PhD Fellowship in Energy Systems: Developing AI Models for Real-Time Optimisation Using Shallow-Geothermal Data

PhD Fellowship in Energy Systems: Developing AI Models for Real-Time Optimisation Using Shallow-Geothermal Data

Quick Decision

How can AI models support reliable decisions in an energy system whose weather, demand and operating conditions change? This three-year UiS doctorate develops and validates models for shallow geothermal and integrated campus energy systems. Its central challenge is to connect prediction quality, computation and physical constraints to useful operational decisions, rather than report accuracy alone.

The living plant and Assadi's operational modelling work provide concrete research context. The scope depends on agreed data, technical support and operational permissions. The position offers NOK 580,000 gross/year, with salary progression and a 2% pension contribution; tax and personal living costs are separate. The deadline is 15 November 2026; the employer materials give the date while an earlier application notice lists 23:59 Europe/Oslo. Confirm the closing time in the application system. A preliminary proposal is required, and the start date remains unconfirmed.

Research Problem and Doctoral Work

The problem is fast, accurate and robust real-time optimisation/control of integrated energy systems, with shallow geothermal energy central. A good forecast alone is insufficient: models must remain useful when weather, load, prices and operating regimes change, and decisions must respect the actual thermal plant. The advertised objectives combine AI-driven modelling, system integration and advanced operational strategies, using weather, electricity prices, fuel costs and emission pricing as external inputs. Employer recruitment PDF.

The researcher would develop and validate digital twins and AI models using monitoring/measurement data from Energy Central, then study techno-economic operation and forecasting. The stated infrastructure includes 120 geothermal boreholes, 10 dedicated research wells with different heat-exchanger designs; two commercial heat pumps of 1.3 MW heating capacity each; a research heat pump using supercritical CO₂; three gas and two electric boilers; a hot-water accumulator; and dry coolers. It is a living energy system rather than an unrestricted student experiment rig. Data access, control permissions and which assets enter the thesis must be agreed.

A defensible possible research angle is comparing a physics-based baseline with data-driven or hybrid surrogates under seasonal and operating-regime shifts, then measuring decision quality as well as prediction error. Separate predictive accuracy, computational latency, constraint compliance and realised cost/emission effects. The advert does not specify an AI architecture, reinforcement learning, specific software or a right to deploy autonomous control. Starting with offline validation and an agreed operational fallback is an interpretation of feasibility, not an extra promised work package.

The IEA's 2022 heat-pump outlook explains the relevance of low-emissions electricity for clean heating and identifies capital-cost/policy barriers. It is an older deployment analysis, not evidence of guaranteed 2026 market expansion. This doctorate's strongest contribution would be reliable operation under realistic economics and physical limits, not another generic black-box accuracy comparison. Heating transition and deployment limits.

School, Programme and Field Context

University of Stavanger, Faculty of Science and Technology, Department of Energy and Petroleum Engineering, is both employer and degree institution, in Stavanger, Norway. The award is the PhD in Science and Technology, with three years of pure research training, rather than a four-year post containing 25% teaching. Doctoral education includes independent research and thesis, courses equivalent to six months, international/national research participation, dissemination, a trial lecture and public defence. Confirm the formal specialisation and individual study plan; do not import another university's programme code. UiS doctoral programme overview; Official institution location.

The required proposal template plans 30 credits of training; the precise energy/AI modules and allocation are agreed with the supervisory team. The exact job promises mentoring, skills development and career planning, but does not name a guaranteed secondment or industrial placement. EERA's account of the 13–15 November 2024 European Geothermal Workshop, hosted at UiS with Assadi's involvement and plant visits, shows an active field network rather than a contractual mobility entitlement. Proposal and training plan; Observed geothermal network activity.

The THE institutional profile lists 2025 World University Rankings 601–800 and 2025 Engineering 601–800. Its menu links 2027 rankings, but the displayed UiS figures remain 2025; do not relabel them. These broad bands are modest and cannot establish the quality of this energy lab. The applied living-laboratory and relevant publications are the stronger project-specific reasons to retain the opportunity. Ranking year and subject scope.

