Riba:
- Ikon amfani da AI azaman taimakon takaddun shaida don tsarin matrix na aikin HAZOP, tambayoyin kalma da gyaran mintuna
- Ƙarfin fahimtar cewa ainihin nazarin haɗari (haɗari na ainihi, tsananin sakamako, isashen matakan kiyayewa) nauyin da ba za a iya ba shi ba ne na ƙungiyar ƙwararrun ƙwararrun ƙwararru.
- Samun damar bambance waccan rarrabuwar haɗari da yanke shawara na taka tsantsan sun dogara da ƙungiyar da aka ba da izini da amincewa, kuma ƙayyadaddun ƙayyadaddun hatsarori ba su wanzu a cikin basirar wucin gadi.
Babban alhakin injiniyan sinadarai shine aminci. Hasashen inda da kuma yadda tsari zai iya yin kuskure da yin taka tsantsan shine mafi mahimmancin aikin injiniya na wurin. Hanyar da ta fi dacewa don wannan ita ce nazarin HAZOP (Hazard and Aperability Study), wanda ke nazarin kullin tsari ta kumburi kuma a tsari yana bincika yiwuwar karkace a cikin kowace siga. HAZOP tsari ne mai ladabi da aiki: ga kowane layi, kowane yanki na kayan aiki, kowane siga, "idan me?" ake tambaya. A cikin wannan rukunin za ku koyi yadda ake amfani da AI azaman taimako wajen haɓaka HAZOP da takaddun tantance haɗari; amma za mu koyi dalilin da ya sa ba za a taɓa iya canja wurin ainihin nazarin haɗari zuwa AI ba.
Mafi ƙaƙƙarfan ƙaƙƙarfan tsarin ya shafi nan: HAZOP aikin ƙwararrun ƙwararrun tsaro ne; AI yana samar da kwarangwal ɗin takaddun kawai, jerin kalmomi da ƙayyadaddun bayanai, ainihin ƙimar haɗari na ƙungiyar ƙwararrun ƙwararru ne. Hadarin da ba a kula da shi yana kashe rayuwar ɗan adam a zahiri, ba akan takarda ba. Babu wani layin karkatacciya da AI ya samar da za a yi la'akari da ingancin aiki ba tare da bitar ƙungiya da amincewar jami'ai ba.
Hankalin HAZOP da iyakacin rawar AI
HAZOP yana raba tsari zuwa nodes (kumburi - wani ɓangare na tsari wanda ke da manufar ƙira ɗaya; misali, layi daga famfo zuwa tanki). A kowane kumburi, ana la'akari da ma'auni (yawan kwarara, matsa lamba, zafin jiki, matakin, maida hankali) kuma ana duba yiwuwar karkatar da su tare da kalmomin jagora (kalmomin da aka yi amfani da su don samar da ɓarna cikin tsari: BABU / BA, MORE, KASA, JAYA, BANGASKIYA, SAI): "Idan yawan kwarara ya fi? Ga kowane karkatacciyar hanya, an tattauna musabbabi, sakamakon, kariyar da ake da su, da taka tsantsan.
A cikin wannan tsari, AI na iya haɓaka abubuwan da ke biyowa: ƙaddamar da matrix node/parameter/matrix matrix, tunawa da misalan ƙetare-sakamakon-sakamako, shirya mintuna tattaunawa a cikin teburin da aka tsara, tunawa da abubuwa masu kama daga HAZOPs na baya. Amma wane karkata ne haxari na gaske a cikin wannan hanya, munin sakamakon da kuma isasshiyar yin taka tsantsan lamari ne na yanke hukunci na ƙungiyoyi masu yawa.
Tukwici: Yi amfani da AI KAFIN taron HAZOP don kwarangwal da tunasarwa, da BAYAN taro don gyara mintoci. A cikin taron da kansa, ana yin kimanta haɗarin haɗari ta hanyar tattaunawa da masana waɗanda suka san ainihin tsari; AI ba maye gurbin waccan tattaunawar ba.
Ƙimar haɗari: yuwuwar × tsanani
Lokacin da aka sami haɗari, ya zama dole a kimanta haɗarin. Hanyar gama gari ita ce matrix ɗin haɗari - tebur wanda ke ƙetare yuwuwar aukuwa tare da tsananin sakamakonsa da launuka matakin haɗarin. AI na iya samar da samfurin matrix na haɗari da kuma shawarwarin farko na wane nau'in abin da zai iya faɗo a ciki; amma yana buƙatar godiyar yuwuwar da tsanani, ilimin filin, da hukuncin injiniya. A cikin bincike na ci gaba kamar LOPA (Layer of Protection Analysis - hanyar da ke bincika ƙididdiga ko matakan kariya masu zaman kansu na yanayin sun isa), AI kawai yana samar da tsarin lissafin kuɗi da takaddun; zato da ƙarewa suna ƙarƙashin ingantaccen kimantawa.
