JP2005003065A - Belt slippage predicting device and belt clamping force control device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To predict the slippage of a belt for a continuously variable transmission before actually slipped while suppressing a cost increase. <P>SOLUTION: A band pass filter 25 extracts the band of a peak frequency fc from sheave pressure Pc2. A threshold value comparison circuit 26 detects a macroslip sign when the filtered sheave pressure Pc2 exceeds a threshold value and predicts the slippage of the belt. A clamping force control circuit 27 controls a pressure control valve 14 so that the filtered sheave pressure Pc2 is kept nearly at the threshold value. As a result, the belt μ for the CVT is held nearly at the maximum value. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

【００４６】
α：調整係数
Ｎｓ：１秒当たりのセカンダリ回転数［Ｈｚ］
φ：セカンダリプーリ上でのベルト掛かり角［ｒａｄ］（減速比γに応じて変化）
（１）式は、ブロックのスティックスリップによるベルト振動がベルトかかり部分全域で伝達され、セカンダリ回転数に応じた周波数で油圧振動を起こしていることを表している。
【００４７】
図５は、各減速比におけるピーク周波数の（１）式による計算値と実験値との比較図である。実験値は、マクロスリップ前兆時の実験データをＦＦＴ解析することで得られた周波数である。図５より、計算値と実験値はほぼ一致し、（１）式が妥当であることを確認することができた。
【００４８】
４．マクロスリップ前兆の検出
本発明は、上述した伝達ゲイン特性に着目することによって、マクロスリップの前兆を検出し、ベルト滑りを予測するものである。マクロスリップ前兆の検出は、以下の手順（１）から手順（４）によって行われる。
【００４９】
手順（１）：セカンダリ回転数に対するセカンダリシーブ圧の伝達ゲインは、図３及び図４に示すように、マクロスリップ発生に近づくほど大きくなる。
【００５０】
手順（２）：伝達ゲイン特性を伝達関数で模擬し、伝達ゲインの変化と伝達関数の係数との関連性を調べる。
【００５１】
手順（３）：セカンダリ回転数とセカンダリシーブ圧とを用いて、伝達関数の係数を最小自乗法によりマッチングさせる。
【００５２】
手順（４）：マッチングした係数と手順（２）の検討結果とを参照して、マクロスリップ前兆を検出する。
【００５３】
ここで、手順（１）については既に説明し、手順（３）及び（４）については後述の具体的な実施形態で説明するので、手順（２）について説明する。
【００５４】
図３及び図４に示す伝達ゲイン特性を（２）式の伝達関数で模擬する。
【００５５】
【数２】

【００５６】
Ｋ：伝達ゲイン係数（Ｋが大きいほどピーク大）
ｆｃ：ピーク周波数［Ｈｚ］
ｓ：ラプラス演算子
（２）式を２次式とした理由は、係数を同定する際の演算規模をできる限り小さくすると共に、演算精度を確保するためである。
【００５７】
図６は、ピーク周波数ｆｃ＝８Ｈｚとし、伝達ゲイン係数Ｋをパラメータとしたときの（２）式の伝達ゲイン周波数特性を示す図である。図６より、伝達ゲイン係数Ｋが大きくなると共にピークが現れた。つまり、（２）式の伝達ゲイン特性は、図３及び図４に示した１次のピーク周波数における伝達ゲイン特性に類似している。
【００５８】
（２）式は連続系である。後述のオンライン同定の計算が容易なように、（２）式を離散系で表現すると、（３）式になる。
【００５９】
【数３】

【００６０】
（３）式のｚはｚ変換の演算子である。係数Ａ１，Ａ２，Ｂ１，Ｂ２は、それぞれ未知係数であり、セカンダリ回転数とセカンダリシーブ圧とによってオンライン同定される。なお、マクロスリップ前兆は、係数Ａ１の値に基づいて検出される。
【００６１】
５．係数Ａ１と伝達ゲイン係数Ｋの関係
（５−１）係数Ａ１に基づくマクロスリップ前兆の検出
係数Ａ１，Ａ２，Ｂ１，Ｂ２は、サンプリング周期ｄｔを用いると、それぞれ（４）式から（７）式で表される。
【００６２】
【数４】

【００６３】
図７は、サンプリング周期ｄｔ＝４ｍｓで（４）式を用いたときの伝達ゲイン係数Ｋ及びピーク周波数ｆｃに対する係数Ａ１の変化を示す図である。図７より、伝達ゲイン係数Ｋが大きくなるに従って係数Ａ１の値が小さくなった。一方、マクロスリップ前兆になると、ピークが現れて、伝達ゲイン係数Ｋが大きくなる。
【００６４】
したがって、係数Ａ１が閾値を超えた（下回った）場合に、マクロスリップ前兆を検出できる。ただし、ピーク周波数ｆｃによって係数Ａ１の値が変化するため、減速比やセカンダリ回転数に応じて閾値を変更する必要がある。
【００６５】
（５−２）伝達ゲイン係数Ｋに基づくマクロスリップ前兆の検出
伝達関数の係数Ａ１は、図７に示すように、伝達ゲイン係数Ｋ及びピーク周波数ｆｃに応じて変化する。このため、減速比やセカンダリ回転数に応じて閾値を変更しなければならない。そこで、閾値を固定するために、マクロスリップ前兆検出のためのパラメータを、係数Ａ１から伝達ゲイン係数Ｋに変更する。（１）式及び（４）式を用いると、伝達ゲイン係数Ｋは（８）式となる。
【００６６】
【数５】

【００６７】
マクロスリップ前兆になると、ピークが現れて、伝達ゲイン係数Ｋが大きくなる。したがって、伝達ゲイン係数Ｋが一定の閾値を超えた（上回った）場合に、マクロスリップ前兆を検出できる。
【００６８】
［具体的な実施形態］
以下、本発明の好ましい実施の形態について図面を参照しながら詳細に説明する。
【００６９】
（第１の実施形態）
図８は、第１の実施形態に係るベルト挟圧力制御装置の構成を示すブロック図である。ベルト挟圧力制御装置は、入力軸側であるプライマリプーリ（以下「プライマリ」と省略する。）、出力軸側であるセカンダリプーリ（以下「セカンダリ」と省略する。）、及びプライマリとセカンダリの間に掛けられたベルトからなる無段階変速機のベルト滑りを予測して、ベルト挟圧力を最適な状態に制御するものである。
【００７０】
ベルト挟圧力制御装置は、プライマリの回転パルス信号を生成するプライマリ回転センサ１１と、セカンダリの回転パルス信号を生成するセカンダリ回転センサ１２と、セカンダリを制御するための油圧であるセカンダリシーブ圧に応じたセンサ信号を生成するシーブ圧センサ１３と、セカンダリシーブ圧を調整する圧力制御弁１４と、各センサからの信号に基づいて、圧力制御弁１４を調整してベルト挟圧力を制御する電子制御ユニット（以下「ＥＣＵ」という。）２０とを備えている。
【００７１】
プライマリ回転センサ１１は、プライマリの回転数に応じた回転パルス信号を生成し、この回転パルス信号をＥＣＵ２０に供給する。セカンダリ回転センサ１２は、セカンダリの回転数に応じた回転パルス信号を生成し、この回転パルス信号をＥＣＵ２０に供給する。
【００７２】
シーブ圧センサ１３は、固定シーブ及び可動シーブで構成されたセカンダリに対して、可動シーブを固定シーブに向けて付勢するための油圧であるセカンダリシーブ圧に応じてセンサ信号を生成して、ＥＣＵ２０に供給する。
【００７３】
図９は、ＥＣＵ２０の構成を示すブロック図である。ＥＣＵ２０は、プライマリ回転数を検出するプライマリ回転数検出回路２１と、セカンダリ回転数を検出するセカンダリ回転数検出回路２２と、セカンダリシーブ圧Ｐｃ２を検出するシーブ圧検出回路２３と、ピーク周波数ｆｃを演算するピーク周波数演算回路２４と、バンドパスフィルタ２５と、閾値と後述する係数Ａ１とを比較する閾値比較回路２６と、ベルト挟圧力を制御する挟圧力制御回路２７と、備えている。
【００７４】
プライマリ回転数検出回路２１は、プライマリ回転センサ１１からの回転パルス信号に基づいてプライマリ回転数Ｎｉｎ［ｒｐｍ］を検出し、プライマリ回転数Ｎｉｎをピーク周波数演算回路２４に供給する。
【００７５】
セカンダリ回転数検出回路２２は、セカンダリ回転センサ１２からの回転パルス信号に基づいてセカンダリ回転数Ｎｏｕｔ［ｒｐｍ］を検出し、セカンダリ回転数Ｎｏｕｔをピーク周波数演算回路２４に供給する。
【００７６】
シーブ圧検出回路２３は、シーブ圧センサ１３からのセンサ信号に基づいてセカンダリシーブ圧Ｐｃ２を検出し、セカンダリシーブ圧Ｐｃ２をバンドパスフィルタ２５に供給する。
【００７７】
ピーク周波数演算回路２４は、プライマリ回転数Ｎｉｎ及びセカンダリ回転数Ｎｏｕｔに基づいて減速比γ（＝Ｎｉｎ／Ｎｏｕｔ）を演算し、減速比γを用いてセカンダリプーリ上でのベルト掛かり角φ［ｒａｄ］を計算する。そして、予め設定された調整係数α、１秒当たりのセカンダリ回転数Ｎｓ［Ｈｚ］（＝Ｎｏｕｔ／６０）を用いて、（９）式に従ってピーク周波数ｆｃを演算する。
【００７８】
【数６】

