CN104282347A

CN104282347A - Wireless sound detection apparatus for nuclear-power-station valve leakage and compression transmission method

Info

Publication number: CN104282347A
Application number: CN201310286249.9A
Authority: CN
Inventors: 邓昌建; 黄志奇; 王泉
Original assignee: Individual
Current assignee: Individual
Priority date: 2013-07-09
Filing date: 2013-07-09
Publication date: 2015-01-14

Abstract

The invention relates to a wireless detection apparatus for monitoring nuclear-power-station valve leakage state, and belongs to the technical field of automatic control and detection of nuclear power equipment. The apparatus comprises a power supply, two wireless sound detectors, two amplifying circuits, a standard sound emitter, a system processor, an RF communication circuit, an antenna and the like; after a received signal is processed by a sound processing circuit, is subjected to A/D conversion, and is subjected to analysis processing by the system processor, and then sound compression is performed according to a sound compression threshold value obtained by sending a reference sound. Sound signals received by two sound detectors are processed by employing an interrupt working manner, so that an effective sound compressed code is obtained, and the code is transmitted by employing a CSMA/CA unidirectional broadcast manner by the system. The state whether the valve is leaked is obtained after the data is processed by using hidden Markov model at a rear terminal.

Description

A kind of nuclear power station valve leak wireless voice pick-up unit and compression transmitting method

Technical field

Patent of the present invention relates to a kind of wireless detection device of nuclear power station valve leak condition monitoring, belongs to that nuclear power generating equipment controls automatically, detection technique field.

Background technology

It is the important means ensureing nuclear power production safety that the valve leak of nuclear power plant equipment detects, but wired monitoring mode exists very large limitation at present, with high costs, and is difficult to expansion.Wireless-aware network is easy to expand monitoring range, improves monitoring level, reduces monitoring cost, can play a significant role in device status monitoring.

In current patented invention, a kind of valve leakage detecting device (patent publication No.: CN102323013A) has been invented by Institutes Of Technology Of Changsha, use Fast Fourier Transform (FFT) signal, and spectrogram when not leaking with the valve of its internal reservoir contrasts, and judges whether valve leaks; This invention adopts cable technology, does not consider the resource-constrained of wireless device, the restriction etc. of communication bandwidth.

Wireless voice transmission system and method (patent publication No.: CN102845080A) have been invented by Phonak AG, be used in transmission unit, in the separation time slot of tdma frame, each packet is transmitted at different frequencies according to frequency hop sequences, and by same audio packet transmission at least twice (one-way broadcast transmissions) in same tdma frame, and without the need to the acknowledge message of expection from receiver unit; But in valve radio detection, general pattern of wants sound array, and wireless voice detection number may be increased and decreased at any time, these situations, which difficult treatment, or need complex process, and limit by scope, be not easy to application.

In current valve leak detects, although have a lot of research based on the measuring technique of wireless voice, bias toward a certain particular problem, as compression algorithm, coding, safety etc., not in conjunction with apparatus design and the sound compression transmission method of embody rule more.

Summary of the invention

The object of patent of the present invention is to provide the wireless detection device of the valve leak that an energy uses under nuclear power plant environment; This device comprises a power supply, two wireless voice detectors and amplifying circuit thereof, a standard voice transmitter, a processor, a telecommunication circuit for RF communication, and antenna.

For realizing the object of valve leak radio detection, the technical solution adopted in the present invention is: the acoustic emitter adopting standard, sends the voice signal of normative reference; Sound is received with two sound-detection gear collections; According to collection normative reference signal, obtain canonical reference sound thresholding; Use sound compression method, reduce the data volume that sound-detection signal sends; The last host-host protocol using a kind of Stochastic accessing in a device, makes multiple device can be connected with network manager, network consisting.

This device is provided with the part compositions such as battery, sound generating circuit, bee bell device, sound treatment circuit, miaow head, processor, radio-frequency (RF) transceiver, antenna, power switch; For judging effective size of leakage signal, device is opened sound generating circuit at the beginning and is sounded, this sound and neighbourhood noise to be measured mixing receive by sound-detection gear, the signal received amplifies after process through sound treatment circuit, the processor being linked into band A/D carries out analyzing and processing, obtains the threshold value of sound compression; Then close sound generator, only obtain sound detection data with sound-detection gear, these data, according to obtaining threshold value above, after compression algorithm process herein, are sent by radio-frequency (RF) transceiver.

The voice signal that two sound-detection gears receive adopts discontinuous mode, and namely in 100ms above, a detector receives voice signal, immediately another detector reception work 100ms, and circulates successively; Sound compression obtains effective peak waveform, adopts and represent compression method by a data after adopting effective thresholding relatively, and the efficient coding of the actual valid data of other bit representation, compresses; In transmission mode, have selected the transmission mode based on CSMA/CA, and adopt One-to-All Broadcast mode to transmit, these data, after hidden Markov model process, obtain valve leak whether situation.