Tsanaki: Rarraba haɗari a matsayin "ƙananan haɗari" yana nufin rashin yin taka tsantsan. Barin wannan rarrabuwa zuwa AI yana buɗe kofa zuwa bala'in da ba a kula da shi ba. Matsayin haɗari da yanke shawara na taka tsantsan na ƙungiyar da aka ba da izini; AI kawai yana ba da zayyanawa da tunatarwa.
uku mini lokuta
Case 1 - Haɗawar mintuna. A cikin taron HAZOP na kwanaki 3, ƙungiyar ta samar da kusan layukan karkata 240. AI ya tsara bayanan da aka warwatse a cikin tebur da aka tsara tare da kumburi / siga / tsarin kalmar jagora; An rage lokacin takardun da kusan 40%. Amma kowane layi ya riga ya kimanta ta ƙungiyar a cikin taron; AI kawai na gyara post dina.
Case 2 - Ƙimar tunatarwa. A wani kumburi, AI ta tuna da yanayin yanayin "taruwa da yiwuwar haɗari mai haɗari a cikin reactor idan famfon ciyarwa ya tsaya" daga irin wannan binciken da ya gabata. Tawagar ta kimanta wannan yanayin bisa ƙayyadaddun tsari kuma sun gano cewa yana da inganci, don haka sun ƙara ma'aunin tsaka-tsaki. Tunasarwar ta taimaka; Matakin ya rage ga kungiyar.
Shari'a ta 3 - Hatsarin karfin gwiwa. Wani sabon injiniya ya yi tunanin cewa daftarin HAZOP da YZ ya samar kusan aiki ne da aka shirya kuma yana so ya rage taron. Babban injiniyan aminci ya tsaya: Daftarin AI bai san wani abin da ke tattare da tsarin ba da kuma ƙayyadaddun ƙayyadaddun kulawa; Waɗannan biyun sun kasance daga cikin manyan haɗari. An gudanar da taron gabaɗaya, kuma a cikin tattaunawar ƙungiyar ne kawai al'amura biyu masu mahimmanci suka bayyana.
Samfura huɗu masu kwafi
1) HAZOP kwarangwal (kafin haduwa):
Matsayinku: HAZOP mataimakin mai gudanarwa. Zan kwatanta maku P&ID node. Saita ni da HAZOP CHART SKELETON don wannan kumburi: sigogi (yawan kwarara, matsa lamba, zafin jiki, matakin, maida hankali) ku CAUSE / RESULT / PRECAUTIONFILLING - ƙwararrun ƙungiyar za su kimanta waɗannan a cikin taron. Node: [ma'anar]
2) Tunasarwar yanayi na al'ada:
Tunatar da ni da SANIN yanayin karkatar da al'amura a cikin wallafe-wallafen / makamantan matakai na wannan kumburi da siga (misali-sakamako).Waɗannan tunatarwa ne kawai don ƙungiyar kada su manta; Ƙwararrun ƙwararrun za su yanke shawara ko an yi amfani da wannan tsari ko a'a. Ka fayyace cewa akwai yuwuwar samun wasu abubuwa na musamman ga tsarin da ba ku sani ba. Node/parameter: [definition]
3) Shirya mintuna na HAZOP (bayan taro):
A ƙasa akwai tarwatsa bayanan HAZOP da ƙungiyar ta samar a taron. Tsara su cikin madaidaicin tebur: Node | Siga | Kalmar Jagora | Karya | Me yasa | Kammalawa | Kariyar yanzu | Shawara | Alhaki. CANZA, KARA KO KARA BAYANI A CIKIN ABUBUWA; Kawai gyara abin da ƙungiyar ta ce. Alama wuraren da kuke gani sun ɓace/masu cin karo da juna kamar "ya kamata ƙungiyar ta tabbatar". Bayanan kula: [manna]
4) Rarraba matrix na farko (shawarwari, ba yanke shawara):
Ba da shawarar matrix na haɗari PRE-rarrabuwa don yanayin mai zuwa: [yanayin]. Rubuta dalilin ku na yuwuwar da tsanani, amma jaddada cewa wannan SHAWARWARI ne kuma cewa yuwuwar/nauyin ƙima na ƙarshe da matakin taka tsantsan ya rataya ne ga ƙungiyar da aka ba da izini. Nuna abin da ƙarin bayani zai iya canza rarrabuwa. Bayyana ainihin matakin haɗari.
