JP3858978B2

JP3858978B2 - Evaluation method and apparatus

Info

Publication number: JP3858978B2
Application number: JP2001327742A
Authority: JP
Inventors: 孝範宮坂; 泰之武藤
Original assignee: NSK Ltd
Current assignee: NSK Ltd
Priority date: 2001-10-25
Filing date: 2001-10-25
Publication date: 2006-12-20
Anticipated expiration: 2021-10-25
Also published as: JP2003130763A

Description

【０００１】
【発明の属する技術分野】
本発明は、１または複数の摺動部材を含む機械設備から発生する音又は振動を検出し、検出した信号を解析して、機械設備に起因する異常の有無を診断する評価方法及び装置に関するもので、詳しくは、検出した信号の解析作業時の負担を軽減して、診断作業の迅速化及び信頼性の向上を実現するための改良にかかるものである。
【０００２】
【従来の技術】
これまで、回転体等の１または複数の摺動部材を含む機械設備の、摺動部材の摩耗や破損等の異常の有無を診断する評価方法として、機械設備から発生する音又は振動を検出し、検出した信号を適宜波形処理によって周波数スペクトルデータに変換し、機械設備の異常に起因した周波数成分と前記した周波数スペクトルデータとの比較照合により、異常の有無を診断する評価方法が知られている。
【０００３】
摺動部材を含む機械設備の異常に起因した周波数成分は、摺動部材の設計諸元及び運用状況によって決まるもので、例えば、前記した回転体の設計諸元や運用状況等の条件から、演算処理によって求めることができる。
従って、機械設備の異常の有無を自動診断する評価装置では、コンピュータ等の演算処理手段を用い、予め、診断対象となる機械設備の異常に起因した周波数成分を算出処理すると共に、機械設備の発生する音又は振動の信号を適宜波形処理して周波数スペクトルデータに変換する演算処理を行い、更には、予め算出して記憶させた機械設備の異常に起因した周波数成分と前記した周波数スペクトルデータとを比較照合する演算処理を行う。
【０００４】
【発明が解決しようとする課題】
ところで、従来の評価方法では、機械設備の特定部位毎に、異常時の周波数成分を１次から多次まで多数求め、これらの多数の周波数成分のそれぞれに対して、実測した機械設備の周波数スペクトルデータ上にピークが表出するか否かを診断する演算処理と、周波数スペクトルデータ上のピーク値が異常を示すピークレベルであるかを診断する演算処理とを繰り返す。
そのため、最終的な診断を下すまでの演算処理が膨大になり、演算処理手段に大きな負担がかかるために、演算処理能力が高い高価なコンピュータが必要となって装置コストの増大を招いたり、また演算処理の所要時間の長大化により診断作業の迅速化が困難になるという問題があった。
【０００５】
本発明は上記事情に鑑みてなされたもので、１または複数の摺動部材を含む機械設備の異常の有無を診断する際の演算処理の負担を軽減して、診断作業の迅速化及び信頼性の向上を実現することのできる評価方法及び装置を提供することを目的とする。
【０００６】
【課題を解決するための手段】
上記目的を達成するための請求項１に記載した本発明に係る評価方法は、摺動部材を含む機械設備から発生する音又は振動を検出し、検出した信号を解析して、前記機械設備に起因する異常の有無を診断する評価方法であって、
前記機械設備から発生した音又は振動から実測周波数スペクトルデータを生成し、
該実測周波数スペクトルデータの実効値を算出すると共に、該実効値に基づいて閾値を設定し、前記機械設備の異常に起因した前記周波数成分の１次、２次、４次値に対する前記実測周波数スペクトルデータ上のピークは前記閾値を超える場合にのみ有効なピークとし、前記機械設備の異常に起因した周波数成分の１次、２次、４次値に対する前記実測周波数スペクトルデータ上の前記有効なピークの有無により、前記機械設備に対する異常の有無の診断を行うことを特徴とする。
【０００７】
上記目的を達成するための請求項５に記載した本発明に係る評価装置は、摺動部材を含む機械設備から発生する音又は振動を検出し、検出した信号を解析して前記機械設備に起因する異常の有無を診断する評価装置であって、
前記機械設備から発生した音又は振動から実測周波数スペクトルデータを生成後、該実測周波数スペクトルデータの実効値を算出し、該実効値に基づいて閾値を設定し、前記機械設備の異常に起因した前記周波数成分の１次、２次、４次値に対する前記実測周波数スペクトルデータ上のピークは前記閾値を超える場合にのみ有効なピークとし、且つ、前記機械設備の異常に起因した周波数成分の１次、２次、４次値に対する前記実測周波数スペクトルデータ上の前記有効なピークの有無により、前記機械設備に対する異常の有無の診断を行う演算処理手段とを備えたことを特徴とする。
【０００８】
そして、請求項１及び請求項５に記載した構成の評価方法及び装置によれば、機械設備の異常に起因した周波数成分に対応する実測周波数スペクトルデータ上の有効なピークの有無を調べる照合処理は、機械設備の異常に起因した周波数成分の１次、２次、４次値の３回に限定されているため、例えば１次から高次の周波成分まで多数の周波数成分に対して照合処理を繰り返す従来技術と比較すると、照合処理時の演算処理量が低減して、演算処理手段への負担が大幅に軽減される。更に、機械設備の異常に起因した周波数成分の１次、２次、４次値の３回に渡って照合処理を実施するため、周波数成分の１次のみで判断する場合と比較して、ノイズ等の影響による誤診断が発生し難く、信頼性の高い診断が可能になる。
【００１２】
【発明の実施の形態】
