一种基于数据和机器学习的智慧测漏手段

A Smart Water Leak Detection Method Based on Data and Machine Learning

在供水行业全面推行“智慧水务”的进程中,漏水探测作为水务工作的重要环节,同样需要提高“智慧”水平。漏水探测一直以来是一项高度依赖于人工的工作,本文尝试探索一种新的声学探测检漏工作模式,通过大量采集量化的漏水噪声以及相关数据,利用这些数据进行机器学习,判断漏水点,使检漏工作逐渐摆脱对人听力、学习、思考和判断的依赖,达到弱人工智能水平。这种数据量化加机器学习的模式,对漏水噪声的判断可以达到远高于人工的准确率,并具有不断自我完善的能力。

In the process of fully implementing the“smart water”in the water supply industry.water leakage detection is an important part of water work,and it is also necessary to raise the level of“smartness”.Water leakage detection has always been a highly human-dependent work.This paper attempts to explore a new acoustic leak detection methodology,by collecting a large amount of quantitative leakage noise and related data.This data is used for machine learning to locate the water leakage point.Make the leak detection work release from human ear,and human judgement,and reach the level of sub-artificial intelligence.The data quantification plus machine learning mode can judge the water leakage noise to be much higher than human,and has the ability to continuously improve itself.