JP3818237B2

JP3818237B2 - Control device for synchronous motor

Info

Publication number: JP3818237B2
Application number: JP2002230881A
Authority: JP
Inventors: 満博正治; 卓明苅込
Original assignee: Nissan Motor Co Ltd
Current assignee: Nissan Motor Co Ltd
Priority date: 2002-08-08
Filing date: 2002-08-08
Publication date: 2006-09-06
Anticipated expiration: 2022-08-08
Also published as: JP2004072931A

Description

【０００１】
【発明の属する技術分野】
本発明は同期電動機の制御装置に関し、例えば回転子に永久磁石を備えた三相同期電動機における制御技術に関する。
【０００２】
【従来の技術】
従来の電動機制御装置としては、例えば特開２０００−２２４８８３号公報に記載されているものがある。この従来例ではトルク指令値を満足するｄ軸電流およびｑ軸電流を発生するためのｄ軸電圧指令値およびｑ軸電圧指令値を算出し、ｄ軸およびｑ軸の実電流とｄ軸およびｑ軸の電流指令値との差から上記電圧指令値をＰＩ（比例積分）制御で補正する構成が記載されている。
【０００３】
【発明が解決しようとする課題】
上記従来技術のように制御した場合には、ｄ軸電流およびｑ軸電流に振動が発生する。これを防止するためにｄ軸電圧指令値を入力し、入力したｄ軸電圧指令値からｄ軸電流の振動成分を除去して出力するフィルタ（例えばローパスフィルタ）およびｑ軸電圧指令値を入力し、入力したｑ軸電圧指令値からｑ軸電流の振動成分を除去して出力するフィルタをそれぞれ設けることが考えられる。しかしながら、ｄ軸電圧指令値が変化した場合には、ｄ軸電流だけでなく、ｑ軸電流にも逆起電力（又は速度起電力とも言う）による変化が発生し、また、ｑ軸電圧指令値が変化した場合ｑ軸電流だけでなく、ｄ軸電流にも逆起電力によって変化が発生する。従って、ｄ軸電圧指令値およびｑ軸電圧指令値の電流振動成分をそれぞれ独立して除去しても、電流振動が発生するという問題が有った。
【０００４】
本発明は上記のごとき従来技術の問題を解決するためになされたものであり、電流応答における振動成分を抑制して電流応答性を向上させた同期電動機の制御装置を提供することを目的とする。
【０００５】
【課題を解決するための手段】
本発明においては、電圧指令値算出手段によって算出されたｄ軸電圧指令値およびｑ軸電圧指令値を入力し、入力されたｄ軸電圧指令値およびｑ軸電圧指令値の変化に起因してｄ軸電流およびｑ軸電流に発生する振動成分を除去するための電圧指令値補正フィルタとして、
ｄ軸電圧指令値の変化に起因してｄ軸電流に発生する振動成分を除去するように、ｄ軸電圧指令値を補正する第一の振動除去手段（Ｃ_１１）と、
ｄ軸電圧指令値の変化に起因してｑ軸電流に発生する振動成分を除去するように、ｄ軸電圧指令値に基づいてｑ軸電圧指令値を補正する第二の振動除去手段（Ｃ_２１）と、
ｑ軸電圧指令値の変化に起因してｑ軸電流に発生する振動成分を除去するように、ｑ軸電圧指令値を補正する第三の振動除去手段（Ｃ_２２）と、
ｑ軸電圧指令値の変化に起因してｄ軸電流に発生する振動成分を除去するように、ｑ軸電圧指令値に基づいてｄ軸電圧指令値を補正する第四の振動除去手段（Ｃ_１２）と、
を備えるように構成している。
【０００６】
【発明の効果】
本発明においては、ｄ軸電圧指令値の変化に起因してｄ軸電流に発生する振動成分とｑ軸電流に発生する振動成分、およびｑ軸電圧指令値の変化に起因してｑ軸電流に発生する振動成分とｄ軸電流に発生する振動成分の全てを除去するようにそれぞれの補正電圧指令値を設定しているので、ｄ軸電圧指令値およびｑ軸電圧指令値から電流振動成分を除去して、電流応答性を確実に向上させることができる。
【０００７】
【発明の実施の形態】
図１は、本発明の第１の実施例を示すブロック図である。
図１において、クロスフィルタ１（電圧指令値補正フィルタ）の各フィルタ（Ｃ_１１）、（Ｃ_１２）、（Ｃ_２１）、（Ｃ_２２）の特性は、下記（数１）式〜（数４）式で示される。
【０００８】
【数１】

【０００９】
【数２】

【００１０】
【数３】

【００１１】
【数４】

ただし、ｇ(ｓ）：電流応答特性
ｓ：ラプラス演算子
Ｒ：電動機の電機子巻線抵抗
Ｌ_ｄｃ、Ｌ_ｑｃ：電動機のインダクタンス推定値
ω：電気角速度
また、クロスフィルタ１の電流応答特性ｇ(ｓ）とフィルタ部２のローパスフィルタの特性ｇ_ｆｌｔ(ｓ）との関係は、下記（数５）式に示すようになる。
【００１２】
【数５】

指令値決定部３では、まず、外部から入力されたトルク指令値Ｔ^＊および現在の回転速度ωを指標として、電流マップ３０１（電流指令値算出手段）を用いてマップ引きによりｄ軸電流指令値ｉ_ｄ ^＊およびｑ軸電流指令値ｉ_ｑ ^＊を求める。そして電圧指令値演算部３０２（電圧指令値算出手段）では入力されたｄ軸電流指令値ｉ_ｄ ^＊およびｑ軸電流ｉ_ｑ ^＊から、下記（数６）式によってｄ軸電圧指令値ｖ_ｄ’およびｑ軸電圧指令値ｖ_ｑ’を演算して出力する。
【００１３】
【数６】

