JP2004125481A - Sonic type method and apparatus for measuring temperature/flow of gas - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enhance the performance and reliability of sonic type instrumentation for the temperature/flow measurement of a gas, based on propagation time detection, by providing a method of identifying misdetection state in a propagation process, caused by superposition of disturbance noises and distortion of transmitted waveform, produced in the propagation process by conditions on physical properties inside a combustion furnace or inside a duct, keeping followability on the temperature/flow states. <P>SOLUTION: This sonic type gas temperature/flow measuring method is provided with a data recording means which stores detected propagation times for a predetermined number of past data; a means for computing differential signal between for each of measured propagation time and the time one measurement period preceding it (last measurement time), with respect to a signal string from a past propagation time up to the present propagation time measured value; a means which performs a wavelet transformation for two kinds of data strings, a data string which does not contain the value measured this time and a data string containing the value measured this time, for evaluation with respect to the computed differential signals; and an error detecting means which determines normality/abnormality of the measured value for this time, to be evaluated by intensity comparison or pattern recognition of wavelet transformation performed for the two kinds of data strings. <P>COPYRIGHT: (C)2004,JPO

Description

このＸｔが図６のマッチドフィルタ２０９の出力である。受信信号中に含まれる送信Ｍ系列ＰＲＫ波形とタイミング（位相）が一致した時刻でマッチドフィルタには鋭いピーク信号が現れ、受信信号と送信信号が同期したことを検知する。図６の伝播時間検出処理部２１０ではマッチドフィルタ出力Ｘｔのピーク検出を行い、ピーク時刻を音波の伝播時間として検出する。
【０００８】
検出した音響センサＡ２０１ａからＢ２０１ｂへの伝播時間をτＡＢ、音響センサＢ２０１ｂからＡ２０１ａへの伝播時間をτＢＡとすると、伝播時間τＡＢ、τＢＡと流速Ｖｆ、ガス温度Ｔにより定まる音速Ｖｓの関係は次の２ａ）、２ｂ）式で表される。ここで、Ｌは音響センサＡ２０１ａとＢ２０１ｂ間の距離、θはＬが流体の流れ方向と成す伝播角である。
【数２】

【０００９】
２ａ）、２ｂ）式から伝播時間検出値τａｂ、τｂａより流速Ｖｆ、音速Ｖｓは次の３ａ）、３ｂ）式で表される。
【数３】

【００１０】
ガス温度Ｔ［℃］は音速の温度依存性より次の４）式で表される。
【数４】

ここでαはガス組成により定まる定数である。
【００１１】
気体温度計測の場合は上記４）式により計測が完了するが、流量計測の場合は、上記３）、４）式のように計測された気体流体の温度、流速から次の５）式より気体流体の流量Ｑ［Ｎｍ^３／ｈ］に換算する。ここでＳは対象ダクトの断面積である。
【００１２】
【数５】