【００７９】
バンドパスフィルタ２５は、（１０）式のバンドパス特性を有している。なお、ｆｃはピーク周波数、ｓはラプラス演算子である。
【００８０】
【数７】

【００８１】
バンドパスフィルタ２５は、セカンダリシーブ圧Ｐｃ２からピーク周波数ｆｃの帯域を抽出して、フィルタリングされたセカンダリシーブ圧Ｐｃ２をベルト挟圧力制御回路２６に供給する。
【００８２】
閾値比較回路２６は、フィルタリングされたセカンダリシーブ圧Ｐｃ２と閾値とを比較して、フィルタリングされたセカンダリシーブ圧Ｐｃ２が閾値を超えたときにマクロスリップ前兆を検出し、ベルト滑りを予測する。すなわち、閾値比較回路２６は、セカンダリシーブ圧Ｐｃ２の振幅が所定値より増大したときを検出して、ベルト滑りを予測することができる。なお、閾値は、ＣＶＴ状態量に応じて学習により設定された値でもよいし、予め設定された定数値でもよく、特に限定されるものではない。
【００８３】
挟圧力制御回路２７は、閾値比較回路２６によるベルト滑り予測結果に基づいて、圧力制御弁１４を調整する。具体的には、フィルタリングされたセカンダリシーブ圧Ｐｃ２が閾値付近の値を維持するように、圧力制御弁１４を制御する。この結果、ＣＶＴのベルトμを最大値付近に維持することができる。
【００８４】
以上のように、第１の実施形態に係るベルト挟圧力制御装置は、セカンダリシーブ圧Ｐｃ２の振幅増大を検出してマクロスリップ前兆をとらえることによって、ベルト滑りを予測することができる。また、ベルト挟圧力制御装置は、プライマリプーリの油圧（プライマリ側でベルト滑り予測をする場合には、セカンダリプーリの油圧）を検出するセンサを用いることなく、ベルト滑りを予測できるので、従来に比べてコストを抑制することもできる。
【００８５】
さらに、ベルト挟圧力制御装置は、ベルト滑り予測結果を用いてベルト挟圧力を制御することで、ＣＶＴのベルトμを最大値付近に維持することができ、その結果、ＣＶＴのベルト伝達効率を向上させることができる。
【００８６】
なお、本実施形態に係るベルト挟圧力制御装置は、バンドパスフィルタ２５及び閾値比較回路２６を用いているが、セカンダリシーブ圧Ｐｃ２の変動振幅の増大を検出することができれば上記構成に限定されるものではない。
【００８７】
また、ピーク周波数演算回路２４は、（９）式に従ってピーク周波数ｆｃを演算したが、ピーク周波数ｆｃを演算することができれば（９）式に限定されるものではない。例えば、ピーク周波数演算回路２４は、セカンダリ回転数や減速比の他に、これらに相関関係のある１つ以上の任意のＣＶＴ状態量（例えば、減速比指令値、セカンダリ上でのベルト掛かり径など）を用いてもよい。
【００８８】
（第２の実施形態）
つぎに、本発明の第２の実施形態について説明する。なお、第１の実施形態と同一の回路については同一の符号を付し、その詳細な説明は省略する。
【００８９】
図１０は、第２の実施の形態に係るＥＣＵ２０Ａの構成を示すブロック図である。第２の実施形態に係るベルト挟圧力制御装置は、第１の実施形態で説明したＥＣＵ２０の代わりに、図１０に示すＥＣＵ２０Ａを適用したものである。
【００９０】
ＥＣＵ２０Ａは、プライマリ回転数検出回路２１、セカンダリ回転数検出回路２２、シーブ圧検出回路２３、ピーク周波数演算回路２４、減速比γを演算する減速比演算回路３１、閾値を生成する閾値生成回路３２、バンドパスフィルタ３３，３４、伝達関数の係数を同定する伝達関数同定回路３５、閾値と所定の係数とを比較する閾値比較回路３６、ベルト挟圧力を制御する挟圧力制御回路３７を備えている。
【００９１】
プライマリ回転数検出回路２１は、プライマリ回転数Ｎｉｎを検出して、このプライマリ回転数Ｎｉｎをピーク周波数演算回路２４及び減速比演算回路３１に供給する。セカンダリ回転数検出回路２２は、セカンダリ回転数Ｎｏｕｔを検出して、このセカンダリ回転数Ｎｏｕｔをピーク周波数演算回路２４、減速比演算回路３１、閾値生成回路３２及びバンドパスフィルタ３３に供給する。シーブ圧検出回路２３は、セカンダリシーブ圧Ｐｃ２を検出し、このセカンダリシーブ圧Ｐｃ２をバンドパスフィルタ３４に供給する。
【００９２】
ピーク周波数演算回路２４は、プライマリ回転数Ｎｉｎ及びセカンダリ回転数Ｎｏｕｔに基づいて減速比γ（＝Ｎｉｎ／Ｎｏｕｔ）を演算し、減速比γを用いてセカンダリプーリ上でのベルト掛かり角φ［ｒａｄ］を計算する。そして、予め設定された調整係数α、１秒当たりのセカンダリ回転数Ｎｓ［Ｈｚ］（＝Ｎｏｕｔ／６０）を用いて、（１１）式に従ってピーク周波数ｆｃを演算する。
【００９３】
【数８】

【００９４】
そして、ピーク周波数演算回路２４は、演算したピーク周波数ｆｃをバンドパスフィルタ３３及び３４に供給する。
【００９５】
一方、減速比演算回路３１は、プライマリ回転数Ｎｉｎ及びセカンダリ回転数Ｎｏｕｔに基づいて減速比γを演算し、減速比γを閾値生成回路３２に供給する。
【００９６】
閾値生成回路３２は、マクロスリップ前兆を検出するための閾値、つまり伝達関数の係数Ａ１と比較するための閾値を生成する。ここで、閾値生成回路３２は、減速比γ、セカンダリ回転数Ｎｏｕｔ、閾値のそれぞれの関係を表した閾値マップを記憶している。そして、閾値生成回路３２は、閾値マップを参照して、現在の減速比γ及びセカンダリ回転数Ｎｏｕｔに対応する閾値を生成する。
【００９７】
また、閾値生成回路３２は、ＣＶＴの定常状態における各信号（例えば、プライマリ回転数、出力トルクなどの信号）を学習し、各信号が最大となるときの閾値を生成してもよい。
【００９８】
バンドパスフィルタ３３及び３４は、上述した（１０）式のバンドパス特性を有している。バンドパスフィルタ３３は、セカンダリ回転数Ｎｏｕｔからピーク周波数ｆｃの帯域を抽出して、フィルタリングされたセカンダリ回転数Ｎｏｕｔを伝達関数同定回路３５に供給する。また、バンドパスフィルタ３４は、セカンダリシーブ圧Ｐｃ２からピーク周波数ｆｃの帯域を抽出して、フィルタリングされたセカンダリシーブ圧Ｐｃ２を伝達関数同定回路３５に供給する。
【００９９】
伝達関数同定回路３５は、フィルタリングされたセカンダリ回転数Ｎｏｕｔ及びセカンダリシーブ圧Ｐｃ２を用いて、上述した（３）式で表された伝達関数の未知係数Ａ１，Ａ２，Ｂ１，Ｂ２を同定する。
【０１００】
ここで、オンライン同定をするために、サンプリング刻みｉを用いて、（３）式を（１２）式のように書き換える。
【０１０１】
【数９】

【０１０２】
（１２）式において、セカンダリ回転数Ｎｏｕｔ及びセカンダリシーブ圧Ｐｃ２は、それぞれバンドパスフィルタでフィルタリングされた後の値である。
【０１０３】
（１２）式の係数Ａ１，Ａ２，Ｂ１，Ｂ２を最小自乗法でオンライン同定するため、（１２）式を次の（１３）式で表す。
【０１０４】
【数１０】

