鍩轰簬
BP
绁炵粡缃戠粶鐨勬粴鍔ㄨ酱鎵挎晠闅滆瘖鏂?/p>
鍒樺厓鏄紝闄堝缓鏀?/p>
锛堣タ鍗椾氦閫氬ぇ瀛︾壍寮曞姩鍔涘浗瀹堕噸鐐瑰疄楠屽锛屾垚閮?/p>
610000
锛?
鎽樿閽堝婊氬姩杞存壙鐨勬晠闅滄崯浼ら毦浠ユ櫤鑳借瘖鏂殑闂锛?/p>
鎻愬嚭浜嗕竴绉嶅熀浜?/p>
Alpha
绋冲畾鍒嗗竷鍙傛暟浼拌鍜岀缁忕綉缁滅殑婊氬姩杞存壙
鏁呴殰璇婃柇鏂规硶銆傞鍏堬紝瀵瑰悇涓姸鎬佹尟鍔ㄤ俊鍙疯繘琛?/p>
绋冲畾鍒嗗竷鐨?/p>
4
鍙傛暟浼拌锛岄�夊彇鏁忔劅鎬у強绋冲畾鎬ф渶濂界殑浜岀鍙傛暟缁?/p>
鎴愪簩缁存晠闅滅壒寰侀噺锛涚劧鍚庯紝鍙皢璇ヤ簩缁存晠闅滅壒寰侀噺浣滀负
BP
绁炵粡缃戠粶鐨勮緭鍏ュ弬鏁版潵璇嗗埆婊氬姩杞存壙鐨勬晠闅滅被鍨嬶紱鐒跺悗锛?/p>
閫氳繃鍙版灦璇曢獙鏁版嵁楠岃瘉浜嗚鏂规硶鐨勬湁鏁堟�э紝骞朵笌
4
涓父鐢ㄧ殑鏃犻噺绾叉椂鍩熺壒寰佸弬鏁颁綔涓?/p>
BP
绁炵粡缃戠粶鐨勮緭鍏ュ弬鏁扮殑鏂规硶
鐨勮瘖鏂粨鏋滆繘琛屾瘮杈冦�?/p>
缁撴灉琛ㄦ槑锛?/p>
鍩轰簬
Alpha
绋冲畾鍒嗗竷鍙傛暟浼拌鏂规硶鍙疄鐜板婊氬姩杞存壙鏁呴殰浣嶇疆鍙婄殑鏅鸿兘璇婃柇锛?/p>
涓庢椂
鍩熺壒寰佸弬鏁扮殑鏂规硶鐩告瘮锛屽湪鏇村皯鐨勫弬鏁颁笅瀹炵幇浜嗘洿鍑嗙‘鏇存湁鏁堢殑鏁呴殰璇嗗埆锛涙渶鍚庯紝搴旂敤宸ョ▼瀹炴祴鏁版嵁瀵硅妯″瀷杩涜
浜嗛獙璇侊紝缁撴灉琛ㄦ槑鍦ㄥ皬鏍锋湰锛屼綆杞�熸潯浠朵笅璇ユā鍨嬩篃鑳藉婊氬姩杞存壙涓嶅悓鐘舵�佹晠闅滆繘琛屾湁鏁堢殑妫�娴嬶紝鏈夋湜鍦ㄥ疄闄呭伐绋?/p>
涓緱鍒板簲鐢ㄣ�?/p>
鍏抽敭璇嶏細婊氬姩杞存壙锛涙晠闅滆瘖鏂紱
绋冲畾鍒嗗竷锛?/p>
BP
绁炵粡缃戠粶
Rolling bearing fault diagnosis based on BP neural network
Liuyuanshi,Chenjianzheng
(Traction Power State Key Laboratory of Southwest Jiaot Tong University,Chengdu 610000,China)
Abstract:
In this paper, rolling bearing fault intelligent diagnosis of damage difficult problem proposed rolling bearing fault
diagnosis method based on Alpha stable distribution parameter estimation and neural networks. First, each state vibration
signal
stable
distribution
four-parameter
estimation,
select
the
best
sensitivity
and
stability
of
two-dimensional
composition
fault
characteristic
parameters;
then,
the
fault
may
be
the
two-dimensional
feature
amount
as
input
parameters BP neural network failure to identify the type of rolling bearings; diagnosis and then by bench test data to
verify the effectiveness of the method, and with the four common dimensionless time domain characteristic parameters as
input parameters BP neural network method were compared. The results show that: based on Alpha stable distribution
parameter estimation method can be realized on Rolling intelligent diagnostics and fault location, compared with the time
domain
method
of
characteristic
parameters
in
fewer
parameters
to
achieve
a
more
accurate
and
efficient
fault
identification; Finally, engineering test data to validate the model results show that in small samples, low-speed conditions
of
the
rolling
bearing
a
different
model
can
effectively
detect
the
fault
state,
it
is
expected
to
be
applied
in
practical
engineering.