The invention has the beneficial effects as follows and can adapt to nuclear power plant environment work, measure the leakage situation of nuclear power station valve, carry out the compression of measurement data, realize collection and the transmission of wireless voice.

Accompanying drawing explanation

Fig. 1 is anatomical connectivity block scheme of the present invention, in figure:

1 is antenna

2 is radio-frequency (RF) transceiver

3 is sound measurement circuit

4 is sound generating circuits

5 is processors

6 is battery and electric power management circuit

Fig. 2 node control sound occurs and receives process flow diagram

Fig. 3 sound measurement circuit

3-1-1 miaow head 1

3-1-2 miaow head interface circuit 1

3-1-3 pre-amplification circuit 1

3-1-4 sound amplifying circuit 1

3-2-1 miaow head 2

3-2-2 miaow head interface circuit 2

3-2-3 pre-amplification circuit 2

3-2-4 sound amplifying circuit 2

Fig. 4 sound transtation mission circuit

4-1 honeybee purine device

4-2 sound control circuit

4-3 sound interface circuit

Fig. 5 sound measurement flow process

The coding method of Fig. 6 sound compression is illustrated

Embodiment

Below in conjunction with accompanying drawing and example, to according to embodiment provided by the invention, structure, feature, details are as follows:

As shown in Figure 1, a kind of nuclear power station valve leak wireless detection device, this device comprises antenna 1, radio-frequency (RF) transceiver 2, processor 3, sound measurement circuit 4(wherein 4-1 is sound measurement circuit 1,4-2 is sound measurement circuit 2), sound generating circuit 5, sound measurement sensor miaow head 6(wherein 6-1 is that miaow is first, 6-2 is miaow head two), bee bell device 7 and battery and power circuit form.

Processor in device and radio-frequency (RF) transceiver, can process device by the MEAGA128A of ATMEL, and CC2420 is rf chip; Also processor and radio-frequency (RF) transceiver can be used as by MC13224V; Rf chip can also be used as with NA5TR1.

Processor circuit in device, with A/D converter interface, can receive the analoging sound signal from sound measurement circuit measuring, change; Miaow head is by miaow head circuit measuring sound, and the signal demand of measurement carries out amplifying through sound measurement circuit, filtering process, and numerical value access processor signal being had mate most with processor A/D input range processes; Processor circuit is accompanied with jtag interface, serial ports UART interface (Transistor-Transistor Logic level), realizes download and the debugging of program.

Battery in device and electric power management circuit provide power supply to sound transducer miaow head, sound measurement circuit, windbell device, sound transtation mission circuit, processor and RF transmit-receive circuit, realize initialization, receive measurement data, manipulate measurement data realized the transmission of data by radio frequency (2.4GHz); This device is the element of nuclear power plant equipment valve leak condition monitoring, forms wireless monitor system, realize equipment health monitoring with telegon (or Network Access Point), gateway and network manager.

For judging effective size of leakage signal, device is opened sound generating circuit at the beginning and is sounded, this sound and neighbourhood noise to be measured mixing receive by sound-detection gear, the signal received amplifies after process through sound treatment circuit, the processor being linked into band A/D carries out analyzing and processing, obtains the threshold value of sound compression; Then close sound generator, only obtain sound detection data with sound-detection gear, the data of measurement, after compression algorithm process herein, are sent by radio-frequency (RF) transceiver.

As shown in Figure 2, the flow process that node control sound sends and receives is as follows: 1 system energization starts, and 2 systems carry out initialization, and 3 judge whether the condition meeting sound transmission, i.e. first time, or whether working time exceedes setting scaled time, if first time, or do not exceed setting-up time, carry out the 4th step, send 3 100ms voice signals, this signal is low frequency (50Hz) the accurate dextrorotation ripple signal of simulation, is low amplitude signal (sensitivity regulates by potentiometer); Open synchronous acoustic reception process at transmission sound, this reception sound and transmission signal contrast, can obtain environmental sound signal size, obtain threshold size simultaneously; If not the 1st time, and not exceeding setting-up time, carrying out sound collection and process with regard to directly utilizing last stored parameter.

As shown in Figure 3, the structure of sound measurement circuit is as follows: sound measurement circuit is made up of two metering circuits in parallel, and 3-1 is metering circuit 1, by 3-1-1 miaow head 1,3-1-2 miaow head interface circuit 1,3-1-3 pre-amplification circuit 1 and 3-1-4 sound amplifying circuit 1 form; 3-2 is metering circuit 2, is made up of 3-2-1 miaow head 2,3-2-2 miaow head interface circuit 2,3-2-3 pre-amplification circuit 2 and 3-2-4 sound amplifying circuit 2; With 3-1, circuit working process is described, miaow head 3-1-1 is converted into the voltage signal of several mV to tens of mV through miaow head interface circuit 3-1-2 voice signal, pre-amplification circuit 3-1-2 via ac amplifier circuit composition amplifies 10-100 doubly signal, become the voltage of hundreds of mV, then amplify 1-10 doubly through sound amplifying circuit, be converted into the voltage signal of 1-3V.