【００１３】
【特許文献１】
特開２０００−２０６１３３号公報
【００１４】
【発明が解決しようとする課題】
音波式流量計の送信波形として用いる疑似ランダム信号であるＭ系列ＰＲＫ波は鋭敏な自己相関性を有しており、ノイズとの弁別、伝播時間の高精度な検出に最適な信号である。しかしながら、炉内やダクト内の圧力変動、極端に偏った温度分布やその変動といった伝播媒体の物性的性状により伝播過程で送信波形が大きな歪みを受ける状態が生じ得る。
【００１５】
この歪みを受けた信号状態において燃焼騒音、ファンや流体騒音等の騒音のノイズや燃焼炉やダクトの構造によって生じるエコー等が畳重すると誤った伝播時間の検出を行い、正しい温度や流量を計測できない問題が生じる。
【００１６】
このような問題に対処するために、従来では評価温度域、すなわち検出する伝播時間範囲を前回計測値又は基準値をもとに限定し、この範囲を逸脱した検出値を誤検出として除外し、前回計測値を維持する等のエラー処理を行っていた。
【００１７】
しかしながら燃焼炉の軌道、停止過程、特に急激な温度変化を生じる停止過程に対して、この方法を汎用的に追従させることが困難であり、計測値がホールド状態のまま復帰しない等の状態を呈し、現象への追従と誤検出の除去を両立させることが困難であった。また、流量計測においてもファンの停止・起動過程では伝播時間が急激に変化し、また火力プラントの燃焼空気ダクト内の空気は燃焼排ガスにより予熱されており、燃焼炉内の温度に応じて温度変化を示し、これに応じて伝播時間も変化を示す。
【００１８】
本発明の課題は、燃焼炉内またはダクト内の物性的条件によって伝播過程において生じる送信波形の歪みや外乱ノイズの畳重によってもたらされる伝播過程の誤検出状態を、温度・流量状態への追従性を維持したまま識別する方法を提供することにより、伝播時間検出に基づく音波式ガス温度・流量計測の計測性能と信頼性を向上させることにある。
【００１９】
【課題を解決するための手段】
上記課題を解決するため本発明では、検出した伝播時間を過去の所定のデータ数を保存するデータ記録手段と、過去の伝播時間から現在の伝播時間計測値までの信号列に対して各々一計測時間前（前回計測値）との差分信号を演算する手段と、演算した差分信号に対して評価する今回計測値を含まないデータ列と今回計測値を含むデータ列の２通りのデータに対してウェーブレット変換を行う手段と、２通りのデータ列に対して行ったウェーブレット振幅の強度比較またはパターン識別により評価する今回計測値の正常・異常を判定するエラー検出手段とを備えた音波式ガス温度・流量計測装置、及び、
検出した伝播時間を過去の所定のデータ数を保存し、過去の伝播時間から現在の伝播時間計測値までの信号列に対して各々一計測時間前（前回計測値）との差分信号を演算し、演算した差分信号に対して評価する今回検出した伝播時間計測値を含まないデータ列と今回の前記計測値を含むデータ列の二通りのデータに対してウェーブレット変換を行い、前記二通りのデータ列に対して行ったウェーブレット振幅の強度比較またはパターン識別により評価する今回計測値の正常・異常を判定してエラーを検出することからなる音波式ガス温度・流量計測方法である。
【００２０】
【作用】