以下、本発明に係る評価方法及び装置の好適な実施の形態を添付図面に基づいて詳細に説明する。
図１は本発明に係る評価方法及び装置の第１の実施の形態の概略構成を示すブロック図、図２は図１に示した評価装置の診断処理の手順を示すフローチャートである。
【００１３】
先ず、第１の実施の形態の評価装置の概略構成を、図１に基づいて説明した後に、この評価装置による評価方法について詳述する。
本実施の形態の評価装置１は、診断対象となる１または複数の摺動部材を含む機械設備３の発生する音又は振動に応じたアナログ信号を出力する振動検出手段５と、この振動検出手段５の出力する信号を増幅する増幅手段７と、増幅手段７によって増幅されたアナログ信号をデジタル信号に変換して実測デジタルデータを生成するＡＤ変換手段９と、このＡＤ変換手段９が出力する実測デジタルデータに基づいて機械設備３の特定部位の異常の有無を診断する演算処理手段１３とを備えた構成である。
【００１４】
本実施の形態の場合、診断対象となる機械設備３の摺動部材としては転がり軸受が適用される。そして、転がり軸受を構成している内外輪、転動体、保持器等の摩耗や損傷を、転がり軸受の駆動時の音又は振動から診断する。
なお、本実施の形態において、転がり軸受の音又は振動とは、転がり軸受の駆動時に現れる超音波振動、所謂ＡＥ（Acoustic Emission ）を含む意味である。
【００１５】
演算処理手段１３は、予め記憶させた処理用の諸データや、ＡＤ変換手段９から受ける実測デジタルデータを、診断用プログラムに基づいて演算処理する診断用コンピュータである。
この演算処理手段１３は、ＡＤ変換手段９が出力する実測デジタルデータに対して周波数分析及びエンベロープ分析等の適宜解析処理を行って実測周波数スペクトルデータを生成すると共に、機械設備３の異常に起因した周波数成分の１次、２次、４次値に対する実測周波数スペクトルデータ上のピークの有無により、機械設備３に対する異常の有無の診断を行う。
【００１６】
以上の評価装置１は、図２に示す手順で、処理を行う。
先ず、振動検出手段５により機械設備３の発生する音又は振動の検出を行い（ステップＳ１０１）、次いで、増幅手段７を経た信号をＡＤ変換手段９によるＡＤ変換によってデジタル信号化して（ステップＳ１０２）、演算処理手段１３に渡す。
演算処理手段１３は、ＡＤ変換手段９から受けた信号を、例えば、ＷＡＶファイル等のファイル形式でデジタルファイル化し（ステップＳ１０３）、必要ならば、フィルタ処理を行って、余分な信号の除去等を行って実測デジタルデータを生成する。
【００１７】
本実施の形態の場合、フィルタ処理は、演算処理手段１３に予め組み込んだフィルタ処理プログラムによって入力信号に所定の処理を行うもので、予めカットする周波数域等の設定を行うフィルタ帯域の選定工程（ステップＳ１０４）と、選定されたフィルタ帯域に従って余分な信号のカットを行うフィルタ処理工程（ステップＳ１０５）とで構成される。
ステップＳ１０４及びステップＳ１０５によるフィルタ処理は、収集してあるデータのＳ／Ｎ比を向上させるために行うもので、入力信号のＳ／Ｎ比が十分であれば、不要である。
【００１８】
次いで、生成した実測デジタルデータに対して、周波数分析及びエンベロープ分析等の解析処理を行って（ステップＳ１０６、Ｓ１０７）、機械設備３から検出した音又は振動を具現した実測周波数スペクトルデータｄ１を得る（ステップＳ１０８）。
ここで得た実測周波数スペクトルデータｄ１は、図３に示す波形ｗ１である。この波形ｗ１は、摺動部材としての転がり軸受において、外輪固定で、毎分１５０回転で内輪を回転させた時のものである。
【００１９】
更に、演算処理手段１３は、機械設備３の特定部位の異常時に発生する周波数成分の１次、２次、４次値に対する実測周波数スペクトルデータｄ１上のピークの有無により、機械設備３の特定部位に対する異常の有無の診断を行う（ステップＳ１０９）。
摺動部材である軸受は、図４に示すように、設計諸元や使用条件に応じて、特定部位の異常時に発生する周波数成分値が決定される。
演算処理手段１３は、機械設備３について、図４に示す特定部位の異常時に発生する周波数成分の１次、２次、４次値を予め基準値として記憶していて、これらの基準値に基づいて、ステップＳ１０９を行う。
【００２０】
ステップＳ１０９では、具体的には、図５に示す手順で、機械設備３の特定部位の異常時に発生する周波数成分の１次、２次、４次値に対する実測周波数スペクトルデータｄ１上のピークの有無をチェックする照合処理を実施し、周波数成分の１次値、２次値の双方が実測周波数スペクトルデータｄ１上のピークに一致した場合（ステップＳ２０１、Ｓ２０２）、あるいは、周波数成分の１次値は実測周波数スペクトルデータｄ１上のピークに一致しないが、２次値、４次値の双方が実測周波数スペクトルデータｄ１上のピークに一致した場合（ステップＳ２１１、Ｓ２１２）など、二つ以上の周波数成分において実測周波数スペクトルデータｄ１上にピークが存在することが確認された場合には、その特定部位について異常有りの診断を下す（ステップＳ２２１）。
一方、実測周波数スペクトルデータｄ１上にピークが存在する周波数成分が一つ以下の場合には、他の部位の異常に起因する振動等がノイズとして影響して、たまたまピークを形成している可能性が高く、異常無しの診断を下す（ステップＳ２３１）。
【００２１】
図６は、摺動部材としての転がり軸受において、外輪固定で、毎分１５０回転で内輪を回転させた時の波形ｗ１に対して、特定部位である外輪の損傷に起因して発生する周波数成分の１次値Ｑ₁ 、２次値Ｑ₂ 、４次値Ｑ₄ の３つの周波数成分を、破線で付記したものである。
前述した図３の場合は、同様の波形ｗ１に対して、外輪の損傷に起因して発生する周波数成分の一次値Ｑ₁ から高次Ｑ_n までの全てのものを、破線により付記している。
【００２２】