Rauni mai ƙarfi / Ƙarfi mai ƙarfi
Rawanin faɗakarwa:
Yi HAZOP don wannan tsari.
Ba tare da sanin ainihin tsarin ba, AI yana samar da abin gaskatawa amma bai cika "HAZOP" ba; ya rasa mafi mahimmanci, ƙayyadaddun hadurran yanar gizo. Wannan yana da haɗari.
Ƙarfi mai ƙarfi:
Matsayinku: HAZOP mataimakin mai gudanarwa. Ba na son ku yi nazarin HAZOP, ina so ku INGANTA ayyukan ƙungiyar mu. Don kumburi mai zuwa (layi daga ciyarwar famfo P-101 zuwa reactor R-101) yana haifar da siga xguideword SKELETON kawai da tambayar da za'a tattauna ga kowane tantanin halitta.Sakamako/tasiri/auna CIKA. A ɗauka akwai abubuwan da ba ku san takamaiman aikin ba. Wannan kwarangwal shine shigarwar zuwa taron ƙungiyar ƙwararru, ba fitarwa ba.
Bambanci a bayyane yake: tambayar AI kawai don kwarangwal, ba binciken kansa ba; ya bar alhakin tsaro ga tawagar.
Rarraba rawa a cikin HAZOP/ binciken haɗari
Mataki
Matsayin AI
Alhakin tawagar kwararru
Node/kwakwalwar siga
matrix shaci
Amincewar iyaka
samar da sabani
Tambayoyin kalmomin jagora
Hukuncin haɗari na gaske
sanadi/tasiri
Misalin tunatarwa na yau da kullun
Tsari takamaiman ƙima
Kariyar yanzu
-
dacewa hukunci
Rarraba haɗari
Shawarwari na farko
Ƙarshen yuwuwar/ tsananin hankali
mintuna
tsari, tsari
Abun ciki da yarda
Kuskuren gama gari
- Kuskure daftarin AI don HAZOP da aka gama. kwarangwal shine shigar da taron masana; Ba fitarwa ba ne.
- Tsammanin cewa AI zai san ƙayyadaddun haɗari na tsari. Halayen mallakar mallaka da ƙayyadaddun ƙayyadaddun wuraren ba su wanzu a cikin AI.
- Barin rarrabuwar haɗari zuwa AI. Ƙimar yiwuwar/nauyin tsanani da shawarar yin taka tsantsan na ƙungiyar da aka ba da izini.
- Bada AI don ƙara tsokaci a cikin mintuna. AI kawai tana gyara abin da ƙungiyar ta ce, ba ta samar da abun ciki.
- Dogaro da AI don rage taron. Haɗari mafi mahimmanci sau da yawa suna tasowa ne kawai yayin tattaunawar ƙungiya.
A takaice
AI a cikin HAZOP da ƙimar haɗari; Taimakon takaddun shaida ne wanda ke kafa tsarin matrix mai aiki, yana tunawa da al'amura na yau da kullun, shirya rahoton kuma yana ba da shawarwarin farko don rarraba haɗari. Amma ma'anar nazarin haɗari - wanda ke haifar da haɗari na gaske, tsananin sakamakonsa, dacewar yin taka tsantsan, da kuma yanke hukunci - shine alhakin da ba za a iya sauke shi ba na ƙungiyar kwararru da yawa. AI yana ba da kwarangwal da sauri; Yana tabbatar da tsaro ta ƙungiya da izini mai izini.
Aikin aikace-aikace
Zaɓi kumburin tsari (misali, layi daga famfo zuwa jirgin ruwa). Sami matrix ɗin jagora × matrix kalmar jagora da tambayoyin tattaunawa daga AI tare da samfurin "HAZOP skeleton" - ba tare da cike sanadi/tasiri/aunawa ba. Sannan gwada kirga wannan kwarangwal tare da ƙwararrun ido (ko simintin ƙungiyar) kuma ƙara aƙalla takamaiman takamaiman tsari guda ɗaya wanda AI zai iya rasa.
jerin abubuwan dubawa
- [ ] Na tambayi AI kawai don kwarangwal / tunatarwa, ba binciken kansa ba.
- [] Na ƙara takamaiman hatsarori waɗanda AI ba za su sani ba.
- [ ] Na bar rabe-raben haɗari da shawarar yin taka tsantsan ga ƙungiyar da aka ba da izini.
- [ ] Na tabbatar da cewa AI ta gyara mintuna ne kawai kuma ba ta ƙara abun ciki ba.
- [ ] Na tabbatar da HAZOP na ƙarshe ta hanyar nazari da amincewar ƙungiyar kwararru.