论文总字数：39007字

摘 要

噪音干扰一直是信号处理中的主要问题之一，目前，为了解决这个问题，平滑滤波算法已开始广泛应用，常见的方法有包括算术平均滤波法，滑动平均滤波法，去极值滤波法，中位值滤波法等等，这些方法有一定的可取之处，但仍存在一些问题。而本次毕业设计所主要研究的对象是基于注意力优先的自动平滑算法，即ASAP，通过将平滑过程与粗糙度、峰度等数据特征相结合，从而实现更有效的算法。实现注意力优先，在对数据的平滑处理中兼顾大规模结构的保留，不仅可以解决噪音等信号干扰问题，同时也可以保留峰度，突出关键问题。

ASAP虽然已经应用广泛，涉及光学，医学等领域，但尚未与热工测量相结合，因此本次毕业设计将尝试把两者结合起来，将ASAP应用于锅炉声信号的处理，计算信号发出端与信号接收端的实际距离，以检测声学测量系统的有效性。由于现场不可避免的噪声干扰，信号发出端与信号接收端数字化后的声信号，在进行互相关的计算后仍与热电偶测出的实际值存在较大的差异，会出现双峰或者多峰现象，因此希望应用ASAP分别平滑互相关前后的信号，以达到突出峰值，找准距离的目的。

关键词：ASAP，粗糙度，分布峰度，锅炉声信号，互相关

ABSTRACT

Noise interference has always been one of the main problems in signal processing. Nowadays, in order to solve this problem, smoothing filtering algorithms have been widely used. Common methods include arithmetic average filtering, sliding average filtering, de-extremum filtering, median filtering and so on. These methods have some advantages, but there are still some problems. The main research object of this graduate design institute is an automatic smoothing algorithm based on attention priority, namely ASAP, which combines smoothing process with data features such as roughness and kurtosis to achieve more effective algorithm. Attention priority and large-scale structure preservation in data smoothing can not only solve the problem of signal interference such as noise, but also retain the kurtosis and highlight key issues.

Although ASAP has been widely used in optics, medicine and other fields, it has not yet been combined with thermal measurement. Therefore, this graduation project will try to combine the two. ASAP will be applied to the processing of boiler acoustic signals, calculating the actual distance between the signal sending end and the signal receiving end, in order to test the effectiveness of acoustic measurement system. Because of the unavoidable noise interference in the field, the digital sound signal from the signal sending end and the signal receiving end is still quite different from the actual value measured by the thermocouple after the cross-correlation calculation, and there will be double or multi-peak phenomena. Therefore, it is hoped that ASAP can be used to smooth the signals before and after cross-correlation respectively, so as to achieve the purpose of highlighting the peak value and locating the distance.

KEY WORDS: ASAP, roughness, kurtosis, boiler sound signal, cross correlation

目 录

摘 要 I

ABSTRACT II

第一章 绪论 1

1.1研究背景 1

1.2文献综述 1

1.2.1信号处理中存在的问题 1

1.2.2一般信号处理平滑方式 2

1.2.3 ASAP的提出 3

1.2.4 预期实现效果 4

1.3工作思路 4

第二章 ASAP相关介绍 5

2.1 粗糙度计算 5

2.2 分布峰度计算 5

2.3 平滑函数 6

2.4 自相关计算 6

2.5 ASAP算法综述 7

2.5.1 搜索周期性数据 7

2.5.2 批量ASAP 8

2.5.3 流ASAP 8

2.5.4 小结 8

第三章 不同平滑方式对比分析 9

3.1 原始数据 9

3.2 算术平均滤波法 10

3.3 滑动平均滤波法 12

3.4 去极值滤波法 15

3.5 中位值滤波法 17

3.6 ASAP 19

第四章 ASAP处理锅炉声信号 22

4.1 系统介绍 22

4.2 处理方式 23

4.2.1 先互相关后平滑 23

4.2.2 先平滑后互相关 23

4.3 处理对象 24

4.3.1 正常情况 24

4.3.2 双峰现象 24

4.3.3 多峰现象 28

4.3.4 其他现象 31

4.4 小结 34

结论 35

参考文献（References） 36

附 录 38

附录A 各函数部分算法 38

A.1 粗糙度算法 38

剩余内容已隐藏，请支付后下载全文，论文总字数：39007字