以上の第１の実施の形態の評価装置１で行う評価方法では、機械設備３の特定部位の異常時に発生する周波数成分に対応する実測周波数スペクトルデータｄ１上のピークの有無を調べる照合処理は、機械設備３の異常に起因した周波数成分の１次、２次、４次値の３回に限定されているため、例えば、図３に示したように１次から高次の周波成分まで多数の周波数成分の全てに対して照合処理を繰り返す従来技術の場合と比較すると、照合処理時の演算処理量が大幅に低減する。
【００２３】
そのため、機械設備３から検出した信号の解析作業時の演算処理手段１３への負担が大幅に軽減され、診断作業の迅速化を図ることができる。また、演算処理量が低減したことで、演算処理手段１３として使用するコンピュータに、演算処理能力が低い安価なコンピュータを使用することが可能になり、装置コストの低減を図ることも可能になる。
【００２４】
更に、異常時に発生する周波数成分の１次のみで判断すると、たまたまノイズ等の影響で、対応する実測周波数スペクトル上のピークにずれが生じたり、逆にピークが増大しているために誤診断が生じる可能性がある。
しかし、上記のように、機械設備３の特定部位の異常時に発生する周波数成分の１次値、２次値、４次値の３回に渡って照合処理を実施する場合には、３回ともノイズの影響を受ける確率は殆どなく、従って、２回の照合処理を実施するだけで、診断に対する信頼性を向上させることができる。
【００２５】
なお、好ましくは、演算処理手段１３は、実測周波数スペクトルデータｄ１の生成後、この実測周波数スペクトルデータｄ１の実効値ｆ１を算出すると共に、この実効値ｆ１に基づいて閾値ｔ１を設定し、機械設備３の特定部位の異常時に発生する周波数成分の１次値Ｑ₁ 、２次値Ｑ₂ 、４次値Ｑ₄ に対する実測周波数スペクトルデータｄ１上のピークは、閾値ｔ１を超える場合にのみ有効なピークとして扱う構成とするとよい。
【００２６】
図７は、摺動部材としての転がり軸受において、外輪固定で、毎分１５０回転で内輪を回転させた時の実測周波数スペクトルデータｄ１の波形ｗ２に対して、実効値ｆ１と、閾値ｔ１とを書き込んだものである。また、摺動部材としての転がり軸受の転動体の傷に起因して発生する周波数成分の１次値ｑ₁ 、２次値ｑ₂ 、４次値ｑ₄ を波形ｗ２上に点線で書き込んでいる。
この場合、実効値ｆ１は、波形ｗ２の振幅の平均レベルを算出したもので、−８．５ｄＢである。
また、閾値ｔ１は、ｔ１＝（ｆ１＋１０ｄＢ） ……（１）
に設定したため、閾値ｔ１は１．５ｄＢとなった。
この例の場合は、転動体の傷に起因して発生する周波数成分（２ｆｂ）の１次値ｑ₁ 、２次値ｑ₂ 、４次値ｑ₄ のうち、２次値ｑ₂ が閾値ｔ１より小さくノイズに埋もれていることから、閾値ｔ１より大きな１次値ｑ₁ 、４次値ｑ₄ のみについて照合処理が必要なことを示している。
【００２７】
このように閾値ｔ１によって有用なピークの選別を可能にすると、例えば、摺動部材の特定部位での異常に起因して発生する周波数成分の１次値、２次値、４次値に対応する実測周波数スペクトルデータ上のピークに対して照合のための演算処理を実施する前に、閾値ｔ１によって有効なピークを抽出する抽出処理をすれば、有意でないピークに対して照合処理を実施する無駄を省くことができ、演算処理量による負担を更に軽減して、診断処理の迅速化を促進することができる。
【００２８】
なお、以上の実施の形態では、特定部位の損傷の有無を診断する場合を示している。
しかし、前述したように、実測デジタルデータに対して周波数分析及びエンベロープ分析等の適宜解析処理を行って実測周波数スペクトルデータｄ１を生成した場合に、例えば、図８に示すように、この実測周波数スペクトルデータｄ１の実効値ｆ１を算出して、算出した実効値ｆ１を基準レベルＬ₀ に設定し、機械設備３の特定部位の異常（本例では外輪損傷を示す）に起因して発生する周波数成分の１次値Ｑ₁ に対する実測周波数スペクトルデータｄ１上のレベルＬ_h と基準レベルＬ₀ とのレベル差ｌの大きさから、異常を起こしている外輪における損傷の大きさを推定できる。
なお、図８は摺動部材である回転体のエンベロープ波形を示す。
【００２９】
図９は、摺動部材としての転がり軸受において、軌道輪の損傷である剥離が生じた場合に、剥離の大きさと、実測周波数スペクトルデータｄ１上に表れるピークと基準レベルとの間のレベル差との関係を示したものである。
このように、一般的に、レベル差は損傷の大きさに比例して増大するため、逆に、実測周波数スペクトルデータｄ１上のピークにおけるレベル差を求めることで、損傷の大きさを推定することが可能である。
しかも、機械設備３の損傷に起因する実測周波数スペクトルデータｄ１上でのピークレベルの増大は、異常に起因する周波数成分の１次値に対応するピークで一番顕著になる。
【００３０】
そのため、既述したように、機械設備３の特定部位での異常に起因して発生する周波数成分の１次値に対する実測周波数スペクトルデータｄ１上のレベルＬ_h とこの実測周波数スペクトルデータｄ１の実効値ｆ１とのレベル差ｌを計算することで、最小限の演算処理で効率よく損傷の大きさを推定でき、推定した損傷の大きさから損傷部品の交換時期を決定することで、過剰な部品交換やメンテナンスを回避して、摺動部材を含む機器や設備における維持コストの削減が可能になる。
【００３１】
なお、前記基準レベルＬ₀ には、実効値ｆ１の代わりに、実測周波数スペクトルデータｄ１の平均値を採用するようにしてもよい。
【００３２】
なお、本発明の評価方法及び装置による診断対象となる摺動部材は、上記の実施の形態で示した転がり軸受に限らない。軸受以外の各種の摺動部材を診断対象とすることができ、例えば、ボールねじ、リニアガイド、モータ等を含めることができる。また、摺動部材は機器や設備に組み込んだ状態のままでも、摺動部材の駆動時に発生する音や振動が所定の振動検出手段によって検出できる状況であれば、機器や設備から取り外して、直接診断可能である。
【００３３】
【発明の効果】
