JP3938276B2 - Flame detector and flame detection method - Google Patents

Description

【００２３】
炎判断手段２において、このようにして求めた平均ピークレベルＭ_avgは、赤外線検出信号のレベルが火災時の炎特有の赤外線強度レベル特性をどの程度有しているかを判断するのに用いられ、また、上記のように求めたピーク度数ＰＣＮＴは、赤外線強度レベルに関するゆらぎ周波数を反映したものとなっていることから、赤外線検出信号のレベルの時間的変化が火災時の炎特有の時間的揺らぎ(ちらつき)特性をどの程度有しているかを判断するのに用いられる。
【００２４】
このような炎判断処理では、平均ピークレベルＭ_avgは、赤外線検出信号が火災時の炎特有の主要な特徴をどの程度有しているかを良好に反映したものとなっており、また、ピーク度数ＰＣＮＴは、赤外線強度レベルに関するゆらぎ周波数を反映したものとなっていることから、炎判断手段２は、基本的には、この平均ピークレベルＭ_avgおよび／またはピーク度数ＰＣＮＴに基づいて、炎か否かの判断を行なうことができる。
【００２５】
さらに、このような炎の判断処理によっても、炎であるか否かの判断が難かしいときには、炎の判断がある程度信頼性良く行なうことができるものとなるまで、所定時間Ｔを設定変更し、炎の判断を行なうこともできる。これにより、基本的な炎判断をより信頼性良く行なうことができる。
【００２６】
上記のような炎判断処理では、例えば、比較的油煙の少ないガス炎、木材，紙類の燃焼炎、初期火災時の炎の場合には、上記のように求めた平均ピークレベルＭ_avgは十分に大きなレベルとなり、また、ピーク度数ＰＣＮＴも大きなものとなるので、これにより、容易に炎と判断することができる。
【００２７】
しかしながら、このように優れた信頼性の高い基本的な炎判断がなされる場合であっても、油煙が非常に多いとき(例えば、ガソリン，重油，石油化学，ゴム製品などの燃焼では、油煙の多い燃焼になり易く、特に燃焼規模が大きく酸素が欠乏気味の火災では、一酸化炭素ＣＯの発生が増え油煙の多い炎となる)、あるいは、他の熱源からのノイズなどが混入するときには、例えば、上記のようにして求めた平均ピークレベルＭ_avgは、炎か否かを判断するのに微妙なレベル、すなわち、あいまいなレベルとなることがあり、このときには、炎か否かを確実に判断することができないことがある。
【００２８】
本願の発明者は、油煙の多い炎の場合でも、炎の判断を正しくかつ確実に行なうことができるよう、炎の燃焼の特徴を調べた。その結果、燃焼炎の場合には、周囲からの気流(酸素供給)の影響を受けるために、放射光に数〜十数Ｈｚの揺らぎが発生すること、また、赤外線センサとして特に微分型の感度特性をもつ焦電型センサを用いて、この放射光を検出すると、焦電型センサからの赤外線検出信号の揺らぎに強弱があり、赤外線検出信号の各ピークのレベル(ピーク電圧値)を調べると、各ピークのレベルがランダム性をもつことがわかった(ランダムに分散している(ばらついている)ことがわかった)。
【００２９】
従って、所定時間内における赤外線検出信号の各ピークレベルについてのヒストグラムをとると、特定のランク(区間)に度数が集中せず、複数の区間に度数がばらついて分布することがわかった。
【００３０】
なお、所定時間内における赤外線検出信号の各ピークレベルについてのヒストグラムは、例えば、次のように作成することができる。
【００３１】
すなわち、赤外線検出信号のレベルを複数の区間に区切り、そのときの区間の数をＮ(例えば“７”)とする。また、所定時間Ｔ内におけるデータ数(ピークの個数)をｍ(例えば、仮に“３５”)とし、ｍ個(３５個)のデータ(ピークレベル)の中で、最大値をＭ_max，最小値をＭ_minとするときに、ヒストグラムの各区間の幅Ｗを、次式により決定する。
【００３２】
【数２】
Ｗ＝(Ｍ_max−Ｍ_min）／Ｎ
【００３３】
このようにして、区間の数Ｎ(例えば“７”)と各区間の幅Ｗとを設定した後、ｍ個(３５個)のデータ(ｍ個のピークのレベル)をＮ個の区間のうちの該当する所定の区間にそれぞれ当てはめて、各区間の度数分布を作成することができる。
【００３４】
図４には、火災時の炎からの赤外線を赤外線センサ(焦電型センサ)で検出した赤外線検出信号のピークレベルの変化の一例が示され、また、図５には、低温熱源(４００℃以下)からの赤外線を赤外線センサ(焦電型センサ)で検出した赤外線検出信号のピークレベルの変化の一例が示されている。また、図６には、図４の赤外線検出信号(火災時の炎による信号)と図５の赤外線検出信号(低温熱源による信号)とのそれぞれの場合におけるピークレベルの度数分布(ヒストグラム)の作成例が示されている。
【００３５】
なお、図４，図５の例では、所定時間Ｔ内のピークレベルの個数ｍが“５０”であり、５０個のピークのレベルのうち、最大のピークレベルが“２４５０ｍＶ”，最小のピークレベルが“０Ｖ”であるので、図６のヒストグラムは、ピークレベルを７個の区間に分割し、各区間の幅Ｗを“３５０ｍＶ”に設定して、５０個のピークレベルを７個の区間の該当する区間にそれぞれ割り振ったときの度数分布として作成されている。
【００３６】
図６からわかるように、火災時の炎による度数分布は、各区間でほぼ均等となる(ランダムなものとなる)のに対し、低温熱源による度数分布は、特定の区間に集中する(ランダムでない)ことがわかる。
【００３７】
従って、上述したような基本的な炎判断処理において、平均ピークレベルＭ_avgおよび／またはピーク度数ＰＣＮＴが、炎か否かを判断する上であいまいなものとなっている場合にも、各ピークレベルがランダム性を有しているか否かを図６のようなヒストグラムを作成して検知することによって、炎か否かを確実に判断することが可能となる。上記例では、火災時の炎による赤外線検出信号は各ピークレベルが十分なランダム性を有しているので、炎と確実に判断でき、また、低温熱源による赤外線検出信号は各ピークレベルがランダム性を有していないので、炎でないと確実に判断できる。
【００３８】
より具体的に、所定時間内のデータ数ｍが“３５”であり、区間数Ｎが“７”である場合の、各区間の度数を例えば次表のようなランク(次表の例では、４種のランクＲ₁，Ｒ₂，Ｒ₃，Ｒ₄)に分類することができる。
【００３９】
【表１】

【００４０】
なお、表１において、平均度数は、全データ数／区間数であり、上記例では、“５”となる。各区間の度数をこのようにランク分けしたときに、各ランクに属する区間の個数により、炎度(炎である蓋然性を表わす指標)を例えば次表のようにして求めることができる。
【００４１】
【表２】

【００４２】
なお、表２には、ランクＲ₁に属する区間の個数Ｕ₁については示されていないが、例えば、(Ｕ₂≦１，Ｕ₃＝０，Ｕ₄＝０)のときには、Ｕ₁は例えば“６”となり、(Ｕ₂≦２，Ｕ₃＝０，Ｕ₄＝０)のときには、Ｕ₁は例えば“５”となる。
【００４３】
表２からわかるように、(Ｕ₂≦１，Ｕ₃＝０，Ｕ₄＝０)のときには、Ｕ₁が最も多く、従って、ピークレベル度数が各区間にほぼ均一に分散(ランダムに分散)しており、このときに、表２から、炎である蓋然性，すなわち、炎度を“１”として割り出すことができる。
【００４４】
このような構成の炎感知器では、例えば、平均ピークレベルＭ_avgおよびピーク度数(ピークの個数)ＰＣＮＴが十分に大きいときには(例えば、平均ピークレベルＭ_avgが０．２Ｖよりも十分に大きく、また、ピーク度数(ピークの個数)ＰＣＮＴが３０個以上のときには)、これのみによって、火災と判断し、その旨を出力することができる。
【００４５】
一方、平均ピークレベルＭ_avgまたはピーク度数(ピークの個数)ＰＣＮＴまたはこれら両方が十分には大きくはなく、これのみによっては火災か否かを判断することが難かしい(微妙な)場合(例えば、平均ピークレベルＭ_avgが０．２Ｖよりもほんの少し大きい程度の場合あるいはピーク度数(ピークの個数)ＰＣＮＴが１０〜２０個程度の場合)には、さらに、例えば、表１，表２に示したようにしてピークレベルのランダム性に基づく炎度を用い、炎度が例えば“１”のとき(すなわちランダム性が十分に大きいとき)、炎と判断し、炎度が例えば“０”のとき、炎ではないと判断することができる。
【００４６】
このように、炎判断手段２において、赤外線検出信号から炎特有の特徴量を抽出し、炎特有の特徴量に基づき炎判断を行なう際、該炎判断があいまいなものとなる場合には、抽出した炎特有の特徴量のランダム性を考慮して炎判断を行なうことにより、在来の炎判断処理において炎であるか否かの判断があいまいなものとなる場合であっても、炎であるか否かの判断を正しくかつ確実に行なうことが可能になる。
【００４７】
上述の例では、１つの赤外線検出手段だけが設けられているとして説明したが、図１のような複数の赤外線検出手段１−１〜１−ｎが設けられる場合にも、上述した炎判断処理の原理を拡張することができる。例えば、ｎが２の場合、あるいは、ｎが３の場合、あるいは、ｎが３よりも大きい炎感知器にも、上述した炎判断処理の原理を拡張することができる。
【００４８】
以下では、図１の炎感知器の構成において、ｎが２であるとして説明する。すなわち、炎感知器が、例えば、赤外線検出手段１−１と赤外線検出手段１−２との２つの赤外線検出手段を有しているものとして説明する。ここで、赤外線検出手段１−１には、炎特有の波長の赤外線を第１波長信号（主波長信号）として検出するものを用い、また、赤外線検出手段１−２には、赤外線検出手段１−１によって検出される赤外線波長とは異なる波長の光を第２波長信号（参照波長信号）として検出するものを用いることができる（２波長式の炎感知器として構成できる）。この場合、炎判断手段２は、赤外線検出手段１−１からの第１波長信号と赤外線検出手段１−２からの第２波長信号とに基づいて炎の判断を行ない、該炎の判断があいまいなものとなる場合であっても、第１波長信号の各ピークのレベルがランダムに分散しているときには炎と判断することができる。
【００４９】
具体的に、赤外線検出手段１−１には、例えば４．３〜４．４μｍ程度の波長の赤外線を検出する赤外線センサ(例えば焦電型センサ)を用いることができる。
【００５０】
また、赤外線検出手段１−２には、炎特有の赤外線波長(４．３〜４．４μｍ程度)とは異なる波長の光として、炎特有の赤外線波長の近傍の波長の赤外線、例えば、３．９μｍ程度または５．０μｍ程度の波長の赤外線を検出する赤外線センサ(例えば焦電型センサ)を用いることができる。すなわち、３．９μｍ程度または５．０μｍ程度の波長では、図２からわかるように、炎のスペクトル強度は小さく(極小値となり)、従って、この場合、赤外線検出手段１−２は、３．９μｍ程度または５．０μｍ程度の波長のところでの赤外線のレベルを、炎の補助的な特徴として検出することができる。
【００５１】
換言すれば、この場合、赤外線検出手段１−１は、基準となる赤外線検出手段として機能し、赤外線検出手段１−２は、補助的な（参照用の）赤外線検出手段として機能するようになっている。
【００５２】
赤外線検出手段１−２に、赤外線検出手段１−１によって検出される赤外線波長の近傍の波長の赤外線を検出するものが用いられる場合には、炎判断手段２は、図７に示すように、赤外線検出手段１−１が検出した赤外線検出信号（第１波長信号（主波長信号））のレベルのピーク時に赤外線検出手段１−２が検出した赤外線検出信号（第２波長信号（参照波長信号（補助検出信号）））のレベルＳ₁，Ｓ₂，…，Ｓ_mを取り込み、所定期間Ｔにわたってこのように取り込んだ赤外線検出手段１−２からの第２波長信号のレベルＳ₁，Ｓ₂，…，Ｓ_mの平均値Ｓ_avgを平均補助レベルとして求め、例えば、赤外線検出手段１−１からの第１波長信号の平均ピークレベルＭ_avgと赤外線検出手段１−２からの第２波長信号の平均補助レベルＳ_avgとの比率(Ｍ_avg／Ｓ_avg)と、第１波長信号の所定期間Ｔにわたるピーク度数(ピークの個数)ＰＣＮＴ(＝ｍ)とに基づき、基本的な炎の判断を行なうようになっている。
【００５３】
また、図７において、所定期間Ｔは、基本的な炎の判断結果に応じて、可変に設定することができる。
【００５４】
例えば、図７の処理において、所定時間Ｔにわたって得られた平均ピークレベルＭ_avgと平均補助レベルＳ_avgとの比Ｍ_avg／Ｓ_avgに基づき炎の判断を行なうことが難しいときには、所定時間Ｔを設定変更し、このように設定変更された所定時間Ｔにわたって、平均ピークレベルＭ_avgと平均補助レベルＳ_avgとの比Ｍ_avg／Ｓ_avgを再度算出し、これに基づき再度炎の判断を行なうというように、炎の判断がある程度信頼性良く行なうことができるものとなるまで、所定時間Ｔを設定変更し、基本的な炎の判断を行なうことができる。
【００５５】
図８は火災時における主波長信号（第１波長信号），参照波長信号（第２波長信号）の一例を示す図であり、また、図９は非火災時(例えば、ハロゲン灯，熱源ノイズなどが加わった場合)における主波長信号（第１波長信号），参照波長信号（第２波長信号）の一例を示す図である。なお、図８，図９の例においては、赤外線検出手段１−２に、赤外線検出手段１−１で検出される赤外線とは異なる波長の赤外線を検出する赤外線センサ(焦電型センサ)が用いられているとしている。
【００５６】
先ず、火災時においては、火災時の炎特有の赤外線の強度レベル(第１波長信号のレベル)は図８(ａ)に示すように大きくなるのに対し、第２波長信号のレベルは図８(ｂ)に示すように差程大きくはならない。従って、所定期間Ｔ(例えば４．５秒)にわたって得られる平均ピークレベルＭ_avgと平均補助レベルＳ_avgとの比率Ｍ_avg／Ｓ_avgは十分に大きくなる。これにより、この場合には、図８(ｃ)，(ｄ)に示すように、所定期間Ｔの更新を行なうことなく、この所定期間Ｔ(＝４．５秒)だけで、炎と判断し、その結果を炎検知信号として出力することができる。
【００５７】
また、非火災時において、ハロゲン灯，熱源ノイズなどが発生したときにも、火災時の炎特有の赤外線波長の強度レベル(第１波長信号のレベル)は、図９(ａ)に示すように大きくなるが、このときには、第２波長信号の強度レベルも図９(ｂ)に示すように大きくなり、従って、所定期間Ｔ(例えば４．５秒)にわたって得られる平均ピークレベルＭ_avgと平均補助レベルＳ_avgとの比率Ｍ_avg／Ｓ_avgは小さくなり、これに基づいて炎の判断を行なうことは難かしい。これにより、この場合には、図９(ｃ)に示すように、炎の判断がある程度信頼性良く行なうことができるものとなるまで、所定時間Ｔを設定変更し、基本的な炎の判断を行なう所定期間Ｔをさらに延長する。図９(ｃ)の例では、さらに９秒延長し、合計でＴ＝１３．５秒に設定変更して、炎の判断を再度行なう。図９の例では、所定期間Ｔを１３．５秒にして炎か否かを判断するとき、炎ではないと判断され、従って、この場合には、図９(ｄ)のように、炎検知信号は出力されない。
【００５８】
このように、基本的な炎判断処理においても信頼性が低いと考えられる時は平均化時間を長くすることで、その積分効果により、より信頼性の高い炎判断が可能となり、誤報の危険を少なくできる。
【００５９】
しかしながら、このように信頼性の高い優れた基本的な炎判断がなされる場合であっても、前述のように、油煙が非常に多いとき、あるいは、他の熱源からのノイズなどが混入するときには、例えば、上記のようにして求めた比率Ｍ_avg／Ｓ_avgが、炎か否かを判断するのに微妙なレベル、すなわち、あいまいなレベルとなることがあり、このときには、炎か否かを確実に判断することができないことがある。
【００６０】
図１０には、低油煙(ノルマル−ヘプタン)炎を発生させたときの第１波長信号(４．３〜４．４μｍ程度)と第２波長信号(例えば５．０μｍ)のレベルの測定結果(時間的変化)が示され、また、図１１には、高油煙(ガソリン)炎を発生させたときの第１波長信号(４．３〜４．４μｍ程度)と第２波長信号(例えば５．０μｍ)のレベルの測定結果(時間的変化)が示され、また、図１２には、高温物体(４００℃以上)をチョッパしたときの第１波長信号(４．３〜４．４μｍ程度)と第２波長信号(例えば５．０μｍ)のレベルの測定結果(時間的変化)が示されている。
【００６１】
