Transformer core noise and vibration prediction methodology

Transcription

1 Transformer core noise and vibration prediction methodology Parmatma Dubey, Manager Technology Sarang Dev, Executive- Technology Atul Daga, Sr. Executive- Technology Date: June 11, 2014

2 CG Background Avantha Group Company CG is a global pioneering leader in the management and application of electrical energy. With more than 15,000 employees across its operations in around 85 countries, CG provides electrical products, systems and services for utilities, power generation, industries, and consumers. The company is organized into four business groups: Power, Industrial, Automation, and Consumer. CG clocks US$ 2.3 billion in revenues from product lines that cover the entire value chain of engineering offerings. here 2 CG Global R&D: was awarded the National Intellectual Property (IP) Award, 2014, in the category of Top Organisation in Designs. The award was presented on April 26th 2014, World IP Day, jointly organised by the Indian Intellectual Property Office and Confederation of Indian Industry (CII).

3 Transformer core noise and vibration prediction methodology Objective of case study High noise and vibration level is reason enough to modify completely a product line. Concerns of the society regarding environmental noise. Demands enforced by standardization entities related to the transformer noise reduction. Demands from customers. Competitors developing the low vibration and noise products. Statement of problem To develop the methodology of noise and vibration prediction of transformer core here 3

4 Main sources of noise in power transformers Core noise, caused by magnetic forces and magnetostriction Load noise, caused by electromagnetic forces in the windings due to leakage flux Noise generated by the cooling circuit like fans and pumps. here 4

5 Transformer core Core build up with laminations Assembled core here 5

6 Transformer core FE modeling Free Tetra Mesh Hexa Brick Mesh here 6

7 Orthotropic material modeling Modulus of Elasticity: Across the grain Along the grain Along the grain: 120E+9 Pa Across the grain: 210E+9 Pa Normal to sheet: 2.1E+6 Pa Poissons Ratio: 0.25 Density: 7850 Kg/m 3 here 7

8 Normal Mode Analysis Across the grain Major Mode in X-Y Plane with Tetra mesh: Hz with Hexa mesh: Hz here 8

9 Normal Mode Validation Normal Modes from LMS Test Lab. at 104 Hz Major Mode in X-Y Plane with Altair Radioss: Hz Across the grain here 9

10 Harmonic Frequency Response Analysis Input Force Spectrum here 10 Load & BC

11 Harmonic Frequency Response Analysis Harmonic Response in Hypergraph here 11

12 Harmonic Frequency Response Analysis Tested Response at 100Hz: 0.65 μm Simulated Response at 100Hz: 0.77 μm Vibration Test Spectrum here 12

13 Vibro-Acoustic Analysis Simulated Sound Pressure Level here 13

14 Vibro-Acoustic Analysis Octave 1/3 Point1 (A) Pa db(a) db Pa Octave 1/3 Hz here Tested Sound Pressure Level : dba Simulated Sound Pressure Level : dba A L

15 Benefits summary Quantitative Benefit- 1. Currently, low noise market share is approx. 10%. With a cost competitive low noise level, the market share could increase to 30-40%. 2. This methodology will help in designing low noise transformers which will contribute to 5M $ of existing CG market. Qualitative Benefit- 1. A good correlation between simulation and test has been established. 2. A complex geometry has been idealized to predict the actual behavior. 3. A parametric model can be developed for correction in existing noise prediction design rules. here 15

16 Challenges faced and solutions presented 1. Modeling Core laminations 2. Extracting input forces from magnetostriction curve 3. Modal correlation and testing 4. Simulating many orthotropic materials like wood and CRGO. here 16

17 Conclusion Altair Hyperworks provides step by step logical flow to simulate forced harmonic response The FE input file can be easily edited for trying various material and design parameters Time taken for simulation was substantially less compared to other solvers. A low noise core design methodology has been established with a good correlation with test data. here 17

18 Acknowledgements/Credits here 18