【０１０５】
ただし、（１３）式においては、次の（１４）式から（１６）式が成り立つ。
【０１０６】
【数１１】

【０１０７】
さらに、次の（１７）式及び（１８）式による重み付け最小自乗法を用いる。
【０１０８】
【数１２】

【０１０９】
（１７）式及び（１８）式において、Ｐは４×４の対象行列である。Ｐの初期値は対角行列とし、Ｐの各要素はパラメータ同定の収束を早めるために大きな値に設定しておく。λは重みである。
【０１１０】
伝達関数同定回路３５は、このような（１３）式から（１８）式に従って、サンプリング周期毎に、セカンダリ回転数Ｎｏｕｔ（ｉ）及びセカンダリシーブ圧Ｐｃ２（ｉ）を用いてτ（ｉ）を同定する。そして、伝達関数同定回路３５は、同定されたτ（ｉ）の要素である係数Ａ１を閾値比較回路３６に供給する。
【０１１１】
閾値比較回路３６は、閾値生成回路３２で生成された閾値と、伝達関数同定回路３５で同定された係数Ａ１とを比較して、４サンプリング周期分連続して係数Ａ１が閾値を超えたとき（下回ったとき）にマクロスリップ前兆を検出する。なお、マクロスリップ前兆を確実に検出するため、「４サンプリング周期分連続」であることを条件としたが、これに限定されないのは勿論である。
【０１１２】
図１１から図１３は、（ａ）定常状態及び（ｂ）マクロスリップ時のプライマリ回転数、出力トルク、セカンダリシーブ圧（セカンダリシーブ圧）、係数Ａ１同定値を示す図である。プライマリ回転数Ｎｉｎ＝１０００ｒｐｍ、定常時の入力トルクＴｉｎ＝５０Ｎｍとした。なお、図１１は減速比γ＝０．６５、図１２は減速比γ＝１．０、図１３は減速比γ＝２．３７９である。しきい値は、ＣＶＴの定常状態における各信号が最大値になるときの値とした。
【０１１３】
挟圧力制御回路３７は、閾値比較回路３６によるベルト滑り予測結果に基づいて、圧力制御弁１４を調整する。具体的には、伝達関数の係数Ａ１が閾値付近の値を維持するように、圧力制御弁１４を制御する。
【０１１４】
以上のように、第２の実施形態に係るベルト挟圧力制御装置は、セカンダリ回転数Ｎｏｕｔに対するセカンダリシーブ圧Ｐｃ２の伝達ゲイン特性を模擬した伝達関数の係数Ａ１を同定し、係数Ａ１と閾値とを比較することで、マクロスリップ前兆をとらえ、ベルト滑りを予測することができる。
【０１１５】
また、ベルト挟圧力制御装置は、第１の実施形態と同様に、プライマリプーリの油圧（プライマリ側でベルト滑り予測をする場合には、セカンダリプーリの油圧）を検出するセンサを必要としないので従来に比べてコストを抑制することができ、さらに、ＣＶＴのベルトμを最大値付近に維持してＣＶＴのベルト伝達効率を向上させることができる。
【０１１６】
なお、ベルト挟圧力制御装置は、上述したバンドパスフィルタ３３，３４の代わりに、それぞれ同じ特性のローパスフィルタを備えてもよい。ローパスフィルタの伝達特性としては、例えば次の（１９）式が好ましい。
【０１１７】
【数１３】

【０１１８】
Ｔ（＝２πｆ）：ローパスフィルタ時定数［ｓ］
ｆ：カットオフ周波数（＞ｆｃ）［Ｈｚ］
ｓ：ラプラス演算子
また、ピーク周波数演算回路２４は、（１１）式に従ってピーク周波数ｆｃを演算したが、ピーク周波数ｆｃを演算することができれば（１１）式に限定されるものではない。例えば、ピーク周波数演算回路２４は、セカンダリ回転数や減速比の他に、これらに相関関係のある任意の１つ以上のＣＶＴ状態量（例えば、減速比指令値、セカンダリ上でのベルト掛かり径など）を用いてもよい。なお、ピーク周波数演算回路２４については、第３の実施形態でも同様である。
【０１１９】
（第３の実施形態）
つぎに、本発明の第３の実施形態について説明する。なお、第２の実施形態と同一の回路については同一の符号を付し、その詳細な説明は省略する。
【０１２０】
上述した第２の実施形態に係るベルト挟圧力制御装置は、伝達関数の係数Ａ１と閾値とを比較してベルト滑りを予測するものである。これに対して、第３の実施形態に係るベルト挟圧力制御装置は、伝達関数の伝達ゲイン係数Ｋと閾値とを比較してベルト滑りを予測するものである。
【０１２１】
図１４は、第３の実施形態に係るＥＣＵ２０Ｂの構成を示すブロック図である。ＥＣＵ２０Ｂは、第２の実施形態に係るＥＣＵ２０Ａから減速比演算回路３１及び閾値生成回路３２を除き、さらに伝達ゲイン係数演算回路３８を備えた構成になっている。
【０１２２】
伝達ゲイン係数演算回路３８は、伝達関数同定回路３５で同定された係数Ａ１、ピーク周波数演算回路２４で演算されたピーク周波数ｆｃ、そしてサンプリング周期ｄｔを用いて、（２０）式に従って伝達ゲイン係数Ｋを演算する。
【０１２３】
【数１４】