音波式計測における伝播時間検出は前記１）式の相関器から構成されるマッチドフィルタにより受信信号と送信信号波形の同期を取ることにより行われる。伝播過程での信号波形の歪みと外乱ノイズの畳重によって生じる誤検出は、送信波形であるＭ系列ＰＲＫ信号の鋭敏な自己相関性がくずれたことにより真の同期時刻以外の相関のピークが優位に生じ、これを検出してしまうために生じる。このため誤検出された伝播時間は真値に対し離散的、突発的に生じる。
【００２１】
ウェーブレット解析は、このような突発的信号の解析に適した時間・周波数解析法であり、また、フーリエ変換のような直接畳み込み演算を行わせる場合に比べて高速に処理させることができ、このようなオンラインで処理が要求されるエラー処理への適用が可能となる。
【００２２】
また、ウェーブレット解析の前処理として、過去の伝播時間から現在の伝播時間計測値までの信号列に対して各々１計測時間前（前回計測値）との差分信号を生成することにより燃焼炉の燃焼状態又はダクト内のファンの運転状態等によって変化する伝播媒体である気体の温度により変動する伝播時間のトレンドを除去し、誤検出の状態をより明確にすることができる。
【００２３】
この場合、急激に温度が変化する瞬間の計測値はエラーと同様に検出されるが、必要とする追従性より早い計測周期を持たせることにより状態への追従性を維持しつつ、計測エラー状態を精度良く検出することができる。これにより誤った計測値の出力を防止し計測信頼性を向上することができる。
【００２４】
ここで問題とする計測エラーは離散的に発生する。一方で燃焼炉内でバーナが添加する等の場合、急激な温度上昇が生じ、音波の伝播時間も急激に変化する。例えば、３秒で数百度の温度変化が生じ、計測周期が３秒とすれば温度変化の過程は分からず、計測値はステップ状に変化する。ここでの処理には１計測時間前の前回値との差分をとるので、このような急激な温度変化は離散的な信号となってエラーと区別できない。
【００２５】
しかし実際には、温度変化等の物理変化はアナログ的に変化するので、このような温度変化過程を検出する（追従できる）計測周期（例えば、３秒で整定する変化に対し、０．５秒周期での計測）を持たせておけば、実際の温度等の物理状態による伝播時間の変化と検出エラーによるものを区別できる。
【００２６】
【発明の実施の形態】
本発明になる音波式温度・流量計測のエラー処理の実施の形態の構成を図１に示す。
図１に示すエラー処理は、検出した伝播時間を過去の所定のデータ数分だけ保存するデータ記録手段１０１と、この記録したデータから前回値との差分信号を演算する手段１０２と、演算した差分信号に対して評価する今回計測値を含まないデータ列と、今回計測値を含むデータ列の２通りのデータに対してウェーブレット変換を行う手段１０３と、２通りのデータ列に対して行ったウェーブレット振幅の強度比較またはパターン識別により評価する今回計測値の正常・異常を判定するエラー検出手段１０４から構成される。
【００２７】
伝播時間の誤検出処理に用いるウェーブレット変換は、基底関数のスケールパラメータ（周波数）と位置パラメータ（時間）とを逐次変えながら対象信号との内積をとり、対象信号波形を時間−周波数スペクトルに変換する手法である。このウェーブレット変換を用いることにより、対象信号の時間波形からどの時刻にどのような周波数成分が発生しているかという情報を得ることができ、特に突発的な状態の検知に有効である。信号ｆ（ｘ）に対するウェーブレット変換は次の６）式により定義されている。
【００２８】
【数６】