低油煙(ノルマル−ヘプタン)炎の場合には、図１０からわかるように、第１波長信号のレベルは第２波長信号のレベルに比べて十分に大きく、従って、これのみで炎であると判断できる。一方、高油煙(ガソリン)炎や高温物体からの熱幅射の場合には、図１１，図１２からわかるように、第１波長信号のレベルは第２波長信号のレベルよりも大きいが十分に大きなものではないので、これに基づく炎判断にあいまい性が生じる。
【００６２】
そこで、図１１，図１２のように、炎判断にあいまい性が生じる場合、炎判断手段２は、前述したと同様に、さらに、第１波長信号の各ピークのレベルがランダムに分散しているか否かを調べ、各ピークのレベルがランダムに分散しているときには炎と判断するようになっている。
【００６３】
このように、炎感知器が上述したような２波長式のものである場合にも、あるいは、３波長式のものである場合にも、その炎判断があいまいなものとなる場合には、炎判断手段２は、さらに、第１波長信号の各ピークのレベルがランダムに分散しているか否かを調べて、炎か否かを最終的に判断することができる。
【００６４】
また、上述の例では、炎の判断を行なう際の特徴量(炎特有の特徴量)として、赤外線検出信号の平均ピークレベルとピーク度数とを用い、また、この際、赤外線検出信号のレベルが所定の閾値レベルＶ_thを越えた時点から所定期間Ｔにわたって得られる赤外線検出信号のレベルのうち、所定の閾値レベルＶ_thを越えたものにのみ着目し、所定の閾値レベルＶ_thを越えた赤外線検出信号についてのみピークを検知して、平均ピークレベル，ピーク度数を求めているが、炎の判断を行なう際の基本的な特徴量としては、場合に応じて、任意のものを用いることができる。例えば、赤外線検出信号の平均ピークレベルのみを炎特有の特徴量として用いることもできるし、あるいは、赤外線検出信号のピーク度数のみを炎特有の特徴量として用いることもできるし、あるいは、赤外線検出信号のレベルが所定の閾値レベルＶ_thを越えた時点から所定期間Ｔにわたって得られる赤外線検出信号について、そのレベルが所定の閾値レベルＶ_thを越えないものをも含めて、レベルのピークを検知して、平均ピークレベルＰＡＶおよび／またはピーク度数を求め、これを炎特有の特徴量とすることもできる。あるいは、平均ピークレベル，ピーク度数に限らず、炎の特徴を抽出したものであれば、これを炎特有の特徴量とすることもできる。
【００６５】
また、上述した例では、ピークレベルのヒストグラムからピークレベルのランクを作成してランダム性を求めるようになっているが、他の方法によってランダム性を求めることもできる。例えば、ピークレベルの標準偏差を求め、その標準偏差の大小によってピークレベルのバラツキ（ランダム性）を第１の条件として判断し、また、所定期間中のピークレベルの周期を求め、その周期について標準偏差を求めて、その標準偏差の大小によって周期のバラツキ（ランダム性）を第２の条件として判断し、そして最終的に第１および第２の両方の条件を満たしたときに炎としてのランダム性を有すると判断することもできる。
【００６６】
ところで、図１のように、複数の赤外線検出手段１−１〜１−ｎでそれぞれ検出された赤外線検出信号から炎特有の特徴量を抽出し、炎特有の特徴量に基づき炎判断を行なう炎感知器において、例えば、基準となる赤外線検出手段（例えば１−１）だけに例えば変則的な光源の入射（揺らいだ光源の入射）がある場合や連続的なノイズが加わるような場合には、炎感知器の炎判断手段２において上述したような炎判断がなされるときにも、誤報が発生する恐れがある。
【００６７】
本願の発明者は、基準となる赤外線検出手段（例えば１−１）だけに例えば変則的な光源の入射（揺らいだ光源の入射）がある場合や連続的なノイズが加わるような場合には、この赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の赤外線検出手段１−２〜１−ｎからの赤外線検出信号（参照波長信号）との間には同期性がなくなる一方、複数の赤外線検出手段１−１〜１−ｎが炎を検出する場合には、これら複数の赤外線検出手段１−１〜１−ｎは、ほぼ同期した信号を出力することに着目し、本発明を完成させた。
【００６８】
すなわち、本発明では、炎判断手段２は、複数の赤外線検出手段１−１〜１−ｎでそれぞれ検出された赤外線検出信号の同期性を判断し、同期性があると判断された赤外線検出信号を用いて炎判断を行なうようになっていることを特徴としている。
【００６９】
換言すれば、本発明では、上述したような炎判断に加えて、複数の赤外線検出手段１−１〜１−ｎが炎を検出した場合、これらがほぼ同期した（同相の）信号を出力することを炎判断の要素に加えるようにしている。
【００７０】
各赤外線検出手段１−１〜１−ｎからの信号の同期性を判断する仕方としては、高速でかつ容易に同期性を判断することの可能な仕方が望まれる。このような仕方として、本発明では、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）のレベルが所定の閾値レベルを超えた時点（立ち上がり時）と、その後にこの赤外線検出信号（主波長信号）のレベルが最大レベル（ピーク）に達した時点（ピーク時）と、その後にこの赤外線検出信号（主波長信号）のレベルが所定の閾値レベルよりも低くなる時点（立ち下がり時）との３つの時点における他の赤外線検出手段１−２〜１−ｎからの赤外線検出信号（参照波長信号）のレベルに基づいて、赤外線検出信号の同期性を判断するようにしている。
【００７１】
図１３（ａ），（ｂ）には、赤外線検出信号の同期性の判断の具体例が示されている。なお、図１３（ａ）は基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）を示す図であり、図１３（ｂ）は他の赤外線検出手段（例えば、１−２）からの赤外線検出信号（参照波長信号）を示す図である。図１３を参照すると、炎判断手段２は、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）の所定の閾値レベルを基準にしたある１つの立ち上がり時，ピーク時，立ち下がり時における他の赤外線検出手段１−２からの赤外線検出信号（参照波長信号）の信号レベルを、それぞれ、ａ_r，ｐ_r，ｂ_rとするとき、他の赤外線検出手段において、ｐ_r＞ａ_r，ｐ_r＞ｂ_rの条件が満たされる場合に、赤外線検出信号に同期性があると判断するようになっている。
【００７２】
図１３の例では、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の１つの赤外線検出手段１−２からの赤外線検出信号（参照波長信号）との間での同期性の判断について説明したが、他の赤外線検出手段が複数設けられている場合には（１−２〜１−ｎの場合には）、炎判断手段２は、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）の所定の閾値レベルを基準にしたある１つの立ち上がり時，ピーク時，立ち下がり時における他の赤外線検出手段からの赤外線検出信号（参照波長信号）の信号レベルを、それぞれ、ａ_r，ｐ_r，ｂ_rとするとき、他の赤外線検出手段１−２〜１−ｎのうちの所定個数以上の赤外線検出手段において、ｐ_r＞ａ_r，ｐ_r＞ｂ_rの条件が満たされる場合に、赤外線検出信号に同期性があると判断するようになっている。
【００７３】
また、図１３の例では、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の１つの赤外線検出手段１−２からの赤外線検出信号（参照波長信号）のうちの１組の立ち上がり，ピーク，立ち下がりに着目したが、火災が発生する場合、所定期間内において、立ち上がり，ピーク，立ち下がりの組が一般に複数生起する。この場合、炎判断手段２は、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）の所定の閾値レベルを基準にしたある１つの立ち上がり時，ピーク時，立ち下がり時における他の赤外線検出手段１−２からの赤外線検出信号（参照波長信号）の信号レベルを、それぞれ、ａ_r，ｐ_r，ｂ_rとするとき、他の赤外線検出手段１−２において、ｐ_r＞ａ_r，ｐ_r＞ｂ_rの条件が、所定期間内に生起する立ち上がり，ピーク，立ち下がりの生起回数のうちの所定回数以上、満たされるときに、赤外線検出信号に同期性があると判断するようになっている。
【００７４】
また、これらを統合して、炎判断手段２は、基準となる赤外線検出手段１−１からの赤外線検出信号の所定の閾値レベルを基準にしたある１つの立ち上がり時，ピーク時，立ち下がり時における他の赤外線検出手段１−２〜１−ｎからの赤外線検出信号の信号レベルを、それぞれ、ａ_r，ｐ_r，ｂ_rとするとき、他の赤外線検出手段１−２〜１−ｎのうちの所定個数以上の赤外線検出手段において、ｐ_r＞ａ_r，ｐ_r＞ｂ_rの条件が、所定期間内に生起する立ち上がり，ピーク，立ち下がりの生起回数のうちの所定回数以上、満たされるときに、赤外線検出信号に同期性があると判断するようになっている。
【００７５】
図１４には、炎が発生したときの基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の１つの赤外線検出手段１−２からの赤外線検出信号（参照波長信号）と他の１つの赤外線検出手段１−３からの赤外線検出信号（参照波長信号）との一例が示されており、図１４の例では、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の１つの赤外線検出手段１−２からの赤外線検出信号（参照波長信号）と他の１つの赤外線検出手段１−３からの赤外線検出信号（参照波長信号）とは同期していることがわかる。この場合、炎判断手段２では、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の１つの赤外線検出手段１−２からの赤外線検出信号（参照波長信号）と他の１つの赤外線検出手段１−３からの赤外線検出信号（参照波長信号）とが同期していると判断し、同期していると判断された赤外線検出信号（主波長信号および／または参照波長信号）を用いて、前述したような炎判断処理を行なうことができる。
【００７６】
また、図１５には、ノイズが混入したときの基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の１つの赤外線検出手段１−２からの赤外線検出信号（参照波長信号）と他の１つの赤外線検出手段１−３からの赤外線検出信号（参照波長信号）との一例が示されており、また、図１６には、変則的な光源が入力したときの基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の１つの赤外線検出手段１−２からの赤外線検出信号（参照波長信号）と他の１つの赤外線検出手段１−３からの赤外線検出信号（参照波長信号）との一例が示されている。図１５，図１６の例では、基準となる赤外線検出手段１−１からの赤外線検出信号（主波長信号）と他の１つの赤外線検出手段１−２からの赤外線検出信号（参照波長信号）と他の１つの赤外線検出手段１−３からの赤外線検出信号（参照波長信号）とは同期しておらず、従って、この場合、炎判断手段２では、赤外線検出手段からの赤外線検出信号がノイズや変則的な光源の影響を受けていると判断し、この赤外線検出信号を炎の判断には用いない。
【００７７】
図１７は本発明の炎感知器の具体例を示す図である。なお、図１７の例では、簡単のため、２つの赤外線検出手段１−１，１−２だけが設けられている場合が示されている。また、炎感知器は火災感知器として構成されている。
【００７８】
図１７を参照すると、この炎感知器の基準となる赤外線検出手段１−１は、炎特有の赤外線(一般にはＣＯ₂共鳴放射の４．３μｍ〜４．４μｍ程度の赤外線)を第１波長信号（主波長信号）として検出する赤外線センサ２１と、該赤外線センサ２１からの第１波長信号(電圧)を増幅する電圧増幅回路２２と、赤外線センサ２１からの第１波長信号のうち、所定の周波数帯域の成分のみを通過させるフィルタ回路２３と、フィルタ回路２３を通過した第１波長信号に対してＤＣレベル変換を施すＤＣレベル変換回路２４とにより構成されている。
【００７９】
また、この炎感知器の赤外線検出手段１−２は、炎特有の赤外線波長とは異なる波長(例えば、３．９μｍまたは５．０μｍの波長)の赤外線を第２波長信号（参照波長信号）として検出する赤外線センサ２５と、該赤外線センサ２５からの第２波長信号(電圧)を増幅する電圧増幅回路２６と、第２波長信号のうち、所定の周波数帯域の成分のみを通過させるフィルタ回路２７と、フィルタ回路２７を通過した第２波長信号に対してＤＣレベル変換を施すＤＣレベル変換回路２８とにより構成されている。
【００８０】
また、この炎感知器の炎判断手段２は、装置全体の制御を行なうマイクロプロセッサ等のＣＰＵ(中央処理装置)３０と、ＤＣレベル変換回路２４の出力電圧(第１波長信号のレベル)と閾値電圧（所定の閾値レベル）Ｖ_thとを比較するコンパレータ（比較回路）２９と、ＣＰＵ３０からの信号Ｐ１に基づいて火災信号を出力する火災出力回路３１とにより構成されている。
【００８１】
ここで、赤外線センサ２１，２５としては、防犯用センサとして広く使用されている焦電型センサ(焦電型素子)を使用することができる。この場合、焦電型センサは、入射光に対し微分の電荷出力を発生するものであり、従って、炎からの熱エネルギーの変化(揺らぎ)に比例した信号を出力するようになっている。また、これに関連させて、フィルタ回路２３，２７は、所定の周波数帯域の成分として、ゆらぎ周波数(炎のちらつき周波数)帯の成分のみを通過させるようになっている(バンドパスフィルタとして構成されている)。
【００８２】
すなわち、赤外線センサ２１，２５で検出される炎の揺らぎは数Ｈｚ〜十数Ｈｚ程度であり、炎特有の第１波長信号，第２波長信号を得るため、図１７の例では、この周波数帯域に透過特性をもつフィルタ回路２３，２７に赤外線センサ２１，２５からの第１波長信号，第２波長信号を通し、第１波長信号，第２波長信号のうち炎の揺らぎ成分のみを保存させた形で(すなわち、４．３〜４．４μｍ程度で捕えた炎の揺らぎ信号，３．９μｍまたは５．０μｍ程度で捕えた炎の揺らぎ信号のものにして)、ＣＰＵ３０に取り込ませるようになっている。
【００８３】
また、ＣＰＵ３０には、所定期間Ｔとして、例えば４．５秒を計時するタイマ機能が内蔵されているとする。また、ＣＰＵ３０には、ＤＣレベル変換回路２４，ＤＣレベル変換回路２８からの第１波長信号，第２波長信号のレベル(強度レベル)をデジタル信号に変換するＡ／Ｄ変換機能が備わっている。なお、このＡ／Ｄ変換機能は、第１波長信号，第２波長信号を所定の時間間隔Δｔ(例えば１０ｍ秒)でサンプリングしデジタル変換して取り込むようになっている。
【００８４】
また、コンパレータ（比較回路）２９は、ＤＣレベル変換回路２４から出力される赤外線検出信号（第１波長信号）のレベル(振幅電圧)が閾値電圧（所定の閾値レベル）Ｖ_th(例えば０．２Ｖ)に達するとき（すなわち、ＤＣレベル変換回路２４から出力される赤外線検出信号の立ち上がり時に）、立ち上がり割込信号を生成して、これをＣＰＵ３０の割り込み端子ＩＮＴ１に加え、また、ＤＣレベル変換回路２４から出力される赤外線検出信号（第１波長信号）のレベル(振幅電圧)が閾値電圧（所定の閾値レベル）Ｖ_th(例えば０．２Ｖ)から低下するとき（すなわち、ＤＣレベル変換回路２４から出力される赤外線検出信号の立ち下がり時に）、立ち下がり割込信号を生成して、これをＣＰＵ３０の割り込み端子ＩＮＴ１に加え、ＣＰＵ３０は、端子ＩＮＴ１への立ち上がり割込信号，立ち下がり割込信号の入力によって、第１波長信号の立ち上がり，立ち下がりを検出するようにしている。そして、ＣＰＵ３０は、第１波長信号の立ち上がりを検出した後、第１波長信号の立ち下がりを検出するまでの間、第１波長信号のピークを検出し、第１波長信号の立ち上がり，ピーク，立ち下がりの３つの時点における第２波長信号のレベルを取り込み、第１波長信号と第２波長信号との間に同期性があるか否かを判断するようになっている。
【００８５】
このような同期性の判断処理とともに、ＣＰＵ３０は、同期性があると判断された赤外線検出信号（第１波長信号，第２波長信号）についてのみ、前述したような炎判断処理を行なうようになっている。
【００８６】
