests at UCB and UCSD – provide results for Tasks 406, 411

Develop linear and nonlinear analytical models

Project Title/ID Number	Improvements to Modeling Substation Equipment—Lifelines 404
Start/End Dates	10/1/02—9/30/03
Project Leader	Gerard Pardoen (UCI/Faculty)
Team Members	Charles Hamilton (UCI/Post Doc), Rick Tavares (UCI/Post Doc), Davis Fetter (UCI/Undergrad Student), Julie Manson UCI/Undergrad Student)

Project goals and objectives
Utilities use different types of substation equipment that have dynamic characteristics that are not well known for many classes of equipment. Manufacturer differences in design, method of structural support, physical dimensions, voltage class differences as well as differences at the component level are just some of the features that introduce variations of dynamic behavior. Although utilities have been specifying that equipment be seismically qualified by analysis for many years, there is uncertainty in the dynamic properties and methods of modeling used in these analyses. This project intends to improve the methods used for modeling electric substation equipment through a combination of experimental and analytical studies of bushings, transformers, and disconnect switches. The experimental program consists of testing substation equipment in laboratory (UCI, UCSD, UCB) and in-situ, field conditions (BPA, SDGE) using force-calibrated hammer and shake table excitation techniques. The analytical program consists of developing linear structural models based upon the experimentally-acquired data that can be integrated with existing finite element models as well as conducting parametric studies for bushing response under various support conditions observed in the field. Using the hysteretic load-deformation test results from a previous research study on bushings, nonlinear structural models have been developed to ascertain the threshold levels of excitation in which linear models of bushings are no longer valid.
Role of this project in supporting PEER’s vision
This program to improve the modeling of substation equipment in field conditions is a user-directed project that features collaboration between several sponsoring lifeline organizations (PG&E, SDGE, BPA) and other PEER researchers within the lifelines program. In particular the vibration testing of bushings at the BPA site has been a direct result of research needs expressed by the lifeline partners whereas the experimental results feed directly into Project 406 (Seismic qualification requirements for transformer bushings). Additionally, the UCI team has been asked to apply the experimental methodology to the disconnect switches being tested at UCB’s RFS as part of Topic 411 (Seismic qualification and fragility testing of 500 kV disconnect switches).
Methodology employed
To date the experimental approach has required a combination of laboratory and field-testing of substation equipment. The laboratory tests have used both force-calibrated hammer and shake table excitations whereas the field-testing has been confined to force-calibrated hammer and sledgehammer excitation. The experimental work has been purposely divided into laboratory and field studies so that the vibration characteristics of the substation equipment can be investigated from two different perspectives. The laboratory test component of the project has provided for more in-depth, iterative investigations using approximate support conditions whereas field-testing has provided the correct support conditions for the substation equipment. Although the field tests at the BPA site have not permitted in-depth investigations, the fundamental vibration characteristics of more than 30 bushings of different sizes, orientations, and support configurations that have been tested significantly adds to the database of such components. The vibration characteristics of the 30+ bushings are currently being interpreted using commercially available software (ME’Scope) in order to define the frequency, mode shape and damping characteristics of the equipment. Since the bushing’s response due to the impact excitation can be characterized as linear, ME’Scope is being used to determine the modal properties of the multiple degree-of-freedom, lumped parameter modal models from the experimentally-derived frequency response functions. The ME’Scope output consists of printed and animated vibration response. A ‘movie’ of the animation (as an *.avi file) can be created for each mode of vibration for ease of visualization.
These modal models can be integrated with standard finite elements to model structural response modifications due to mass or stiffness changes resulting from support structures. The linear analytic studies will recommend methods for modeling equipment and major components. The significance of the experimentally-derived, modal model serves at least four purposes. First, the analytical models for representative substation equipment should be relatively simple compared to those that would have been derived from standard finite element techniques. Secondly, the effects of boundary conditions can be easily assessed by posing ‘what if’ questions to the ME’Scope Structural Dynamics Modification (SDM) software. Thirdly, the lumped parameter model of the substation equipment should be easily integrated with more elaborate finite element models of the support structure. Lastly, the modal models for the response of the substation equipment due to low- and moderate-level shake table excitation should provide some guidance regarding the appropriateness of assuming a linear analytical model.		Figure 1. Bushing Vibration Tests at Bonneville Power Larger View
Brief description of past year’s accomplishments and more detail on expected Year 6 accomplishments