【０１２４】
閾値比較回路３６は、予め設定された閾値と、伝達ゲイン係数演算回路３８で演算された伝達ゲイン係数Ｋとを比較して、４サンプリング周期分連続して伝達ゲイン係数Ｋが閾値を超えたとき（上回ったとき）にマクロスリップ前兆を検出する。なお、第２の実施形態と同様に、「４サンプリング周期分連続」に限定されるものではない。
【０１２５】
図１５は、（ａ）減速比γ＝０．６５、（ｂ）減速比γ＝１．０、（ｃ）減速比γ＝２．３７９の場合に、伝達ゲイン係数Ｋ及び係数Ａ１を用いてそれぞれマクロスリップ前兆を検出した場合の比較結果を示す図である。伝達ゲイン係数Ｋの閾値は、定常状態における各々の減速比γの伝達ゲイン係数Ｋの中で最大値となったときの値である。図１５より、閾値を一定値に設定することができると共に、伝達ゲイン係数Ｋを用いてマクロスリップ前兆を検出できることが分かった。
【０１２６】
挟圧力制御回路３７は、閾値比較回路３６によるベルト滑り予測結果に基づいて、圧力制御弁１４を調整する。具体的には、伝達ゲイン係数Ｋが閾値付近の値を維持するように、圧力制御弁１４を制御する。
【０１２７】
以上のように、第３の実施形態に係るベルト挟圧力制御装置は、セカンダリ回転数Ｎｏｕｔに対するセカンダリシーブ圧Ｐｃ２の伝達ゲイン特性を模擬した伝達関数の伝達ゲイン係数Ｋを同定し、伝達ゲイン係数Ｋと閾値とを比較することで、マクロスリップ前兆をとらえ、ベルト滑りを予測することができる。また、閾値を予め一定値に設定しておくことができるので、ＣＶＴの状態に応じて閾値を随時変更する手間を省くことができる。
【０１２８】
また、ベルト挟圧力制御装置は、第１及び第２の実施形態と同様に、プライマリプーリに供給される油圧（プライマリ側でベルト滑り予測をする場合には、セカンダリプーリの油圧）を検出するセンサを必要としないので従来に比べてコストを抑制することができ、さらに、ＣＶＴのベルトμを最大値付近に維持してＣＶＴのベルト伝達効率を向上させることができる。
【０１２９】
なお、本実施形態に係るベルト挟圧力制御装置も、第２の実施形態と同様に、バンドパスフィルタ３３，３４の代わりにそれぞれローパスフィルタを設けてもよい。また、伝達ゲイン係数Ｋと比較するための閾値は、図１５に示した値に限定されるものではないのは勿論である。
【０１３０】
【発明の効果】
本発明に係るベルト滑り予測装置は、入力側、出力側２つのシーブ圧を用いることなく、いずれか一方の側のシーブ圧の変動に基づいて、低コストでベルト滑りを予測することができる。
【０１３１】
本発明に係るベルト挟圧力制御装置ば、ベルト滑り予測装置の予測結果に基づいて、ベルト滑りが予測される前後の状態を保持するように前記無段階変速機のベルト挟圧力を制御することにより、無段階変速装置のベルト伝達効率を向上させることができる。
【図面の簡単な説明】
【図１】ＣＶＴの（ａ）定常試験及び（ｂ）マクロスリップ試験を行ったときの波形図である。
【図２】図１の試験条件及びマクロスリップ発生方法を示す表図である。
【図３】減速比γ＝０．６５でマクロスリップを発生させたときの（ａ）実験波形及び（ｂ）伝達ゲイン周波数特性を示す図である。
【図４】減速比γ＝１．０でマクロスリップを発生させたときの（ａ）実験波形及び（ｂ）伝達ゲイン周波数特性を示す図である。
【図５】各減速比におけるピーク周波数の計算値と実験値との比較図である。
【図６】ピーク周波数ｆｃ＝８Ｈｚとし、伝達ゲイン係数Ｋをパラメータとしたときの（２）式の伝達ゲイン周波数特性を示す図である。
【図７】サンプリング周期ｄｔ＝４ｍｓで（４）式を用いたときの伝達ゲイン係数Ｋ及びピーク周波数ｆｃに対する係数Ａ１の変化を示す図である。
【図８】第１の実施形態に係るベルト挟圧力制御装置の構成を示すブロック図である。
【図９】ＥＣＵの構成を示すブロック図である。
【図１０】第２の実施の形態に係るＥＣＵの構成を示すブロック図である。
【図１１】減速比γ＝０．６５において（ａ）定常状態及び（ｂ）マクロスリップ時のプライマリ回転数、出力トルク、セカンダリシーブ圧（セカンダリシーブ圧）、係数Ａ１同定値を示す図である。
【図１２】減速比γ＝１．０において（ａ）定常状態及び（ｂ）マクロスリップ時のプライマリ回転数、出力トルク、セカンダリシーブ圧（セカンダリシーブ圧）、係数Ａ１同定値を示す図である。
【図１３】減速比γ＝２．３７９において（ａ）定常状態及び（ｂ）マクロスリップ時のプライマリ回転数、出力トルク、セカンダリシーブ圧（セカンダリシーブ圧）、係数Ａ１同定値を示す図である。
【図１４】第３の実施形態に係るＥＣＵの構成を示すブロック図である。
【図１５】（ａ）減速比γ＝０．６５、（ｂ）減速比γ＝１．０、（ｃ）減速比γ＝２．３７９の場合に、伝達ゲイン係数Ｋ及び係数Ａ１を用いてそれぞれマクロスリップ前兆を検出した場合の比較結果を示す図である。
【符号の説明】
１１ プライマリ回転センサ
１２ セカンダリ回転センサ
１３ シーブ圧センサ
１４ 圧力制御弁
２０，２０Ａ，２０Ｂ ＥＣＵ
２１ プライマリ回転数検出回路
２２ セカンダリ回転数検出回路
２３ シーブ圧検出回路
２４ ピーク周波数演算回路
２５，３３，３４ バンドパスフィルタ
２６，３６ 閾値比較回路
２７，３７ 挟圧力制御回路
３１ 減速比演算回路
３２ 閾値生成回路
３５ 伝達関数同定回路
３８ 伝達ゲイン係数演算回路[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a belt slip prediction device and a belt clamping pressure control device, and more particularly to a belt slip prediction device that predicts belt slip before belt slip actually occurs in a continuously variable transmission, and a belt clamping pressure using the same. The present invention relates to a belt clamping pressure control device that controls the belt.
[0002]
[Prior art]
Conventionally, a hydraulic control device for a continuously variable transmission that reduces engine load and prevents belt slip has been proposed (see, for example, Patent Document 1).
[0003]
As shown in FIG. 1 of the publication, the virtual gear ratio calculation unit 28 described in Patent Document 1 performs first-order lag filter processing on the actual gear ratio γ detected by the CVT gear ratio detection unit 22. The virtual reduction ratio γe is calculated. Here, the filter time constant is set to a sufficiently large value so as not to follow the change of γ when the belt slips.
[0004]
The belt slip detection unit 27 generates a slip detection signal DS when the slip state of the V belt 5c is detected based on the comparison between the transmission gear ratio GR and the virtual transmission gear ratio GRe.
[0005]
Here, the belt slip detection unit 27 validates the slip state detection process when the line pressure deviation ΔP2 between the target line pressure Po2 and the line pressure P2 (second actual oil pressure) indicates a predetermined amount β or more. . Further, the belt slip detection unit 27 detects a slip state when the primary pressure deviation ΔP1 between the target primary pressure Po1 and the actually detected primary pressure P1 (first actual hydraulic pressure) is equal to or greater than a predetermined amount α. Enable processing.
[0006]
[Patent Document 1]
Japanese Patent Laid-Open No. 2001-349418 (FIG. 1, paragraphs 58 and 59)
[0007]
[Problems to be solved by the invention]
The technique described in Patent Document 1 detects belt slip when the belt actually slips and a change in the number of rotations of the pulley occurs. Therefore, the belt slip detection timing is late, and belt damage cannot be avoided. was there.
[0008]
Furthermore, as shown in FIG. 7 of the above publication, the technique described in Patent Document 1 requires two expensive sensors, that is, the primary pressure sensor 11 and the line pressure sensor 12, and is therefore costly. There was also a problem.
[0009]
The present invention has been proposed to solve the above-described problems, and a belt slip prediction device that predicts belt slip before the belt of the continuously variable transmission actually slips while suppressing cost, and using the same. An object of the present invention is to provide a belt clamping pressure control device that controls the belt clamping pressure.
[0010]
[Means for Solving the Problems]
A belt slip prediction device according to claim 1 is a belt slip prediction device for a continuously variable transmission including an input pulley, an output pulley, and a belt, and detects a sheave pressure. And belt slip prediction means for predicting the occurrence of belt slip based on the change in sheave pressure detected by the sheave pressure detection means.
[0011]
The sheave pressure detecting means detects a sheave pressure that is a hydraulic pressure for urging the movable sheave toward the fixed sheave with respect to the output-side pulley that is configured by the fixed sheave and the movable sheave. Here, the sheave pressure has a characteristic that fluctuates immediately before the macro slip of the continuously variable transmission. The reason is considered to be that the pulley is vibrated by stick slip of the belt block. Therefore, the belt slip prediction means predicts belt slip based on the change in sheave pressure.
[0012]
Therefore, according to the first aspect of the present invention, since the macro slip precursor is detected based on the variation of the sheave pressure without using the input-side sheave pressure, the belt slip can be predicted at a low cost.
[0013]
A belt slip prediction apparatus according to a second aspect of the present invention is the belt slip prediction device according to the first aspect, wherein the belt slip prediction means predicts the occurrence of belt slip based on the fluctuation amplitude of the sheave pressure. This is because the sheave pressure has a characteristic that vibration increases immediately before the macro slip of the continuously variable transmission.
[0014]
A belt slip prediction device according to a third aspect of the present invention is the belt slip prediction device according to the first aspect, wherein the belt slip prediction means uses a sheave pressure in a predetermined band from a sheave pressure detected by the sheave pressure detection means. And the occurrence of belt slip is predicted based on the change in the sheave pressure in the predetermined band.
[0015]
When the sheave pressure is just before the macro slip, the fluctuation in the predetermined band increases. Therefore, according to the third aspect of the invention, since the sheave pressure in the predetermined band is extracted from the sheave pressure and the macro slip precursor can be detected from the sheave pressure in the predetermined band, the occurrence of the belt slip can be predicted.
[0016]
A belt slip prediction device according to a fourth aspect of the present invention is the belt slip prediction device according to the third aspect, wherein the output-side pulley rotation speed detecting means for detecting the output-side pulley rotation speed, and the output-side pulley rotation speed detection. Band calculation for calculating the predetermined band by using at least one of the output pulley rotation speed detected by the means, the reduction ratio of the continuously variable transmission, the reduction ratio command value, and the belt engagement diameter on the output pulley. Means.
[0017]
When at least one of the operation state of the continuously variable transmission changes, for example, the output pulley speed, the reduction ratio of the continuously variable transmission, the reduction ratio command value, and the belt engagement diameter on the output pulley, a sign of macro slip As shown in FIG.
[0018]
Therefore, according to the invention described in claim 4, the output pulley speed, the reduction ratio of the continuously variable transmission, the reduction ratio command value, and the belt engagement diameter on the output pulley are used. By calculating the predetermined band, even if the operating state of the continuously variable transmission changes, the belt slip can be reliably predicted.
[0019]
A belt slip prediction device according to a fifth aspect of the present invention is the belt slip prediction device according to the first aspect of the present invention, further comprising output-side pulley rotational speed detection means for detecting the output-side pulley rotational speed. The means predicts the occurrence of belt slip based on a value at a peak frequency of the transmission characteristic of the sheave pressure with respect to the output side pulley rotational speed.
[0020]
The sheave pressure transmission characteristic with respect to the output pulley rotation speed has a peak at a predetermined frequency. When the macro slip sign is reached, the value (peak value) of the transfer characteristic at the peak frequency has a property of increasing.
[0021]
Therefore, according to the fifth aspect of the present invention, it is possible to detect the macro slip based on the value at the peak frequency of the transfer characteristic of the sheave pressure with respect to the output side pulley rotation speed, thereby predicting the occurrence of the belt slip. can do.
[0022]
A belt slip prediction device according to a sixth aspect of the present invention is the belt slip prediction device according to the fifth aspect of the present invention, wherein the belt slip prediction means is configured to perform belt movement when the value of the transfer characteristic at the peak frequency exceeds a threshold value. Predict the occurrence of slip.
[0023]
A belt slip prediction device according to a seventh aspect of the present invention is the belt slip prediction device according to the fifth or sixth aspect, wherein the output-side pulley rotational speed detected by the output-side pulley rotational speed detection means, the stepless speed change. The apparatus further includes frequency calculating means for calculating the peak frequency of the transmission characteristic using at least one of a reduction ratio of the machine, a reduction ratio command value, and a belt engagement diameter on the output pulley.
[0024]
The peak frequency of the transfer characteristic changes according to the reduction ratio. The reduction ratio also changes depending on changes in the output pulley rotation speed and the belt engagement diameter on the output pulley.
[0025]
Therefore, according to the seventh aspect of the present invention, transmission is performed using at least one of the output side pulley rotational speed, the continuously variable transmission reduction ratio, the reduction ratio command value, and the belt engagement diameter on the output side pulley. By calculating the characteristic peak frequency, even if the operating state of the continuously variable transmission changes, the belt slip can be reliably predicted.
[0026]
The belt slip prediction device according to an eighth aspect of the present invention is the invention according to any one of the fifth to seventh aspects, wherein the belt slip prediction means has a predetermined function expression representing the transfer characteristic. A coefficient is identified, and occurrence of belt slip is predicted when the predetermined coefficient exceeds a threshold value.
[0027]
The belt slip prediction means uses a functional expression representing a transfer characteristic having a peak frequency. This function expression is preferably an expression that can reduce the calculation scale when identifying the coefficient as much as possible while ensuring the calculation accuracy. The predetermined coefficient among the various coefficients of the functional expression representing the transfer characteristics changes according to an increase in gain at the peak frequency.
[0028]
Therefore, according to the eighth aspect of the invention, the predetermined coefficient of the functional expression representing the transfer characteristic is identified, and the gain at the peak frequency increases when the predetermined coefficient exceeds the threshold value. Can be predicted.
[0029]
The belt slip prediction device according to claim 9 is the invention according to claim 8, wherein the belt slip prediction means determines a transfer gain at a peak frequency of a functional expression representing the transfer characteristic. And the occurrence of belt slip is predicted when the transfer gain coefficient exceeds a threshold value.
[0030]
The transfer gain coefficient is a coefficient that determines the transfer gain at the peak frequency of the functional expression representing the transfer characteristics. For example, the transfer gain increases as the transfer gain coefficient increases, and the transfer gain decreases as the transfer gain coefficient decreases.
[0031]
Therefore, according to the ninth aspect of the present invention, the transfer gain coefficient is identified, and when the transfer gain coefficient exceeds a threshold value, the transfer gain at the peak frequency increases, so that occurrence of belt slip can be predicted. it can.
[0032]
A belt clamping pressure control device according to a tenth aspect of the present invention is based on the belt slip prediction device according to any one of claims 1 to 9 and a prediction result of the belt slip prediction device. Belt clamping pressure control means for controlling the belt clamping pressure of the continuously variable transmission so as to maintain the predicted state before and after.
[0033]
When the state before and after the belt slip predicted by the belt slip prediction device is reached, the belt friction force coefficient is close to the maximum value. In this state, the belt transmission efficiency of the continuously variable transmission is the best.
[0034]
Therefore, according to the invention of claim 10, the belt clamping pressure of the continuously variable transmission is controlled based on the prediction result of the belt slip prediction device so as to maintain the state before and after the belt slip is predicted. As a result, the belt transmission efficiency of the continuously variable transmission can be improved.
[0035]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. First, a prediction principle of a belt slip of a continuously variable transmission (CVT) will be described, and then a specific embodiment will be described.
[0036]
[Prediction principle]
As a result of analyzing the CVT test data, it was found that the vibration of the sheave pressure of the secondary pulley (hereinafter abbreviated as “secondary”) increases before the macro slip occurs.
[0037]
When the transmission characteristics from the secondary rotational speed to the secondary sheave pressure were determined, a characteristic was obtained that had a resonance peak and the transmission gain at the peak frequency increased as the macro slip occurred. The same can be said for the primary side.
[0038]
The present invention uses this transfer gain characteristic to detect a macro slip precursor and predict belt slip. The following description shows the secondary side, but the same can be said for the primary side.
[0039]
1. Analysis of experimental data
FIG. 1 is a waveform diagram when a CVT (a) steady test and (b) macro slip test are performed (primary rotational speed Nin = 1000 rpm, reduction ratio γ = 0.65). FIG. 2 is a table showing the test conditions and macro slip generation method of FIG.
[0040]
As shown in FIG. 1, the amplitude of the secondary sheave pressure increased immediately before the macro slip (near the maximum value of the belt μ). The increase in the amplitude of the secondary sheave pressure is considered to be because the pulley was vibrated by the stick slip of the belt block.
[0041]
2. Transfer characteristics of secondary sheave pressure against secondary rotation speed
FIG. 3 is a diagram showing (a) an experimental waveform and (b) a transfer gain frequency characteristic when a macro slip is generated at a reduction ratio γ = 0.65. FIG. 4 is a diagram showing (a) experimental waveforms and (b) transfer gain frequency characteristics when a macro slip is generated at a reduction ratio γ = 1.0. The transfer gain here is a transfer characteristic from the secondary rotational speed Nout to the secondary sheave pressure Pc2.
[0042]
The transfer gain at the time of the macro slip sign is obtained by FFT processing of data in each FFT (fast Fourier transform) processing range (2.5 seconds) in FIGS. 3 and 4. In addition, the transmission gain in the steady state is also obtained by performing FFT processing on data for 2.5 seconds. 3 and 4, the input torque is an average value of the input torque within the FFT processing range.
[0043]
3 and 4, the transmission gain has a peak at different frequencies (around 10 Hz when the reduction ratio γ = 0.65, and around 8 Hz when the reduction ratio γ = 1.0), and the transmission gain increases at the sign of a macro slip. I understand that.
[0044]
3. Peak frequency of transfer gain
As shown in FIGS. 3 and 4, the peak frequency fc changes according to the reduction ratio γ. Therefore, equation (1) is assumed as the peak frequency fc.
[0045]
[Expression 1]