ここでψ（ｔ）は基底関数、ａはスケールパラメータ、ｂは位置パラメータであり、それぞれ時間−周波数解析の周波数と時間に対応するパラメータである。信号ｆ（ｘ）が、本実施の形態では正常・異常の判定を行う伝播時間検出信号波形である。また基底関数には種々の方式があるが、本実施例ではガボールウェーブレットを用いた。
【００２９】
図２に正常に伝播時間を検出している状態の伝播時間検出値信号波形に対するウェーブレット変換後の信号強度波形を示す。これに対して図３は誤検出データを含む伝播時間検出値信号波形に対して行ったウェーブレット信号波形を示す。
【００３０】
図３の中央の時刻が誤検出データの位置する時刻である。図２、図３を比較して明らかなように、突発的に生じる誤検出データに対してウェーブレット変換後の信号強度は優れた弁別性を示し、正常・異常の判定を精度良く行うことができる。
【００３１】
図４に実際の強度の温度分布を有し、外乱騒音の著しく、またエコーの影響を強く受ける構造の燃焼炉内での伝播時間検出値信号例を示す。これに対し図５は、本発明になる図１のエラー処理を用い誤検出値を除去した出力信号である。この例では、図５の伝播時間検出値から燃焼炉内温度を計測し、制御、監視に供される。
【００３２】
【発明の効果】
本発明によれば、音波式温度・流量計測において離散的、突発的に生じる伝播時間の誤検出に対し、このような信号状態の検出に優れる時間−周波数解析法であるウェーブレット変換を伝播時間検出値の差分信号に対して適用することにより、伝播時間の真の変動に追従しつつ、誤検出状態を精度良く判定することができる。
【００３３】
またこれにより、燃焼炉内、ダクト内の物性的条件によって伝播過程に於いて生じる送信波形の歪みや外乱ノイズの影響を強い計測環境において、伝播時間の誤検出状態を、温度・流量状態への追従性を維持したまま除去でき、伝播時間検出に基づく音波式ガス温度・流量計測の計測性と信頼性を向上させることができる。
【図面の簡単な説明】
【図１】音波式温度・流量計測の本発明になる実施の形態のエラー処理の構成を示した図。
【図２】正常な伝播時間検出状態の伝播時間検出信号波形に対するウェーブレット変換後の信号波形を示した図。
【図３】誤検出値を含む伝播時間検出信号波形に対するウェーブレット変換後の信号波形を示した図。
【図４】検出性の著しく悪い燃焼炉での伝播時間検出値の例を示した図。
【図５】図４の伝播時間検出値に対し本発明になるエラー処理を施し温度換算に用いる伝播時間信号を示した図である。
【図６】音波式ガス流量計の構成を示す図である。
【図７】従来の音波式ガス流量計の伝播時間検出処理のフローを示す。
【符号の説明】
１０１　伝播時間データ記録手段
１０２　前回値との差分信号演算手段
１０３　ウェーブレット変換を行う手段
１０４　今回計測値のエラー検出手段[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an apparatus for measuring the temperature and flow rate of a gas using sound waves, and particularly to measurement of high temperature in a combustion furnace, a combustion air duct of a large thermal power plant, an exhaust gas duct, etc., and temperature gas and air in a large structure. More specifically, the present invention relates to a method and apparatus for measuring temperature and flow rate which are suitable for the present invention.
[0002]
[Prior art]
In general, temperature control in a combustion furnace of a large thermal power plant such as a boiler for thermal power generation has been performed based on the measurement values of thermocouples and suction pyrometers. Such as clogging, the need for a long probe when applied to a large furnace (cross section 15m x 30m), and the durability when used continuously in a high temperature environment in a combustion furnace up to a thousand and hundred degrees. There was a problem and it was difficult to measure constantly.
[0003]
For measuring the flow rate of high-temperature air and gas fluid in large ducts such as thermal power plants, differential pressure type measuring devices such as orifices have been used. There are problems such as the necessity of a straight section, pressure loss caused by a differential pressure mechanism, and measurement clogging caused by dust clogging caused in a dirty environment such as exhaust gas.
[0004]
The temperature measurement in these combustion furnaces and the measurement of the flow rate of high-temperature air and gas fluid in large ducts, which were conventionally the mainstays of measurement devices of the differential pressure type such as orifice, were compared with the sound waves sent into the combustion furnace or duct. There is a sonic measuring device that detects the propagation time of a sound wave and measures the temperature and the flow velocity from the measurement of the propagation time of the sound wave in the gas expressed as a function of the temperature or the function of the flow velocity. As an example of this sonic measuring device, the configuration of a sonic gas flow meter disclosed in JP-A-2000-206133 is shown below.
[0005]
The sonic gas temperature / flow meter having the configuration shown in FIG. 6 is an M-sequence PRK (Phase Reverse Keying) waveform having sharp autocorrelation at an audible frequency of 20 kHz or less that can be followed by a gas having low damping characteristics. Transmission signal generation means 207 for generating a pseudo-random signal waveform, a D / A converter 205 for converting the signal generated by the transmission signal generation means 207 into an analog signal, Acoustic sensors A201a and B201b that transmit and receive sound waves propagated in a gaseous fluid, A / D converters 206 that digitally convert signals received by acoustic sensors A201a and B201b, and are captured by A / D converter 206. High-speed synchronization between the reception signal waveform and the transmission signal waveform generated by the transmission signal generation means 207 is achieved. Matched filter 209 includes a propagation time detection processing unit 210 for detecting the propagation time from output of the matched filter 209, the detected flow rate from the propagation time, is configured to be converted into the gas temperature and flow rate.
[0006]
FIG. 7 shows a flow of a propagation time detecting process of a conventional sonic gas temperature / flow meter. An M-sequence PRK wave as a transmission signal is generated as a digital signal rj (j = 0, 1,..., Q) of q + 1 data pieces, and this digital signal is generated by the D / A converter 205 in FIG. The signal is converted into an analog signal capable of driving a speaker, and drives a transmission mechanism of the acoustic sensor. Reception signals received by the paired acoustic sensors A201a and B201b after propagating in the flow path are sampled by the A / D converter 206 in FIG. 6, and n + 1 digital signals Sk (k = 0, 1, ..., N). Here, n> q.
The transmission signal r and the reception signal S are input to the correlator of the following formula 1) constituting the matched filter, and the correlation Xt between the reception signal and the transmission signal at each time t is calculated.
[0007]
(Equation 1)

