温州市城区管网漏损噪声监测数据

通过将噪声监测设备吸附在管网上，监测管网2:00~4:00时段内某段时间内的供水音频，分析音频频谱频率，计算漏损可能性，生成特征值，帮助检漏人员快速确定漏损位置，提高检漏效率，为城市供水公司提供低成本、高效率的管网漏损检测解决方案，协助优化管网运行，降低漏损率，节约水资源。1、数据获取
通过金卡物联网平台及漏损管控平台获取噪声监测设备的设备编号、安装位置、日期、幅值强度（db）、频宽（Hz）、频率（Hz）。
2、规则设定
将音频数据的幅值强度（db）、频宽（Hz）、频率（Hz）按照傅里叶变换（FFT）进行标准处理提取，计算特征值，根据特征值结果，输出报警状态，如特征值>20,报警状态输出为漏损，其余为正常。
计算公式如下:
①如果 幅值强度（db）＜15，则特征值为1
②如果 幅值强度（db）≥15，则
特征值=幅值强度（db）*频宽（Hz）/频率（Hz）*10
3、结果输出
输出特征值、报警状态，筛选出存在报警的设备编号和日期，进一步判断是否存在漏损，并根据安装位置信息，委派人员现场核实。

By attaching noise monitoring devices to water supply pipelines, this solution collects and monitors water supply audio segments during the 2:00–4:00 time window of the pipeline network, analyzes the audio spectrum frequencies, calculates the likelihood of pipeline leakage, and generates feature values. It helps leakage detection personnel quickly identify leakage locations, improves detection efficiency, provides low-cost and high-efficiency pipeline leakage detection solutions for urban water supply companies, assists in optimizing pipeline network operations, reduces leakage rates, and conserves water resources.
1. Data Acquisition
Relevant data including device ID, installation location, recording date, amplitude intensity (dB), bandwidth (Hz), and frequency (Hz) of the noise monitoring devices are collected via the Golden Card IoT Platform and Leakage Control Platform.
2. Rule Setting
The amplitude intensity (dB), bandwidth (Hz), and frequency (Hz) of the audio data are extracted through standard processing via Fast Fourier Transform (FFT) to calculate feature values. The alarm status is output based on the feature values: if the feature value > 20, the alarm status is set to "leakage detected"; otherwise, it is set to "normal".
The calculation formulas are as follows:
① If the amplitude intensity (dB) < 15, the feature value is 1.
② If the amplitude intensity (dB) ≥ 15, the feature value is calculated as: Feature Value = (Amplitude Intensity (dB) × Bandwidth (Hz)) / Frequency (Hz) × 10.
3. Result Output
The feature values and alarm status are output. Device IDs and recording dates with alarm triggers are screened out to further determine whether leakage exists. On-site verification will be arranged for relevant personnel based on the installation location information.

搜集汇总

背景与挑战

背景概述

该数据集包含温州市城区管网漏损噪声监测数据，共700条记录，每日更新，用于分析供水音频频谱频率，计算漏损可能性，帮助检漏人员快速定位漏损位置。数据由金卡水务科技有限公司自行产生，包含设备编号、安装位置、日期、幅值强度、频宽、频率、特征值和报警状态等字段。