以上述べたように、請求項１及び請求項５に記載の本発明の評価方法及び装置によれば、摺動部材を含む機械設備の特定部位の異常時に発生する周波数成分に対応する実測周波数スペクトルデータ上の有効なピークの有無を調べる照合処理は、異常に起因した周波数成分の１次値、２次値、４次値の３回に限定されているため、例えば、１次から高次の周波数成分まで多数の周波数成分に対して照合処理を繰り返す従来技術の場合と比較すると、照合処理時の演算処理量が大幅に低減する。そのため、機械設備から検出した振動信号の解析作業時の演算処理手段への負担が大幅に軽減され、診断作業の迅速化を図ることができる。また、演算処理量が低減したことで、演算処理手段として使用するコンピュータに、演算処理能力が低い安価なコンピュータを使用することが可能になり、装置コストの低減を図ることも可能になる。
【００３４】
更に、異常時に発生する周波数成分の１次値のみで判断すると、たまたまノイズ等の影響で、対応する実測周波数スペクトル上のピークにずれが生じたり、逆にピークが増大しているために誤診断を生じる可能性がある。
しかし、上記のように、異常に起因して発生する周波数成分の１次値、２次値、４次値の３回に渡って照合処理を実施する場合には、３回ともノイズの影響を受ける確率は殆どなく、診断に対する信頼性を向上させることができる。
【図面の簡単な説明】
【図１】本発明に係る評価方法を実現する評価装置の第１の実施の形態の構成を示すブロック図である。
【図２】図１に示した評価装置による処理手順を示すフローチャートである。
【図３】転がり軸受の外輪の傷で異常振動が発生している時の実測周波数スペクトルを示す波形図である。
【図４】転がり軸受における傷の箇所と周波数との関係を示す図である。
【図５】異常時の周波数成分と実測周波数スペクトルデータのピーク箇所との照合処理の処理手順を示すフローチャートである。
【図６】外輪の傷により異常振動が発生している時の実測周波数スペクトル上の周波数成分の照合箇所を示す波形図である。
【図７】転動体の傷により異常振動が発生している時の実測周波数スペクトル上の周波数成分の照合箇所を示す波形図である。
【図８】異常に起因した周波数成分と実測周波数スペクトルとのエンベロープ波形図である。
【図９】転動体表面の剥離の大きさと実測周波数スペクトルに表れるピークの平均レベルのレベル差との相関図である。
【符号の説明】
１評価装置
３１または複数の摺動部材を含む機械設備
５振動検出手段
７増幅手段
９ＡＤ変換手段
１３演算処理手段[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an evaluation method and apparatus for detecting sound or vibration generated from mechanical equipment including one or a plurality of sliding members, analyzing the detected signal, and diagnosing the presence or absence of abnormality caused by the mechanical equipment. More specifically, the present invention relates to an improvement for reducing the burden of analyzing the detected signal and realizing a quick diagnosis work and an improvement in reliability.
[0002]
[Prior art]
Up to now, as an evaluation method for diagnosing the presence or absence of abnormalities such as wear or breakage of a sliding member in a mechanical facility including one or more sliding members such as a rotating body, sound or vibration generated from the mechanical facility has been detected. An evaluation method is known in which a detected signal is appropriately converted into frequency spectrum data by waveform processing, and the presence / absence of an abnormality is diagnosed by comparing and comparing the frequency component resulting from the abnormality of the mechanical equipment with the frequency spectrum data described above. .
[0003]
The frequency component resulting from the abnormality of the mechanical equipment including the sliding member is determined by the design specifications and operation status of the sliding member. For example, it is calculated from the conditions such as the design specifications and operation status of the rotating body described above. It can be determined by processing.