This Xt is the output of the matched filter 209 in FIG. At the time when the timing (phase) matches the transmission M-sequence PRK waveform included in the reception signal, a sharp peak signal appears in the matched filter, and it is detected that the reception signal and the transmission signal are synchronized. The propagation time detection processing unit 210 in FIG. 6 detects the peak of the matched filter output Xt, and detects the peak time as the propagation time of the sound wave.
[0008]
Assuming that the detected propagation time from the acoustic sensors A201a to B201b is τAB and the propagation time from the acoustic sensors B201b to A201a is τBA, the relationship between the propagation times τAB, τBA, the flow velocity Vf, and the sound velocity Vs determined by the gas temperature T is 2a ), 2b). Here, L is the distance between the acoustic sensors A201a and B201b, and θ is the propagation angle formed by L with the flow direction of the fluid.
(Equation 2)

[0009]
From the equations 2a) and 2b), the flow velocity Vf and the sound velocity Vs are expressed by the following equations 3a) and 3b) from the detected propagation time values τab and τba.
[Equation 3]

[0010]
The gas temperature T [° C.] is expressed by the following equation 4) based on the temperature dependence of the speed of sound.
(Equation 4)

Here, α is a constant determined by the gas composition.
[0011]
In the case of gas temperature measurement, the measurement is completed by the above equation 4), but in the case of flow rate measurement, the gas is obtained from the following equation 5) based on the temperature and flow rate of the gas fluid measured as in the above equations 3) and 4). It is converted into a fluid flow rate Q [Nm ³ / h]. Here, S is the cross-sectional area of the target duct.
[0012]
(Equation 5)

