Modelling Intern

Randstad (Schweiz) AG; Paris; Hybride; EUR 40 000 - 60 000; Plein temps

Résumé du poste

A leading energy solutions firm is looking for an engineering student for a 6-month internship in Paris. The intern will work on improving power generation efficiency through the application of advanced modeling techniques with Python. Responsibilities include conducting literature reviews, building models, and developing workflows for fault detection. The ideal candidate is dynamic, autonomous, and fluent in French or English. This role offers flexible work options and various perks, including transportation and meal support.

Prestations

75% coverage of Navigo Pass or mobility budget

Lunch card with 60% coverage

Healthcare coverage

Yoga and sports classes

Team buildings & offsites

Qualifications

• Candidate should be in an engineering program seeking a 6-month internship.
• Prior experience with modeling is preferred.
• Knowledge and experience in Python coding required.

Responsabilités

• Conduct literature reviews on failure modes of CCGT.
• Compare detection results from both modules.
• Build models for selected use cases.
• Combine inferential engine with the AD models.
• Develop workflow for promoting fault detections.

Connaissances

Modeling

Python

Data Science Basics

Statistics

Teamwork

Autonomy

Fluent in French or English

Formation

Current engineering school student

Outils

Modelica

Description du poste

Context

Metroscope software assists operators in maintaining the efficiency and availability of power generation assets, including combined cycle gas turbines (CCGT) and nuclear power plants (NPP).

Two technical solutions are developed by Metroscope today:

• A “Diagnostics” module [1], based on a physics‑based digital twin predicting the best performance, using plant sensor measurements to estimate deviation from the best performance and an algorithm inferring the most probable failure modes explaining the deviations.

• An “Anomaly Detection” (AD) module, based on Advanced Pattern Recognition (APR) models, using machine learning algorithms trained on healthy plant data, that detect deviations and generate alerts.

The two solutions cover different but complementary scopes: the Diagnostics module monitors the full thermodynamic cycle of the power plant, detecting the root cause of the main sources of efficiency loss. The Anomaly Detection module offers more flexibility: any kind of sensor can be used to monitor any kind of system. However, it is not able to identify the root cause of the deviations.

Decision tree‑based diagnosis is a long‑known and proven method to determine the root cause of deviations ([2], [4]) on rotating machines as well as on energy generation assets. The inferential engine proposed by the Diagnostic module is only an evolution of the decision tree methodology with a more quantitative approach, but that was deployed primarily on a physical model ([7]).

The deviations found by APR models could directly be used to identify specific failure modes, based on industrial knowledge and Metroscope’s diagnosis expertise.

The goal of the internship is to leverage pre‑defined APR models and Metroscope’s inferential engine to predict failure modes with a greater scope than the actual physical models. The systems monitored are the main systems of a CCGT power plant, as well as some critical auxiliaries. The domains covered go from thermodynamics to vibration analysis.

Missions

Once introduced to Metroscope solution and tools for modeling industrial installations, you will perform the following tasks:

• Conduct a literature review on the different failure modes of a CCGT and the Root Cause Analysis (RCA) methods
• Conduct a comparative study of the detections we have on both modules for similar systems
• Build models on both modules on selected use cases
• Combine the inferential engine with the AD models, using the deviations predicted to infer the root cause of the issue
• Develop a workflow to automatically promote faults detections in the physical Diagnosis module based on the alerts of the AD module

Preferred Profile

You are currently in an engineering school and are looking for a 6 month end‑of‑study internship. You have some prior experience with modeling, ideally with Modelica or Python, and at least knowledge & experience in Python coding. You are interested in the energy sector, and in studying in details industrial processes. You are dynamic, autonomous and want to work in a team. You have some basics in data science and statistics. You speak French or English fluently.

Workplace

📌 This position is based in our Paris office, boulevard Haussmann, 3 minutes from the Saint‑Lazare train station. 👩💻 You will be able to work from home up to 2 days per week (days determined at your convenience). Joining Metroscope, you’ll benefit from: 🌱 Making an Impact: Contribute to the energy transition by improving efficiency and reliability in power generation. 💡 Opportunity to innovate. 🤝 Collaborative & Supportive Culture: Be part of a dynamic and caring team that values innovation, teamwork, and professional growth.

Perks & Benefits

🚇 Pass Navigo 75% covered by Metroscope OR a soft mobility budget (up to 700€/year). 🍽️ Lunch card (10€ / day – 60% covered by Metroscope). 🐻❄️ Healthcare Alan (70% covered by Metroscope). 🌻 Yoga and sport classes. 🛶 Team buildings & offsites.