As shown in Figure 4, sound transtation mission circuit sends the interface circuit 4-3 with single-chip microcomputer by sound, the honeybee purine device composition of sound control circuit 4-2 and transmission sound; Voice signal produces DDS(digital synthesis signal by single-chip microcomputer) signal generates; It is generally low frequency (50Hz) the accurate dextrorotation ripple signal of simulation, and send 5 times that sound time length is signal acquisition periods, sampled data is about 512 bytes, and arranging sample frequency is 5KHz; Be convenient to data processing like this.

As shown in Figure 5, when acquisition is with reference to after voice data, the measuring process of sound is as follows: start through 1,2 measure initialization, and 3 open Measurement channel Isosorbide-5-Nitrae measurement data, carry out compressing and stores processor, 5 judge whether to have processed 255 byte datas, as do not continued process, if processed, 6 according to coding, carries out wireless transmission, 7 send after, be switched to passage 2, repeat the process of 3-6.

One of object that two-way is measured improves its precision, makes its uncertainty obtain decline 1.414 times.

The another one object that two-way is measured improves shown in reliability (increasing by 1 times), makes it to be more suitable for nuclear power station valve detection and use.

As shown in Figure 6, sound compression process is as follows, first according to the threshold value that measurement one judges; Then each data measured are compared with this threshold value; If 4 are less than threshold value, just its value are set to 0, and determine compress mode; If 3 are greater than threshold value, just compress according to PCM or other compress mode, high 4 bit data as used it maximum are encoded as the size of this signal; When conversion is complete, 6 arrange compress mode code, and line sampling of going forward side by side is selected; 7 obtain sending sequence, carry out wireless transmission.

The encoding example that absolute value is less than threshold value is as follows:

00 34 compress mode codes arrange F1, and wherein, F represents pcm encoder, have 1 zero before 1 representative;

00 00 34 compress mode codes arrange F2, and wherein, F represents pcm encoder, have 2 zero before 2 representatives;

00 00 00 34 compress mode codes arrange F3, and wherein, F represents pcm encoder, have 3 zero before 3 representatives;

Be greater than 15 zero situations if had above, FF puts in compression side, F represents pcm encoder, has 3 zero before 3 representatives;

The encoding example that absolute value is greater than threshold value is as follows:

Actual measurement voice data subtracts 16, is converted into 16 complement on n ns, for sending data.

The coding of such wireless voice, transmits according to ' frame head '-' data length '-' protocol prefix '-' sound compression type '-' voice data ', selects the byte once transmitted to be 255 bytes.

In transmission mode, have selected the transmission mode based on CSMA/CA, adopt One-to-All Broadcast mode to transmit.

On clear channel assessment (CCA), realize based on measurable RSSI value and a programmable threshold values, Carrier Sense Multiple Access as clear channel assessment (CCA) function on CC2420 avoids conflict for performing CSMA-CA() function, when receiver has been enabled at least 8 character cycles time, CCA becomes effectively, carrier sense threshold values is programmed by RSSI.CCA_THR, and threshold values can be programmed with the stepping of 1dB.The execution of CCA model selection 1, CSMA-CA can be accomplished the simplest by using STXONCCA order, and transmitting channel does not have be activated during other signal, TX_ACTIVE mode bit can be used to the result detecting CCA.

The example of One-to-All Broadcast mode as, uncertain clear and definite receiver address, protocol prefix is marked as the modes such as broadcast frame and realizes.

Valve leak state be identified by voice recognition to realize, be made up of four parts such as characteristic parameter extraction, HMM model training, pattern-recognition and result treatment; Characteristic parameter selection Mel frequency cepstral coefficient, uses hidden Markov model (Hidden Markov Model HMM) add up it and set up parameterized model, probability inference, and dynamic mode classification, realizes the detection to valve leak state; Hidden Markov Models comprises two steps, and a step is model learning and training, and another step is pattern-recognition.

Claims

1. a wireless detection device for nuclear power station valve leak condition monitoring, is characterized in that comprising a power supply, two wireless voice detectors and amplifying circuit thereof, a standard voice transmitter, a processor, a telecommunication circuit etc. for RF communication.

2. two-way sound-detection circuit adopts discontinuous mode collected sound signal, processes.

3. one kind is suitable for the sound compression method of wireless valve leak detecting device, after its feature adopts effective thresholding relatively, obtain effective peak waveform, accordingly 4 bit data are converted to the data of useful signal and represent, realize compression, and effectively represent Various types of data by data encoding.

4. in transmission mode, have selected the media access control mode of CSMA/CA, adopt One-to-All Broadcast mode to carry out Wireless Data Transmission.

5. the method for effective thresholding is obtained by sound dispensing device and receiving trap.