[0013]
[Patent Document 1]
JP 2000-206133 A
[Problems to be solved by the invention]
An M-sequence PRK wave, which is a pseudo-random signal used as a transmission waveform of an acoustic flowmeter, has a sharp autocorrelation and is an optimal signal for discrimination from noise and highly accurate detection of a propagation time. However, the transmission waveform may be greatly distorted in the propagation process due to physical properties of the propagation medium, such as pressure fluctuations in the furnace or duct, and extremely biased temperature distributions and fluctuations thereof.
[0015]
In the signal state subjected to this distortion, if the noise of combustion noise, noise such as fan or fluid noise, or the echo generated by the structure of the combustion furnace or duct overlaps, the wrong propagation time is detected and the correct temperature and flow rate are measured. The problem that cannot be done arises.
[0016]
In order to cope with such a problem, conventionally, the evaluation temperature range, that is, the propagation time range to be detected is limited based on the previous measurement value or the reference value, and the detection value that deviates from this range is excluded as erroneous detection, Previously, error processing such as maintaining measured values was performed.
[0017]
However, it is difficult to make this method generally follow the orbit of the combustion furnace, the stopping process, especially the stopping process that causes a rapid temperature change, and the measured value does not return to the hold state and does not return. It has been difficult to achieve both the following of the phenomenon and the elimination of erroneous detection. Also, in the flow measurement, the propagation time changes rapidly during the stop / start process of the fan, and the air in the combustion air duct of the thermal power plant is preheated by the combustion exhaust gas, and the temperature changes according to the temperature in the combustion furnace. And the propagation time also changes accordingly.
[0018]
An object of the present invention is to make it possible to follow a temperature / flow rate state by detecting a transmission waveform distortion generated in a propagation process due to physical conditions in a combustion furnace or a duct or a false detection state of a propagation process caused by a superposition of disturbance noise. It is an object of the present invention to improve the measurement performance and reliability of the sonic gas temperature and flow rate measurement based on the detection of the propagation time by providing a method for identifying the signal while maintaining the measurement time.
[0019]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, the present invention provides a data recording means for storing a predetermined number of past data of a detected propagation time, and one measurement for each signal sequence from a past propagation time to a current propagation time measurement value. Means for calculating a difference signal from a previous time (previous measurement value); and two data: a data string not including the current measurement value and a data string including the current measurement value to evaluate the calculated difference signal. A sonic gas temperature and temperature detecting means comprising a means for performing a wavelet transform and an error detecting means for judging whether or not a current measurement value is normal or abnormal, which is evaluated by comparing the intensity of a wavelet amplitude performed on two kinds of data strings or by pattern identification. Flow measurement device, and
The detected propagation time is stored as a predetermined number of data in the past, and a difference signal is calculated for each signal sequence from the past propagation time to the current propagation time measurement value with respect to one measurement time before (previous measurement value). Performing a wavelet transform on two types of data, a data sequence that does not include the currently detected propagation time measurement value and a data sequence that includes the current measurement value, to evaluate the calculated difference signal, and the two types of data This is a sonic gas temperature and flow rate measurement method that determines whether the current measurement value is normal or abnormal and evaluates an error by comparing the intensity of wavelet amplitudes performed on the columns or by pattern identification.
[0020]
[Action]
The detection of the propagation time in the acoustic wave measurement is performed by synchronizing the waveform of the received signal and the waveform of the transmitted signal by a matched filter composed of the correlator of the above formula (1). The false detection caused by the distortion of the signal waveform in the propagation process and the superposition of disturbance noise is caused by the fact that the sharp autocorrelation of the M-sequence PRK signal, which is the transmission waveform, is lost, and the correlation peak other than the true synchronization time is superior. And this is detected because it is detected. Therefore, the erroneously detected propagation time occurs discretely and suddenly with respect to the true value.
[0021]
Wavelet analysis is a time-frequency analysis method suitable for analyzing such sudden signals, and can be processed at a higher speed than when direct convolution operation such as Fourier transform is performed. It can be applied to error processing that requires processing online.
[0022]
In addition, as a pre-process of the wavelet analysis, a difference signal is generated for each signal sequence from the past propagation time to the current propagation time measurement value, which is one measurement time ago (previous measurement value), thereby producing a combustion furnace combustion. It is possible to remove the trend of the propagation time that fluctuates depending on the state or the temperature of the gas serving as the propagation medium, which fluctuates depending on the operation state of the fan in the duct, and the like, so that the erroneous detection state can be made clearer.
[0023]
In this case, the measurement value at the moment when the temperature suddenly changes is detected in the same way as an error, but the measurement error state is maintained while maintaining the followability to the state by providing a measurement cycle faster than the required followability. Can be accurately detected. This prevents the output of erroneous measurement values and improves measurement reliability.
[0024]
The measurement error in question here occurs discretely. On the other hand, when a burner is added in the combustion furnace or the like, a sharp rise in temperature occurs, and the propagation time of the sound wave also changes rapidly. For example, a temperature change of several hundred degrees occurs in 3 seconds, and if the measurement cycle is 3 seconds, the process of the temperature change is unknown, and the measured value changes stepwise. In this process, a difference from the previous value one measurement time before is obtained, and thus such a rapid temperature change becomes a discrete signal and cannot be distinguished from an error.
[0025]
However, in practice, since a physical change such as a temperature change changes in an analog manner, a measurement cycle for detecting (following) such a temperature change process (for example, 0.5 seconds for a change settled in 3 seconds) (Measurement in a cycle), it is possible to distinguish between a change in propagation time due to a physical state such as an actual temperature and a detection error.
[0026]
BEST MODE FOR CARRYING OUT THE INVENTION
FIG. 1 shows the configuration of an embodiment of the error processing of the sonic temperature / flow rate measurement according to the present invention.
The error processing shown in FIG. 1 includes a data recording unit 101 for storing the detected propagation time for a predetermined number of past data, a unit 102 for calculating a difference signal from the recorded data and a previous value, Means 103 for performing a wavelet transform on two types of data, ie, a data string that does not include the current measurement value to be evaluated for the signal, and a data string that includes the current measurement value, and a wavelet that is performed on the two data strings. An error detecting means 104 for judging whether the current measurement value is normal or abnormal, which is evaluated by comparing the amplitude intensity or pattern identification.
[0027]
The wavelet transform used for the erroneous detection of the propagation time takes the inner product of the target signal while sequentially changing the scale parameter (frequency) and the position parameter (time) of the basis function, and converts the target signal waveform into a time-frequency spectrum. Method. By using this wavelet transform, it is possible to obtain information as to what frequency component is generated at which time from the time waveform of the target signal, which is particularly effective for detecting a sudden state. The wavelet transform for the signal f (x) is defined by the following equation 6).
[0028]
(Equation 6)

