JPH06109858A - Underground pipe line abnormality detecting method - Google Patents

Abstract

PURPOSE:To automatically detect an abnormality of a pipe line by executing an input of a measurement result from an output of a microphone arranged in an underground pipe line into a neural net which is made to study a spectral distribution at a normal state and at an abnormal one. CONSTITUTION:During a predeterminermined time (about one second), a sampling is executed with respect to an output of a microphone 1 arranged with close contact to an underground pipe line with an interval of about 0.1ms and its wave shape is recorded after A/D conversion 2. This signal is subjected to a Fourier conversion by an auxiliary microprocessor 3 of a pipe part device and a spectral distribution is required. And, into a neutral net weight function which is made to study preliminarily a spectral distribution at a normal state and at an abnormal state, an input of a measured spectral distribution is executed, judgement is executed with respect to abnormality or normality, at an abnormality, a modem 5 is passed through, and an input is executed into a personal computer 7 at an observation place through a modem 6 by using a telephone line and the like. By this, without a limitation of an observation place or an observation time, an automatic detection of an abnormality such as an access of a work machine to a pipe line can be executed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the detection of abnormalities in underground pipes.

[0002]

2. Description of the Related Art At present, many pipelines for gas, water and electric power are buried in the ground. These can be accidentally destroyed during ground excavation work, in which case it takes a great deal of time to restore, and in the case of gas and electricity there is a great risk. For this reason, when the machine tool approaches the pipeline, most of the means to prevent the self are to look around and rely on human power.

[0003] In such a patrol that mainly involves human power,
There was a big problem that continuous monitoring was not possible because there were restrictions on the monitoring place and time.

[0004]

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned drawbacks of the prior art and to provide an underground pipeline abnormality detecting method for automatically detecting abnormality such as approach of a construction machine to the pipeline. To provide.

[0005]

The above object can be achieved by the underground conduit abnormality detecting method of the present invention.

That is, in an abnormality detecting device for detecting an abnormality in a pipeline from the output of a microphone installed in an underground pipeline,
Waveform analysis of the microphone output, measured as a spectrum distribution, input the measurement results to a neural network that has previously learned the spectrum distribution of abnormal and normal time, to determine whether abnormal or normal This is a method for detecting an abnormality in the middle duct.

The microphone is a sensor for detecting the vibration of the pipe in the underground pipe.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The configuration of an example of a suitable apparatus used in the underground conduit abnormality detection method of the present invention will be shown below with reference to FIG. 1 and its operation will be described. However, the description does not limit the invention.

In FIG. 1, a microphone 1 which is in close contact with an underground pipe and detects vibration of the pipe, and an analog / digital converter (A / D converter) 2, A for converting an analog output of the microphone 1 into a digital signal A memory unit 4 for storing the / D-converted signal, a microprocessor 3 for processing information such as a judgment of a received signal or a memory signal by a Fourier transform or a learned neural network, and a modem 5 for transmitting the judgment result. And the modem 6 for receiving information from the pipeline device and the personal computer 7 for displaying information constitute a monitoring station device.

Next, the operations of the conduit unit and the monitoring station unit will be described.

The microphone output is sampled at an interval of about 0.1 ms for a fixed time (about 1 second), the waveform is A / D converted, and the (A / D converted microphone output) signal is recorded in the memory 4. . This signal is Fourier-transformed by the microprocessor 3 attached to the conduit unit to obtain the spectrum distribution.

A neural network weighting function in which an abnormal spectrum distribution and a normal spectrum distribution are learned is prepared in advance, and the measured spectrum distribution is input to the neural network weighting function to determine whether the distribution is abnormal or normal. If there is an abnormality, a telephone line or the like is used through the modem 5 to notify the monitoring station through the modem 6 of the abnormality.

[0013]

[Effects of the Invention] The presence or absence of abnormality in the pipeline is automatically detected by installing a microphone in the pipeline and determining the presence or absence of abnormality from the spectral distribution obtained by processing the output using a neural network. It became possible to do.

[Brief description of drawings]

FIG. 1 is a device configuration diagram showing an example of a detection device used in an underground conduit abnormality detection method of the present invention.

[Explanation of symbols]

 1 Microphone 2 A / D Converter 3 Microprocessor 4 Memory 5 Modem 6 Modem 7 Personal Computer

Claims (1)

[Claims] 1. An abnormality detection device for detecting a pipeline abnormality from a microphone output installed in an underground pipeline, analyzes the waveform of the microphone output, measures it as a spectrum distribution, and learns the spectrum distribution in the normal and abnormal times in advance. A method for detecting an abnormality in an underground conduit, characterized in that the measurement result is input to the neural network to determine whether it is abnormal or normal.