炎判断処理は、具体的には、次のようにしてなされる。すなわち、ＣＰＵ３０は、所定期間Ｔ(例えば４．５秒間)にわたり第１波長信号のレベルを所定の時間間隔(サンプリング周期)Δｔ(例えば１０ｍ秒の時間間隔)でデジタルデータとして取り込み、所定期間Ｔにわたり時間間隔Δｔ(＝１０ｍ秒)ごとに取り込んだ第１波長信号のレベルデータ(デジタルデータ)のうち、例えば、所定の閾値レベルＶ_thを越えた第１波長信号のレベルについてだけピークを検知して、平均ピークレベルＭ_avgとピーク度数ＰＣＮＴとを例えば数１に従って算出する。また、上述のように赤外線検出手段１−１のＤＣレベル変換回路２４からの第１波長信号のピークを検知したとき、ＣＰＵ３０は、この抽出時点でのＤＣレベル変換回路２８からの第２波長信号のレベルを取り込み、上記所定期間Ｔにわたって取り込んだ第２波長信号のレベルの平均をとって平均補助レベルＳ_avgを算出する。そして、ＣＰＵ３０は、平均ピークレベルＭ_avgと平均補助レベルＳ_avgとの比率Ｍ_avg／Ｓ_avgとを求める。このようにして、ＣＰＵ３０は、所定期間(Ｔ＝４．５秒)にわたって赤外線検出信号の平均ピークレベルＭ_avg，ピーク度数ＰＣＮＴ，並びに、比率Ｍ_avg／Ｓ_avgの３つのパラメータを求め、これらの３つのパラメータに基づき炎の判断(例えば火災の程度の判断)を行なうことができる。
【００８７】
そして、ＣＰＵ３０は、同期性があると判断された第１波長信号と第２波長信号とにより上述の炎判断処理で火災であると判断したときに、火災出力回路３１に火災信号を出力するようになっている。
【００８８】
次に、このような構成の炎感知器の処理動作，主にＣＰＵ３０の処理動作について図１８のフローチャートを用いて説明する。図１８を参照すると、ステップＳ１では、ＣＰＵ３０は、先ず、ピークサーチ（ピーク探索）中であることを示すフラグＦＬＧを“０”に初期設定する。次いで、ステップＳ２では、ＣＰＵ３０は、コンパレータ２９から立ち上がり割込み入力があったか否かを判断する。この結果、立ち上がり割込みがあったときには、ステップＳ３に進み、ステップＳ３では、この時点が立ち上がり時であるとして、このときの第２波長信号（参照波長信号）のレベルを取得して記憶する。そして、ステップＳ４でフラグＦＬＧを“１”に設定し、再び、ステップＳ２に戻る。
【００８９】
ステップＳ２において、立ち上がり割込み入力がないと判断されると、ステップＳ５に進む。ステップＳ５では、立ち下がり割込み入力があったか否かを判断する。この結果、立ち下がり割込み入力がないと判断されると、ステップＳ６に進む。ステップＳ６では、フラグＦＬＧが“１”であるか否かを判断し、フラグＦＬＧが“１”でないときには（すなわち、“０”のときには）、再びステップＳ２に戻る。一方、ステップＳ６でフラグＦＬＧが“１”であるときには、ステップＳ７に進む。ステップＳ７では、第１波長信号（主波長信号）がピークとなったかを判断する。この結果、ピークでないときには、ステップＳ８に進み、ステップＳ８では、信号取得期間中か否かを判断する。ステップＳ８において、信号取得期間中と判断されると、再びステップＳ２に戻り、上述の処理を繰り返し行なう。そして、ステップＳ７において、第１波長信号（主波長信号）がピークとなったときには、ステップＳ９において、このときの第１波長信号（主波長信号）のレベルと第２波長信号（参照波長信号）のレベルを取得して記憶し、再びステップＳ２に戻り、上述の処理を繰り返し行なう。
【００９０】
そして、ステップＳ５において、立ち下がり割込み入力があったときには、ステップＳ１０に進み、ステップＳ１０では、この時点が立ち下がり時であるとして、このときの第２波長信号（参照波長信号）のレベルを取得して記憶する。そして、ステップＳ１１では、フラグＦＬＧを“０”に設定し、再びステップＳ２に戻り、上述の処理を繰り返し行なう。そして、ステップＳ８において、信号取得期間中でないと判断されると、ステップＳ１２では、ステップＳ３，Ｓ９，Ｓ１０において取得，記憶した立ち上がり時，ピーク時，立ち下がり時の第２波長信号（参照波長信号）のレベルに基づいて、第１波長信号（主波長信号）と第２波長信号（参照波長信号）とが同期しているか否かを判断する。この結果、同期しているときには、ステップＳ１３に進み、炎判断処理を行ない、炎と判断されたときには、ステップＳ１４において火災信号を出力する。これに対し、ステップＳ１２において、同期性がないと判断されると、炎判断処理を行なわずに、処理を終了する。
【００９１】
このように、本発明では、炎判断手段２は、各赤外線検出手段１−１〜１−ｎからの各赤外線検出信号において、非同期の信号成分をノイズと検知し、同期している信号成分だけに基づいて炎の判断を行なう。
【００９２】
また、上述の例では、基準となる赤外線検出手段１１−１と他の赤外線検出手段１１−２〜１１−ｎとは、検出波長が互いに異なるものであるとしたが、同期性の判断を行なうに関しては、基準となる赤外線検出手段１１−１と他の赤外線検出手段１１−２〜１１−ｎとは、検出波長が同じものを用いることもできる。換言すれば、基準となる赤外線検出手段１１−１と他の赤外線検出手段１１−２〜１１−ｎとに同じ種類のものを用いることもできる。
【００９３】
すなわち、図１の構成において、各赤外線検出手段１−１〜１−ｎとして、各赤外線検出信号に重畳するノイズに関して等価な構造のものを用いることができる。
【００９４】
このような構成の炎感知器の具体例は、ｎが２のとき、赤外線検出手段１−１と赤外線検出手段１−２とを等価な構造のものにし、また、ＣＰＵ３０に、炎判断手段２の機能、すなわち赤外線検出手段１−１，赤外線検出手段１−２からの各赤外線検出信号(第１波長信号，第２波長信号)において、非同期の信号成分をノイズと検知し、同期している信号成分だけに基づいて炎の判断を行なう機能をもたせたものとして構成できる。
【００９５】
より詳細には、図１７の赤外線検出手段１−１の赤外線センサ２１と赤外線検出手段１−２の赤外線センサ２５とに同一構造の焦電型素子を用い、また、赤外線検出手段１−１の電圧増幅回路２２，フィルタ回路２３，ＤＣレベル変換回路２４と赤外線検出手段１−２の電圧増幅回路２６，フィルタ回路２７，ＤＣレベル変換回路２８とに、それぞれ等価に設計された回路構造のものを用いたものとして、具体化できる。
【００９６】
一般に、ポップコーンノイズは、急激に温度が変化したとき、あるいは、部品(例えば赤外センサ)に機械的ストレスが加わったときに発生するものであるが、複数の部品(例えば、複数の赤外センサ)が同じ環境下に互いに近接して置かれている場合であっても、これらにポップコーンノイズが同時に発生する確率は極めて少ない。一方、互いに等価な構造の複数の赤外線検出手段が同じ環境下に互いに近接して置かれている場合、これらの赤外線検出手段によって検知される火災信号，非火災信号には、信号の同期性がある。
【００９７】
従って、図１７のような炎感知器において、赤外線検出手段１−１，赤外線検出手段１−２が等価な構造となっている場合、赤外線検出手段１−１，赤外線検出手段１−２によってそれぞれ検知された信号が互いに同期している場合、この信号は、火災信号または非火災信号であり、ノイズではないと判断でき、また、赤外線検出手段１−１，赤外線検出手段１−２によってそれぞれ検知された信号が非同期のものである場合、この信号はノイズであると判別できる。
【００９８】
具体的に、高温物体，低温物体(４００℃以下)，低油煙炎，高油煙炎から放射される赤外線については、赤外線検出手段１−１で検出される第１波長信号と、赤外線検出手段１−２で検出される第２波長信号とは、大小の差はあるものの、互いに同期している。
【００９９】
図１９は低油煙炎から放射された赤外線を、互いに等価な構造の赤外線検出手段１−１，赤外線検出手段１−２でそれぞれ検知した結果を示す図である。また、図２０は手でチョッピングされた低温物体(４００℃以下)から放射された赤外線を、互いに等価な構造の赤外線検出手段１−１，赤外線検出手段１−２でそれぞれ検知した結果を示す図である。
【０１００】
これらの結果からもわかるように、ノイズでない信号の場合には、これが火災によるものであるか、非火災によるものであるかを問わず、赤外線検出手段１−１で検出される第１波長信号と、赤外線検出手段１−２で検出される第２波長信号とは、互いに同期している。これに対し、例えば、赤外線検出手段１−１にポップコーンノイズが発生したときには、赤外線検出手段１−１で検出される第１波長信号のレベルと、赤外線検出手段１−２で検出される第２波長信号のレベルとは、同期しない。
【０１０１】
図２１，図２２は赤外線検出手段１−１にポップコーンノイズが発生するときの第１波長信号，第２波長信号をそれぞれ示す図である。なお、図２２には、図２１を時間軸方向に拡大し、第１波長信号のレベルのピークが検知されたときの第２波長信号のレベルの取り込みが示されている。図２１，図２２からもわかるように、第１波長信号と第２波長信号は、一般に、互いに非同期となる。
【０１０２】
この特徴を利用し、図１の炎感知器の炎判断手段２においては、各赤外線検出手段１−１〜１−ｎからの各赤外線検出信号が非同期の信号成分については、これをノイズと検知し、同期している信号成分だけに基づいて炎の判断がなされる。
【０１０３】
なお、上記説明では、図１において、ｎが“２”であり、炎感知器が図１７に示したような２波長式のものとなっているとしたが、図１において、ｎが例えば“３”である場合、すなわち３波長式のものであるような場合にも、各赤外線検出手段を等価な構造のものとし、各赤外線検出手段で検出された各赤外線検出信号の同期，非同期を調べることによって、ポップコーンノイズが発生しているか否かを検知することができる。さらに、ポップコーンノイズのみならず、ホワイトノイズや、電気ノイズ，電波ノイズ，振動ノイズ等、さらには、変則的な光源からの入射時をも検知でき、これらの影響を抑えて、あるいは受けずに、炎判断，火災判断を行なうことができる。すなわち、内部雑音(ポップコーンノイズ，ホワイトノイズ)や外部雑音(電気、電波、光、振動ノイズ)の影響を抑えて、あるいは、除去して、炎判断，火災判断を行なうことができる。
【０１０４】
また、図１における炎判断手段２としては、ランダム性を考慮した炎判断処理機能に、同期性判断機能を備えたものでも良いし、あるいは、ランダム性を考慮していない在来の炎判断処理機能に、同期性判断機能を備えたものであっても良い。但し、ランダム性を考慮した炎判断処理機能に、同期性判断機能を備えたものである場合には、炎判断を極めて高性能に行なうことが可能となる。
【０１０５】
また、上述の実施形態では、炎感知器内において、複数の赤外線検出信号間の同期性の判断を行ない、炎判断処理を行なって、炎か否かの判断結果を出力するものとなっているが、本発明は、１つの炎感知器の構成に限定されず、任意の炎監視システムに適用可能である。すなわち、本発明の適用対象としては、単体の炎感知器に限らず、任意の炎監視システム(オン・オフ型監視システム，アナログ型監視システム)に適用可能であり、アナログ型監視システムに適用するとき、炎判断処理を例えば受信機側で行なわせることも可能である。
【０１０６】
すなわち、図１の構成例では、炎感知器自体に炎判断手段２が設けられ、同期性判断，炎判断，火災判断をこの炎感知器内で行なうようになっている(すなわち、オン・オフ型感知器として構成されている)が、炎感知器をアナログ型感知器として構成することもできる。図２３は炎感知器が受信機からアドレスポーリングされるアナログ型感知器として構成されている場合の一例を示す図であり、この場合、炎感知器の信号処理手段において、同期性の判断等を行ない、信号から非同期の部分をノイズとして除去した上で、この信号を受信機に返送し、受信機において炎判断処理を行なわせることができる。
【０１０７】
【発明の効果】
以上に説明したように、請求項１乃至請求項８記載の発明によれば、炎特有の赤外線波長の光を赤外線検出信号としてそれぞれ検出する複数の赤外線検出手段と、複数の赤外線検出手段でそれぞれ検出された赤外線検出信号から炎特有の特徴量を抽出し、炎特有の特徴量に基づき炎判断を行なう炎判断手段とを有し、前記炎判断手段は、前記複数の赤外線検出手段でそれぞれ検出された赤外線検出信号の同期性を判断し、同期性があると判断された赤外線検出信号に基づいて炎判断を行なうようになっているので、変則的な光源の入射や連続的なノイズの影響などを排除でき、炎であるか否かの判断を正確に行なうことができる。
【０１０８】
特に、請求項１乃至請求項８記載の発明では、前記複数の赤外線検出手段のうちの１つを、基準となる赤外線検出手段とするときに、前記炎判断手段は、基準となる赤外線検出手段で検出された赤外線検出信号が所定の閾値レベルよりも大きくなったかを判断し、基準となる赤外線検出手段で検出された赤外線検出信号が所定の閾値レベルよりも大きくなったときに、基準となる赤外線検出手段からの赤外線検出信号の所定の閾値レベルを基準にした立ち上がり時，ピーク時，立ち下がり時における他の赤外線検出手段からの赤外線検出信号の信号レベルに基づいて、赤外線検出信号の同期性を判断するようになっており、立ち上がり，ピーク，立ち下がり時の３点のみの信号で簡単に同期性を判断できるため、処理の高速化を図ることができる。
【０１０９】
また、請求項５記載の発明では、請求項１乃至請求項４のいずれか一項に記載の炎感知器において、前記炎判断手段は、基準となる赤外線検出手段からの赤外線検出信号の所定の閾値レベルを基準にした立ち上がりと立ち下がりを検出するための比較回路と、同期性の判断を含めた炎判断処理を行なうＣＰＵとを有し、基準となる赤外線検出手段からの赤外線検出信号の所定の閾値レベルを基準とした立ち上がりと立ち下がりが比較回路で検出されたとき、比較回路で検出された立ち上がりと立ち下がりは、ＣＰＵに割り込みとして通知されるようになっているので（すなわち、比較回路を用い、基準となる主波長を検出するセンサからの信号の立ち上がりと立ち下がりの検出をＣＰＵの割り込み機能を利用することで）、ＣＰＵの負担を大幅に軽減させることができる。
【図面の簡単な説明】
【図１】本発明に係る炎感知器の構成例を示す図である。
【図２】一般的な炎のスペクトル強度を示す図である。
【図３】火災時の炎から放射される赤外線の強度レベルの時間的変化の一例を示す図である。
【図４】火災時の炎からの赤外線を赤外線センサ(焦電型センサ)で検出した赤外線検出信号のピークレベルの変化の一例を示す図である。
【図５】低温熱源(４００℃以下)からの赤外線を赤外線センサ(焦電型センサ)で検出した赤外線検出信号のピークレベルの変化の一例を示す図である。
【図６】図４の赤外線検出信号（火災時の炎による信号）と図５の赤外線検出信号（低温熱源による信号）とのそれぞれの場合におけるピークレベルの度数分布（ヒストグラム）の作成例を示す図である。
【図７】図１の炎感知器において、赤外線検出手段１−２に、赤外線検出手段１−１によって検出される赤外線波長の近傍の波長の赤外線を検出するものが用いられる場合の第２波長信号のレベルの取り込みを説明するための図である。
【図８】火災時における第１波長信号，第２波長信号の一例を示す図である。
【図９】非火災時(例えば、ハロゲン灯，熱源ノイズなどが加わった場合)における第１波長信号，第２波長信号の一例を示す図である。
【図１０】低油煙(ノルマル−ヘプタン)炎を発生させたときの第１波長信号(４．３〜４．４μｍ程度)と第２波長信号(例えば５．０μｍ)のレベルの測定結果(時間的変化)を示す図である。
【図１１】高油煙(ガソリン)炎を発生させたときの第１波長信号(４．３〜４．４μｍ程度)と第２波長信号(例えば５．０μｍ)のレベルの測定結果(時間的変化)を示す図である。
【図１２】高温物体(４００℃以上)をチョッパしたときの第１波長信号(４．３〜４．４μｍ程度)と第２波長信号(例えば５．０μｍ)のレベルの測定結果(時間的変化)を示す図である。
【図１３】本発明における赤外線検出信号の同期性の判断の具体例を示す図である。
【図１４】炎が発生したときの基準となる赤外線検出手段からの赤外線検出信号（主波長信号）と他の赤外線検出手段からの赤外線検出信号（参照波長信号）との一例を示す図である。
【図１５】ノイズが混入したときの基準となる赤外線検出手段からの赤外線検出信号（主波長信号）と他の赤外線検出手段からの赤外線検出信号（参照波長信号）との一例を示す図である。
【図１６】変則的な光源が入力したときの基準となる赤外線検出手段からの赤外線検出信号（主波長信号）と他の赤外線検出手段からの赤外線検出信号（参照波長信号）との一例を示す図である。