The experimental studies have been a combination of (a) laboratory experimental modal analyses, (b) laboratory shake table tests of specific, stand alone equipment as well as (c) in-situ field measurements of this equipment at the Bonneville Power Administration (BPA) site in Vancouver, Washington. Additional field measurements will be conducted at the San Diego Gas & Electric, UCSD, and UCB Richmond Field Station facilities in support of Tasks 406 (UCSD) and 411 (UCB). The analytical studies have been devoted to linear and nonlinear system responses. The linear analytical studies have interpreted the experimental data using commercially available software (ME’Scope) in order to define the frequency, mode shape and damping characteristics of the equipment. Multiple degree-of-freedom, lumped parameter modal models have been developed for a variety of bushings that are assumed to respond in the linear range. These modal models can be integrated with standard finite elements to model structural response modifications due to mass or stiffness changes resulting from support structures.
The nonlinear analytical studies have been devoted primarily to single-degree-of-freedom (SDOF) models of bushings exhibiting the hysteric load-deformation behavior defined in earlier bushing tests conducted at UCB’s Richmond Field Station. The purpose of these studies has been to quantify the excitation/response levels at which a linear model can no longer adequately predict the dynamic behavior of a bushing. The basic approach has been to calculate the dynamic time history response of the nonlinear SDOF system and then to calculate the system’s frequency response function (FRF) under different levels of excitation and/or different levels of stiffness non-linearity. Although the FRF is only theoretically defined for a linear system, the graphical changes in the FRF for various nonlinear excitation/response levels have been tracked. One study has noted that for a given level of excitation, the FRF’s of the nonlinear SDOF tend to be characterized by a SDOF system with a lower fundamental frequency with increased damping compared to its linear SDOF system counterpart. Preliminary FRF studies of a nonlinear two-degree-of-freedom system exhibit similar conclusions as those of the nonlinear SDOF system.		Figure 2. Bushing #1 Larger View
Analyses of the experimentally acquired data have noted that the 15%-30% difference in fundamental frequencies among identical bushings have been attributed to the subtle differences in mounting configurations. Furthermore, the test results of more than 30 bushings at the BPA site have shown that for a seemingly axi-symmetric bushing, the fundamental frequency of the bushing in the transverse and longitudinal planes can easily differ by 25% even though an analyst would only assume a single frequency for the bushing. Note in the figure below, for example, that the fundamental frequencies of a 230 kV bushing on transformer T896 at the BPA site range from 9.47 Hz in the transverse (front-to-back) direction to 12.03 Hz in the longitudinal (left-to-right) direction. A major focus of the analytical modeling has been, and will continue to be, the bushing’s connection to the transformer. An Appendix of the project’s draft report has over 50 photos denoting the various support conditions of bushings on representative transformers.
Other similar work being conducted within and outside PEER and how this project differs
Vibration studies of 230 kV and 500 kV bushings were conducted at the Richmond field station. The modal properties of the bushings supported on a specially constructed test stand were obtained from shake table excitation. Project 404 has been complementary in that a significant component of the experimental program has considered the vibration characteristic of bushings under field conditions. The load-deformation results reported in the earlier lifelines research project have been used in the numerical simulations for the nonlinear behavior of bushings.
Plans for Year 7 if this project is expected to be continued
There are no plans to continue this two-year project into Year 7.
Describe any instances where you are aware that your results have been used in industry
Although there have not been any instances where industry may have used the experimental results yet, it is anticipated that the range of frequency and damping values for in-situ bushings will be of interest to engineers conducting dynamic analyses.
Expected milestones
Define substation equipment to be tested in UCI laboratory Define substations to conduct in-situ tests Conduct modal analysis tests of substation equipment – provide results Conduct shake table tests of substation equipment – provide preliminary results Conduct in-situ tests at substation – provide preliminary results Conduct in-situ tests at UCB and UCSD – provide results for Tasks 406, 411 Develop linear and nonlinear analytical models
Deliverables
CD of the frequency response function data for the substation equipment (in Standard Data Format) from the force-calibrated hammer and shake table excitations; the CD will also include the ASCII data files of the modal parameters as well as *.AVI files of the animated mode shapes. preliminary report following each significant experimental investigation that interprets the vibration data in terms of frequency, mode shape and damping characteristics of the substation equipment. series of analytical models from the experimentally-acquired laboratory and field data will be developed with ME’Scope. final report that compiles the experimental results as well as providing the analytic studies that recommend methods for modeling equipment and major components.