[0046]
α: Adjustment coefficient
Ns: Secondary rotation speed per second [Hz]
φ: Belt engagement angle [rad] on secondary pulley (varies according to reduction ratio γ)
The expression (1) indicates that belt vibration due to the stick-slip of the block is transmitted throughout the belt-engaged portion, and hydraulic vibration is generated at a frequency corresponding to the secondary rotational speed.
[0047]
FIG. 5 is a comparison diagram between the calculated value by the equation (1) and the experimental value of the peak frequency at each reduction ratio. The experimental value is a frequency obtained by performing FFT analysis on experimental data at the time of a macro slip sign. From FIG. 5, the calculated value and the experimental value almost coincided, and it was confirmed that the expression (1) is valid.
[0048]
4). Macroslip sign detection
The present invention detects a precursor of macro slip and predicts belt slip by paying attention to the above-described transfer gain characteristics. The detection of the macro slip precursor is performed by the following procedure (1) to procedure (4).
[0049]
Procedure (1): As shown in FIGS. 3 and 4, the transfer gain of the secondary sheave pressure with respect to the secondary rotational speed increases as the macro slip occurs.
[0050]
Procedure (2): The transfer gain characteristic is simulated with a transfer function, and the relationship between the change in transfer gain and the coefficient of the transfer function is examined.
[0051]
Step (3): Using the secondary rotational speed and the secondary sheave pressure, the transfer function coefficients are matched by the method of least squares.
[0052]
Procedure (4): A macro slip precursor is detected with reference to the matched coefficient and the examination result of procedure (2).
[0053]
Here, since the procedure (1) has already been described and the procedures (3) and (4) will be described in a specific embodiment described later, the procedure (2) will be described.
[0054]
The transfer gain characteristics shown in FIGS. 3 and 4 are simulated by the transfer function of equation (2).
[0055]
[Expression 2]