【図１７】本発明に係る炎感知器の具体例を示す図である。
【図１８】本発明の炎感知器の処理動作を説明するためのフローチャートである。
【図１９】低油煙炎から放射された赤外線を、互いに等価な構造の赤外線検出手段１−１，赤外線検出手段１−２でそれぞれ検知した結果を示す図である。
【図２０】手でチョッピングされた低温物体(４００℃以下)から放射された赤外線を、互いに等価な構造の赤外線検出手段１−１，赤外線検出手段１−２でそれぞれ検知した結果を示す図である。
【図２１】赤外線検出手段１−１にポップコーンノイズが発生するときの第１波長信号，第２波長信号を示す図である。
【図２２】赤外線検出手段１−１にポップコーンノイズが発生するときの第１波長信号，第２波長信号を示す図である。
【図２３】炎感知器がアナロク型感知器として構成されている場合の一例を示す図である。
【符号の説明】
１−１〜１−ｎ 赤外線検出手段
２ 炎判断手段
２１，２５ 赤外線センサ
２２，２６ 電圧増幅回路
２３，２７ フィルタ回路
２４，２８ ＤＣレベル変換回路
２９ コンパレータ
３０ ＣＰＵ
３１ 火災出力回路[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a flame detector and a flame detection method for detecting a flame generated during a fire or the like.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, there is known a flame detector that performs flame determination (fire determination) by detecting infrared light having a predetermined wavelength emitted from a flame.
[0003]
[Problems to be solved by the invention]
However, the conventional flame detector as described above has a problem that, for example, it is not a flame but it is erroneously detected as a flame due to the influence of an irregular light source or continuous noise. It was.
[0004]
More specifically, in general, a flame detector senses the characteristics of infrared light emitted from a flame and makes a fire judgment. However, a flame detector is required to have high sensitivity in the same manner as a general fire detector. And malfunction is not allowed. In particular, an infrared flame detector using a pyroelectric element as a sensor is susceptible to various noises because it uses a high-amplification amplifier to detect flickering of flame. In particular, popcorn noise occurs when electronic components such as operational amplifiers (operational amplifiers), capacitors, and pyroelectric sensors are subjected to rapid temperature changes and mechanical stress. Among these, the pyroelectric type is the first stage of the amplifier. Popcorn noise generated by the element has the greatest influence.
[0005]
An object of the present invention is to provide a flame detector and a flame detection method capable of avoiding malfunctions caused by irregular light source incidence or continuous noise effects.
[0006]
[Means for Solving the Problems]
In order to achieve the above-mentioned object, the invention according to claim 1 is characterized in that a plurality of infrared detecting means for detecting light having an infrared wavelength peculiar to a flame as an infrared detection signal, and infrared detections respectively detected by the plurality of infrared detecting means. Flame determination means for extracting a flame-specific characteristic amount from the signal and making a flame determination based on the flame-specific characteristic amount; When one of the plurality of infrared detection means is used as a reference infrared detection means, the flame determination means is configured such that the infrared detection signal detected by the reference infrared detection means is lower than a predetermined threshold level. When the infrared detection signal detected by the reference infrared detection means becomes larger than a predetermined threshold level, the predetermined threshold level of the infrared detection signal from the reference infrared detection means is determined. Based on the signal level of the infrared detection signal from other infrared detection means at the time of rising, peaking, falling at the reference, It is characterized in that the synchronism of the infrared detection signal is determined, and the flame determination is performed using the infrared detection signal determined to be synchronic.
[0008]
Also, Claim 2 The described invention Claim 1 The flame detector described above is characterized in that data is taken in at predetermined time intervals over a predetermined time from the time when the reference infrared detection level exceeds a predetermined threshold level, and the maximum value is peaked.
[0009]
Also, Claim 3 The described invention Claim 1 Or Claim 2 In the flame detector described above, the flame determination unit may detect other infrared rays at one rising, peaking, or falling time with reference to a predetermined threshold level of the infrared detection signal from the infrared detection unit serving as a reference. The signal levels of the infrared detection signals from the means are respectively a _r , P _r , B _r When the predetermined number or more of infrared detecting means among other infrared detecting means, _r > A _r , P _r > B _r When the above condition is satisfied, the infrared detection signal is determined to be synchronized.
[0010]
Also, Claim 4 The described invention Claim 1 Thru Claim 3 The flame detector according to any one of the above, wherein the flame determination means has one rising, peak, falling edge based on a predetermined threshold level of an infrared detection signal from an infrared detection means serving as a reference. The signal levels of infrared detection signals from other infrared detection means at the time _r , P _r , B _r In other infrared detection means, p _r > A _r , P _r > B _r The infrared detection signal is determined to be synchronized when the above condition is satisfied more than a predetermined number of rising, peaking, and falling occurrences occurring within a predetermined period. It is said.
[0011]
Also, Claim 5 The described invention Claim 1 Or Claim 2 In the flame detector described above, the flame determination means may detect other infrared rays at one rising, peaking, and falling times based on a predetermined threshold level of an infrared detection signal from the infrared detecting means serving as a reference. The signal levels of the infrared detection signals from the means are respectively a _r , P _r , B _r When the predetermined number or more of infrared detecting means among other infrared detecting means, _r > A _r , P _r > B _r The infrared detection signal is determined to be synchronized when the above condition is satisfied more than a predetermined number of rising, peaking, and falling occurrences occurring within a predetermined period. It is said.
[0012]
Also, Claim 6 The described invention Claim 1 Thru Claim 5 The flame detector according to any one of the preceding claims, wherein the flame determination means detects a rising edge and a falling edge with reference to a predetermined threshold level of an infrared detection signal from an infrared detection means as a reference. And a CPU that performs flame determination processing including determination of synchronism, and the comparison circuit detects rising and falling with reference to a predetermined threshold level of the infrared detection signal from the infrared detection means serving as a reference. The rising and falling edges detected by the comparison circuit are notified to the CPU as interrupts.
[0013]
Also, Claim 7 The invention described in claims 1 to Claim 6 In the flame detector according to any one of the above, a pyroelectric element is used as an infrared sensor in at least one of the plurality of infrared detection means.