Therefore, in an evaluation apparatus for automatically diagnosing the presence or absence of an abnormality in a mechanical facility, a calculation component such as a computer is used in advance to calculate and process a frequency component resulting from the abnormality of the mechanical facility to be diagnosed, and to generate a mechanical facility. The sound or vibration signal is appropriately waveform processed and converted into frequency spectrum data. Further, the frequency component resulting from the abnormality of the mechanical equipment calculated and stored in advance and the frequency spectrum data described above are stored. Computation processing for comparison and collation is performed.
[0004]
[Problems to be solved by the invention]
By the way, in the conventional evaluation method, for each specific part of the mechanical equipment, a large number of frequency components at the time of abnormality are obtained from the first order to the multi-order, and the frequency spectrum of the mechanical equipment measured for each of these many frequency components An arithmetic process for diagnosing whether or not a peak appears on the data and an arithmetic process for diagnosing whether or not the peak value on the frequency spectrum data is a peak level indicating abnormality are repeated.
For this reason, the calculation processing until final diagnosis is enormous, and the calculation processing means is heavily loaded. Therefore, an expensive computer with high calculation processing capability is required, resulting in an increase in apparatus cost. There is a problem that it is difficult to speed up the diagnostic work due to the length of time required for the arithmetic processing.
[0005]
The present invention has been made in view of the above circumstances, and reduces the burden of arithmetic processing when diagnosing the presence or absence of an abnormality in mechanical equipment including one or a plurality of sliding members, thereby speeding up and improving the reliability of diagnostic work. It is an object of the present invention to provide an evaluation method and apparatus capable of realizing the improvement of the above.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, the evaluation method according to the present invention described in claim 1 detects sound or vibration generated from mechanical equipment including a sliding member, analyzes the detected signal, An evaluation method for diagnosing the presence or absence of an abnormality caused by
Generate measured frequency spectrum data from sound or vibration generated from the mechanical equipment,
An effective value of the actually measured frequency spectrum data is calculated, a threshold is set based on the effective value, and the actually measured frequency spectrum with respect to the first, second, and fourth values of the frequency component caused by the abnormality of the mechanical equipment. A peak on the data is an effective peak only when the threshold value is exceeded, and the effective peak on the measured frequency spectrum data with respect to the first, second, and fourth order values of the frequency component caused by the abnormality of the mechanical equipment is determined . The presence or absence of an abnormality in the mechanical equipment is diagnosed based on the presence or absence.
[0007]
In order to achieve the above object, the evaluation apparatus according to the present invention described in claim 5 detects sound or vibration generated from mechanical equipment including a sliding member, analyzes the detected signal, and originates from the mechanical equipment. An evaluation device for diagnosing the presence or absence of abnormalities,
After generating the measured frequency spectrum data from the sound or vibration generated from the mechanical equipment , calculate the effective value of the measured frequency spectrum data, set a threshold based on the effective value, the said attributed to the abnormality of the mechanical equipment The peak on the actually measured frequency spectrum data for the first, second, and fourth order values of the frequency component is an effective peak only when the threshold value is exceeded, and the first order of the frequency component due to the abnormality of the mechanical equipment, Computational processing means for diagnosing whether there is an abnormality in the mechanical equipment based on the presence / absence of the effective peak on the measured frequency spectrum data with respect to secondary and quaternary values is provided.
[0008]
And according to the evaluation method and apparatus of the structure described in Claim 1 and Claim 5 , the collation process which investigates the presence or absence of the effective peak on the measured frequency spectrum data corresponding to the frequency component resulting from abnormality of a mechanical installation is Because the frequency component is limited to three times, the first, second, and fourth values of the frequency component due to the abnormality of the machine equipment, for example, the verification processing is performed on a large number of frequency components from the first to the higher order frequency components. Compared with the conventional technique to be repeated, the amount of calculation processing at the time of collation processing is reduced, and the burden on the calculation processing means is greatly reduced. Furthermore, since the verification process is performed three times, the primary, secondary, and quaternary values of the frequency component due to the abnormality of the mechanical equipment, the noise is compared with the case where the determination is made only with the primary of the frequency component. It is difficult for misdiagnosis due to the influence of the above and the like, and a highly reliable diagnosis is possible.
[0012]
DETAILED DESCRIPTION OF THE INVENTION
DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of an evaluation method and apparatus according to the invention will be described in detail with reference to the accompanying drawings.
FIG. 1 is a block diagram showing a schematic configuration of a first embodiment of an evaluation method and apparatus according to the present invention, and FIG. 2 is a flowchart showing a procedure of diagnostic processing of the evaluation apparatus shown in FIG.
[0013]
First, after describing the schematic configuration of the evaluation apparatus according to the first embodiment with reference to FIG. 1, the evaluation method using the evaluation apparatus will be described in detail.
The evaluation apparatus 1 according to the present embodiment includes a vibration detection unit 5 that outputs an analog signal corresponding to sound or vibration generated by the mechanical equipment 3 including one or a plurality of sliding members to be diagnosed, and the vibration detection unit. 5, an amplifying means 7 for amplifying the signal output, an AD converting means 9 for converting the analog signal amplified by the amplifying means 7 into a digital signal and generating measured digital data, and an actual measurement output from the AD converting means 9 It is the structure provided with the arithmetic processing means 13 which diagnoses the presence or absence of the abnormality of the specific site | part of the mechanical equipment 3 based on digital data.
[0014]
In the case of the present embodiment, a rolling bearing is applied as a sliding member of the mechanical equipment 3 to be diagnosed. Then, the wear and damage of the inner and outer rings, rolling elements, cages, and the like constituting the rolling bearing are diagnosed from sound or vibration during driving of the rolling bearing.
