본문

서브메뉴

Data-Driven Approaches for Predicting Polymer Solution Phase Behavior: Integrating High-Throughput Experimentation and Machine Learning
Data-Driven Approaches for Predicting Polymer Solution Phase Behavior: Integrating High-Th...
Data-Driven Approaches for Predicting Polymer Solution Phase Behavior: Integrating High-Throughput Experimentation and Machine Learning

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260202105529
ISBN  
9798263340445
DDC  
547.14
저자명  
Amrihesari, Mona.
서명/저자  
Data-Driven Approaches for Predicting Polymer Solution Phase Behavior: Integrating High-Throughput Experimentation and Machine Learning
발행사항  
[Sl] : Georgia Institute of Technology, 2025
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2025
형태사항  
338 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
주기사항  
Advisor: Brettmann, Blair.
학위논문주기  
Thesis (Ph.D.)--Georgia Institute of Technology, 2025.
초록/해제  
요약As artificial intelligence (AI) and machine learning (ML) continue to advance, datadriven approaches are increasingly used to uncover complex phenomena across variousfields, including polymer science. However, the effectiveness of these methodsfundamentally depends on the availability, structure, and quality of data. In polymerscience, the lack of comprehensive datasets presents a significant challenge, as polymersare highly sensitive to chemical, physical, and measurement parameters. Among polymerproperties, solubility is of particular interest, as homogeneous polymer solutions serve asthe foundation for numerous industrial applications, including pharmaceuticals, plasticrecycling, and membrane science. However, polymer solubility is influenced by multiplefactors-such as molecular weight, concentration, morphology, and temperature-whichare not always reported in existing datasets. To fully harness the potential of AI/ML in thisdomain, this thesis develops experimental methods to systematically measure polymersolubility with high precision and detail. By generating a comprehensive dataset thatcaptures key influencing factors, this work bridges the existing data gap and advancessolubility prediction using fully experimental data and precipitation kinetics-an area thathas not previously been explored through data-driven approaches using machine learning.The integration of well-structured experimental data with predictive modeling advancespolymer informatics while providing a valuable framework for accelerating materialsdesign and optimizing industrial applications.This thesis first explores the design of experimental methods for measuringpolymer solubility using a parallel crystallizer, focusing on the effects of key experimental parameters-including temperature ramp rate, hold temperature, hold period, and mixingspeed-on turbidity measurements. Through the development of a standardized method,this work demonstrates the potential for high-throughput data collection. While thistechnique may not be universally optimal, it provides a balance between speed, structuralinsights, and precise temperature control, making it particularly useful for industrialapplications where rapid and reliable solubility measurements are essential.Using machine learning (ML) techniques, this thesis demonstrates the impact ofwell-structured experimental data on predicting polymer solubility as a function ofconcentration and temperature, leveraging state-of-the-art ML models. This workhighlights how detailed data enhances solubility predictions beyond a simple binaryclassification of "soluble" or "insoluble," enabling more nuanced, quantitative insights.Furthermore, integrating kinetic information provides a valuable framework for capturingsolubility behavior over time, facilitating data collection under dynamic conditions.Additionally, targeted data collection across a broader range of polymers and solventssignificantly improves the generalizability of ML models, ensuring more robust andreliable predictions across diverse polymer-solvent systems.As research moves toward fully automated laboratories for materials design, thisthesis also presents the development of an extraction model utilizing two experimentallydefined parameters to study precipitation kinetics in polymer solutions. The effectivenessof such models relies not only on their structural design-requiring interdisciplinaryexpertise-but also on an iterative validation process to continuously refine the quality ofextracted data.The findings of this thesis demonstrate that standardizing experimental methods forpolymer solubility measurements enables the generation of high-quality datasets forpredicting both solubility and precipitation in polymer-solvent systems. This work alsocontributes a comprehensive turbidity dataset comprising 1,000 unique polymer-solventcombinations, along with an extraction model for precipitation kinetics, providing avaluable foundation for future studies in polymer informatics.
일반주제명  
Polymer solubility
일반주제명  
User interface
일반주제명  
Cooling
일반주제명  
Temperature effects
일반주제명  
Solvents
일반주제명  
Kinetics
일반주제명  
Molecular weight
일반주제명  
Structured Query Language-SQL
일반주제명  
Polyethylene glycol
일반주제명  
Reproducibility
일반주제명  
Polymer chemistry
기타저자  
Georgia Institute of Technology.
기본자료저록  
Dissertations Abstracts International. 87-05B.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260126s2025        us                              c    eng  d
■001000017360454
■00520260202105529
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798263340445
■035    ▼a(MiAaPQ)AAI32309803
■035    ▼a(MiAaPQ)GeorgiaTech77823
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a547.14
