| 摘要: | 本研究針對迴轉式壓縮機之電磁噪音預測技術，建立一套結合多軟體模擬與實驗並聯驗證的分析流程，顯著提升模擬準確性與可靠度。研究採用磁-固-聲多重物理耦合分析方法，其中馬達電磁分析使用電磁有限元素分析軟體2D，自然頻率與結構振動模態分析採用有限元素結構分析軟體，噪音輻射分析則由聲場有限元素分析軟體執行，並將模擬結果與實際運轉量測數據進行比對，以驗證整體模擬技術的有效性與應用潛力。
本研究的學術貢獻在於：（1）建立可應用於旋轉機械之高精度電磁噪音預測模型，證實模擬與實測數據高度一致，為未來電磁噪音預測奠定分析基礎；（2）導入工程傳遞路徑分析（TPA）理論，首度將電磁噪音明確區分為電磁激振源、結構自然頻率耦合路徑與空氣聲場傳遞等三項主要影響因素，提供深入理解與控制機制的新觀點；（3）探討轉子動偏心對邊頻噪音的影響，發現壓縮機軸彎曲與轉子安裝不對稱為主要成因，進一步指出傳統假設轉子完全同心的建模方式存在侷限性，對未來建模策略提出具體修正依據。
在模型準確性驗證方面，電磁力有限元素模擬之頓轉轉矩、表面磁通、反電動勢與單體效率等關鍵參數均與實驗數據誤差控制在 5% 以內，結構模態分析亦展現優異對應，平均誤差為 -1.83%、均方根誤差為 4.8%。電磁噪音分析結果顯示，在 6、12、18、24 倍頻處具有明顯振動聲峰值，且高轉速下（6000 rpm 與 7200 rpm）更易激發，模擬與實測之主要頻率誤差皆低於1%，顯示所提出之分析方法具高度準確性與重現性。
綜合而言，本研究建立之模擬技術可有效應用於旋轉電機設計階段，提前預測聲學表現、篩選不良設計、降低試作成本並加速開發流程，具高度學術價值與實務應用潛力。
;This study focuses on the electromagnetic noise prediction technology for rotary compressors and establishes an analysis framework that integrates multi-software simulations with parallel experimental validation, significantly enhancing the accuracy and reliability of the simulation results. A multi-physics coupling approach involving magnetic-structural-acoustic analysis is adopted: ANSYS Maxwell is used for motor electromagnetic analysis, ANSYS Mechanical for natural frequency and structural vibration analysis, and Actran for acoustic radiation analysis. Simulation outcomes are compared against experimental measurements under actual operating conditions to verify the effectiveness and applicability of the proposed method.
The academic contributions of this study are as follows: (1) a high-precision electromagnetic noise prediction model for rotating machinery is developed, demonstrating strong agreement between simulation and experimental data, laying a solid foundation for future research in this field; (2) the Theory of Transfer Path Analysis (TPA) is introduced to clearly decompose electromagnetic noise into three major influencing paths—electromagnetic excitation sources, structural natural frequency coupling, and airborne acoustic transmission—offering new insights into noise mechanisms and control strategies; (3) rotor dynamic eccentricity is investigated, revealing that sideband noise phenomena are mainly induced by shaft bending and rotor misalignment, thus highlighting the limitations of conventional concentric rotor assumptions and providing guidance for future modeling improvements.
Regarding model accuracy validation, finite element simulations of electromagnetic torque ripple, surface flux density, back-EMF, and unit efficiency exhibit deviations within 5% of the measured values. Structural modal analysis also shows excellent correlation, with an average frequency error of -1.83% and a root mean square error of 4.8%. The results of electromagnetic noise analysis reveal prominent vibration noise peaks at 6, 12, 18, and 24 times the fundamental frequency, particularly under high-speed conditions (6000 rpm and 7200 rpm), where the main peak frequency prediction errors are within 1%, demonstrating the high accuracy and reproducibility of the proposed method.
In summary, the simulation technology developed in this study can be effectively applied during the design phase of electric rotating machines, enabling early-stage acoustic performance prediction, elimination of inferior designs, and reduction of prototyping costs and development time. The proposed methodology possesses significant academic value and practical application potential. |