In the present embodiment, the sound or vibration of the rolling bearing includes ultrasonic vibration that appears when the rolling bearing is driven, so-called AE (Acoustic Emission).
[0015]
The arithmetic processing means 13 is a diagnostic computer that performs arithmetic processing on various processing data stored in advance and actually measured digital data received from the AD conversion means 9 based on a diagnostic program.
This arithmetic processing means 13 generates measured frequency spectrum data by performing appropriate analysis processing such as frequency analysis and envelope analysis on the measured digital data output from the AD conversion means 9, and is caused by an abnormality in the mechanical equipment 3. The presence / absence of an abnormality in the mechanical equipment 3 is diagnosed based on the presence / absence of a peak on the actually measured frequency spectrum data for the first, second, and fourth order values of the frequency component.
[0016]
The above evaluation apparatus 1 performs processing in the procedure shown in FIG.
First, the vibration detection means 5 detects the sound or vibration generated by the mechanical equipment 3 (step S101), and then the signal passed through the amplification means 7 is converted into a digital signal by AD conversion by the AD conversion means 9 (step S102). To the arithmetic processing means 13.
The arithmetic processing unit 13 converts the signal received from the AD conversion unit 9 into a digital file in a file format such as a WAV file (step S103), and performs a filtering process to remove an extra signal if necessary. Go to generate measured digital data.
[0017]
In the case of the present embodiment, the filter process is a process for performing a predetermined process on the input signal by a filter processing program incorporated in advance in the arithmetic processing means 13, and a filter band selecting step for setting a frequency band to be cut in advance ( Step S104) and a filter processing step (Step S105) for cutting off an extra signal in accordance with the selected filter band.
The filtering process in steps S104 and S105 is performed to improve the S / N ratio of the collected data, and is unnecessary if the S / N ratio of the input signal is sufficient.
[0018]
Next, analysis processing such as frequency analysis and envelope analysis is performed on the generated actual measurement digital data (steps S106 and S107) to obtain actual frequency spectrum data d1 that embodies the sound or vibration detected from the mechanical equipment 3 ( Step S108).
The actually measured frequency spectrum data d1 obtained here is a waveform w1 shown in FIG. This waveform w1 is obtained when the inner ring is rotated at 150 revolutions per minute with the outer ring fixed in the rolling bearing as the sliding member.
[0019]
Further, the arithmetic processing means 13 determines the specific part of the mechanical equipment 3 based on the presence or absence of a peak on the measured frequency spectrum data d1 with respect to the first, second, and fourth order values of the frequency component generated when the specific part of the mechanical equipment 3 is abnormal. The presence or absence of abnormality is diagnosed (step S109).
As shown in FIG. 4, the frequency component value generated when a specific part is abnormal is determined for a bearing that is a sliding member according to design specifications and use conditions.
The arithmetic processing means 13 stores the primary, secondary, and quaternary values of the frequency components generated when the specific part shown in FIG. 4 is abnormal for the mechanical equipment 3 as reference values in advance, and based on these reference values. Then, step S109 is performed.
[0020]
In step S109, specifically, the presence or absence of a peak on the measured frequency spectrum data d1 with respect to the primary, secondary, and quadratic values of the frequency component generated when the specific part of the mechanical equipment 3 is abnormal is performed according to the procedure shown in FIG. In the case where both the primary value and the secondary value of the frequency component coincide with the peak on the actually measured frequency spectrum data d1 (steps S201 and S202), or the primary value of the frequency component is In two or more frequency components, such as when the peak on the measured frequency spectrum data d1 does not match the peak on the measured frequency spectrum data d1, but both the quadratic value and the quaternary value match the peak on the measured frequency spectrum data d1 (steps S211, S212). If it is confirmed that a peak exists in the measured frequency spectrum data d1, a diagnosis of abnormality is made for the specific part. (Step S221).
On the other hand, when the frequency component having a peak on the actually measured frequency spectrum data d1 is less than one, vibrations caused by abnormalities in other parts may be affected as noise and may form a peak by chance. Is high and a diagnosis of no abnormality is made (step S231).
[0021]
FIG. 6 shows a frequency component generated due to damage to the outer ring, which is a specific part, with respect to the waveform w1 when the inner ring is rotated at 150 rotations per minute with the outer ring fixed in a rolling bearing as a sliding member. The three frequency components of the primary value Q ₁ , the secondary value Q ₂ , and the quaternary value Q ₄ are appended with broken lines.
In the case of FIG. 3 described above, all the components from the primary value Q ₁ to the high-order Q _n of the frequency components generated due to the damage of the outer ring are appended to the same waveform w1 by broken lines. .
[0022]
In the evaluation method performed by the evaluation apparatus 1 according to the first embodiment described above, the verification process for examining the presence or absence of a peak on the measured frequency spectrum data d1 corresponding to the frequency component generated when the specific part of the mechanical equipment 3 is abnormal is as follows: Since the frequency component is limited to three times of the primary, secondary, and quaternary values due to the abnormality of the mechanical equipment 3, for example, as shown in FIG. Compared to the case of the prior art in which the matching process is repeated for all frequency components, the amount of calculation processing during the matching process is greatly reduced.
[0023]
Therefore, the burden on the arithmetic processing means 13 at the time of the analysis work of the signal detected from the mechanical equipment 3 is greatly reduced, and the diagnosis work can be speeded up. In addition, since the amount of arithmetic processing is reduced, it is possible to use an inexpensive computer with low arithmetic processing capability as the computer used as the arithmetic processing means 13, and it is also possible to reduce the apparatus cost.