[0014]
Also, Claim 8 In the described invention, in a plurality of infrared detection means, light having an infrared wavelength peculiar to a flame is detected as an infrared detection signal, and feature quantities specific to the flame are extracted from the infrared detection signals respectively detected by the plurality of infrared detection means. A flame detection method that performs flame judgment based on flame-specific features, When one of the plurality of infrared detection means is used as a reference infrared detection means, the flame determination means is configured such that the infrared detection signal detected by the reference infrared detection means is lower than a predetermined threshold level. When the infrared detection signal detected by the reference infrared detection means becomes larger than a predetermined threshold level, the predetermined threshold level of the infrared detection signal from the reference infrared detection means is determined. Based on the signal level of the infrared detection signal from other infrared detection means at the time of rising, peaking, falling at the reference, It is characterized in that the synchronism of the infrared detection signal is determined, and the flame determination is performed using the infrared detection signal determined to be synchronic.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing a configuration example of a flame detector according to the present invention. Referring to FIG. 1, the flame detector includes a plurality of infrared detection units 1-1 to 1-n that detect light having an infrared wavelength peculiar to a flame as an infrared detection signal, and a plurality of infrared detection units 1-1 to 1-1. Flame determining means 2 that extracts flame-specific feature values from the infrared detection signals respectively detected at 1-n and performs flame determination based on the flame-specific feature values.
[0016]
Fig. 2 shows the spectral intensities of common flames, such as gas flames with relatively little oil smoke, wood and paper combustion flames, or flames at the time of the initial fire. Flames that occur at times etc. are at an infrared wavelength of about 4.3 to 4.4 μm, and CO ₂ It has the largest spectral intensity due to resonance radiation (a clear peak spectrum appears). Note that FIG. 2 shows the spectrum intensity of a flame with relatively little oil smoke, but even in the case of a flame with much oil smoke such as gasoline, the largest spectrum is obtained at an infrared wavelength of about 4.3 to 4.4 μm. Has strength. Accordingly, at least one infrared detection means among the plurality of infrared detection means 1-1 to 1-n has an infrared wavelength specific to the flame to be detected, that is, as a main feature of the flame, for example, 4.3. An infrared sensor that detects infrared light having a wavelength of about 4.4 μm can be used.
[0017]
Now, in order to simplify the description, a general flame determination process in the flame determination unit 2 will be described assuming that only one infrared detection unit is provided. FIG. 3 is a diagram showing an example of a temporal change in the level of infrared rays radiated from the flame at the time of fire (intensity level of the infrared detection signal output from the infrared sensor when an infrared sensor is used as one infrared detection means). It is. Referring to FIG. 3, the intensity level of the infrared rays emitted from the flame at the time of fire is a predetermined threshold level V _th The intensity level of infrared rays radiated from a flame at the time of a fire has a temporal fluctuation of about 8 Hz, that is, a flicker frequency characteristic, for example.
[0018]
Focusing on this, in the flame determination means 2, the infrared intensity level (for example, the level of the infrared detection signal from one infrared detection means) is a predetermined threshold level V. _th When reaching (for example, 0.2 V), it is determined that there is a possibility of flame, and from this point, the level of the infrared detection signal is captured at a predetermined sampling interval.
[0019]
Then, the flame determination means 2 determines that the level of the infrared detection signal is a predetermined threshold level V _th In order to investigate how much the infrared detection signal from the point of time reaches the characteristic (flame specific intensity level characteristic, temporal fluctuation (flickering characteristic)) of the flame, a predetermined threshold level V _th The peak of the level of the infrared detection signal (for example, the predetermined threshold level V out of the levels of the infrared detection signal over the predetermined period T over a predetermined period T from _th The average peak level M is obtained by averaging the levels of the peaks over a predetermined period T. _avg Further, the number of peaks over a predetermined period T is obtained as the peak frequency (that is, the number of fluctuations) PCNT.
[0020]
More specifically, as shown in FIG. 3, the level of the infrared detection signal is a predetermined threshold level V. _th The infrared detection signal is captured (sampled) as data at a predetermined time interval Δt (for example, 10 msec) from a point in time exceeding a predetermined period T. Therefore, in the predetermined period T, about T / Δt pieces of data are captured. (Sampling), for example, the flame determination means 2 is set to a predetermined threshold level V out of about T / Δt data (infrared detection signal level) acquired in this way. _th Focusing only on those that exceed the threshold level V _th An increase / decrease in the level of the infrared detection signal exceeding the threshold is detected, and the largest portion (mountain portion) is detected as a level peak. That is, when the level of the infrared detection signal is captured as digital data at every predetermined time interval Δt, for example, among the captured level data (digital data) of the infrared detection signal, for example, a predetermined threshold level V _th Compare the previous data with the current data for the level of infrared detection signal that exceeds, and determine whether the trend is an increase or decrease, and the point of change from increase to decrease (when there are multiple such points, The point with the highest level) can be extracted as the peak point. In other words, data is taken in at predetermined time intervals for a predetermined time from the time when the reference infrared detection level exceeds the predetermined threshold level, and the maximum value is taken as a peak.
[0021]
Then, in the predetermined period T, the number of level peaks is detected as, for example, m, and the level of each peak is determined as M. ₁ , M ₂ , ..., M _m Mean peak level M _avg The peak frequency PCNT is obtained by the following equation, for example.
[0022]
[Expression 1]

[0023]
The average peak level M determined in this way in the flame determination means 2 _avg Is used to determine how much the level of the infrared detection signal has an infrared intensity level characteristic peculiar to a flame at the time of a fire, and the peak frequency PCNT obtained as described above relates to the infrared intensity level. Because it reflects the fluctuation frequency, it is used to determine how much the temporal change in the level of the infrared detection signal has the characteristic of the temporal fluctuation (flicker) characteristic of flames in a fire. .
[0024]
In such flame determination processing, the average peak level M _avg Is a good reflection of how much the infrared detection signal has the main characteristics peculiar to flames at the time of fire, and the peak frequency PCNT reflects the fluctuation frequency related to the infrared intensity level Therefore, the flame determination means 2 basically has the average peak level M _avg And / or based on the peak frequency PCNT, it can be determined whether or not it is a flame.
[0025]
Further, when it is difficult to determine whether or not the flame is determined by such a flame determination process, the predetermined time T is set and changed until the determination of the flame can be performed with some reliability. It is also possible to make a flame judgment. Thereby, basic flame judgment can be performed more reliably.
[0026]
In the flame determination process as described above, for example, in the case of a gas flame with relatively little oil smoke, a combustion flame of wood or paper, or a flame at the time of initial fire, the average peak level M obtained as described above is used. _avg Becomes a sufficiently large level, and the peak frequency PCNT also becomes large, so that it can be easily determined as a flame.
[0027]
However, even when such an excellent and reliable basic flame judgment is made, when there is a large amount of smoke (for example, in the combustion of gasoline, heavy oil, petrochemicals, rubber products, etc.) (In a fire with a large combustion scale and a lack of oxygen, the generation of carbon monoxide CO is increased and a flame with a lot of oily smoke) or when noise from other heat sources is mixed, for example, The average peak level M obtained as described above _avg May be at a subtle level, i.e., an ambiguous level, in determining whether or not it is a flame, and at this time, it may not be possible to reliably determine whether or not it is a flame.
[0028]
The inventor of the present application investigated the characteristics of flame combustion so that the flame can be judged correctly and reliably even in the case of a flame with a lot of oily smoke. As a result, in the case of a combustion flame, since it is affected by the airflow (oxygen supply) from the surroundings, fluctuations of several to several tens of Hz are generated in the emitted light. When this emitted light is detected using a pyroelectric sensor with characteristics, the fluctuation of the infrared detection signal from the pyroelectric sensor is strong and weak, and the level (peak voltage value) of each peak of the infrared detection signal is examined. , It was found that the level of each peak was random (it was found to be randomly distributed (scattered)).
[0029]
Therefore, when a histogram is taken for each peak level of the infrared detection signal within a predetermined time, it has been found that the frequencies are not concentrated in a specific rank (section) and the frequencies are distributed in a plurality of sections.
[0030]
A histogram for each peak level of the infrared detection signal within a predetermined time can be created as follows, for example.
[0031]
That is, the level of the infrared detection signal is divided into a plurality of sections, and the number of sections at that time is N (for example, “7”). Also, the number of data (number of peaks) within a predetermined time T is m (for example, “35”), and the maximum value among M (35) data (peak level) is M. _max , The minimum value is M _min , The width W of each section of the histogram is determined by the following equation.
[0032]
[Expression 2]
W = (M _max -M _min ) / N
[0033]
In this way, after setting the number N of sections (for example, “7”) and the width W of each section, m (35) data (m peak levels) are selected from the N sections. The frequency distribution of each section can be created by applying to each of the predetermined sections.
[0034]
FIG. 4 shows an example of a change in the peak level of an infrared detection signal obtained by detecting an infrared ray from a flame at the time of a fire with an infrared sensor (pyroelectric sensor), and FIG. 5 shows a low-temperature heat source (400 ° C. An example of a change in the peak level of an infrared detection signal in which infrared rays from the following are detected by an infrared sensor (pyroelectric sensor) is shown. In addition, FIG. 6 shows the creation of a frequency distribution (histogram) of peak levels in each case of the infrared detection signal of FIG. 4 (signal due to flame at the time of fire) and the infrared detection signal of FIG. 5 (signal by a low temperature heat source). An example is shown.
[0035]
In the example of FIGS. 4 and 5, the number m of peak levels within a predetermined time T is “50”, and among the 50 peak levels, the maximum peak level is “2450 mV” and the minimum peak level. 6 is divided into seven sections, the peak level is divided into seven sections, the width W of each section is set to "350 mV", and the 50 peak levels are divided into seven sections. It is created as a frequency distribution when assigned to the corresponding section.
[0036]
As can be seen from FIG. 6, the frequency distribution due to the flame at the time of the fire is almost equal (random) in each section, whereas the frequency distribution due to the low-temperature heat source is concentrated in a specific section (not random) I understand)
[0037]
Therefore, in the basic flame determination process as described above, the average peak level M _avg And / or when the peak frequency PCNT is ambiguous in judging whether or not it is a flame, a histogram as shown in FIG. 6 is created to determine whether or not each peak level has randomness. Thus, it is possible to reliably determine whether or not the flame is detected. In the above example, each peak level of the infrared detection signal due to the flame at the time of fire has sufficient randomness, so it can be reliably determined that it is flame, and each peak level of the infrared detection signal from the low temperature heat source is random. Since it does not have, it can be judged reliably that it is not a flame.
[0038]
More specifically, when the number of data m within a predetermined time is “35” and the number of sections N is “7”, the frequency of each section is expressed as a rank as shown in the following table (in the example of the following table, 4 ranks R ₁ , R ₂ , R _Three , R _Four ).
[0039]
[Table 1]

[0040]
In Table 1, the average frequency is the total number of data / number of sections, and is “5” in the above example. When the frequency of each section is ranked in this way, the flame degree (an index indicating the probability of being a flame) can be obtained, for example, as shown in the following table, based on the number of sections belonging to each rank.
[0041]
[Table 2]

[0042]
In Table 2, rank R ₁ Number U of sections belonging to ₁ Is not shown, but for example (U ₂ ≦ 1, U _Three = 0, U _Four = 0), U ₁ Becomes "6", for example (U ₂ ≦ 2, U _Three = 0, U _Four = 0), U ₁ Becomes, for example, “5”.
[0043]
As can be seen from Table 2, (U ₂ ≦ 1, U _Three = 0, U _Four = 0), U ₁ Therefore, the peak level frequency is almost uniformly distributed (randomly distributed) in each section. At this time, the probability of being a flame, that is, the flame level is calculated as “1”. Can do.
[0044]
In the flame detector having such a configuration, for example, the average peak level M _avg And when the peak frequency (number of peaks) PCNT is sufficiently large (for example, the average peak level M _avg Is sufficiently larger than 0.2 V, and the peak frequency (the number of peaks) PCNT is 30 or more), it is possible to determine that there is a fire and output that fact.
[0045]
Meanwhile, average peak level M _avg Or the peak frequency (number of peaks) PCNT or both are not sufficiently large, and it is difficult to determine whether or not a fire is caused by this alone (for example, average peak level M _avg In the case where the peak frequency (the number of peaks) is about 10 to 20), for example, as shown in Table 1 and Table 2, When the flame degree is “1” (ie, when the randomness is sufficiently large), it is judged as flame, and when the flame degree is “0”, it is judged as not flame. can do.
[0046]
In this way, when the flame determination means 2 extracts a flame-specific feature amount from the infrared detection signal and makes a flame determination based on the flame-specific feature amount, if the flame determination becomes ambiguous, extraction is performed. Even if the judgment of whether or not the flame is ambiguous in the conventional flame judgment processing by performing the flame judgment in consideration of the randomness of the characteristic amount peculiar to the flame, This makes it possible to make a correct and reliable determination.
[0047]
In the above-described example, it has been described that only one infrared detection unit is provided, but the above-described flame determination process is also performed when a plurality of infrared detection units 1-1 to 1-n as illustrated in FIG. 1 are provided. The principle of can be extended. For example, the principle of the flame determination process described above can be extended to a flame detector in which n is 2, n is 3, or n is larger than 3.
[0048]
In the following description, it is assumed that n is 2 in the configuration of the flame detector of FIG. That is, the flame detector will be described as having two infrared detection means, for example, an infrared detection means 1-1 and an infrared detection means 1-2. Here, the infrared detection means 1-1 uses an infrared ray having a wavelength specific to a flame as a first wavelength signal (main wavelength signal), and the infrared detection means 1-2 includes an infrared detection means 1. -1 can be used to detect light having a wavelength different from the infrared wavelength detected by -1 as a second wavelength signal (reference wavelength signal) (can be configured as a two-wavelength flame detector). In this case, the flame determination means 2 makes a flame determination based on the first wavelength signal from the infrared detection means 1-1 and the second wavelength signal from the infrared detection means 1-2, and the flame determination is ambiguous. Even in such a case, when the level of each peak of the first wavelength signal is randomly dispersed, it can be determined that the flame is present.