[0056]
K: Transfer gain coefficient (The larger the K, the larger the peak)
fc: Peak frequency [Hz]
s: Laplace operator
The reason why the equation (2) is a quadratic equation is to reduce the operation scale when identifying the coefficient as much as possible and to ensure the operation accuracy.
[0057]
FIG. 6 is a diagram showing the transfer gain frequency characteristic of the equation (2) when the peak frequency fc = 8 Hz and the transfer gain coefficient K is used as a parameter. As shown in FIG. 6, a peak appears as the transfer gain coefficient K increases. That is, the transfer gain characteristic of the equation (2) is similar to the transfer gain characteristic at the primary peak frequency shown in FIGS.
[0058]
Equation (2) is a continuous system. When equation (2) is expressed in a discrete system so that calculation of online identification described later is easy, equation (3) is obtained.
[0059]
[Equation 3]

[0060]
In the equation (3), z is an operator for z conversion. The coefficients A1, A2, B1, and B2 are unknown coefficients, and are identified online by the secondary rotation speed and the secondary sheave pressure. The macro slip precursor is detected based on the value of the coefficient A1.
[0061]
5. Relation between coefficient A1 and transfer gain coefficient K
(5-1) Detection of macro slip precursor based on coefficient A1
The coefficients A1, A2, B1, and B2 are expressed by equations (4) to (7), respectively, using the sampling period dt.
[0062]
[Expression 4]

[0063]
FIG. 7 is a diagram showing changes in the coefficient A1 with respect to the transfer gain coefficient K and the peak frequency fc when the expression (4) is used at the sampling period dt = 4 ms. From FIG. 7, the value of the coefficient A1 decreased as the transfer gain coefficient K increased. On the other hand, when a macro slip sign is reached, a peak appears and the transfer gain coefficient K increases.
[0064]
Therefore, when the coefficient A1 exceeds (below) the threshold value, a macro slip sign can be detected. However, since the value of the coefficient A1 varies depending on the peak frequency fc, it is necessary to change the threshold according to the reduction ratio and the secondary rotation speed.
[0065]
(5-2) Detection of macro slip precursor based on transfer gain coefficient K
As shown in FIG. 7, the transfer function coefficient A1 changes according to the transfer gain coefficient K and the peak frequency fc. For this reason, the threshold value must be changed according to the reduction ratio and the secondary rotational speed. Therefore, in order to fix the threshold, the parameter for detecting the macro slip precursor is changed from the coefficient A1 to the transfer gain coefficient K. When the expressions (1) and (4) are used, the transfer gain coefficient K becomes the expression (8).
[0066]
[Equation 5]

[0067]
When a macro slip sign is reached, a peak appears and the transfer gain coefficient K increases. Therefore, when the transfer gain coefficient K exceeds (becomes) a certain threshold, a macro slip precursor can be detected.
[0068]
[Specific Embodiment]
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.
[0069]
(First embodiment)
FIG. 8 is a block diagram illustrating a configuration of the belt clamping pressure control device according to the first embodiment. The belt clamping pressure control device includes a primary pulley on the input shaft side (hereinafter abbreviated as “primary”), a secondary pulley on the output shaft side (hereinafter abbreviated as “secondary”), and a primary pulley and a secondary pulley. A belt slip of a continuously variable transmission composed of a belt that has been applied is predicted to control the belt clamping pressure to an optimum state.
[0070]
The belt clamping pressure control device corresponds to a primary rotation sensor 11 that generates a primary rotation pulse signal, a secondary rotation sensor 12 that generates a secondary rotation pulse signal, and a secondary sheave pressure that is a hydraulic pressure for controlling the secondary. A sheave pressure sensor 13 that generates a sensor signal, a pressure control valve 14 that adjusts the secondary sheave pressure, and an electronic control unit that controls the belt clamping pressure by adjusting the pressure control valve 14 based on signals from each sensor ( (Hereinafter referred to as “ECU”) 20.
[0071]
The primary rotation sensor 11 generates a rotation pulse signal corresponding to the primary rotation speed, and supplies this rotation pulse signal to the ECU 20. The secondary rotation sensor 12 generates a rotation pulse signal corresponding to the secondary rotation speed and supplies the rotation pulse signal to the ECU 20.
[0072]
The sheave pressure sensor 13 generates a sensor signal according to a secondary sheave pressure, which is a hydraulic pressure for urging the movable sheave toward the fixed sheave, with respect to the secondary composed of the fixed sheave and the movable sheave. To supply.
[0073]
FIG. 9 is a block diagram showing a configuration of the ECU 20. The ECU 20 calculates a primary rotational speed detection circuit 21 that detects a primary rotational speed, a secondary rotational speed detection circuit 22 that detects a secondary rotational speed, a sheave pressure detection circuit 23 that detects a secondary sheave pressure Pc2, and a peak frequency fc. A peak frequency calculation circuit 24, a band pass filter 25, a threshold comparison circuit 26 that compares a threshold with a coefficient A1 described later, and a clamping pressure control circuit 27 that controls the belt clamping pressure.
[0074]
The primary rotation speed detection circuit 21 detects the primary rotation speed Nin [rpm] based on the rotation pulse signal from the primary rotation sensor 11 and supplies the primary rotation speed Nin to the peak frequency calculation circuit 24.
[0075]
The secondary rotation speed detection circuit 22 detects the secondary rotation speed Nout [rpm] based on the rotation pulse signal from the secondary rotation sensor 12 and supplies the secondary rotation speed Nout to the peak frequency calculation circuit 24.
[0076]
The sheave pressure detection circuit 23 detects the secondary sheave pressure Pc2 based on the sensor signal from the sheave pressure sensor 13, and supplies the secondary sheave pressure Pc2 to the bandpass filter 25.
[0077]
The peak frequency calculation circuit 24 calculates a reduction ratio γ (= Nin / Nout) based on the primary rotation speed Nin and the secondary rotation speed Nout, and uses the reduction ratio γ to make a belt engagement angle φ [rad] on the secondary pulley. Calculate Then, the peak frequency fc is calculated according to the equation (9) using the preset adjustment coefficient α and the secondary rotational speed Ns [Hz] (= Nout / 60) per second.
[0078]
[Formula 6]

[0079]
The bandpass filter 25 has a bandpass characteristic of equation (10). Note that fc is a peak frequency and s is a Laplace operator.
[0080]
[Expression 7]

[0081]
The bandpass filter 25 extracts the band of the peak frequency fc from the secondary sheave pressure Pc2, and supplies the filtered secondary sheave pressure Pc2 to the belt clamping pressure control circuit 26.
[0082]
The threshold comparison circuit 26 compares the filtered secondary sheave pressure Pc2 with the threshold, detects a macro slip precursor when the filtered secondary sheave pressure Pc2 exceeds the threshold, and predicts belt slip. That is, the threshold value comparison circuit 26 can predict the belt slip by detecting when the amplitude of the secondary sheave pressure Pc2 has increased from a predetermined value. The threshold value may be a value set by learning according to the CVT state quantity, or may be a preset constant value, and is not particularly limited.
[0083]
The clamping pressure control circuit 27 adjusts the pressure control valve 14 based on the belt slip prediction result by the threshold comparison circuit 26. Specifically, the pressure control valve 14 is controlled such that the filtered secondary sheave pressure Pc2 maintains a value near the threshold. As a result, the belt μ of the CVT can be maintained near the maximum value.
[0084]
As described above, the belt clamping pressure control device according to the first embodiment can predict the belt slip by detecting the amplitude increase of the secondary sheave pressure Pc2 and capturing the macro slip precursor. In addition, the belt clamping pressure control device can predict belt slip without using a sensor that detects the hydraulic pressure of the primary pulley (if the belt slip is predicted on the primary side, the hydraulic pressure of the secondary pulley). Cost.
[0085]
Furthermore, the belt clamping pressure control device can control the belt clamping pressure using the belt slip prediction result, so that the belt μ of the CVT can be maintained near the maximum value, thereby improving the belt transmission efficiency of the CVT. Can be made.
[0086]
The belt clamping pressure control device according to the present embodiment uses the bandpass filter 25 and the threshold comparison circuit 26. However, the belt clamping pressure control device is limited to the above configuration as long as the fluctuation amplitude of the secondary sheave pressure Pc2 can be detected. It is not a thing.
[0087]
Further, the peak frequency calculation circuit 24 calculates the peak frequency fc according to the equation (9), but is not limited to the equation (9) as long as the peak frequency fc can be calculated. For example, in addition to the secondary rotation speed and the reduction ratio, the peak frequency calculation circuit 24 has one or more arbitrary CVT state quantities (for example, a reduction ratio command value, a belt engagement diameter on the secondary, etc.) correlated with these. ) May be used.
[0088]
(Second Embodiment)
Next, a second embodiment of the present invention will be described. The same circuits as those in the first embodiment are denoted by the same reference numerals, and detailed description thereof is omitted.
[0089]
FIG. 10 is a block diagram showing a configuration of the ECU 20A according to the second embodiment. The belt clamping pressure control apparatus according to the second embodiment is an application of an ECU 20A shown in FIG. 10 instead of the ECU 20 described in the first embodiment.
[0090]
The ECU 20A includes a primary rotation speed detection circuit 21, a secondary rotation speed detection circuit 22, a sheave pressure detection circuit 23, a peak frequency calculation circuit 24, a reduction ratio calculation circuit 31 that calculates a reduction ratio γ, a threshold generation circuit 32 that generates a threshold value, Band pass filters 33 and 34, a transfer function identification circuit 35 for identifying a coefficient of the transfer function, a threshold comparison circuit 36 for comparing the threshold with a predetermined coefficient, and a clamping pressure control circuit 37 for controlling the belt clamping pressure.
[0091]
The primary rotational speed detection circuit 21 detects the primary rotational speed Nin and supplies the primary rotational speed Nin to the peak frequency calculation circuit 24 and the reduction ratio calculation circuit 31. The secondary rotation speed detection circuit 22 detects the secondary rotation speed Nout and supplies the secondary rotation speed Nout to the peak frequency calculation circuit 24, the reduction ratio calculation circuit 31, the threshold value generation circuit 32, and the bandpass filter 33. The sheave pressure detection circuit 23 detects the secondary sheave pressure Pc2, and supplies the secondary sheave pressure Pc2 to the bandpass filter 34.
[0092]
The peak frequency calculation circuit 24 calculates a reduction ratio γ (= Nin / Nout) based on the primary rotation speed Nin and the secondary rotation speed Nout, and uses the reduction ratio γ to make a belt engagement angle φ [rad] on the secondary pulley. Calculate Then, the peak frequency fc is calculated according to the equation (11) using the preset adjustment coefficient α and the secondary rotational speed Ns [Hz] (= Nout / 60) per second.
[0093]
[Equation 8]