Here, ψ (t) is a basis function, a is a scale parameter, and b is a position parameter, which are parameters corresponding to the frequency and time in the time-frequency analysis, respectively. The signal f (x) is a propagation time detection signal waveform for determining normal / abnormal in the present embodiment. There are various types of basis functions. In this embodiment, Gabor wavelets are used.
[0029]
FIG. 2 shows a signal intensity waveform after wavelet transform for a propagation time detection value signal waveform in a state where the propagation time is normally detected. On the other hand, FIG. 3 shows a wavelet signal waveform performed on a propagation time detection value signal waveform including erroneously detected data.
[0030]
The time at the center in FIG. 3 is the time at which the erroneously detected data is located. As is clear from the comparison between FIGS. 2 and 3, the signal strength after the wavelet transform shows excellent discrimination against suddenly generated erroneously detected data, and the normality / abnormality can be accurately determined. .
[0031]
FIG. 4 shows an example of a detected signal of a propagation time in a combustion furnace having a structure having an actual temperature distribution of intensity, a remarkable disturbance noise, and a strong influence of an echo. On the other hand, FIG. 5 shows an output signal from which an erroneous detection value has been removed by using the error processing of FIG. 1 according to the present invention. In this example, the temperature in the combustion furnace is measured from the detected propagation time in FIG. 5, and is used for control and monitoring.
[0032]
【The invention's effect】
According to the present invention, a wavelet transform, which is a time-frequency analysis method excellent in detecting such a signal state, is used to detect a propagation time in erroneous detection of a propagation time which is discrete and suddenly generated in a sound wave type temperature / flow rate measurement. By applying to the difference signal of the value, the erroneous detection state can be accurately determined while following the true fluctuation of the propagation time.
[0033]
In addition, this allows the erroneous detection of the propagation time to be changed to the temperature / flow rate condition in a measurement environment where the effect of the transmission waveform distortion and disturbance noise generated in the propagation process due to the physical conditions in the combustion furnace and duct is strong. It can be removed while maintaining the followability, and the measurement performance and reliability of the sonic gas temperature / flow rate measurement based on the detection of the propagation time can be improved.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration of an error processing according to an embodiment of the present invention for sonic temperature / flow rate measurement.
FIG. 2 is a diagram showing a signal waveform after wavelet transform with respect to a propagation time detection signal waveform in a normal propagation time detection state.
FIG. 3 is a diagram showing a signal waveform after a wavelet transform for a propagation time detection signal waveform including an erroneous detection value.
FIG. 4 is a diagram showing an example of a propagation time detection value in a combustion furnace having extremely poor detectability.
5 is a diagram showing a propagation time signal used for temperature conversion by performing error processing according to the present invention on the propagation time detection value of FIG. 4;
FIG. 6 is a diagram showing a configuration of a sound wave type gas flow meter.
FIG. 7 shows a flow of a propagation time detection process of a conventional sonic gas flow meter.
[Explanation of symbols]
101 Propagation time data recording means 102 Difference signal calculation means from the previous value 103 Wavelet transform means 104 Error detection means of current measurement value

Claims (2)

A sonic gas temperature / flow rate measuring device that detects the propagation time of sound waves sent into a combustion furnace or duct, and measures gas temperature / flow rate by detecting sound propagation time in gas expressed as a function of temperature or flow velocity At
Data recording means for storing the detected propagation time in the past predetermined number of data,
Means for calculating a difference signal from the previous propagation time to the current propagation time measurement value and a signal sequence one measurement time before each,
Means for performing a wavelet transform on two types of data: a data string that does not include the currently detected propagation time measurement value and a data string that includes the current measurement value to be evaluated for the calculated difference signal;
Sonic gas temperature and flow rate characterized by comprising error detection means for judging the normality / abnormality of the current measurement value evaluated by comparing the intensity of the wavelet amplitude or pattern identification performed on the two data strings. Measuring device. A sonic gas temperature / flow rate measurement method that detects the propagation time of sound waves sent into a combustion furnace or duct and measures gas temperature / flow rate from sound wave propagation time detection in gas expressed as a function of temperature or flow velocity In addition,
The detected propagation time is stored for a predetermined number of past data,
Calculate the difference signal from one measurement time before for each signal sequence from the past propagation time to the current propagation time measurement value,
Perform a wavelet transform on two types of data, a data sequence that does not include the currently detected propagation time measurement value and a data sequence that includes the current measurement value, to evaluate the calculated difference signal,
The sonic gas temperature and the sonic gas temperature characterized by determining whether the current measurement value is normal or abnormal to be evaluated by comparing the intensity of the wavelet amplitude or pattern identification performed on the two data strings and detecting an error. Flow measurement method.

Images