[0049]
Specifically, an infrared sensor (for example, a pyroelectric sensor) that detects infrared rays having a wavelength of about 4.3 to 4.4 μm, for example, can be used as the infrared detection unit 1-1.
[0050]
In addition, the infrared detection means 1-2 uses infrared light having a wavelength near the infrared wavelength specific to the flame as light having a wavelength different from the infrared wavelength specific to the flame (about 4.3 to 4.4 μm), for example, 3. An infrared sensor (for example, a pyroelectric sensor) that detects infrared rays having a wavelength of about 9 μm or about 5.0 μm can be used. That is, at a wavelength of about 3.9 μm or 5.0 μm, as can be seen from FIG. 2, the spectrum intensity of the flame is small (becomes a minimum value). Therefore, in this case, the infrared detecting means 1-2 is 3.9 μm. An infrared level at a wavelength of about or about 5.0 μm can be detected as an auxiliary feature of the flame.
[0051]
In other words, in this case, the infrared detection unit 1-1 functions as a reference infrared detection unit, and the infrared detection unit 1-2 functions as an auxiliary (reference) infrared detection unit. ing.
[0052]
In the case where an infrared detecting unit 1-2 that detects infrared rays having a wavelength near the infrared wavelength detected by the infrared detecting unit 1-1 is used as the infrared detecting unit 1-2, as shown in FIG. Infrared detection signal (second wavelength signal (reference wavelength signal (reference wavelength signal)) detected by the infrared detection means 1-2 at the peak level of the infrared detection signal (first wavelength signal (main wavelength signal)) detected by the infrared detection means 1-1. Level S of auxiliary detection signal))) ₁ , S ₂ , ..., S _m , And the level S of the second wavelength signal from the infrared detection means 1-2 thus captured over a predetermined period T ₁ , S ₂ , ..., S _m Mean value S _avg For example, the average peak level M of the first wavelength signal from the infrared detecting means 1-1. _avg And the average auxiliary level S of the second wavelength signal from the infrared detecting means 1-2 _avg Ratio to (M _avg / S _avg ) And the peak frequency (number of peaks) PCNT (= m) over a predetermined period T of the first wavelength signal, a basic flame determination is performed.
[0053]
In FIG. 7, the predetermined period T can be variably set according to the basic flame determination result.
[0054]
For example, the average peak level M obtained over a predetermined time T in the process of FIG. _avg And average support level S _avg Ratio M _avg / S _avg When it is difficult to determine the flame based on the above, the predetermined time T is set and the average peak level M is set over the predetermined time T thus changed. _avg And average support level S _avg Ratio M _avg / S _avg Is calculated again, and based on this, the flame is determined again. The predetermined time T is changed until the flame can be determined with a certain degree of reliability. Can be done.
[0055]
FIG. 8 is a diagram illustrating an example of a main wavelength signal (first wavelength signal) and a reference wavelength signal (second wavelength signal) at the time of fire, and FIG. 9 is a non-fire time (for example, halogen lamp, heat source noise, etc.) FIG. 6 is a diagram illustrating an example of a main wavelength signal (first wavelength signal) and a reference wavelength signal (second wavelength signal) in the case of (a) added. In the examples of FIGS. 8 and 9, an infrared sensor (pyroelectric sensor) that detects infrared rays having a wavelength different from the infrared rays detected by the infrared detector 1-1 is used as the infrared detector 1-2. It is said that
[0056]
First, in the event of a fire, the infrared intensity level (the level of the first wavelength signal) peculiar to the flame at the time of the fire is increased as shown in FIG. 8A, whereas the level of the second wavelength signal is that of FIG. As shown in (b), the difference is not so large. Therefore, the average peak level M obtained over a predetermined period T (for example, 4.5 seconds) _avg And average support level S _avg Ratio M _avg / S _avg Will be big enough. As a result, in this case, as shown in FIGS. 8C and 8D, it is determined that the flame is in the predetermined period T (= 4.5 seconds) without updating the predetermined period T. The result can be output as a flame detection signal.
[0057]
In addition, when a halogen lamp, heat source noise, or the like occurs during a non-fire, the intensity level of the infrared wavelength peculiar to the flame during fire (the level of the first wavelength signal) is as shown in FIG. At this time, the intensity level of the second wavelength signal also increases as shown in FIG. 9B, and therefore the average peak level M obtained over a predetermined period T (for example, 4.5 seconds). _avg And average support level S _avg Ratio M _avg / S _avg It becomes difficult to make a flame judgment based on this. Accordingly, in this case, as shown in FIG. 9C, the predetermined time T is set and changed until the flame determination can be performed with a certain degree of reliability. The predetermined period T to be performed is further extended. In the example of FIG. 9C, the setting is further extended by 9 seconds, the setting is changed to T = 13.5 seconds in total, and the determination of the flame is performed again. In the example of FIG. 9, when it is determined whether or not the flame is set at a predetermined period T of 13.5 seconds, it is determined that it is not a flame. Therefore, in this case, as shown in FIG. No signal is output.
[0058]
In this way, when the reliability is considered to be low even in basic flame judgment processing, the averaging time is increased, so that the integration effect enables more reliable flame judgment and reduces the risk of false alarms. Less.
[0059]
However, even when such a reliable and excellent basic flame judgment is made, as described above, when there is a lot of oily smoke or when noise from other heat sources is mixed in For example, the ratio M obtained as described above _avg / S _avg However, there may be a subtle level, i.e., an ambiguous level, in determining whether or not it is a flame. In this case, it may not be possible to reliably determine whether or not it is a flame.
[0060]
FIG. 10 shows the measurement results of the levels of the first wavelength signal (about 4.3 to 4.4 μm) and the second wavelength signal (for example, 5.0 μm) when a low oil smoke (normal-heptane) flame is generated ( FIG. 11 shows a first wavelength signal (approximately 4.3 to 4.4 μm) and a second wavelength signal (for example, 5.4) when a high oil smoke (gasoline) flame is generated. 0 μm) level measurement results (temporal changes) are shown, and FIG. 12 shows the first wavelength signal (approximately 4.3 to 4.4 μm) when a high temperature object (400 ° C. or higher) is chopped. The measurement result (temporal change) of the level of the second wavelength signal (for example, 5.0 μm) is shown.
[0061]
In the case of a low oil smoke (normal-heptane) flame, as can be seen from FIG. 10, the level of the first wavelength signal is sufficiently larger than the level of the second wavelength signal. it can. On the other hand, in the case of thermal spray from a high oil smoke (gasoline) flame or a hot object, the level of the first wavelength signal is sufficiently larger than the level of the second wavelength signal, as can be seen from FIGS. Since it is not a big thing, the flame judgment based on this produces ambiguity.
[0062]
Therefore, when ambiguity occurs in the flame determination as shown in FIGS. 11 and 12, the flame determination means 2 further determines whether the levels of the respective peaks of the first wavelength signal are randomly distributed, as described above. It is determined whether or not it is a flame when the level of each peak is randomly dispersed.
[0063]
As described above, when the flame detector is of the two-wavelength type as described above, or when the flame detector is of the three-wavelength type, the flame judgment is ambiguous. The determination unit 2 can further determine whether or not it is a flame by examining whether or not the level of each peak of the first wavelength signal is randomly dispersed.
[0064]
In the above-described example, the average peak level and the peak frequency of the infrared detection signal are used as the feature amount (flame-specific feature amount) at the time of determining the flame. At this time, the level of the infrared detection signal is Predetermined threshold level V _th Out of the levels of infrared detection signals obtained over a predetermined period T from the time when the threshold value is exceeded, a predetermined threshold level V _th Focusing only on those exceeding the predetermined threshold level V _th The peak is detected only for the infrared detection signal that exceeds, and the average peak level and peak frequency are obtained. Can be used. For example, only the average peak level of the infrared detection signal can be used as a feature quantity specific to the flame, or only the peak frequency of the infrared detection signal can be used as a feature quantity specific to the flame, or the infrared detection signal Is a predetermined threshold level V _th The level of the infrared detection signal obtained over a predetermined period T from a point in time exceeding the predetermined threshold level V _th It is also possible to detect level peaks including those that do not exceed, determine the average peak level PAV and / or peak frequency, and use this as a characteristic quantity specific to flames. Alternatively, not only the average peak level and the peak frequency, but also if a flame feature is extracted, it can be used as a feature amount unique to the flame.
[0065]
In the above-described example, the peak level rank is created from the peak level histogram to obtain the randomness, but the randomness can be obtained by other methods. For example, the standard deviation of the peak level is obtained, and the variation (randomness) of the peak level is determined as the first condition depending on the magnitude of the standard deviation, and the period of the peak level in a predetermined period is obtained, and the standard for the period is obtained. The deviation is obtained, the variation of the period (randomness) is judged as the second condition based on the standard deviation, and finally the randomness as the flame when both the first and second conditions are satisfied. It can also be determined that
[0066]
By the way, as shown in FIG. 1, a flame-specific feature value is extracted from the infrared detection signals respectively detected by the plurality of infrared detection means 1-1 to 1-n, and a flame is determined based on the flame-specific feature value. In a sensor, for example, when there is an irregular light source incidence (incident light source) on only the reference infrared detection means (eg 1-1) or when continuous noise is added, Even when the flame determination means 2 of the flame detector makes a flame determination as described above, a false alarm may occur.
[0067]
The inventor of the present application, for example, when there is an irregular light source incidence (incidence of a fluctuating light source) only in the reference infrared detection means (for example, 1-1) or when continuous noise is added, While there is no synchronization between the infrared detection signal (main wavelength signal) from the infrared detection means 1-1 and the infrared detection signals (reference wavelength signals) from the other infrared detection means 1-2 to 1-n. When the plurality of infrared detection means 1-1 to 1-n detect a flame, the plurality of infrared detection means 1-1 to 1-n output substantially synchronized signals. Completed the invention.
[0068]
That is, in the present invention, the flame determination unit 2 determines the synchronism of the infrared detection signals detected by the plurality of infrared detection units 1-1 to 1-n, and the infrared detection signal determined to have synchronism. It is characterized in that flame judgment is performed using.
[0069]
In other words, in the present invention, in addition to the flame determination as described above, when the plurality of infrared detection units 1-1 to 1-n detect a flame, they output a substantially synchronized (in-phase) signal. This is added to the element of flame judgment.
[0070]
As a method of determining the synchronism of the signals from the infrared detecting units 1-1 to 1-n, a method capable of easily determining the synchronism at high speed is desired. As such a method, in the present invention, when the level of the infrared detection signal (main wavelength signal) from the infrared detection means 1-1 serving as a reference exceeds a predetermined threshold level (at the time of rising), and thereafter this infrared ray When the level of the detection signal (main wavelength signal) reaches the maximum level (peak) (at the peak), and thereafter when the level of the infrared detection signal (main wavelength signal) becomes lower than the predetermined threshold level (rise) The synchronism of the infrared detection signal is determined on the basis of the level of the infrared detection signal (reference wavelength signal) from the other infrared detection means 1-2 to 1-n at the three points in time. .
[0071]
FIGS. 13A and 13B show specific examples of determination of synchronism of infrared detection signals. FIG. 13A is a diagram showing an infrared detection signal (main wavelength signal) from the infrared detection unit 1-1 as a reference, and FIG. 13B is another infrared detection unit (for example, 1-2). It is a figure which shows the infrared detection signal (reference wavelength signal) from). Referring to FIG. 13, the flame determination unit 2 has one rising time, peak time, and rising time based on a predetermined threshold level of the infrared detection signal (main wavelength signal) from the infrared detection unit 1-1 as a reference. The signal levels of the infrared detection signals (reference wavelength signals) from the other infrared detection means 1-2 at the time of falling are respectively represented by a _r , P _r , B _r In other infrared detection means, p _r > A _r , P _r > B _r When the above condition is satisfied, it is determined that the infrared detection signal is synchronized.
[0072]
In the example of FIG. 13, between the infrared detection signal (main wavelength signal) from the infrared detection unit 1-1 serving as a reference and the infrared detection signal (reference wavelength signal) from another infrared detection unit 1-2. In the case where a plurality of other infrared detection means are provided (in the case of 1-2 to 1-n), the flame determination means 2 is a reference infrared detection means. 1-1 Infrared detection signals (reference wavelength signals) from other infrared detection means at one rising, peaking, and falling at a predetermined threshold level of the infrared detection signal (main wavelength signal) from 1-1 The signal levels of _r , P _r , B _r When the predetermined number or more of infrared detecting means among the other infrared detecting means 1-2 to 1-n, _r > A _r , P _r > B _r When the above condition is satisfied, it is determined that the infrared detection signal is synchronized.
[0073]
In the example of FIG. 13, the infrared detection signal (main wavelength signal) from the infrared detection unit 1-1 serving as a reference and the infrared detection signal (reference wavelength signal) from another infrared detection unit 1-2. Focusing on one set of rise, peak, and fall, when a fire occurs, a plurality of sets of rise, peak, and fall generally occur within a predetermined period. In this case, the flame determination means 2 is at a certain rise time, peak time, and fall time based on a predetermined threshold level of the infrared detection signal (main wavelength signal) from the reference infrared detection means 1-1. The signal level of the infrared detection signal (reference wavelength signal) from the other infrared detection means 1-2 is set to a, _r , P _r , B _r In the other infrared detection means 1-2, p _r > A _r , P _r > B _r When the above condition is satisfied for a predetermined number of times of rising, peaking, and falling occurring within a predetermined period, it is determined that the infrared detection signal is synchronous.
[0074]
Also, by integrating these, the flame determination means 2 is in a certain rise time, peak time, and fall time based on a predetermined threshold level of the infrared detection signal from the infrared detection means 1-1 as a reference. The signal levels of the infrared detection signals from the other infrared detection means 1-2 to 1-n are respectively represented as a _r , P _r , B _r When the predetermined number or more of infrared detecting means among the other infrared detecting means 1-2 to 1-n, _r > A _r , P _r > B _r When the above condition is satisfied for a predetermined number of times of rising, peaking, and falling occurring within a predetermined period, it is determined that the infrared detection signal is synchronous.