[0094]
Then, the peak frequency calculation circuit 24 supplies the calculated peak frequency fc to the band pass filters 33 and 34.
[0095]
On the other hand, the reduction ratio calculation circuit 31 calculates the reduction ratio γ based on the primary rotation speed Nin and the secondary rotation speed Nout, and supplies the reduction ratio γ to the threshold value generation circuit 32.
[0096]
The threshold generation circuit 32 generates a threshold for detecting a macro slip precursor, that is, a threshold for comparison with the transfer function coefficient A1. Here, the threshold value generation circuit 32 stores a threshold value map representing the relationship among the reduction ratio γ, the secondary rotation speed Nout, and the threshold value. Then, the threshold generation circuit 32 refers to the threshold map and generates a threshold corresponding to the current reduction ratio γ and the secondary rotation speed Nout.
[0097]
Further, the threshold value generation circuit 32 may learn each signal (for example, a signal such as a primary rotation speed and an output torque) in the steady state of the CVT, and generate a threshold value when each signal becomes maximum.
[0098]
The bandpass filters 33 and 34 have the bandpass characteristic of the above-described equation (10). The band pass filter 33 extracts the band of the peak frequency fc from the secondary rotation speed Nout, and supplies the filtered secondary rotation speed Nout to the transfer function identification circuit 35. Further, the band pass filter 34 extracts the band of the peak frequency fc from the secondary sheave pressure Pc2, and supplies the filtered secondary sheave pressure Pc2 to the transfer function identification circuit 35.
[0099]
The transfer function identification circuit 35 identifies the unknown coefficients A1, A2, B1, and B2 of the transfer function expressed by the above-described equation (3) using the filtered secondary rotation speed Nout and the secondary sheave pressure Pc2.
[0100]
Here, in order to perform online identification, the expression (3) is rewritten as the expression (12) using the sampling interval i.
[0101]
[Equation 9]

[0102]
In the equation (12), the secondary rotation speed Nout and the secondary sheave pressure Pc2 are values after being filtered by the band-pass filter, respectively.
[0103]
In order to identify the coefficients A1, A2, B1, and B2 of the equation (12) online by the method of least squares, the equation (12) is expressed by the following equation (13).
[0104]
[Expression 10]

[0105]
However, in the equation (13), the following equations (14) to (16) hold.
[0106]
[Expression 11]

[0107]
Further, a weighted least square method according to the following equations (17) and (18) is used.
[0108]
[Expression 12]

[0109]
In Expressions (17) and (18), P is a 4 × 4 target matrix. The initial value of P is a diagonal matrix, and each element of P is set to a large value in order to speed up the convergence of parameter identification. λ is a weight.
[0110]
The transfer function identification circuit 35 identifies τ (i) using the secondary rotation speed Nout (i) and the secondary sheave pressure Pc2 (i) for each sampling period in accordance with the equations (13) to (18). To do. Then, the transfer function identification circuit 35 supplies a coefficient A1 that is an element of the identified τ (i) to the threshold comparison circuit 36.
[0111]
The threshold comparison circuit 36 compares the threshold generated by the threshold generation circuit 32 with the coefficient A1 identified by the transfer function identification circuit 35, and when the coefficient A1 exceeds the threshold continuously for four sampling periods ( Macroslip precursor is detected when In order to detect the macro slip precursor with certainty, the condition is “continuous for four sampling periods”. However, the present invention is not limited to this.
[0112]
FIGS. 11 to 13 are diagrams showing (a) the steady state and (b) the primary rotation speed, the output torque, the secondary sheave pressure (secondary sheave pressure), and the coefficient A1 identification value during macro slip. Primary rotation speed Nin = 1000 rpm, steady-state input torque Tin = 50 Nm. 11 shows the reduction ratio γ = 0.65, FIG. 12 shows the reduction ratio γ = 1.0, and FIG. 13 shows the reduction ratio γ = 2.379. The threshold value was a value at which each signal in the steady state of the CVT becomes the maximum value.
[0113]
The clamping pressure control circuit 37 adjusts the pressure control valve 14 based on the belt slip prediction result by the threshold comparison circuit 36. Specifically, the pressure control valve 14 is controlled so that the coefficient A1 of the transfer function maintains a value near the threshold value.
[0114]
As described above, the belt clamping pressure control apparatus according to the second embodiment identifies the transfer function coefficient A1 that simulates the transfer gain characteristic of the secondary sheave pressure Pc2 with respect to the secondary rotation speed Nout, and sets the coefficient A1 and the threshold value. By comparing, it is possible to predict a macro slip and predict belt slip.
[0115]
Further, the belt clamping pressure control device does not require a sensor for detecting the hydraulic pressure of the primary pulley (when the belt slippage is predicted on the primary side) as in the first embodiment, it does not require a sensor. The cost can be reduced as compared with the above, and the belt transmission efficiency of the CVT can be improved by maintaining the belt μ of the CVT near the maximum value.
[0116]
Note that the belt clamping pressure control device may include low-pass filters having the same characteristics instead of the band-pass filters 33 and 34 described above. As a transfer characteristic of the low-pass filter, for example, the following equation (19) is preferable.
[0117]
[Formula 13]

[0118]
T (= 2πf): Low-pass filter time constant [s]
f: Cut-off frequency (> fc) [Hz]
s: Laplace operator
Further, the peak frequency calculation circuit 24 calculates the peak frequency fc according to the equation (11). However, the peak frequency calculation circuit 24 is not limited to the equation (11) as long as the peak frequency fc can be calculated. For example, in addition to the secondary rotation speed and the reduction ratio, the peak frequency calculation circuit 24 may include any one or more CVT state quantities (for example, a reduction ratio command value, a belt engagement diameter on the secondary, etc.) correlated with these. ) May be used. The peak frequency calculation circuit 24 is the same in the third embodiment.
[0119]
(Third embodiment)
Next, a third embodiment of the present invention will be described. The same circuits as those of the second embodiment are denoted by the same reference numerals, and detailed description thereof is omitted.
[0120]
The belt clamping pressure control device according to the second embodiment described above predicts belt slip by comparing the coefficient A1 of the transfer function with a threshold value. On the other hand, the belt clamping pressure control apparatus according to the third embodiment predicts belt slip by comparing the transfer gain coefficient K of the transfer function with a threshold value.
[0121]
FIG. 14 is a block diagram showing a configuration of an ECU 20B according to the third embodiment. The ECU 20B is configured to further include a transfer gain coefficient calculation circuit 38 except for the reduction ratio calculation circuit 31 and the threshold value generation circuit 32 from the ECU 20A according to the second embodiment.
[0122]
The transfer gain coefficient calculation circuit 38 uses the coefficient A1 identified by the transfer function identification circuit 35, the peak frequency fc calculated by the peak frequency calculation circuit 24, and the sampling period dt according to the equation (20). Is calculated.
[0123]
[Expression 14]