Keywords:
Rolling; fault diagnosis; stable distribution; BP neural network
寮曡█
婊氬姩杞存壙鍦ㄦ満姊板伐涓氳澶囦腑搴旂敤闈炲父鏅亶锛屼絾鍙堟槸闈炲父瀹规槗鎹熷潖銆傛粴鍔ㄨ酱鎵跨殑杩愯鐘舵�佽兘鐩存帴褰卞搷鏁翠釜鏈哄櫒璁惧鐨?/p>
鐘舵�併�傚ぇ绾︽湁涓夊垎涔嬩竴鐨勬棆杞満姊拌澶囨晠闅滄槸鐢辫酱鎵挎崯鍧忓紩璧风殑銆備紶缁熺殑璇婃柇鏂规硶濡傚寘缁滄娉㈠拰鍏辨尟瑙ｈ皟娉曡櫧鐒跺疄
鐜颁簡璇婃柇绮惧害鐨勬瀬澶ф彁楂橈紝浣嗘槸浜哄伐杈呭姪璇婃柇渚濈劧涓嶅彲缂哄皯锛屽浜庡鏉傜幆澧冧笅婊氬姩杞存壙绮剧‘璇婃柇浠荤劧闅句互瀹炵幇銆傚洜
姝わ紝寰堟湁蹇呰浣跨敤鏅鸿兘鐨勬柟娉曞婊氬姩杞存壙杩涜鐘舵�佹娴嬪拰鏁呴殰璇婃柇銆?/p>
浠庢湰璐ㄤ笂鏉ヨ锛屾粴鍔ㄨ酱鎵跨殑鏁呴殰璇婃柇灏辨槸瀵逛簬鍚勭妯″紡鐨勮瘑鍒繃绋嬨�?/p>
BP
绁炵粡缃戠粶鎷ユ湁姣旇緝濂界殑鑷涔犺嚜閫傚簲鑳藉姏锛?/p>
闈炵嚎鎬ф槧灏勮兘鍔涳紝瀹归敊鑳藉姏浠ュ強娉涘寲鑳藉姏锛岃兘澶熷湪婊氬姩杞存壙鐨勬晠闅滆瘖鏂腑鍙戞尌寰堝ソ鐨勪綔鐢ㄣ�?/p>
浣嗘槸锛屽湪婊氬姩杞存壙涓晠闅滅壒寰佸弬鏁扮殑鎻愬彇鏂规硶鍦ㄥ緢澶х殑绋嬪害涓婂喅瀹氫簡璇婃柇鐨勫噯纭�с�傝繍鐢ㄥ父鐢ㄧ殑鏃犻噺绾茬殑鏃跺煙鐗瑰緛
鍙傛暟浣滀负鏁呴殰鐗瑰緛鍙傛暟锛?/p>
鍙互姣旇緝濂界殑瀵规粴鍔ㄨ酱鎵胯繘琛屾晠闅滆瘑鍒紝
浣嗕篃瀛樺湪鐫�闇�瑕佸弬鏁拌繃澶氫互鍙婂噯纭巼涓嶉珮绛夐棶棰樸�?/p>
绋冲畾鍒嗗竷鏄竴绉嶉潪甯搁噸瑕佺殑闈為珮鏂垎甯冿紝
鍦ㄦ弿杩板皷宄拌剦鍐茬姸鍜屾湁涓ラ噸鎷栧熬鐗规�х殑闈為珮鏂俊鍙风殑鏂归潰鏈変紭鑹殑椴佹
鎬с�?/p>
绋冲畾鍒嗗竷鏂规硶鍘熸湰涓昏搴旂敤浜庤闊充俊鍙峰鐞嗐�佺敓鐗╁尰瀛︿俊鍙峰鐞嗐�侀�氫俊璇嗗埆绛夐鍩燂紝杩戜簺骞存潵涔熻閫愭笎杩愮敤
鍒版棆杞満姊扮殑鏁呴殰璇婃柇褰撲腑銆?/p>
浣嗘槸灏?/p>
绋冲畾鍒嗗竷鏂规硶涓庣缁忕綉缁滃垎绫绘柟娉曠粨鍚堬紝
浠庤�屽缓绔嬭捣鏅鸿兘鐨勬粴鍔ㄨ酱鎵挎晠闅滆瘖
鏂ā鍨嬬殑鐮旂┒杩樺崄鍒嗗皯瑙併�?/p>
鏂囩珷瀵?/p>
绋冲畾鍒嗗竷鍥涚被鍙傛暟杩涜瀵规瘮鐮旂┒鍒嗘瀽锛?/p>
骞堕�夊彇浜嗗叾涓晱鎰熸�у拰绋冲畾鎬ч兘闈炲父浼樺紓鐨勪袱绫诲弬鏁颁綔涓轰簩缁?/p>
鏁呴殰鐗瑰緛閲忎笌
BP
绁炵粡缃戠粶鐩哥粨鍚堬紝寤虹珛浜嗕竴绉嶆粴鍔ㄨ酱鎵跨殑鏁呴殰鏅鸿兘璇婃柇妯″瀷锛屽苟涓庡父鐢ㄧ殑鏃犻噺绾叉椂鍩熺壒寰佸弬鏁颁綔
涓烘晠闅滅壒寰佸弬鏁拌繘琛屽姣旂爺绌躲�傝瘯楠屽垎鏋愯〃鏄庯紝姝ゆ柟娉曞彲瀹炵幇婊氬姩杞存壙涓嶅悓鐨勬晠闅滀綅缃紙姝ｅ父銆佸鍦堛�佸唴鍦堛�佹粴鍔?/p>
浣擄級鐨勬櫤鑳借瘖鏂紝涓斿熀浜?/p>
绋冲畾鍒嗗竷鐨勬柟娉曡兘澶熷疄鐜颁娇鐢ㄨ緝灏戝弬鏁板苟寰楀埌鏇撮珮鐨勮瘖鏂噯纭巼锛屾湁鏈涘湪宸ョ▼瀹為檯涓緱
飦?/p>
飦?/p>
飦?/p>
飦?/p>
飦?/p>
飦?/p>
飦?/p>