[0075]
FIG. 14 shows an infrared detection signal (main wavelength signal) from the infrared detection means 1-1 as a reference when a flame occurs, and an infrared detection signal (reference wavelength signal) from another infrared detection means 1-2. ) And an infrared detection signal (reference wavelength signal) from another infrared detection means 1-3. In the example of FIG. 14, infrared detection from the infrared detection means 1-1 serving as a reference is shown. The signal (main wavelength signal), the infrared detection signal (reference wavelength signal) from the other one infrared detection means 1-2, and the infrared detection signal (reference wavelength signal) from the other one infrared detection means 1-3 You can see that they are synchronized. In this case, the flame determination means 2 uses the infrared detection signal (main wavelength signal) from the infrared detection means 1-1 as a reference and the infrared detection signal (reference wavelength signal) from the other infrared detection means 1-2. It is determined that the infrared detection signal (reference wavelength signal) from the other one infrared detection means 1-3 is synchronized, and the infrared detection signal (main wavelength signal and / or reference wavelength) determined to be synchronized. The flame determination process as described above can be performed using the signal.
[0076]
FIG. 15 also shows an infrared detection signal (main wavelength signal) from the infrared detection means 1-1 that becomes a reference when noise is mixed, and an infrared detection signal (reference) from another infrared detection means 1-2. An example of a wavelength signal) and an infrared detection signal (reference wavelength signal) from another infrared detection means 1-3 is shown, and FIG. 16 shows a reference when an irregular light source is input. An infrared detection signal (main wavelength signal) from the infrared detection means 1-1, an infrared detection signal (reference wavelength signal) from the other infrared detection means 1-2, and another infrared detection means 1-3 An example of an infrared detection signal (reference wavelength signal) from is shown. In the example of FIGS. 15 and 16, the infrared detection signal (main wavelength signal) from the infrared detection unit 1-1 serving as a reference and the infrared detection signal (reference wavelength signal) from the other infrared detection unit 1-2 It is not synchronized with the infrared detection signal (reference wavelength signal) from the other one infrared detection means 1-3. Therefore, in this case, in the flame determination means 2, the infrared detection signal from the infrared detection means is noise or It is determined that the light source is affected by an irregular light source, and this infrared detection signal is not used for determining the flame.
[0077]
FIG. 17 is a diagram showing a specific example of the flame detector of the present invention. In the example of FIG. 17, for simplicity, a case where only two infrared detection units 1-1 and 1-2 are provided is shown. The flame detector is configured as a fire detector.
[0078]
Referring to FIG. 17, the infrared detector 1-1 serving as a reference for this flame detector is an infrared ray peculiar to a flame (generally CO 2 ₂ Infrared sensor 21 that detects resonance radiation (infrared rays of about 4.3 μm to 4.4 μm) as a first wavelength signal (main wavelength signal), and voltage amplification that amplifies the first wavelength signal (voltage) from the infrared sensor 21 Among the first wavelength signals from the circuit 22 and the infrared sensor 21, a filter circuit 23 that passes only components in a predetermined frequency band, and a DC that performs DC level conversion on the first wavelength signal that has passed through the filter circuit 23 And a level conversion circuit 24.
[0079]
The infrared detector 1-2 of the flame detector uses infrared light having a wavelength (for example, a wavelength of 3.9 μm or 5.0 μm) different from the flame specific infrared wavelength as the second wavelength signal (reference wavelength signal). An infrared sensor 25 for detection, a voltage amplification circuit 26 for amplifying a second wavelength signal (voltage) from the infrared sensor 25, and a filter circuit 27 for passing only a component in a predetermined frequency band of the second wavelength signal; The DC level conversion circuit 28 performs DC level conversion on the second wavelength signal that has passed through the filter circuit 27.
[0080]
The flame detector 2 of the flame detector includes a CPU (central processing unit) 30 such as a microprocessor that controls the entire apparatus, an output voltage (level of the first wavelength signal) of the DC level conversion circuit 24, and a threshold value. Voltage (predetermined threshold level) V _th And a fire output circuit 31 that outputs a fire signal based on a signal P1 from the CPU 30.
[0081]
Here, as the infrared sensors 21 and 25, pyroelectric sensors (pyroelectric elements) widely used as security sensors can be used. In this case, the pyroelectric sensor generates a differential charge output with respect to the incident light, and therefore outputs a signal proportional to the change (fluctuation) of the thermal energy from the flame. In relation to this, the filter circuits 23 and 27 are configured to pass only the component of the fluctuation frequency (flaming flicker frequency) band as the component of the predetermined frequency band (configured as a band pass filter). ing).
[0082]
That is, the fluctuation of the flame detected by the infrared sensors 21 and 25 is about several Hz to several tens of Hz. In order to obtain the first wavelength signal and the second wavelength signal peculiar to the flame, in the example of FIG. The first wavelength signal and the second wavelength signal from the infrared sensors 21 and 25 are passed through the filter circuits 23 and 27 having transmission characteristics, and only the flame fluctuation component of the first wavelength signal and the second wavelength signal is stored. In the form (that is, a flame fluctuation signal captured at about 4.3 to 4.4 μm, a flame fluctuation signal captured at about 3.9 μm or 5.0 μm), and is incorporated into the CPU 30. Yes.
[0083]
Further, it is assumed that the CPU 30 has a built-in timer function for measuring, for example, 4.5 seconds as the predetermined period T. Further, the CPU 30 has an A / D conversion function for converting the levels (intensity levels) of the first wavelength signal and the second wavelength signal from the DC level conversion circuit 24 and the DC level conversion circuit 28 into digital signals. In this A / D conversion function, the first wavelength signal and the second wavelength signal are sampled at a predetermined time interval Δt (for example, 10 msec) and digitally converted to be captured.
[0084]
The comparator (comparison circuit) 29 is configured such that the level (amplitude voltage) of the infrared detection signal (first wavelength signal) output from the DC level conversion circuit 24 is a threshold voltage (predetermined threshold level) V. _th (For example, 0.2V) (that is, when the infrared detection signal output from the DC level conversion circuit 24 rises), a rising interrupt signal is generated and added to the interrupt terminal INT1 of the CPU 30, and The level (amplitude voltage) of the infrared detection signal (first wavelength signal) output from the DC level conversion circuit 24 is the threshold voltage (predetermined threshold level) V. _th When falling from (for example, 0.2 V) (that is, when the infrared detection signal output from the DC level conversion circuit 24 falls), a falling interrupt signal is generated and applied to the interrupt terminal INT1 of the CPU 30. The CPU 30 detects the rising edge and the falling edge of the first wavelength signal by inputting the rising interrupt signal and the falling interrupt signal to the terminal INT1. Then, after detecting the rising edge of the first wavelength signal, the CPU 30 detects the peak of the first wavelength signal until the falling edge of the first wavelength signal is detected, and the rising edge, peak, and rising edge of the first wavelength signal are detected. The level of the second wavelength signal at the three falling times is taken in, and it is determined whether or not there is synchronization between the first wavelength signal and the second wavelength signal.
[0085]
In addition to the synchronization determination process, the CPU 30 performs the flame determination process as described above only for the infrared detection signals (first wavelength signal and second wavelength signal) determined to have synchronization. ing.
[0086]
Specifically, the flame determination process is performed as follows. That is, the CPU 30 captures the level of the first wavelength signal as digital data at a predetermined time interval (sampling period) Δt (for example, a time interval of 10 milliseconds) over a predetermined period T (for example, 4.5 seconds), and over the predetermined period T. Of the level data (digital data) of the first wavelength signal captured every time interval Δt (= 10 milliseconds), for example, a predetermined threshold level V _th The peak is detected only for the level of the first wavelength signal exceeding the average peak level M _avg And the peak frequency PCNT are calculated according to Equation 1, for example. When the peak of the first wavelength signal from the DC level conversion circuit 24 of the infrared detecting means 1-1 is detected as described above, the CPU 30 detects the second wavelength signal from the DC level conversion circuit 28 at the time of extraction. The average auxiliary level S is obtained by averaging the levels of the second wavelength signals acquired over the predetermined period T. _avg Is calculated. And CPU30 is average peak level M _avg And average support level S _avg Ratio M _avg / S _avg And ask. In this way, the CPU 30 determines the average peak level M of the infrared detection signal over a predetermined period (T = 4.5 seconds). _avg , Peak frequency PCNT, and ratio M _avg / S _avg These three parameters are obtained, and flame determination (for example, determination of the degree of fire) can be performed based on these three parameters.
[0087]
Then, the CPU 30 outputs a fire signal to the fire output circuit 31 when it is determined that there is a fire in the above-described flame determination process based on the first wavelength signal and the second wavelength signal that are determined to be synchronized. It has become.
[0088]
Next, the processing operation of the flame detector having such a configuration, mainly the processing operation of the CPU 30, will be described with reference to the flowchart of FIG. Referring to FIG. 18, in step S1, the CPU 30 first initializes a flag FLG indicating that a peak search (peak search) is being performed to “0”. Next, in step S <b> 2, the CPU 30 determines whether there has been a rising interrupt input from the comparator 29. As a result, when there is a rising interrupt, the process proceeds to step S3. In step S3, the level of the second wavelength signal (reference wavelength signal) at this time is acquired and stored, assuming that this time is a rising time. In step S4, the flag FLG is set to “1”, and the process returns to step S2.
[0089]
If it is determined in step S2 that there is no rising interrupt input, the process proceeds to step S5. In step S5, it is determined whether there has been a falling interrupt input. As a result, if it is determined that there is no falling interrupt input, the process proceeds to step S6. In step S6, it is determined whether or not the flag FLG is “1”. When the flag FLG is not “1” (that is, when it is “0”), the process returns to step S2. On the other hand, when the flag FLG is “1” in step S6, the process proceeds to step S7. In step S7, it is determined whether the first wavelength signal (main wavelength signal) has reached a peak. As a result, when it is not a peak, the process proceeds to step S8, and in step S8, it is determined whether or not it is during the signal acquisition period. If it is determined in step S8 that it is during the signal acquisition period, the process returns to step S2 again, and the above processing is repeated. In step S7, when the first wavelength signal (main wavelength signal) reaches a peak, in step S9, the level of the first wavelength signal (main wavelength signal) at this time and the second wavelength signal (reference wavelength signal). The level is acquired and stored, and the process returns to step S2 again to repeat the above process.
[0090]
In step S5, when there is a falling interrupt input, the process proceeds to step S10. In step S10, the level of the second wavelength signal (reference wavelength signal) at this time is acquired assuming that this time is a falling time. And remember. In step S11, the flag FLG is set to “0”, the process returns to step S2 again, and the above-described processing is repeated. If it is determined in step S8 that it is not in the signal acquisition period, in step S12, the second wavelength signal (reference wavelength signal) at the rise time, peak time, and fall time acquired and stored in steps S3, S9, and S10. ) To determine whether the first wavelength signal (main wavelength signal) and the second wavelength signal (reference wavelength signal) are synchronized. As a result, when it is synchronized, the process proceeds to step S13, flame determination processing is performed, and when it is determined that a flame is detected, a fire signal is output in step S14. On the other hand, if it is determined in step S12 that there is no synchronization, the process is terminated without performing the flame determination process.
[0091]
As described above, in the present invention, the flame determination unit 2 detects an asynchronous signal component as noise in each infrared detection signal from each infrared detection unit 1-1 to 1-n, and only the synchronized signal component is detected. The flame is judged based on the above.
[0092]
In the above-described example, the infrared detection unit 11-1 serving as a reference and the other infrared detection units 11-2 to 11-n have different detection wavelengths, but the synchronization is determined. As for the reference infrared detecting means 11-1 and other infrared detecting means 11-2 to 11-n, those having the same detection wavelength can be used. In other words, the same type of infrared detecting means 11-1 as a reference and the other infrared detecting means 11-2 to 11-n can be used.
[0093]
That is, in the configuration of FIG. 1, each of the infrared detection units 1-1 to 1-n can have an equivalent structure with respect to noise superimposed on each infrared detection signal.
[0094]
A specific example of the flame detector having such a configuration is such that when n is 2, the infrared detection means 1-1 and the infrared detection means 1-2 have equivalent structures, and the CPU 30 has a flame determination means 2 as well. In other words, in each infrared detection signal (first wavelength signal, second wavelength signal) from the infrared detection means 1-1 and 1-2, the asynchronous signal component is detected as noise and synchronized. It can be configured to have a function of judging the flame based only on the signal component.
[0095]
More specifically, pyroelectric elements having the same structure are used for the infrared sensor 21 of the infrared detection means 1-1 and the infrared sensor 25 of the infrared detection means 1-2 in FIG. The voltage amplifying circuit 22, the filter circuit 23, the DC level converting circuit 24, and the voltage amplifying circuit 26, the filter circuit 27, and the DC level converting circuit 28 of the infrared detecting means 1-2 have circuit structures designed to be equivalent to each other. As used, it can be embodied.
[0096]
In general, popcorn noise is generated when the temperature changes suddenly or when mechanical stress is applied to a component (for example, an infrared sensor). ) Are placed in close proximity to each other in the same environment, the probability of popcorn noise occurring in them is very low. On the other hand, when a plurality of infrared detection means having an equivalent structure are placed in close proximity to each other in the same environment, the signal synchronism is detected between the fire signal and the non-fire signal detected by these infrared detection means. is there.
[0097]
Therefore, in the flame detector as shown in FIG. 17, when the infrared detection means 1-1 and the infrared detection means 1-2 have an equivalent structure, the infrared detection means 1-1 and the infrared detection means 1-2 respectively. When the detected signals are synchronized with each other, this signal is a fire signal or a non-fire signal and can be determined not to be noise, and is detected by the infrared detection means 1-1 and the infrared detection means 1-2, respectively. If the received signal is asynchronous, it can be determined that this signal is noise.
[0098]
Specifically, for infrared rays emitted from a high-temperature object, a low-temperature object (400 ° C. or less), a low oil smoke flame, and a high oil smoke flame, the first wavelength signal detected by the infrared detector 1-1 and the infrared detector 1 The second wavelength signal detected at -2 is synchronized with each other, though there is a difference in magnitude.