ただし、φ：磁石による巻線鎖交磁束数
フィルタ部２（補正値算出手段）は、例えば１次遅れのローパスフィルタで構成し、ｄ軸電流指令値ｉ_ｄ ^＊およびｑ軸電流指令値ｉ_ｑ ^＊の変化速度を緩和し、緩和後のｄ軸電流指令値ｉ_ｄ’およびｑ軸電流指令値ｉ_ｑ’として出力する。
フィルタ部２のローパスフィルタの特性ｇ_ｆｌｔ(ｓ）は下記（数７）式で示される。
【００１４】
【数７】

ただし、α：ｇ(ｓ）を一次遅れとした場合のカットオフ周波数
クロスフィルタ１では、指令値決定部３からのｄ軸電圧指令値ｖ_ｄ’およびｑ軸電圧指令値ｖ_ｑ’を入力し、前記（数１）式〜（数４）式の特性を有するクロスフィルタを通すことにより、下記（数８）式、（数９）式に示す補正電圧指令値（ｖ_ｄ１、ｖ_ｑ１）を出力する。
【００１５】
【数８】

【００１６】
【数９】

一方、座標変換器１１は、電流検出器８で検出した二相の電流ｉ_ｕ、ｉ_ｖから求めた三相の検出電流（実電流）ｉ_ｕ、ｉ_ｖ、ｉ_ｗと、位置検出器１０で検出した検出位置（回転子の回転位相）とに基づいて、ｄ軸電流（磁束電流）ｉ_ｄとｑ軸電流（トルク電流）ｉ_ｑを求める。なお、電流検出器８、位置検出器１０および座標変換器１１の部分が実電流検出手段に相当する。
【００１７】
ＰＩ制御部４（補正値算出手段）では、フィルタ部２から送られた緩和後のｄ軸電流指令値ｉ_ｄ’およびｑ軸電流指令値ｉ_ｑ’と座標変換器１１から送られたｄ軸電流ｉ_ｄとｑ軸電流ｉ_ｑとの差分を増幅し、下記（数１０）式、（数１１）式に示す電圧補正値（ｖ_ｄ２、ｖ_ｑ２）を生成する。ただしパラメータ誤差がなく、電流応答特性はｇ_ｆｌｔ(ｓ）が得られている場合、電流の差分は０なので（ｖ_ｄ２、ｖ_ｑ２）は０である。
【００１８】
【数１０】

【００１９】
【数１１】

ただし、ｇ_ｐｉｄ、ｇ_ｐｉｑ：ＰＩ制御部４におけるＰＩゲイン
上記のようにして求められたクロスフィルタ処理後の補正電圧指令値（ｖ_ｄ１、ｖ_ｑ１）は電圧補正値（ｖ_ｄ２、ｖ_ｑ２）との和を求めることによってさらに補正を行い、補正後の補正電圧指令値（ｖ_ｄ、ｖ_ｑ）として電圧座標変換器５へ送られる。そして電圧座標変換器５で三相交流ｖ_ｕ、ｖ_ｖ、ｖ_ｗに変換され、ＰＷＭ変換器６でＰＷＭ信号に変換される。そのＰＷＭ信号でインバータ７を制御することにより、三相の交流電力を作り、それによって電動機９を駆動する。なお、電圧座標変換器５、ＰＷＭ変換器６およびインバータ７の部分が電圧制御手段に相当する。
上記補正後の補正電圧指令値（ｖ_ｄ、ｖ_ｑ）は下記（数１２）式、（数１３）式で示される。
【００２０】
【数１２】

【００２１】
【数１３】

次に、作用を説明する。
定常状態ではクロスフィルタ１の特性は下記（数１４）式のように表すことができる。
【００２２】
【数１４】

（数１４）式において、ｃ_１２とｃ_２１は、ｄ軸電圧指令値ｖ_ｄ’およびｑ軸電圧指令値ｖ_ｑ’が変化する過渡的な状態にのみ適用されるフィルタとなっているため、この場合は０となる。
過渡状態では、ｄ軸電圧指令値ｖ_ｄ’のみが変化した場合、クロスフィルタ１の特性は下記（数１５）式のように表すことができる。
【００２３】
【数１５】

ただし、Ｌ_ｄｃ、Ｌ_ｑｃ：制御に用いる電動機のインダクタンス値（測定値又は推定値）
補正電圧指令値ｖ_ｄのみを操作すると、電流応答には必ず振動が起こるため、補正電圧指令値ｖ_ｄを操作する場合には必ず補正電圧指令値ｖ_ｑも操作する必要がある。そのためｃ_１１によって生成されるｄ軸の補正電圧指令値ｖ_ｄのほかにｃ_２１によるｑ軸の補正電圧指令値ｖ_ｑも変化させ、ｄ軸電圧指令値ｖ_ｄ’の変化およびこれに起因するｄｑ軸間の干渉成分によりｄ軸電流およびｑ軸電流に発生する電動機に固有の電気的振動を抑制している。
また、ｑ軸電圧指令値ｖ_ｑ’が変化した場合、クロスフィルタ１の特性は下記（数１６）式のように表すことができ、上記のｄ軸と同様にして振動を抑制している。
【００２４】
【数１６】

以上の作用をまとめて数式で表すと、以下のようになる。
パラメータ誤差が無い場合、電圧座標変換器５に入力する補正電圧指令値（ｖ_ｄ、ｖ_ｑ）は、下記（数１７）式、（数１８）式に示すようになる。
【００２５】
【数１７】

