4 présentations

• 
  10h30 - 10h55

  Cold load pick-up management based on advanced service restoration in electricity distribution networks

  • Sanja Dzeletovic, prés., McGill University
  • François Bouffard, McGill University

  A methodology for service restoration with active management of cold load pick-up is developed. Re-connection schedules for feeder customers are calculated through the solution of a mixed-integer linear programming problem. Results of the proposed approach are compared against full feeder restoration. Machine learning approaches applied to historical post-outage data are being explored for cold load pick-up modelling with an aim to improve the accuracy of service restoration decisions.

  Keywords: mixed integer linear programming, service restoration, cold load pick-up
• 
  10h55 - 11h20

  Flexibility management considering electricity transmission network limits and correlation of wind power

  • Yuchong Huo, prés., McGill University
  • François Bouffard, McGill University

  Uncertainty associated with wind power generation in bulk power system operations requires increasingly systematic attention. As practical levels of wind power generation penetration keep on deepening in many power systems, it is now widely accepted that adequate supply of flexibility resources is essential to managing this uncertainty. This paper extends the notion of flexibility envelopes to network-constrained power systems and to systems with geographic wind power output correlation. Preliminary results indicate that flexibility envelopes are effective means of achieving flexibility adequacy.
• 
  11h20 - 11h45

  Incorporating a PARMA statistical model in SDDP for hydroelectric system management.

  • Yves Mbeutcha, prés., Polytechnique Montréal
  • Michel Gendreau, Polytechnique Montréal
  • Grégory Émiel, Hydro-Québec
  • jasson Pina, PhD candidate

  Résumé:

  Stochastic Dual Dynamic Programming (SDDP) is a widely used method to optimize the operations of a large hydrothermal system. The uncertainty on the future inflows in SDDP is usually represented by a linear time series like the family of Periodic Autoregressive (PAR) models. A model with a moving average component like the class of Periodic Autoregressive and Moving Average (PARMA) models can be more advantageous than pure autoregressive higher order models. We will compare the performance of the two-time series models on a hydroelectric system of Hydro-Quebec using the SDDP approach.

  Mots-clés: hydroelectric system management, Stochastic Dual Dynamic Programming, PARMA models.
• 
  11h45 - 12h10

  Using stochastic optimization for policy analysis in the electricity sector

  • Jesus Rodriguez, prés., HEC Montreal
  • Pierre-Olivier Pineau, HEC Montréal
  • Michel Denault, GERAD - HEC Montréal

  In several regions the electricity sector will require major investments to achieve decarbonization targets and to upgrade and replace ageing power infrastructure. As transmission and generation expansion projects typically require several years or even decades to be completed, capacity investment decisions should be made well in advance, in the face of significant uncertainties related to electricity demand, environmental policies and investment and operation costs. We present a multi-stage stochastic program for capacity expansion in the electricity sector, considering the short-term variability of intermittent renewable energies and the long-term uncertainty of decarbonization policies, load profiles and fossil fuel costs. With reference to the North American Northeast Electricity sector, we show how this model can support the analysis of energy policy decisions.
  Keywords: Capacity expansion planning, stochastic programming, energy policy, power systems