■1001  ▼aAmrihesari,  Mona.
■24510▼aData-Driven  Approaches  for  Predicting  Polymer  Solution  Phase  Behavior:  Integrating  High-Throughput  Experimentation  and  Machine  Learning
■260    ▼a[Sl]▼bGeorgia  Institute  of  Technology▼c2025
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2025
■300    ▼a338  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-05,  Section:  B.
■500    ▼aAdvisor:  Brettmann,  Blair.
■5021  ▼aThesis  (Ph.D.)--Georgia  Institute  of  Technology,  2025.
■520    ▼aAs  artificial  intelligence  (AI)  and  machine  learning  (ML)  continue  to  advance,  datadriven  approaches  are  increasingly  used  to  uncover  complex  phenomena  across  variousfields,  including  polymer  science.  However,  the  effectiveness  of  these  methodsfundamentally  depends  on  the  availability,  structure,  and  quality  of  data.  In  polymerscience,  the  lack  of  comprehensive  datasets  presents  a  significant  challenge,  as  polymersare  highly  sensitive  to  chemical,  physical,  and  measurement  parameters.  Among  polymerproperties,  solubility  is  of  particular  interest,  as  homogeneous  polymer  solutions  serve  asthe  foundation  for  numerous  industrial  applications,  including  pharmaceuticals,  plasticrecycling,  and  membrane  science.  However,  polymer  solubility  is  influenced  by  multiplefactors-such  as  molecular  weight,  concentration,  morphology,  and  temperature-whichare  not  always  reported  in  existing  datasets.  To  fully  harness  the  potential  of  AI/ML  in  thisdomain,  this  thesis  develops  experimental  methods  to  systematically  measure  polymersolubility  with  high  precision  and  detail.  By  generating  a  comprehensive  dataset  thatcaptures  key  influencing  factors,  this  work  bridges  the  existing  data  gap  and  advancessolubility  prediction  using  fully  experimental  data  and  precipitation  kinetics-an  area  thathas  not  previously  been  explored  through  data-driven  approaches  using  machine  learning.The  integration  of  well-structured  experimental  data  with  predictive  modeling  advancespolymer  informatics  while  providing  a  valuable  framework  for  accelerating  materialsdesign  and  optimizing  industrial  applications.This  thesis  first  explores  the  design  of  experimental  methods  for  measuringpolymer  solubility  using  a  parallel  crystallizer,  focusing  on  the  effects  of  key  experimental  parameters-including  temperature  ramp  rate,  hold  temperature,  hold  period,  and  mixingspeed-on  turbidity  measurements.  Through  the  development  of  a  standardized  method,this  work  demonstrates  the  potential  for  high-throughput  data  collection.  While  thistechnique  may  not  be  universally  optimal,  it  provides  a  balance  between  speed,  structuralinsights,  and  precise  temperature  control,  making  it  particularly  useful  for  industrialapplications  where  rapid  and  reliable  solubility  measurements  are  essential.Using  machine  learning  (ML)  techniques,  this  thesis  demonstrates  the  impact  ofwell-structured  experimental  data  on  predicting  polymer  solubility  as  a  function  ofconcentration  and  temperature,  leveraging  state-of-the-art  ML  models.  This  workhighlights  how  detailed  data  enhances  solubility  predictions  beyond  a  simple  binaryclassification  of  "soluble"  or  "insoluble,"  enabling  more  nuanced,  quantitative  insights.Furthermore,  integrating  kinetic  information  provides  a  valuable  framework  for  capturingsolubility  behavior  over  time,  facilitating  data  collection  under  dynamic  conditions.Additionally,  targeted  data  collection  across  a  broader  range  of  polymers  and  solventssignificantly  improves  the  generalizability  of  ML  models,  ensuring  more  robust  andreliable  predictions  across  diverse  polymer-solvent  systems.As  research  moves  toward  fully  automated  laboratories  for  materials  design,  thisthesis  also  presents  the  development  of  an  extraction  model  utilizing  two  experimentallydefined  parameters  to  study  precipitation  kinetics  in  polymer  solutions.  The  effectivenessof  such  models  relies  not  only  on  their  structural  design-requiring  interdisciplinaryexpertise-but  also  on  an  iterative  validation  process  to  continuously  refine  the  quality  ofextracted  data.The  findings  of  this  thesis  demonstrate  that  standardizing  experimental  methods  forpolymer  solubility  measurements  enables  the  generation  of  high-quality  datasets  forpredicting  both  solubility  and  precipitation  in  polymer-solvent  systems.  This  work  alsocontributes  a  comprehensive  turbidity  dataset  comprising  1,000  unique  polymer-solventcombinations,  along  with  an  extraction  model  for  precipitation  kinetics,  providing  avaluable  foundation  for  future  studies  in  polymer  informatics.
■590    ▼aSchool  code:  0078.
■650  4▼aPolymer  solubility
■650  4▼aUser  interface
■650  4▼aCooling
■650  4▼aTemperature  effects
■650  4▼aSolvents
■650  4▼aKinetics
■650  4▼aMolecular  weight
■650  4▼aStructured  Query  Language-SQL
■650  4▼aPolyethylene  glycol
■650  4▼aReproducibility
■650  4▼aPolymer  chemistry
■690    ▼a0800
■690    ▼a0495
■71020▼aGeorgia  Institute  of  Technology.
■7730  ▼tDissertations  Abstracts  International▼g87-05B.
■790    ▼a0078
■791    ▼aPh.D.
■792    ▼a2025
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360454▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF17294 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.