【００２６】
【数１８】

また、過渡応答を含んだ電動機の電圧方程式は下記（数１９）式に示すようになる。
【００２７】
【数１９】

ただし、Ｌ_ｄｍ、Ｌ_ｑｍ：電動機の実際のインダクタンス値
ここで、（数６）式、（数１９）式を下記（数２０）式、（数２１）式のように次元を拡張し、クロスフィルタ１の入出力の関係を下記（数２２）式のように整理する。
【００２８】
【数２０】

【００２９】
【数２１】

【００３０】
【数２２】

（数２２）式において、両辺に左からＢの逆行列Ｂ^−１をかけると下記（数２３）式、（数２４）式となる。
【００３１】
【数２３】

【００３２】
【数２４】

入力する補正電圧指令値（ｖ_ｄ、ｖ_ｑ）にかかるＢ^−１の要素は分母に二次遅れ成分１／Ｇ(ｓ）が含まれており、クロスフィルタ１がない場合にはｄ軸電圧指令値ｖ_ｄ’およびｑ軸電圧指令値ｖ_ｑ’に対する電波応答は振動的になる。
これに対して、クロスフィルタ１を用いた場合、ｄ軸電圧指令値ｖ_ｄ’およびｑ軸電圧指令値ｖ_ｑ’とｄ軸電流およびｑ軸電流の応答の関係は、下記（数２５）式、（数２６）式に示すようになり、二次遅れ成分１／Ｇ（ｓ）は除去されるので振動的にはならない。
【００３３】
【数２５】

【００３４】
【数２６】

また、ｄ軸電流指令値ｉ_ｄ ^＊およびｑ軸電流指令値ｉ_ｑ ^＊応答の関係は、下記（数２７）式、（数２８）式、（数２９）式に示すようになる。
【００３５】
【数２７】