[0024]
Further, if only the first order of the frequency component generated at the time of abnormality is determined, the peak on the corresponding actually measured frequency spectrum is shifted due to the influence of noise or the like. Can occur.
However, as described above, in the case where the verification process is performed over three times of the primary value, the secondary value, and the quaternary value of the frequency component generated when the specific part of the mechanical equipment 3 is abnormal, all three times There is almost no probability of being affected by noise, and therefore the reliability for diagnosis can be improved by performing only two matching processes.
[0025]
Preferably, the arithmetic processing means 13 calculates the effective value f1 of the actually measured frequency spectrum data d1 after generating the actually measured frequency spectrum data d1, and sets a threshold value t1 based on the effective value f1, The peak on the measured frequency spectrum data d1 with respect to the primary value Q ₁ , the secondary value Q ₂ , and the quaternary value Q ₄ of the frequency component generated when the specific part 3 is abnormal is a peak that is effective only when the threshold value t1 is exceeded. It is good to have a configuration to handle as
[0026]
FIG. 7 shows an effective value f1 and a threshold value t1 for the waveform w2 of the actually measured frequency spectrum data d1 when the inner ring is rotated at 150 rotations per minute with the outer ring fixed in the rolling bearing as the sliding member. It is written. Further, the primary value q ₁ , the secondary value q ₂ , and the quaternary value q ₄ of the frequency component generated due to scratches on the rolling elements of the rolling bearing as the sliding member are written on the waveform w2 by dotted lines. .
In this case, the effective value f1 is obtained by calculating the average level of the amplitude of the waveform w2, and is −8.5 dB.
The threshold t1 is t1 = (f1 + 10 dB) (1)
Therefore, the threshold value t1 is 1.5 dB.
In this example, rolling primary value q ₁ wound on due to generate frequency components of the moving object (2 fb), of the secondary value q _2, 4-order value q _4, secondary value q ₂ is the threshold t1 Since it is smaller and buried in noise, it indicates that the collation process is necessary only for the primary value q ₁ and the quaternary value q _{4 which} are larger than the threshold value t1.
[0027]
When the useful peak can be selected based on the threshold value t1, the frequency component primary value, the secondary value, and the quaternary value that are generated due to an abnormality in a specific portion of the sliding member can be used. If an extraction process for extracting an effective peak with the threshold value t1 is performed before performing the calculation process for matching on the peak on the actually measured frequency spectrum data, the waste of performing the matching process on the insignificant peak is eliminated. It can be omitted, and the burden due to the amount of calculation processing can be further reduced, and the speeding up of the diagnostic processing can be promoted.
[0028]
In the above embodiment, the case where the presence or absence of damage of a specific part is diagnosed is shown.
However, as described above, when the measured frequency spectrum data d1 is generated by performing appropriate analysis processing such as frequency analysis and envelope analysis on the measured digital data, for example, as shown in FIG. The effective value f1 of the data d1 is calculated, the calculated effective value f1 is set to the reference level L _0, and a frequency component generated due to an abnormality in a specific part of the mechanical equipment 3 (indicating outer ring damage in this example) From the magnitude of the level difference l between the level L _h and the reference level L ₀ on the actually measured frequency spectrum data d1 with respect to the primary value Q ₁ , the magnitude of damage in the outer ring causing the abnormality can be estimated.
FIG. 8 shows an envelope waveform of a rotating body that is a sliding member.
[0029]
FIG. 9 shows the magnitude of separation and the level difference between the peak appearing on the actually measured frequency spectrum data d1 and the reference level when separation, which is damage to the raceway ring, occurs in a rolling bearing as a sliding member. This shows the relationship.
Thus, since the level difference generally increases in proportion to the magnitude of damage, conversely, the magnitude of damage is estimated by obtaining the level difference at the peak on the actually measured frequency spectrum data d1. Is possible.
Moreover, the increase in the peak level on the actually measured frequency spectrum data d1 due to the damage of the mechanical equipment 3 becomes most significant at the peak corresponding to the primary value of the frequency component due to the abnormality.
[0030]
Therefore, as described above, mechanical equipment 3 of the effective value of the measured frequency spectrum data d1 level L _h Toko on the measured frequency spectrum data d1 to the primary value of the abnormality due to occurrence frequency components at a specific site By calculating the level difference l from f1, it is possible to efficiently estimate the magnitude of damage with a minimum of arithmetic processing, and by determining the replacement timing of damaged parts from the estimated magnitude of damage, excessive parts replacement And maintenance can be avoided, and the maintenance cost of equipment and facilities including sliding members can be reduced.
[0031]
Note that the average value of the actually measured frequency spectrum data d1 may be adopted as the reference level L ₀ instead of the effective value f1.
[0032]
The sliding member to be diagnosed by the evaluation method and apparatus of the present invention is not limited to the rolling bearing shown in the above embodiment. Various sliding members other than the bearing can be targeted for diagnosis, and for example, a ball screw, a linear guide, a motor, and the like can be included. In addition, even if the sliding member is in the state where it is incorporated in the device or equipment, if the sound or vibration generated when the sliding member is driven can be detected by a predetermined vibration detecting means, it can be removed from the device or equipment and directly Diagnosis is possible.