Supervisor, Recent Papers and Resources

The exact employer advert names Professor Mohsen Assadi as academic contact (mohsen.assadi@uis.no), with Rosa Andrade, HR advisor, rekruttering@uis.no, +47 51831191 for recruitment. His official departmental profile lists geothermal prediction, energy management and AI publications, and doctoral supervision records including Fadnes, Banihabib and Weerakoon. This demonstrates relevant research/training activity, not a current student count or guaranteed supervision capacity. Confirm the formal co-supervisor and day-to-day support arrangement. Official staff and publications.

Three recent primary papers involving Assadi inform different parts of the proposed research:

• Fadnes, Olsen and Assadi, Frontiers in Energy Research 11:1078603, 22 February 2023, studies a municipal thermal plant combining sewage heat pumps, solar heat, biogas and grey-water recycling. Comparing design targets with operating experience identifies practical integration problems and data-driven improvement opportunities. It establishes industry–municipality–university working experience, not proof that AI already delivered prospective closed-loop savings or that this is today's campus plant. Holistic plant-management primary study

• Fadnes, Banihabib and Assadi, Energies 16:3875, 3 May 2023, uses measured operating data for ANN models in an existing municipal building cluster. Reported outlet-temperature MAPE is 2–5%, energy prediction generally below 10%, and hourly aggregated energy 3% or less, with low-demand cooling limitations. It is a strong modelling precedent, but test accuracy at that system does not guarantee transfer to new regimes or beneficial autonomous control. The complete 34-page publisher PDF was read. Operational heat-pump ANN primary study

• Fadnes and Assadi, Energies 17:4832, 26 September 2024, combines domain knowledge and ANN monitoring for wastewater-tunnel heat exchange, reporting stable observed exchanger performance over 2020–2024 and anomaly-detection utility. This shows long-term operational interpretation; the wastewater reservoir differs from geothermal boreholes, and a case's low fouling is not a universal maintenance guarantee. The complete 35-page publisher PDF was read. Reservoir and anomaly-monitoring primary study

The Energy Central page updated 14 August 2026 documents operation since September 2024, full operation from January 2025, alternative collectors and fibre-optic temperature monitoring. It also lists Fredrik Skaug Fadnes as Associate Professor II alongside Assadi; that does not appoint him as this candidate's supervisor. Different dated facility descriptions give 119/117 boreholes and two 650 m research wells, whereas the vacancy description says 120 and 10 research wells. Confirm the actual inventory allocated to the project and keep the differing scopes visible; the counts alone do not establish the vacancy status. Energy Central page updated 14 August 2026; 2024 opening scope.

Practical feasibility depends on usable historical time series, missing-data/calibration history, seasonal coverage, operational constraints, operator cooperation and permission to publish data/code. Confirm usable data, technical staff time, software licences, publication permissions and the individual research budget with the team. Industrial collaboration is an opportunity, not a promised placement or unrestricted command of a working plant.

Employment, Funding and Living Cost

This is a three-year salaried position, State code 1017, NOK 580,000 gross/year, with progression during employment and 2% statutory public-service pension deduction. It is not a tax-free stipend. Confirm the applicable annual progression and payroll calculation with HR. Benefits include pension/insurance, relocation support, health/parental arrangements, Lifekeys, sports discounts and Hjem-jobb-hjem commuting support; cash limits and eligibility need HR confirmation. Free Norwegian courses are offered if Norwegian/Swedish/Danish is below A2 at employment. Exact salary and benefit terms.

The official national study portal says doctoral candidates are normally exempt from public-university tuition, including where non-EU/EEA degree students ordinarily pay fees. Confirm UiS enrolment and semester-fee treatment, and the research/travel/software budget. The general public-university tuition rule does not settle an individual admission or residence decision; security/export-control procedures and personal eligibility require separate confirmation. Doctoral fee category.