【００３６】
【数２８】

【００３７】
【数２９】

以上説明したように、（数１）式〜（数４）式によるクロスフィルタ１を用いることによって、ｄ軸電圧指令値ｖ_ｄ’およびｑ軸電圧指令値ｖ_ｑ’に対する電流応答の振動成分は除去され、かつ、電流応答特性はｇ(ｓ）となり、ｇ(ｓ）を（数７）式のように設計した場合は応答周波数α［ｒａｄ／ｓ］の滑らかな一次遅れの電流応答が得られる。
【００３８】
図２は、本発明の第２の実施例を示すブロック図である。
図１と異なる点は、図１の電圧指令値演算部３０２の代わりに、電圧マップ３０３を備え、トルク指令値Ｔ^＊と回転速度ωからマップ引きにより電圧指令値（ｖ_ｄ’、ｖ_ｑ’）を求める点である。
図２において、電圧マップ３０３は、電動機の評価試験の際に、トルク指令値Ｔ^＊と回転速度ωを指標とした、定常運転時のｄ軸電圧ｖ_ｄおよびｑ軸電圧ｖ_ｑによって作成する。このマップを用いてｄ軸電圧指令値ｖ_ｄ’およびｑ軸電圧指令値ｖ_ｑ’を求める。
次に、作用を説明する。
電圧マップ３０３には、電流によるインダクタンス値の変動分が含まれるので、電流応答はパラメータ誤差の影響を受けにくくなる。
【００３９】
以下、第１および第２の実施例における制御の実例について説明する。
図３〜図８は、それぞれステップ入力時の電流応答の一例を示す図であり、図３〜図６は本発明における特性、図７、図８は比較のために示した従来例の特性である。
従来例においては、図７に示すように、ｄ軸電流指令値ｉ_ｄ ^＊およびｑ軸電流ｉ_ｑ ^＊が実線で示すようにステップ状に変化した場合、実際のｄ軸電流ｉ_ｄとｑ軸電流ｉ_ｑは破線で示すように大きく振動する。また、ｄ軸電圧ｖ_ｄおよびｑ軸電圧ｖ_ｑは図示のようにステップ状に変化し、かつ振動する。
また、図８は、従来例において、ＰＩフィードバックループのゲインを出来るだけ大きくした場合の特性である。この場合には図７よりは少ないが、やはり振動が残っている。従来例において、電動機に固有の電気的振動成分のような、周波数が大きい電流誤差を補正するためには、ＰＩフィードバックループのゲインを充分に大きくしなくてはならないが、ＰＩゲインを大きくするに従って、ノイズまで増幅してしまう。実際の制御は離散系なので、電流応答速度には制御周期による限界があり、図７、図８からも判るように、充分に振動を抑制できない。
【００４０】
これに対して、本発明においては、図３に示すように、ｄ軸電流指令値ｉ_ｄ ^＊およびｑ軸電流ｉ_ｑ ^＊が実線で示すようにステップ状に変化した場合、実際のｄ軸電流ｉ_ｄとｑ軸電流ｉ_ｑは破線で示すように時間遅れをもって滑らかに変化し、振動することはない。また、ｄ軸電圧ｖ_ｄおよびｑ軸電圧ｖ_ｑも滑らかに変化し、振動することはない。つまり、電流応答から振動成分が完全に除去され、滑らかな１次遅れの応答となる。
また、図４は、本発明において、カットオフ周波数αを図７の３倍の値に設定した場合の特性である。この場合には、実際のｄ軸電流ｉ_ｄとｑ軸電流ｉ_ｑの変化が、かなりｄ軸電流指令値ｉ_ｄ ^＊およびｑ軸電流ｉ_ｑ ^＊の変化に近づいているが、やはり振動は生じていない。
上記のように、電流の応答周波数はα［ｒａｄ／ｓ］となり、制御周期による電流応答速度の限界の範囲内あるいはノイズの増幅による悪影響が出ない範囲内で自由に決めることができる。
【００４１】
また、図５、図６は、本発明において、制御に用いる推定インダクタンス値（Ｌ_ｄｃ、Ｌ_ｑｃ）が、実際のインダクタンス値（Ｌ_ｄｍ、Ｌ_ｑｍ）に対して誤差を含んでいる場合の特性例であり、図５は「Ｌ_ｄｍ＝Ｌ_ｄｃ×１.１、Ｌ_ｑｍ＝Ｌ_ｑｃ×０.９」の誤差を含んでいる場合、図６は「Ｌ_ｄｍ＝Ｌ_ｄｃ×１.１、Ｌ_ｑｍ＝Ｌ_ｑｃ×１.１」の誤差を含んでいる場合の特性を示す。
図５、図６から判るように、制御に用いる推定インダクタンス値（Ｌ_ｄｃ、Ｌ_ｑｃ）が、実際のインダクタンス値（Ｌ_ｄｍ、Ｌ_ｑｍ）に対して誤差を含んでいる場合でも、電流応答が振動的になったり、発振することはない。
【００４２】
以上説明したように、本発明においては、クロスフィルタ１（電圧指令値補正フィルタ）として、ｄ軸電圧指令値の変化に起因してｄ軸電流に発生する振動成分を除去するように、ｄ軸電圧指令値を補正する第一の振動除去手段（Ｃ_１１）と、ｄ軸電圧指令値の変化に起因してｑ軸電流に発生する振動成分を除去するように、ｄ軸電圧指令値に基づいてｑ軸電圧指令値を補正する第二の振動除去手段（Ｃ_２１）と、ｑ軸電圧指令値の変化に起因してｑ軸電流に発生する振動成分を除去するように、ｑ軸電圧指令値を補正する第三の振動除去手段（Ｃ_２２）と、ｑ軸電圧指令値の変化に起因してｄ軸電流に発生する振動成分を除去するように、ｑ軸電圧指令値に基づいてｄ軸電圧指令値を補正する第四の振動除去手段（Ｃ_１２）と、を備え、そしてクロスフィルタ１の特性はモータ定数を用いて（数１）式〜（数４）式のように設計している。そのため電流応答から電動機に固有の電気的な振動成分が除去される。
また、ｇ(ｓ）を（数７）式に示した一次遅れ成分とすることにより、電流応答から振動成分が除去されると共に、電流応答は応答周波数α［ｒａｄ／ｓ］の滑らかな一次遅れとなり、制御周期による電流応答速度の限界の範囲内で自由に決めることができる。
また、制御に用いる推定インダクタンス値（Ｌ_ｄｃ、Ｌ_ｑｃ）が、実際のインダクタンス値（Ｌ_ｄｍ、Ｌ_ｑｍ）に対して誤差を含んでいた場合でも、電流応答が不安定になったり、発振することはない。
また、電動機の評価試験の際、出力トルクと回転速度を指標とした、定常運転時のｄ軸電圧およびｑ軸電圧より作成した電圧マップを用いて電圧指令値を求めることにより、パラメータ誤差による電流応答への影響をより少なくすることができる。
また、フィルタ部２のローパスフィルタから出力されたｄ軸指令値およびｑ軸電流指令値と、ｄ軸およびｑ軸の実電流値とに基づいて、クロスフィルタ１から出力されたｄ軸補正電圧指令値およびｑ軸補正電圧指令値をさらに補正することにより、実電流と電流指令値の差を補正して、更に確実に電流応答性を向上させることができる。
また、フィルタ部２のローパスフィルタの特性ｇ_ｆｌｔ(ｓ）を（数７）式とし、クロスフィルタ１の電流応答特性ｇ(ｓ）をｇ(ｓ）＝ｇ_ｆｌｔ(ｓ）とすることによって、電流応答周波数をαとして、所望の応答特性を得ることができる。
【図面の簡単な説明】
【図１】本発明の第１の実施例のブロック図。
【図２】本発明の第２の実施例のブロック図。
【図３】本発明におけるステップ入力時の電流応答その１。
【図４】本発明におけるステップ入力時の電流応答その２。
【図５】本発明におけるステップ入力時の電流応答その３。
【図６】本発明におけるステップ入力時の電流応答その４。
【図７】従来技術におけるステップ入力時の電流応答。
【図８】従来技術におけるＰＩゲインを大きくした場合のステップ入力時の電流応答。
【符号の説明】
１…クロスフィルタ２…フィルタ部
３…指令値決定部４…ＰＩ制御部
５…電圧座標変換器６…ＰＷＭ変換器
７…インバータ８…電流検出器
９…電動機１０…位置検出器
１１…座標変換器１２…速度演算器
３０１…電流マップ３０２…電圧指令値演算部
３０３…電圧マップ[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a control device for a synchronous motor, for example, a control technique in a three-phase synchronous motor having a permanent magnet in a rotor.
[0002]
[Prior art]
As a conventional electric motor control device, for example, there is one described in Japanese Patent Application Laid-Open No. 2000-224883. In this conventional example, the d-axis voltage command value and the q-axis voltage command value for generating the d-axis current and the q-axis current satisfying the torque command value are calculated, and the actual currents of the d-axis and the q-axis are calculated from A configuration is described in which the voltage command value is corrected by PI (proportional integration) control from the difference from the shaft current command value.
[0003]
[Problems to be solved by the invention]
When controlled as in the above prior art, vibrations occur in the d-axis current and the q-axis current. In order to prevent this, a d-axis voltage command value is input, a filter (for example, a low-pass filter) that removes a vibration component of the d-axis current from the input d-axis voltage command value, and a q-axis voltage command value are input. It is conceivable to provide a filter for removing the vibration component of the q-axis current from the input q-axis voltage command value and outputting it. However, when the d-axis voltage command value changes, not only the d-axis current but also the q-axis current changes due to the back electromotive force (also referred to as speed electromotive force), and the q-axis voltage command value Changes in the q-axis current as well as the d-axis current due to the back electromotive force. Therefore, there has been a problem that even if the current vibration components of the d-axis voltage command value and the q-axis voltage command value are independently removed, current vibration occurs.
[0004]
The present invention has been made to solve the problems of the prior art as described above, and an object of the present invention is to provide a synchronous motor control device that improves the current response by suppressing the vibration component in the current response. .
[0005]
[Means for Solving the Problems]
In the present invention, the d-axis voltage command value and the q-axis voltage command value calculated by the voltage command value calculation means are input, and d is caused by changes in the input d-axis voltage command value and q-axis voltage command value. As a voltage command value correction filter for removing vibration components generated in the shaft current and the q-axis current,
first vibration removing means (C ₁₁ ) for correcting the d-axis voltage command value so as to remove the vibration component generated in the d-axis current due to the change in the d-axis voltage command value;
Second vibration removing means (C ₂₁ ) for correcting the q-axis voltage command value based on the d-axis voltage command value so as to remove the vibration component generated in the q-axis current due to the change in the d-axis voltage command value. )When,
third vibration removing means (C ₂₂ ) for correcting the q-axis voltage command value so as to remove the vibration component generated in the q-axis current due to the change in the q-axis voltage command value;
Fourth vibration removing means (C ₁₂ ) for correcting the d-axis voltage command value based on the q-axis voltage command value so as to remove the vibration component generated in the d-axis current due to the change in the q-axis voltage command value. )When,
It comprises so that it may be equipped with.
[0006]
【The invention's effect】
In the present invention, the vibration component generated in the d-axis current due to the change in the d-axis voltage command value, the vibration component generated in the q-axis current, and the q-axis current due to the change in the q-axis voltage command value. Since each correction voltage command value is set so as to remove all of the vibration component generated and the vibration component generated in the d-axis current, the current vibration component is removed from the d-axis voltage command value and the q-axis voltage command value. Thus, the current response can be improved with certainty.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a block diagram showing a first embodiment of the present invention.
In FIG. 1, the characteristics of the respective filters (C ₁₁ ), (C ₁₂ ), (C ₂₁ ), and (C ₂₂ ) of the cross filter 1 (voltage command value correction filter) are expressed by the following equations (1) to (4). ).
[0008]
[Expression 1]

