赣州市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 affixing noise monitoring devices to water supply pipelines, this dataset collects water supply audio signals during a specific segment of the 2:00–4:00 period, analyzes audio spectrum frequencies, calculates the likelihood of pipeline leakage, and generates feature values. It enables leak detection personnel to quickly pinpoint leakage locations, enhances detection efficiency, provides low-cost and high-efficiency pipeline leakage detection solutions for urban water supply companies, assists in optimizing pipeline operation, reduces leakage rates, and conserves water resources.

1. Data Acquisition
Relevant parameters including device ID, installation location, monitoring date, amplitude intensity (dB), frequency bandwidth (Hz), and signal frequency (Hz) of noise monitoring devices are collected via the Jinka IoT Platform and the Leakage Control Platform.

2. Rule Setting
The amplitude intensity (dB), frequency bandwidth (Hz), and signal frequency (Hz) of the collected audio data are standardly processed and extracted using Fast Fourier Transform (FFT) to compute feature values. The alarm status is output based on the calculated feature values: if the feature value exceeds 20, the alarm status is set to "leakage detected"; otherwise, it is set to "normal".
The calculation formulas are as follows:
① If amplitude intensity (dB) < 15, feature value = 1
② If amplitude intensity (dB) ≥ 15, feature value = (amplitude intensity (dB) × frequency bandwidth (Hz) / signal frequency (Hz)) × 10

3. Result Output
Output the feature values and alarm status, filter out the device IDs and monitoring dates corresponding to active alarms, further determine whether a leakage exists, and arrange on-site verification by 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700条管网漏损噪声监测数据，每日更新，用于分析音频频谱频率以检测管网漏损。数据通过傅里叶变换处理，计算特征值并输出报警状态，帮助检漏人员快速定位漏损位置。