[0124]
The threshold comparison circuit 36 compares a preset threshold with the transfer gain coefficient K calculated by the transfer gain coefficient calculation circuit 38, and when the transfer gain coefficient K exceeds the threshold continuously for four sampling periods. Detect macroslip precursor (when exceeded). As in the second embodiment, the present invention is not limited to “continuous for four sampling periods”.
[0125]
FIG. 15 shows the case where (a) reduction ratio γ = 0.65, (b) reduction ratio γ = 1.0, and (c) reduction ratio γ = 2.379, using transfer gain coefficient K and coefficient A1. It is a figure which shows the comparison result at the time of detecting a macro slip precursor, respectively. The threshold value of the transfer gain coefficient K is a value when the transfer gain coefficient K of each reduction ratio γ in the steady state becomes the maximum value. From FIG. 15, it was found that the threshold value can be set to a constant value and that the macro slip precursor can be detected using the transfer gain coefficient K.
[0126]
The clamping pressure control circuit 37 adjusts the pressure control valve 14 based on the belt slip prediction result by the threshold comparison circuit 36. Specifically, the pressure control valve 14 is controlled so that the transfer gain coefficient K maintains a value near the threshold value.
[0127]
As described above, the belt clamping pressure control device according to the third embodiment identifies the transfer gain coefficient K of the transfer function that simulates the transfer gain characteristic of the secondary sheave pressure Pc2 with respect to the secondary rotation speed Nout, and transfers the transfer gain coefficient K. By comparing the threshold value with the threshold value, it is possible to capture the sign of macro slip and predict belt slip. In addition, since the threshold value can be set to a predetermined value in advance, it is possible to save the trouble of changing the threshold value as needed according to the state of the CVT.
[0128]
Also, the belt clamping pressure control device is a sensor that detects the hydraulic pressure supplied to the primary pulley (when the belt slip is predicted on the primary side) as in the first and second embodiments. Therefore, the cost can be reduced compared to the conventional case, and the belt transmission efficiency of the CVT can be improved by maintaining the belt μ of the CVT near the maximum value.
[0129]
Note that the belt clamping pressure control device according to the present embodiment may also be provided with a low-pass filter instead of the band-pass filters 33 and 34, as in the second embodiment. Of course, the threshold for comparison with the transfer gain coefficient K is not limited to the values shown in FIG.
[0130]
【The invention's effect】
The belt slip prediction apparatus according to the present invention can predict belt slip at a low cost based on the variation of the sheave pressure on one side without using the two sheave pressures on the input side and the output side.
[0131]
The belt clamping pressure control device according to the present invention controls the belt clamping pressure of the continuously variable transmission so as to maintain the state before and after the belt slip is predicted based on the prediction result of the belt slip prediction device. The belt transmission efficiency of the continuously variable transmission can be improved.
[Brief description of the drawings]
FIG. 1 is a waveform diagram when a (a) steady test and (b) macro slip test of CVT were performed.
FIG. 2 is a table showing the test conditions and macro slip generation method of FIG. 1;
FIG. 3 is a diagram showing (a) an experimental waveform and (b) a transfer gain frequency characteristic when a macro slip is generated at a reduction ratio γ = 0.65.
FIG. 4 is a diagram showing (a) an experimental waveform and (b) a transfer gain frequency characteristic when a macro slip is generated at a reduction ratio γ = 1.0.
FIG. 5 is a comparison diagram between a calculated value and an experimental value of a peak frequency at each reduction ratio.
FIG. 6 is a diagram showing the transfer gain frequency characteristic of the equation (2) when the peak frequency fc = 8 Hz and the transfer gain coefficient K is used as a parameter.
FIG. 7 is a diagram showing changes in a coefficient A1 with respect to a transfer gain coefficient K and a peak frequency fc when the expression (4) is used at a sampling period dt = 4 ms.
FIG. 8 is a block diagram showing a configuration of a belt clamping pressure control device according to the first embodiment.
FIG. 9 is a block diagram showing a configuration of an ECU.
FIG. 10 is a block diagram showing a configuration of an ECU according to a second embodiment.
FIG. 11 is a diagram showing (a) steady state and (b) primary rotational speed, output torque, secondary sheave pressure (secondary sheave pressure), and coefficient A1 identification value at the time of reduction ratio γ = 0.65. .
FIG. 12 is a diagram showing (a) steady state and (b) primary rotational speed, output torque, secondary sheave pressure (secondary sheave pressure), and coefficient A1 identification value at the time of a reduction ratio γ = 1.0. .
FIG. 13 is a diagram showing (a) steady state and (b) primary rotational speed, output torque, secondary sheave pressure (secondary sheave pressure), and coefficient A1 identification value at the time of macro slip in a reduction ratio γ = 2.379. .
FIG. 14 is a block diagram showing a configuration of an ECU according to a third embodiment.
FIGS. 15A and 15B show the transmission gain coefficient K and coefficient A1 when (a) reduction ratio γ = 0.65, (b) reduction ratio γ = 1.0, and (c) reduction ratio γ = 2.379. It is a figure which shows the comparison result at the time of detecting a macro slip precursor, respectively.
[Explanation of symbols]
11 Primary rotation sensor
12 Secondary rotation sensor
13 Sheave pressure sensor
14 Pressure control valve
20, 20A, 20B ECU
21 Primary rotational speed detection circuit
22 Secondary rotational speed detection circuit
23 Sheave pressure detection circuit
24 Peak frequency calculation circuit
25, 33, 34 Band pass filter
26, 36 Threshold comparison circuit
27, 37 Clamping pressure control circuit
31 Reduction ratio calculation circuit
32 threshold generation circuit
35 Transfer function identification circuit
38 Transfer gain coefficient calculation circuit

Claims (10)

A belt slip prediction device for a continuously variable transmission including an input pulley, an output pulley, and a belt,
A sheave pressure detecting means for detecting the sheave pressure;
Belt slip prediction means for predicting the occurrence of belt slip based on fluctuations in the sheave pressure detected by the sheave pressure detection means;
Belt slip prediction device with The belt slip prediction device according to claim 1, wherein the belt slip prediction unit predicts occurrence of belt slip based on a fluctuation amplitude of the sheave pressure. The belt slip prediction means extracts a change in sheave pressure in a predetermined band from a sheave pressure detected by the sheave pressure detection means, and predicts occurrence of belt slip based on the change in sheave pressure in the predetermined band. The belt slip prediction apparatus according to 1. Output-side pulley rotation speed detecting means for detecting the output-side pulley rotation speed;
Using at least one of the output-side pulley rotation speed detected by the output-side pulley rotation speed detection means, the reduction ratio of the continuously variable transmission, the reduction ratio command value, and the belt engagement diameter on the output-side pulley, A band calculating means for calculating a predetermined band;
The belt slip prediction apparatus according to claim 3, further comprising: An output side pulley rotation number detecting means for detecting the output side pulley rotation number;
The belt slip prediction device according to claim 1, wherein the belt slip prediction means predicts occurrence of belt slip based on a value at a peak frequency of a transfer characteristic of the sheave pressure with respect to the output pulley rotation speed. 6. The belt slip prediction device according to claim 5, wherein the belt slip prediction means predicts occurrence of belt slip when a value of the transfer characteristic at the peak frequency exceeds a threshold value. Using at least one of the output-side pulley rotation speed detected by the output-side pulley rotation speed detection means, the reduction ratio of the continuously variable transmission, the reduction ratio command value, and the belt engagement diameter on the output-side pulley, The belt slip prediction apparatus according to claim 5 or 6, further comprising frequency calculation means for calculating the peak frequency of the transfer characteristic. The belt slip prediction means identifies a predetermined coefficient of a functional expression representing the transfer characteristic, and predicts occurrence of belt slip when the predetermined coefficient exceeds a threshold value. The belt slip prediction apparatus according to the item. The belt slip prediction means identifies a transfer gain coefficient for determining a transfer gain at a peak frequency of a functional expression representing the transfer characteristic, and predicts occurrence of belt slip when the transfer gain coefficient exceeds a threshold value. The belt slip prediction apparatus according to claim 8. A belt slip prediction device according to any one of claims 1 to 9,
Belt clamping pressure control means for controlling the belt clamping pressure of the continuously variable transmission so as to maintain the state before and after the belt slip is predicted based on the prediction result of the belt slip prediction device;
A belt clamping pressure control device.

Images