[0033]
【The invention's effect】
As described above, according to the evaluation method and apparatus of the present invention described in claims 1 and 5 , the actually measured frequency spectrum corresponding to the frequency component generated at the time of abnormality of a specific part of the mechanical equipment including the sliding member. The verification process for checking the presence or absence of a valid peak on the data is limited to three times of the primary value, the secondary value, and the quaternary value of the frequency component caused by the abnormality. compared to the prior art to repeat the verification process to a number of frequency components to frequency components, the arithmetic processing amount at the time of collation processing can be greatly reduced. Therefore, the burden on the arithmetic processing means at the time of the analysis work of the vibration signal detected from the mechanical equipment is greatly reduced, and the diagnosis work can be speeded up. In addition, since the amount of calculation processing is reduced, it is possible to use an inexpensive computer with low calculation processing capability as the computer used as the calculation processing means, and it is also possible to reduce the apparatus cost.
[0034]
Furthermore, if only the primary value of the frequency component generated at the time of abnormality is judged, it will happen that the peak on the corresponding measured frequency spectrum is shifted due to the influence of noise, etc. May occur.
However, as described above, when the verification process is performed three times, ie, the primary value, the secondary value, and the quaternary value of the frequency component generated due to the abnormality, the influence of noise is affected all three times. There is almost no probability of receiving, and the reliability for diagnosis can be improved.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a first embodiment of an evaluation apparatus that implements an evaluation method according to the present invention.
FIG. 2 is a flowchart showing a processing procedure performed by the evaluation apparatus shown in FIG.
FIG. 3 is a waveform diagram showing an actually measured frequency spectrum when abnormal vibration occurs due to scratches on the outer ring of the rolling bearing.
FIG. 4 is a diagram showing a relationship between a flawed location and a frequency in a rolling bearing.
FIG. 5 is a flowchart showing a processing procedure for collation processing between an abnormal frequency component and a peak portion of measured frequency spectrum data.
FIG. 6 is a waveform diagram showing a location where frequency components are compared on an actually measured frequency spectrum when abnormal vibration is generated due to scratches on the outer ring.
FIG. 7 is a waveform diagram showing a location where frequency components are compared on an actually measured frequency spectrum when abnormal vibration occurs due to scratches on rolling elements.
FIG. 8 is an envelope waveform diagram of a frequency component caused by an abnormality and an actually measured frequency spectrum.
FIG. 9 is a correlation diagram between the magnitude of peeling on the surface of a rolling element and the level difference of the average level of peaks appearing in the actually measured frequency spectrum.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Evaluation apparatus 3 Mechanical equipment 5 including one or more sliding members Vibration detection means 7 Amplification means 9 AD conversion means 13 Arithmetic processing means

Claims

It is an evaluation method for detecting sound or vibration generated from mechanical equipment including a sliding member, analyzing the detected signal, and diagnosing the presence or absence of an abnormality caused by the mechanical equipment,
Generate measured frequency spectrum data from sound or vibration generated from the mechanical equipment,
An effective value of the actually measured frequency spectrum data is calculated, a threshold is set based on the effective value, and the actually measured frequency spectrum with respect to the first, second, and fourth values of the frequency component caused by the abnormality of the mechanical equipment. A peak on the data is an effective peak only when the threshold value is exceeded, and the effective peak on the measured frequency spectrum data with respect to the first, second, and fourth order values of the frequency component caused by the abnormality of the mechanical equipment is determined . An evaluation method comprising diagnosing the presence / absence of an abnormality in the mechanical equipment based on presence / absence.

An abnormality diagnosis is performed when two or more of the effective peaks on the measured frequency spectrum data match the first, second, and fourth order values of the frequency component caused by the abnormality of the mechanical equipment. The evaluation method according to claim 1.

The effective peak is obtained by extracting a peak on the measured frequency spectrum data exceeding the threshold value, and then, with respect to the extracted peak, the primary, secondary, and quadratic frequency components due to the abnormality of the mechanical equipment. 3. The evaluation method according to claim 1, wherein the values are collated.

The evaluation method according to claim 1, wherein the sliding member is any one of a rolling bearing, a ball screw, and a linear guide.

An evaluation device that detects sound or vibration generated from mechanical equipment including a sliding member, analyzes the detected signal, and diagnoses the presence or absence of an abnormality caused by the mechanical equipment,
After generating the measured frequency spectrum data from the sound or vibration generated from the mechanical equipment , calculate the effective value of the measured frequency spectrum data, set a threshold based on the effective value, the said attributed to the abnormality of the mechanical equipment The peak on the actually measured frequency spectrum data for the first, second, and fourth order values of the frequency component is an effective peak only when the threshold value is exceeded, and the first order of the frequency component due to the abnormality of the mechanical equipment, An evaluation apparatus comprising: arithmetic processing means for diagnosing the presence / absence of an abnormality in the mechanical equipment based on the presence / absence of the effective peak on the measured frequency spectrum data with respect to secondary and quaternary values.

An abnormality diagnosis is performed when two or more effective peaks on the measured frequency spectrum data coincide with the primary, secondary, and quaternary values of the frequency components caused by the abnormality of the mechanical equipment. The evaluation apparatus according to claim 5.

The effective peak is obtained by extracting a peak on the measured frequency spectrum data exceeding the threshold value, and then, with respect to the extracted peak, the primary, secondary, and quadratic frequency components due to the abnormality of the mechanical equipment. The evaluation device according to claim 5 or 6, wherein values are collated.

The evaluation apparatus according to claim 5, wherein the sliding member is one of a rolling bearing, a ball screw, and a linear guide.