[0009]
[Expression 2]

[0010]
[Equation 3]

[0011]
[Expression 4]

Where g (s): current response characteristic s: Laplace operator R: motor armature winding resistance L _dc , L _qc : estimated motor inductance ω: electrical angular velocity and current response characteristic g ( The relationship between s) and the characteristic g _flt (s) of the low-pass filter of the filter unit 2 is as shown in the following equation (5).
[0012]
[Equation 5]

In the command value determination unit 3, first, the d-axis current command value is obtained by map drawing using the current map 301 (current command value calculation means) using the torque command value T ^* input from the outside and the current rotational speed ω as indices. Find i _d ^* and q-axis current command value i _q ^* . Then, the voltage command value calculation unit 302 (voltage command value calculation means) calculates the d-axis voltage command value v _d ′ from the input d-axis current command value i _d ^* and q-axis current i _q ^* by the following equation (6). And q-axis voltage command value v _q 'is calculated and output.
[0013]
[Formula 6]

However, φ: winding interlinkage magnetic flux number filter unit 2 (correction value calculation means) is constituted by, for example, a first-order lag low-pass filter, and includes a d-axis current command value i _d ^* and a q-axis current command value i _q. The rate of change of ^* is relaxed and output as a d-axis current command value i _d ′ and q-axis current command value i _q ′ after relaxation.
The characteristic g _flt (s) of the low-pass filter of the filter unit 2 is expressed by the following equation (7).
[0014]
[Expression 7]

However, in the cut-off frequency cross filter 1 when α: g (s) is a first order lag, the d-axis voltage command value v _d ′ and the q-axis voltage command value v _q ′ from the command value determination unit 3 are input. The corrected voltage command values (v _d1 , v _q1 ) shown in the following (Equation 8) and (Equation 9) are obtained by passing through a cross filter having the characteristics of the above (Equation 1) to (Equation 4). Output.
[0015]
[Equation 8]

[0016]
[Equation 9]

On the other hand, the coordinate converter 11 includes three-phase detection currents (actual currents) i _u , i _v , i _w obtained from the two-phase currents i _u , i _v detected by the current detector 8, and the position detector 10. The d-axis current (magnetic flux current) _id and the q-axis current (torque current) i _q are obtained based on the detection position (rotation phase of the rotor) detected in (1). The current detector 8, the position detector 10, and the coordinate converter 11 correspond to the actual current detecting means.
[0017]
In the PI control unit 4 (correction value calculating means), the relaxed d-axis current command value i _d ′ and q-axis current command value i _q ′ sent from the filter unit 2 and the d-axis sent from the coordinate converter 11 The difference between the current i _d and the q-axis current i _q is amplified to generate voltage correction values (v _d2 , v _q2 ) represented by the following formulas (10) and (11). However, when there is no parameter error and the current response characteristic g _flt (s) is obtained, the difference in current is 0, so (v _d2 , v _q2 ) is 0.
[0018]
[Expression 10]

[0019]
[Expression 11]

However, g _pid , g _piq : PI gain in the PI control unit 4 The corrected voltage command values (v _d1 , v _q1 ) obtained as described above are the voltage correction values (v _d2 , v _q2 ). Further correction is performed by obtaining the sum of and the corrected voltage command values (v _d , v _q ) after correction are sent to the voltage coordinate converter 5. The voltage coordinate converter 5 converts the signals into three-phase alternating currents v _u , v _v , and v _w , and the PWM converter 6 converts them into PWM signals. By controlling the inverter 7 with the PWM signal, three-phase AC power is generated, and thereby the electric motor 9 is driven. The voltage coordinate converter 5, the PWM converter 6, and the inverter 7 correspond to voltage control means.
The corrected voltage command values (v _d , v _q ) after the correction are expressed by the following formulas (12) and (13).
[0020]
[Expression 12]