For comparison only, NOK 580,000/12=NOK 48,333 gross/month. The 2026 PAYE guidance lists 25% PAYE for eligible new foreign workers, with a NOK 725,050 income ceiling and eligibility/time limits; ordinary taxation may apply instead. Subtracting 25% tax and 2% of gross pension yields an illustrative annual-average NOK 35,283/month. This is not an individual payroll quote or a rate guaranteed for all three years; holiday pay, tax residence and future rules matter. 2026 PAYE guidance.

Statistics Norway's 2025 survey gives NOK 11,400/month for a two-room Stavanger dwelling and NOK 13,630 for three rooms; one-room figures are unpublished. These are weighted averages of surveyed tenancy contracts, not 2027 offers, student-room prices or a housing guarantee. Rooms include living rooms and exclude kitchens/bathrooms; the averages do not uniformly remove electricity/heating costs. An analytical one-person planning scenario is NOK 19,000–23,000/month total, or NOK 228,000–276,000 for twelve months, using the two-room benchmark and an assumed NOK 7,600–11,600 for food, transport, utilities actually excluded from the tenancy and personal reserves. Check whether electricity/heating is already included to avoid double counting. This non-rent range is not an official measured basket; relocation/deposit, dependants and individual extras are separate. The conditional 2026 tax illustration cannot replace individual payroll for a future start year: obtain actual housing and relevant-year salary calculations before projecting a surplus. Stavanger rental evidence; SSB table 09895.

The national 2026/27 student benchmark is NOK 15,488/month or NOK 170,368 for 11 months; 12 months at that monthly rate is NOK 185,856. It is not an employee budget for a private flat, and student-welfare housing eligibility must be checked rather than assumed. The post requires living within regular on-campus attendance range, and relocation support does not establish an apartment allocation. National benchmark and its scope.

Requirements and Application

The employer requires a relevant five-year master's or equivalent, preferably recently obtained, with both master's thesis and weighted master's average at B or better. Listed fields are mechanical, energy, thermal/process engineering and computer science particularly with energy applications, or closely related disciplines. Required capabilities are strong energy-system knowledge, solid AI/ML/data-driven understanding, and documented mathematical modelling, programming/software development, analysis and English communication. Generic ML experience without energy-system competence does not replace those requirements.

Preferred, rather than universally mandatory, experience includes AI energy optimisation/control, digital twins, forecasting, energy markets/decision support, techno-economic assessment and real industrial/demonstration projects. Integrity, quality, curiosity, independent work and constructive interdisciplinary teamwork enter selection. Provide official grade-system documentation; the linked conversion page explicitly describes guiding scales for local master's admission, with individual assessment for unlisted scales. It does not establish a universal Chinese percentage-to-PhD-B conversion. Grade-conversion limitations.

Unless exempt, English evidence is TOEFL iBT 90/4.5; IELTS Academic 6.5; Cambridge CAE/CPE; or PTE Academic 62. The required IELTS overall score is 6.5; confirm whether any component-score condition applies. Exemptions cover at least one year of university study in Australia, Canada, Ireland, New Zealand, UK or US; an English-taught master's in EU/EEA or India; a relevant HK-dir GSU exemption; or five years' Norwegian residence outside the other categories. An English-taught Chinese master's is not automatically a listed exemption. HK-dir's GSU framework is general entrance guidance, not itself this candidate's PhD admission decision. Referenced national qualification framework.

Apply through the employer's online route by 15 November 2026. The employer webpage and PDF list 15 November 2026, while an earlier application notice gives 23:59 Europe/Oslo; confirm which closing time applies. Submit motivation/research-interest letter, CV, master's diploma/transcript or current transcript if unfinished, Diploma Supplement/equivalent and grade-conversion evidence if required, English proof, relevant publications/research, references, and the preliminary project proposal. Materials must be English or Scandinavian; attachments over 100 MB must be compressed. Only portal submissions are considered, and public applicant-list disclosure may occur even if exemption is requested.

