I-252927 i-Automatic Transmission AL4 DPO Switch Pressure Sensor

1. Iindlela zokuxilongwa kwempazamo eqhelekileyo yenzwa

Ngophuhliso lwenzululwazi kunye nobuchwepheshe, iindlela zokuxilongwa kwempazamo yenzwa zininzi kwaye zininzi, ezinokuhlangabezana neemfuno zokusetyenziswa kwansuku zonke. Ngokukodwa, iindlela eziqhelekileyo zokuxilongwa kwe-sensor ye-sensor zibandakanya ezi zilandelayo:

1.1 Ukuxilongwa kwempazamo esekwe kwimodeli

Imodeli yokuqala ephuhliswe i-sensor ye-fault diagnosis iteknoloji ithatha i-analytical redundancy endaweni yokunciphisa umzimba njengengcamango yayo ephambili, kwaye ifumana ulwazi olunempazamo ngokuyithelekisa kunye nemveliso yexabiso elilinganisiweyo yinkqubo yoqikelelo. Okwangoku, le teknoloji yokuxilongwa inokwahlulwa ibe ngamacandelo amathathu: iparameter estimation-based fault diagnosis method, indlela yokuxilongwa kwesiphoso esekwe kurhulumente kunye nendlela yokuxilongwa kwendawo efanayo. Ngokubanzi, sichaza iiparameters zeempawu zamacandelo ezenza inkqubo yomzimba njengeeparamitha zemiba, kunye nokwahlula okanye ukungafani kwee-equations ezichaza inkqubo yokulawula njengeeparamitha zemodyuli. Xa inzwa kwisistim ihluleka ngenxa yomonakalo, ukungaphumeleli okanye ukuthotywa kwentsebenzo, inokubonakaliswa ngokuthe ngqo njengotshintsho lweeparitha zezinto eziphathekayo, ezibangela ukuba utshintsho lweeparamitha zemodulus, eziqulethe yonke ingcaciso ephosakeleyo. Ngokuchasene noko, xa iiparameters zemodyuli ziyaziwa, utshintsho lweparameter lunokubalwa, ukwenzela ukuba unqume ubungakanani kunye neqondo lesiphoso senzwa. Okwangoku, iteknoloji yokuxilongwa kwe-sensor esekelwe kwimodeli isetyenziswe ngokubanzi, kwaye iziphumo zayo zophando zijolise kwiinkqubo ezihambelanayo, kodwa uphando kwiinkqubo ezingabonakaliyo kufuneka ziqiniswe.

1.2 Ukuxilongwa kweempazamo ezisekelwe kulwazi

Ngokwahlukileyo kwiindlela zokuxilongwa kwesiphoso ezikhankanywe ngasentla, ukuxilongwa kwempazamo esekelwe kulwazi akufuneki ukuseka imodeli yemathematika, eyoyisa iintsilelo okanye iziphene zokuxilongwa kweempazamo ezisekelwe kwimodeli, kodwa ayinayo isethi yenkxaso yethiyori eqolileyo. Phakathi kwabo, indlela yenethiwekhi ye-neural yokwenziwa ngummeli wokuxilongwa kwempazamo esekelwe kulwazi. Inethiwekhi ebizwa ngokuba yi-artificial neural network ifinyezwa njenge-ANN ngesiNgesi, esekelwe ekuqondeni komntu inethiwekhi ye-neural yengqondo kwaye iqonda umsebenzi othile ngokwakhiwa kokwenziwa. Inethiwekhi ye-neural ye-Artificial neural inokugcina ulwazi ngendlela esasazwa ngayo, kwaye iqonde inguqu engabonakaliyo kunye nemephu ngoncedo lwe-topology yenethiwekhi kunye nokusabalalisa ubunzima. Ngokwahlukileyo, indlela yenethiwekhi ye-neural eyenziweyo yenza ukusilela kokuxilongwa kwempazamo esekwe kwimodeli kwiinkqubo ezingezizo. Nangona kunjalo, indlela yenethiwekhi ye-neural eyenziweyo ayifezekanga, kwaye ixhomekeke kuphela kwiimeko ezithile eziphathekayo, ezingasebenzisi ngokufanelekileyo amava aqokelelweyo kwiinkalo ezikhethekileyo kwaye ziphenjelelwa ngokulula kukhetho lwesampulu, ngoko ke izigqibo zokuxilonga ezithathwe kuyo azikho. inokutolika.