[0099]
FIG. 19 is a diagram showing the results of detecting the infrared rays radiated from the low oil smoke flame by the infrared detection means 1-1 and the infrared detection means 1-2 having structures equivalent to each other. FIG. 20 is a diagram showing a result of detecting infrared rays radiated from a low-temperature object (400 ° C. or lower) chopped by hand by the infrared detection means 1-1 and the infrared detection means 1-2 having a structure equivalent to each other. It is.
[0100]
As can be seen from these results, in the case of a non-noise signal, the first wavelength signal detected by the infrared detecting means 1-1 regardless of whether the signal is caused by fire or non-fire. And the second wavelength signal detected by the infrared detecting means 1-2 are synchronized with each other. On the other hand, for example, when popcorn noise is generated in the infrared detection unit 1-1, the level of the first wavelength signal detected by the infrared detection unit 1-1 and the second level detected by the infrared detection unit 1-2. It is not synchronized with the level of the wavelength signal.
[0101]
21 and 22 are diagrams showing a first wavelength signal and a second wavelength signal when popcorn noise is generated in the infrared detecting means 1-1. FIG. 22 is an enlarged view of FIG. 21 in the direction of the time axis, and shows the level of the second wavelength signal when the peak of the first wavelength signal level is detected. As can be seen from FIGS. 21 and 22, the first wavelength signal and the second wavelength signal are generally asynchronous with each other.
[0102]
By utilizing this feature, in the flame determination means 2 of the flame detector of FIG. 1, each infrared detection signal from each of the infrared detection means 1-1 to 1 -n is detected as noise for an asynchronous signal component. The flame is determined based on only the synchronized signal components.
[0103]
In the above description, in FIG. 1, n is “2” and the flame detector is a two-wavelength type as shown in FIG. 17, but in FIG. Even in the case of 3 ″, that is, in the case of a three-wavelength type, each infrared detecting means has an equivalent structure, and the synchronization and asynchronousness of each infrared detection signal detected by each infrared detecting means are examined. Thus, it can be detected whether or not popcorn noise is generated. Furthermore, not only popcorn noise, but also white noise, electrical noise, radio noise, vibration noise, and even incidents from irregular light sources can be detected, with or without these effects. Can make flame judgment and fire judgment. That is, flame judgment and fire judgment can be performed while suppressing or eliminating the influence of internal noise (popcorn noise, white noise) and external noise (electricity, radio waves, light, vibration noise).
[0104]
Further, as the flame determination means 2 in FIG. 1, a flame determination processing function considering randomness may be provided with a synchronization determination function, or conventional flame determination processing not considering randomness The function may be provided with a synchronism determination function. However, if the flame determination processing function considering the randomness is provided with a synchronization determination function, the flame determination can be performed with extremely high performance.
[0105]
Further, in the above-described embodiment, in the flame detector, synchronism between a plurality of infrared detection signals is determined, flame determination processing is performed, and a determination result as to whether or not the flame is output. However, the present invention is not limited to the configuration of one flame detector, and can be applied to any flame monitoring system. That is, the application object of the present invention is not limited to a single flame detector, but can be applied to any flame monitoring system (on / off type monitoring system, analog type monitoring system), and applied to an analog type monitoring system. In some cases, it is possible to perform flame determination processing on the receiver side, for example.
[0106]
That is, in the configuration example of FIG. 1, the flame detector itself is provided with the flame determination means 2, and the synchronization determination, the flame determination, and the fire determination are performed in the flame detector (that is, on / off). It is also possible to configure the flame sensor as an analog sensor. FIG. 23 is a diagram showing an example of a case where the flame detector is configured as an analog type sensor that is address polled from the receiver. In this case, in the signal processing means of the flame detector, the synchronism is judged. After removing the asynchronous portion from the signal as noise, this signal is returned to the receiver, and the receiver can perform flame determination processing.
[0107]
【The invention's effect】
As explained above, claims 1 to Claim 8 According to the described invention, a plurality of infrared detection means for detecting light having infrared wavelengths specific to the flame as infrared detection signals, and a feature quantity specific to the flame is extracted from the infrared detection signals respectively detected by the plurality of infrared detection means Flame determining means for making a flame determination based on a characteristic quantity unique to the flame, wherein the flame determining means determines the synchronism of the infrared detection signals respectively detected by the plurality of infrared detecting means. Since the flame is determined based on the infrared detection signal that is determined to be present, it is possible to eliminate the effects of irregular light source incidents and continuous noise, and to determine whether or not it is a flame. Can be done accurately.
[0108]
In particular, Claim 1 Thru Claim 8 In the described invention ,in front When one of the plurality of infrared detection means is used as a reference infrared detection means, the flame determination means is configured such that the infrared detection signal detected by the reference infrared detection means is lower than a predetermined threshold level. When the infrared detection signal detected by the reference infrared detection means becomes larger than a predetermined threshold level, the predetermined threshold level of the infrared detection signal from the reference infrared detection means is determined. The synchronization of the infrared detection signal is judged based on the signal level of the infrared detection signal from the other infrared detection means at the time of rising, peaking, and falling based on the reference. Since the synchronicity can be easily determined from the signals of only three points at the time of falling, the processing speed can be increased.
[0109]
Also, Claim 5 In the described invention, Claim 1 Thru Claim 4 The flame detector according to any one of the preceding claims, wherein the flame determination means detects a rising edge and a falling edge with reference to a predetermined threshold level of an infrared detection signal from an infrared detection means as a reference. And a CPU that performs flame determination processing including determination of synchronism, and the comparison circuit detects rising and falling with reference to a predetermined threshold level of the infrared detection signal from the infrared detection means serving as a reference. The rise and fall detected by the comparison circuit is notified to the CPU as an interrupt (that is, the rise of the signal from the sensor that detects the reference main wavelength using the comparison circuit). By using the interrupt function of the CPU to detect the falling edge, the burden on the CPU can be greatly reduced.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration example of a flame detector according to the present invention.
FIG. 2 is a diagram showing the spectral intensity of a general flame.
FIG. 3 is a diagram showing an example of temporal change in the intensity level of infrared rays emitted from a flame at the time of a fire.
FIG. 4 is a diagram illustrating an example of a change in peak level of an infrared detection signal in which infrared rays from a flame at the time of a fire are detected by an infrared sensor (pyroelectric sensor).
FIG. 5 is a diagram showing an example of a change in peak level of an infrared detection signal in which infrared rays from a low-temperature heat source (400 ° C. or lower) are detected by an infrared sensor (pyroelectric sensor).
6 shows an example of creation of a peak level frequency distribution (histogram) in each case of the infrared detection signal of FIG. 4 (signal due to a flame in a fire) and the infrared detection signal of FIG. 5 (signal by a low-temperature heat source). FIG.
FIG. 7 shows a second wavelength in the case where the flame detector shown in FIG. 1 uses an infrared detector having a wavelength in the vicinity of the infrared wavelength detected by the infrared detector 1-1 as the infrared detector 1-2. It is a figure for demonstrating taking in of the level of a signal.
FIG. 8 is a diagram illustrating an example of a first wavelength signal and a second wavelength signal during a fire.
FIG. 9 is a diagram illustrating an example of a first wavelength signal and a second wavelength signal when there is no fire (for example, when a halogen lamp or heat source noise is added).
FIG. 10 shows measurement results (time) of the levels of the first wavelength signal (about 4.3 to 4.4 μm) and the second wavelength signal (for example, 5.0 μm) when a low oil smoke (normal-heptane) flame is generated. FIG.
FIG. 11 shows the measurement results of the level of the first wavelength signal (approximately 4.3 to 4.4 μm) and the second wavelength signal (for example, 5.0 μm) when a high oil smoke (gasoline) flame is generated (change over time). FIG.
FIG. 12 shows the measurement results of the level of the first wavelength signal (approximately 4.3 to 4.4 μm) and the second wavelength signal (for example, 5.0 μm) when a high-temperature object (400 ° C. or higher) is chopped (change over time). FIG.
FIG. 13 is a diagram showing a specific example of determination of synchronism of infrared detection signals in the present invention.
FIG. 14 is a diagram showing an example of an infrared detection signal (main wavelength signal) from an infrared detection means serving as a reference when a flame occurs and an infrared detection signal (reference wavelength signal) from another infrared detection means; .
FIG. 15 is a diagram showing an example of an infrared detection signal (main wavelength signal) from an infrared detection unit serving as a reference when noise is mixed and an infrared detection signal (reference wavelength signal) from another infrared detection unit; .
FIG. 16 shows an example of an infrared detection signal (main wavelength signal) from an infrared detection means serving as a reference when an irregular light source is input and an infrared detection signal (reference wavelength signal) from another infrared detection means; FIG.
FIG. 17 is a diagram showing a specific example of a flame detector according to the present invention.
FIG. 18 is a flowchart for explaining the processing operation of the flame detector of the present invention.
FIG. 19 is a diagram showing the results of detecting the infrared rays emitted from the low oil smoke flame by the infrared detecting means 1-1 and the infrared detecting means 1-2 having a structure equivalent to each other.
FIG. 20 is a diagram showing the results of detecting infrared rays radiated from a low-temperature object (400 ° C. or lower) chopped by hand by infrared detecting means 1-1 and infrared detecting means 1-2 having a structure equivalent to each other. is there.
FIG. 21 is a diagram showing a first wavelength signal and a second wavelength signal when popcorn noise is generated in the infrared detecting means 1-1.
FIG. 22 is a diagram showing a first wavelength signal and a second wavelength signal when popcorn noise is generated in the infrared detecting means 1-1.
FIG. 23 is a diagram showing an example when the flame sensor is configured as an analog sensor.
[Explanation of symbols]
1-1 to 1-n Infrared detecting means
2 Flame judgment means
21, 25 Infrared sensor
22, 26 Voltage amplification circuit
23, 27 Filter circuit
24, 28 DC level conversion circuit
29 Comparator
30 CPU
31 Fire output circuit

Claims (8)

Multiple infrared detection means for detecting light of infrared wavelength specific to flame as an infrared detection signal, and extracting flame-specific feature quantities from the infrared detection signals respectively detected by the multiple infrared detection means, and flame-specific feature quantities Flame determination means for making a flame determination based on the above, and when one of the plurality of infrared detection means is a reference infrared detection means, the flame determination means is a reference infrared detection means It is determined whether or not the infrared detection signal detected in step 1 is larger than a predetermined threshold level, and becomes a reference when the infrared detection signal detected by the reference infrared detection means becomes larger than the predetermined threshold level. Infrared detection signals from other infrared detection means at the time of rising, peaking and falling based on a predetermined threshold level of the infrared detection signal from the infrared detecting means Based on the signal level, it is determined synchronicity of the infrared detection signal, the flame detector, characterized in that is adapted to perform flame determined using an infrared detection signal which is determined as being synchronous. 2. The flame detector according to claim 1 , wherein data is taken in at predetermined time intervals over a predetermined time from the time when the reference infrared detection level exceeds a predetermined threshold level, and the maximum value is peaked. Flame detector. 3. The flame detector according to claim 1 or 2 , wherein the flame judgment means is at one rising, peaking, or rising time based on a predetermined threshold level of an infrared detection signal from an infrared detection means as a reference. When the signal levels of the infrared detection signals from the other infrared detection means at the time of falling are a _r , p _r , and b _r , respectively, in a predetermined number or more of the other infrared detection means, p _{when r>} a r, the condition of _{p r>} b _r is satisfied, flame detector, characterized by being adapted to determine that there is a synchronous to the infrared detection signal. In flame detector according to any one of claims 1 to 3, wherein the flame determining means, a primary one certain rising relative to the predetermined threshold level of the infrared detection signal from the infrared detector When the signal levels of the infrared detection signals from the other infrared detection means at the time, the peak time, and the fall time are a _r , p _r , and b _r , respectively, p _r > a _r , P _r > br so that the infrared detection signal is synchronous when the condition of _{r r} , p _r > b _r is satisfied for a predetermined number of times of rising, peaking, and falling occurring within a predetermined period. A flame detector characterized by 3. The flame detector according to claim 1 or 2 , wherein the flame judgment means is at one rising, peaking, or rising time based on a predetermined threshold level of an infrared detection signal from an infrared detection means as a reference. When the signal levels of the infrared detection signals from the other infrared detection means at the time of falling are a _r , p _r , and b _r , respectively, in a predetermined number or more of the other infrared detection means, p _r> a r, the condition of _{p r>} b _r is rising that occurs within a predetermined time period, the peak, a predetermined number of times or more of the occurrence frequency of falling, when filled, when there is a synchronous to the infrared detection signal A flame detector characterized by being made to judge. In flame detector according to any one of claims 1 to 5, wherein the flame determining means, the rising and falling a predetermined threshold level of the infrared detection signal from the infrared detecting means as a reference to the reference And a CPU for performing flame determination processing including determination of synchronism, and rising and rising with reference to a predetermined threshold level of an infrared detection signal from a reference infrared detection means A flame detector characterized in that when a fall is detected by a comparison circuit, the rise and fall detected by the comparison circuit are notified to the CPU as an interrupt. The flame detector according to any one of claims 1 to 6 , wherein a pyroelectric element is used as an infrared sensor in at least one infrared detection means of the plurality of infrared detection means. A flame detector. In a plurality of infrared detection means, light of infrared wavelength peculiar to flame is detected as an infrared detection signal, and a feature quantity peculiar to flame is extracted from each of the infrared detection signals detected by the plurality of infrared detection means. A flame detection method for making a flame determination based on an amount, wherein when one of the plurality of infrared detection means is a reference infrared detection means, the flame determination means is a reference infrared detection means It is determined whether or not the infrared detection signal detected in step 1 is larger than a predetermined threshold level, and becomes a reference when the infrared detection signal detected by the reference infrared detection means becomes larger than the predetermined threshold level. Infrared detection from other infrared detection means at the rise, peak, and fall times based on a predetermined threshold level of the infrared detection signal from the infrared detection means Based on the signal level of the signal, flame detection method to determine the synchronization of the infrared detection signal, and performs a flame decision by using the infrared detection signal is determined as being synchronous.

Images