[0021]
[Formula 13]

Next, the operation will be described.
In the steady state, the characteristic of the cross filter 1 can be expressed as the following (Equation 14).
[0022]
[Expression 14]

In the equation (14), c ₁₂ and c ₂₁ are filters that are applied only to a transient state in which the d-axis voltage command value v _d ′ and the q-axis voltage command value v _q ′ change. In this case, it is 0.
In the transient state, when only the d-axis voltage command value v _d ′ changes, the characteristics of the cross filter 1 can be expressed by the following equation (Equation 15).
[0023]
[Expression 15]

However, L _dc , L _qc : Inductance value (measured value or estimated value) of motor used for control
When only the correction voltage command value v _d is operated, vibration always occurs in the current response. Therefore, when the correction voltage command value v _d is operated, the correction voltage command value v _q must also be operated. Therefore, in addition to the d-axis correction voltage command value v _d generated by c ₁₁ , the q-axis correction voltage command value v _q by c ₂₁ is also changed, resulting in a change in the d-axis voltage command value v _d ′ and this. The electric vibration inherent to the motor generated in the d-axis current and the q-axis current is suppressed by the interference component between the dq axes.
Further, when the q-axis voltage command value v _q ′ changes, the characteristics of the cross filter 1 can be expressed by the following (Equation 16), and vibration is suppressed in the same manner as the d-axis.
[0024]
[Expression 16]

The above operations can be expressed by mathematical expressions as follows.
When there is no parameter error, the corrected voltage command values (v _d , v _q ) input to the voltage coordinate converter 5 are as shown in the following formulas (17) and (18).
[0025]
[Expression 17]

[0026]
[Formula 18]

Further, the voltage equation of the motor including the transient response is as shown in the following (Equation 19).
[0027]
[Equation 19]

However, L _dm , L _qm : Actual inductance value of the motor Here, the equation (6) and the equation (19) are expanded to the following equations (20) and (21), and the cross The input / output relationship of the filter 1 is arranged as shown in the following equation (22).
[0028]
[Expression 20]

[0029]
[Expression 21]

[0030]
[Expression 22]

In the formula (22), when both sides are multiplied by the inverse matrix B- ¹ of B from the left, the following formula (23) and formula (24) are obtained.
[0031]
[Expression 23]

[0032]
[Expression 24]

The element of B ⁻¹ relating to the input correction voltage command value (v _d , v _q ) includes the second-order lag component 1 / G (s) in the denominator. The radio wave response to the command value v _d ′ and the q-axis voltage command value v _q ′ becomes oscillatory.
On the other hand, when the cross filter 1 is used, the relationship between the d-axis voltage command value v _d ′ and the q-axis voltage command value v _q ′ and the response of the d-axis current and the q-axis current is expressed by the following equation (25). Since the second-order lag component 1 / G (s) is removed, it should not vibrate.
[0033]
[Expression 25]

[0034]
[Equation 26]

Further, the relationship between the d-axis current command value i _d ^* and the q-axis current command value i _q ^* response is as shown in the following formulas (27), (28), and (29).
[0035]
[Expression 27]

[0036]
[Expression 28]

[0037]
[Expression 29]

