Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/92202
Title: | 刀具磨耗與振動訊號之頻域相關分析 Correlation Analysis between Tool Wear and Frequency-Domain Vibration Signals |
Authors: | 范瀚仁 Hen-Ren Fan |
Advisor: | 廖運炫 Yunn-Shiuan Liao |
Keyword: | 刀具磨耗,剩餘使用壽命,振動訊號,切屑型態,銑削, tool wear,remaining useful life,vibration signal,chip morphology,milling, |
Publication Year : | 2024 |
Degree: | 碩士 |
Abstract: | 刀具剩餘壽命的預測一直以來都是切削加工中一個極為重要的議題,旨在確保在滿足要求的情況下最大程度地發揮刀具價值,精確的壽命預測對於加工規劃、良率、成本和效率等方面都有正面影響。由於其涉及的方面相當多,從加工參數、機台振動特性、受力情形、溫度到材料特性等,彼此又互相影響,也因此有多種多樣的標準存在。
本論文以應變率變化為切入點,研究切削中由磨耗引起的刀具幾何形狀變化,造成不同剪切角與溫升的影響。在切削過程中,切削所形成的切屑的塑性應變也隨之變化,展現出不同的形態,也受到溫度變化引起的氧化和其他因素的影響,可作為評估壽命的參考。實驗中觀察到切屑厚度隨著磨耗程度的變化,且隨著磨耗增加,切屑由完整一片轉變為破碎狀態。切屑的形成受到切削條件和刀具幾何的變化影響,因此,本研究嘗試使用相關頻率分析方法預測刀具壽命。 基於切削理論,金屬材料的應變速率應會在相當高頻的頻段,因此在訊號處理階段,將濾除低頻訊號再進行壽命指標的運算。指標的計算基於高頻頻段的能量總和,並乘上該頻段的偏度(skewness)修正值。修正值取自切削當下與振動值變異數處於低值時的兩個偏度相差百分比絕對值,將上述以累積和(CUSUM)方式處理,超過能量限時判斷刀具已嚴重磨耗。研究中使用動力計量測的力與振動作為嚴重磨耗的參照標準,以加速度規信號計算所提出之指標進行刀具壽命預測,結果顯示提前預測效果良好,平均誤差在7%以下,並且適用於不同切削條件。 The prediction of tool remaining life has always been a crucial issue in machining processes, aiming to maximize the value of tools while ensuring compliance with specified requirements. Accurate life prediction has positive impacts on machining planning, yield, costs, and efficiency. Due to the multifaceted nature of this topic, it involves various aspects such as machining parameters, machine vibration characteristics, force conditions, temperature, and material properties, all of which are interconnected. Consequently, multiple standards exist to address these complexities. This paper takes the variation in strain rate as a starting point to investigate the geometric changes in tools caused by wear during cutting processes, resulting in different shear angles and temperature rises. During the cutting process, the plastic strain of the chips formed also changes, exhibiting various forms. It is influenced by factors such as oxidation due to temperature changes, serving as a reference for assessing tool life. The experiments observed variations in chip thickness with the degree of wear, and as wear increased, the chips transitioned from a complete piece to a fragmented state. The formation of chips is influenced by changes in cutting conditions and tool geometry. Therefore, this study attempts to predict tool life using the method of frequency analysis. According to cutting theory, the strain rate of metallic materials during machining is expected to fall within a considerably high-frequency range. During signal processing, low-frequency signals are filtered out before computing the tool life indicator. The indicator is based on the cumulative energy in the high-frequency range, multiplied by the skewness correction value for that range. This correction value is derived from the absolute percentage difference in skewness between the cutting moment and the variance of vibration values when they are in a low state. The tool is considered severely worn when the cumulative sum (CUSUM) exceeds the energy limit. Force and vibration measurements obtained through dynamic sensing are utilized as reference standards for severe tool wear. The proposed indicator, calculated from accelerometer signals, demonstrates early prediction capabilities with an average error of below 7%, proving its applicability across various cutting conditions. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/92202 |
DOI: | 10.6342/NTU202400730 |
Fulltext Rights: | 同意授權(全球公開) |
Appears in Collections: | 機械工程學系 |
Files in This Item:
File | Size | Format | |
---|---|---|---|
ntu-112-1.pdf | 5.48 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.