The linked proposal template caps the description at 3,000 words including footnotes, schedule and references, covering question, originality, prior work, theory, methods/material, ethics, semester milestones, 30-credit training, possible overseas study and outputs. Its generic examples do not require interviews or qualitative research for this engineering topic. If appointed, the proposal/timeline is developed with supervisors in the first three months, forming the final plan. Required template and word limit.

UiS's research-fellow recruitment route normally uses an internal two- or three-member committee, interview and references, and often takes about three months, not a promised appointment date. The immigrant-background interview provision requires actual residence in Norway; it is not an automatic interview entitlement for every overseas Chinese applicant. The exact advert requires regular on-campus participation, security-conscious conduct and applicable export-control compliance; it states English prevails if language versions differ. No automatic nationality prohibition is stated. Correct recruitment route and conditions.

Career Directions

Likely transferable areas are energy-system modelling, building/industrial thermal optimisation, digital-twin software and research engineering. IEA 2025 reports 76 million energy jobs in 2024 and 2.2% growth, but also slower expected growth and skills bottlenecks; much shortage evidence concerns applied technical roles, not specifically AI doctorates. A group placement rate, an employer commitment and a salary premium remain unconfirmed. Sector workforce evidence and its limits.

Next Actions and Questions

Before applying: contact Assadi about the initial energy-system scope, available data and the proposal's novelty relative to the 2023–2024 operational-model papers. Clarify Chinese degree/B-grade equivalence and English exemption/test evidence. Prepare the 3,000-word-limit proposal and all portal documents for 15 November, after confirming the supplied 23:59 Oslo clock and allowing time for submission; do not wait for a generic admission window.

At interview: ask which boreholes/heat pumps and data are actually available, how the differing facility counts relate, and whether the thesis is offline modelling, supervised operational trials or both. Establish operator/technical support, seasonal validation, uncertainty and safety constraints, publication/IP permissions, co-supervision, meeting frequency and a fallback if live access is delayed. No particular AI method, experimental deployment right or industrial placement is guaranteed.

Before accepting: obtain salary progression, payroll/tax/pension calculation, three-year start/end dates, final admission/specialisation, tuition/semester-fee treatment, research budget, relocation amount and affordable housing. HR must confirm the appropriate residence/work procedure and any individual security requirements. Retain this as an applied energy/AI opportunity with substantial physical infrastructure; the main uncertainty is access and scope within three years, rather than an evidenced lack of salary funding.

Evidence and Sources

The recruitment information carries a 21 October 2026 date; that date alone does not establish whether the fellowship remains available. The full 34-page publisher PDF was read for the 2023 ANN study, and the full 35-page PDF for the 2024 monitoring study. Their results remain specific to the reported systems and data and should not be assumed to transfer to Energy Central. Facility descriptions retain their stated dates and counting scopes. Payroll and individual admission arrangements remain to be confirmed.

Employer, school and application

• UiS exact employer recruitment 309678 PDF

• UiS doctoral education overview

• UiS required indicative proposal template

• UiS recruitment-position assessment and applicant rights

• UiS local grade-conversion guidance

• THE University of Stavanger ranking profile

• HK-dir GSU education and language guidance

• UiS official location and contact addresses

Supervisor, primary papers and facilities

• Mohsen Assadi official staff and publications

• Fadnes Olsen Assadi 2023 holistic thermal plant management

• Fadnes Banihabib Assadi 2023 neural-network heat-pump models

• Fadnes Assadi 2024 wastewater reservoirs and ANN monitoring

• UiS 2024 official EnergyCentral opening release

• UiS EnergyCentral facility page, updated 14 August 2026

• EERA European Geothermal Workshop 2024

Finance, field and career

• Norwegian Tax Administration PAYE guidance

• Statistics Norway 2025 rental market survey

• HK-dir Study in Norway costs and doctoral fee exemption

• IEA Future of Heat Pumps 2022

• IEA World Energy Employment 2025 executive summary