As described above, the vibration component of the current response with respect to the d-axis voltage command value v _d ′ and the q-axis voltage command value v _q ′ is obtained by using the cross filter 1 according to the equations (1) to (4). In addition, the current response characteristic is g (s), and when g (s) is designed as shown in equation (7), a smooth first-order lag current response with a response frequency α [rad / s] is obtained. It is done.
[0038]
FIG. 2 is a block diagram showing a second embodiment of the present invention.
A difference from FIG. 1 is that a voltage map 303 is provided instead of the voltage command value calculation unit 302 of FIG. 1, and voltage command values (v _d ′, v _q ′) are obtained by subtracting the map from the torque command value T ^* and the rotational speed ω. ).
In FIG. 2, a voltage map 303 is created based on the d-axis voltage v _d and the q-axis voltage v _q during steady operation, using the torque command value T ^* and the rotational speed ω as indices during the motor evaluation test. Using this map, the d-axis voltage command value v _d ′ and the q-axis voltage command value v _q ′ are obtained.
Next, the operation will be described.
Since the voltage map 303 includes the variation of the inductance value due to the current, the current response is less affected by the parameter error.
[0039]
Hereinafter, actual examples of control in the first and second embodiments will be described.
3 to 8 are diagrams showing examples of current responses at the time of step input. FIGS. 3 to 6 are characteristics of the present invention, and FIGS. 7 and 8 are characteristics of a conventional example shown for comparison. is there.
In the conventional example, as shown in FIG. 7, when the d-axis current command value i _d ^* and the q-axis current i _q ^* change stepwise as indicated by the solid line, the actual d-axis current _id and the q-axis The current _iq oscillates greatly as shown by the broken line. Further, the d-axis voltage v _d and the q-axis voltage v _q change stepwise as shown and vibrate.
FIG. 8 shows characteristics when the gain of the PI feedback loop is increased as much as possible in the conventional example. In this case, the vibration still remains, although it is less than in FIG. In the conventional example, in order to correct a current error with a large frequency, such as an electric vibration component inherent to the motor, the gain of the PI feedback loop must be sufficiently increased, but as the PI gain is increased, Amplify up to noise. Since the actual control is a discrete system, the current response speed is limited by the control cycle, and as can be seen from FIGS. 7 and 8, the vibration cannot be sufficiently suppressed.
[0040]
On the other hand, in the present invention, as shown in FIG. 3, when the d-axis current command value i _d ^* and the q-axis current i _q ^* change stepwise as shown by the solid line, the actual d-axis current The _id and the q-axis current i _q change smoothly with a time delay as shown by the broken line, and do not vibrate. Also, the d-axis voltage v _d and the q-axis voltage v _q change smoothly and do not vibrate. That is, the vibration component is completely removed from the current response, resulting in a smooth first-order lag response.
FIG. 4 shows characteristics when the cutoff frequency α is set to a value three times that of FIG. 7 in the present invention. In this case, the actual change in the d-axis current _id and the q-axis current _iq is very close to the change in the d-axis current command value _id ^* and the q-axis current _iq ^* , but vibration still occurs. Not.
As described above, the current response frequency is α [rad / s], and can be freely determined within the range of the limit of the current response speed due to the control cycle or within the range where no adverse effects are caused by noise amplification.
[0041]
5 and 6 show characteristics when the estimated inductance values (L _dc , L _qc ) used for control in the present invention include errors with respect to the actual inductance values (L _dm , L _qm ). For example, FIG. 5 includes an error of “L _dm = L _dc × 1.1, L _qm = L _qc × 0.9”, and FIG. 6 shows “L _dm = L _dc × 1.1, The characteristic when an error of “L _qm = L _qc × 1.1” is included is shown.
As can be seen from FIGS. 5 and 6, even when the estimated inductance values (L _dc , L _qc ) used for the control include an error with respect to the actual inductance values (L _dm , L _qm ), the current response is It will not vibrate or oscillate.
[0042]
As described above, in the present invention, as the cross filter 1 (voltage command value correction filter), the d-axis so as to remove the vibration component generated in the d-axis current due to the change in the d-axis voltage command value. Based on the first vibration removing means (C ₁₁ ) for correcting the voltage command value and the d-axis voltage command value so as to remove the vibration component generated in the q-axis current due to the change in the d-axis voltage command value. A second vibration removing means (C ₂₁ ) for correcting the q-axis voltage command value, and a q-axis voltage command so as to remove a vibration component generated in the q-axis current due to a change in the q-axis voltage command value. Third vibration removing means (C ₂₂ ) for correcting the value and d based on the q-axis voltage command value so as to remove the vibration component generated in the d-axis current due to the change in the q-axis voltage command value. fourth Bei vibration removing means and (C _12), of correcting the axis voltage value And the characteristics of cross filter 1 is designed as with motor constant (number 1) through (4) below. Therefore, the electric vibration component specific to the motor is removed from the current response.
Further, by using g (s) as the first-order lag component shown in Equation (7), the vibration component is removed from the current response, and the current response is a smooth first-order lag of the response frequency α [rad / s]. Thus, it can be freely determined within the limit of the current response speed due to the control cycle.
Further, even when the estimated inductance values (L _dc , L _qc ) used for the control include errors with respect to the actual inductance values (L _dm , L _qm ), the current response becomes unstable or oscillates. There is nothing.
In addition, during an electric motor evaluation test, a voltage command value is obtained using a voltage map created from the d-axis voltage and the q-axis voltage during steady operation using the output torque and the rotational speed as indices, thereby obtaining a current due to parameter error. The influence on the response can be reduced.
Further, the d-axis correction voltage command output from the cross filter 1 based on the d-axis command value and the q-axis current command value output from the low-pass filter of the filter unit 2 and the actual current values of the d-axis and the q-axis. By further correcting the value and the q-axis correction voltage command value, the difference between the actual current and the current command value can be corrected, and the current response can be improved more reliably.
Further, by setting the characteristic g _flt (s) of the low-pass filter of the filter unit 2 to Equation (7) and the current response characteristic g (s) of the cross filter 1 to g (s) = g _flt (s), A desired response characteristic can be obtained by setting the current response frequency to α.
[Brief description of the drawings]
FIG. 1 is a block diagram of a first embodiment of the present invention.
FIG. 2 is a block diagram of a second embodiment of the present invention.
FIG. 3 shows current response at the time of step input according to the present invention.
FIG. 4 shows a current response at the time of step input according to the present invention.
FIG. 5 shows a current response at the time of step input according to the present invention, part 3;
FIG. 6 shows a current response at the time of step input according to the present invention, part 4;
FIG. 7 is a current response at the time of step input in the prior art.
FIG. 8 shows a current response at the time of step input when the PI gain is increased in the prior art.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Cross filter 2 ... Filter part 3 ... Command value determination part 4 ... PI control part 5 ... Voltage coordinate converter 6 ... PWM converter 7 ... Inverter 8 ... Current detector 9 ... Electric motor 10 ... Position detector 11 ... Coordinate conversion 12: Speed calculator 301 ... Current map 302 ... Voltage command value calculator 303 ... Voltage map

Claims

Voltage command value calculating means for calculating a d-axis voltage command value and a q-axis voltage command value based on the torque command value;
The d-axis voltage command value and the q-axis voltage command value calculated by the voltage command value calculation means are input, and the d-axis current and q are caused by changes in the input d-axis voltage command value and q-axis voltage command value. Voltage command value correction filter that corrects the d-axis voltage command value and the q-axis voltage command value and outputs them as a d-axis correction voltage command value and a q-axis correction voltage command value so as to remove vibration components generated in the shaft current When,
Voltage control means for controlling the voltage for each phase of the motor based on the d-axis correction voltage command value and the q-axis correction voltage command value output from the voltage command value correction filter;
With
The voltage command value correction filter is:
first vibration removing means (C ₁₁ ) for correcting the d-axis voltage command value so as to remove the vibration component generated in the d-axis current due to the change in the d-axis voltage command value;
Second vibration removing means (C ₂₁ ) for correcting the q-axis voltage command value based on the d-axis voltage command value so as to remove the vibration component generated in the q-axis current due to the change in the d-axis voltage command value. )When,
third vibration removing means (C ₂₂ ) for correcting the q-axis voltage command value so as to remove the vibration component generated in the q-axis current due to the change in the q-axis voltage command value;
Fourth vibration removing means (C ₁₂ ) for correcting the d-axis voltage command value based on the q-axis voltage command value so as to remove the vibration component generated in the d-axis current due to the change in the q-axis voltage command value. )When,
A control apparatus for a synchronous motor, comprising:

The first vibration removing means (C ₁₁ ), the second vibration removing means (C ₂₁ ), the third vibration removing means (C ₂₂ ), and the fourth vibration removing means (C ₁₂ ) are each represented by the following (Equation 1): 2. The control apparatus for a synchronous motor according to claim 1, wherein the control apparatus has characteristics shown in equations (1) to (4).

However, g (s): current response characteristic s: Laplace operator R: armature winding resistance _L dc _{motor, L qc:} inductance estimate of the motor omega: electrical angular speed

Current command value calculating means for calculating a d-axis current command value and a q-axis current command value based on the torque command value;
A low-pass that inputs the d-axis current command value and the q-axis current command value calculated by the current command value calculation means and relaxes and outputs the change speeds of the input d-axis current command value and q-axis current command value, respectively. Filters,
Real current detection means for detecting current flowing in each phase of the motor and detecting actual current values of the d-axis and the q-axis;
Based on the d-axis current command value and the q-axis current command value calculated by the current command value calculation means, and the d-axis and q-axis actual current values detected by the actual current detection means, the voltage command value Correction value calculation means for calculating a correction value for further correcting the d-axis correction voltage command value and the q-axis correction voltage command value output from the correction filter;
Correction means for correcting the d-axis correction voltage command value and the q-axis correction voltage command value output from the voltage command value correction filter based on the correction value calculated by the correction value calculation means;
The control apparatus for a synchronous motor according to claim 1 or 2, further comprising:

A current map for obtaining a d-axis current command value and a q-axis current command value based on the torque command value and the rotation speed of the motor;
Voltage command value calculating means for calculating a d-axis voltage command value and a q-axis voltage command value based on the d-axis current command value and the q-axis current command value from the current map;
The d-axis voltage command value and the q-axis voltage command value calculated by the voltage command value calculation means are input, and the d-axis current and q are caused by changes in the input d-axis voltage command value and q-axis voltage command value. Voltage command value correction filter that corrects the d-axis voltage command value and the q-axis voltage command value and outputs them as a d-axis correction voltage command value and a q-axis correction voltage command value so as to remove vibration components generated in the shaft current When,
A low-pass filter that inputs the d-axis current command value and the q-axis current command value obtained from the current map, and that relaxes and outputs the input d-axis current command value and the q-axis current command value, respectively;
Real current detection means for detecting current flowing in each phase of the motor and detecting actual current values of the d-axis and the q-axis;
Based on the d-axis current command value and the q-axis current command value output from the low-pass filter, and the actual current values of the d-axis and the q-axis detected by the actual current detecting means, the proportional integral control is used. Correction value calculation means for calculating a correction value for further correcting the d-axis correction voltage command value and the q-axis correction voltage command value output from the voltage command value correction filter;
Correction means for correcting the d-axis correction voltage command value and the q-axis correction voltage command value output from the voltage command value correction filter based on the correction value calculated by the correction value calculation means;
Voltage control means for controlling the voltage of each phase of the motor based on the corrected d-axis correction voltage command value and q-axis correction voltage command value;
A control apparatus for a synchronous motor, comprising:

A current map for obtaining a d-axis current command value and a q-axis current command value based on the torque command value and the rotation speed of the motor;
A voltage map for obtaining a d-axis voltage command value and a q-axis voltage command value based on the torque command value and the rotation speed of the motor;
The d-axis voltage command value and the q-axis voltage command value obtained from the voltage map are input, and generated in the d-axis current and the q-axis current due to changes in the input d-axis voltage command value and q-axis voltage command value. A voltage command value correction filter that corrects the d-axis voltage command value and the q-axis voltage command value and outputs them as a d-axis correction voltage command value and a q-axis correction voltage command value so as to remove the vibration component
A low-pass filter that inputs the d-axis current command value and the q-axis current command value obtained from the current map, and that relaxes and outputs the input d-axis current command value and the q-axis current command value, respectively;
Real current detection means for detecting current flowing in each phase of the motor and detecting actual current values of the d-axis and the q-axis;
Based on the d-axis current command value and the q-axis current command value output from the low-pass filter, and the actual current values of the d-axis and the q-axis detected by the actual current detecting means, the proportional integral control is used. Correction value calculation means for calculating a correction value for further correcting the d-axis correction voltage command value and the q-axis correction voltage command value output from the voltage command value correction filter;
Correction means for correcting the d-axis correction voltage command value and the q-axis correction voltage command value output from the voltage command value correction filter based on the correction value calculated by the correction value calculation means;
Voltage control means for controlling the voltage of each phase of the motor based on the corrected d-axis correction voltage command value and q-axis correction voltage command value;
A control apparatus for a synchronous motor, comprising:

The low-pass filter has a characteristic g _flt (s) of g _flt (s) = α / (s + α).
Represented by
The current response characteristic g (s) of the voltage command value correction filter is g (s) = g _flt (s).
6. The control apparatus for a synchronous motor according to claim 3, wherein the control apparatus is a synchronous motor.
However,
s: Laplace operator α: Cut-off frequency of low-pass filter