978-1-4244-7467-7/10/$26.00 ©2010 IEEE 1 2010 IREP Symposium- Bulk Power System Dynamics and Control – VIII (IREP), August 1-6, 2010, Buzios, RJ, Brazil Coordination of Day-Ahead Scheduling with a Stochastic Weekly Unit Commitment for the efficient scheduling of slow-start thermal units Pandelis N. Biskas, Costas G. Baslis, Christos K. Simoglou, Anastasios G. Bakirtzis School of Electrical & Computer Engineering, Aristotle University of Thessaloniki, Greece E-mail: pbiskas@auth.gr, cbaslis@ee.auth.gr, chsimoglou@ee.auth.gr, bakiana@eng.auth.gr Abstract This paper addresses the problem of the coordination of the day-ahead scheduling with a stochastic weekly unit commitment for the efficient scheduling of slow-start thermal units. The solution of the 24-hour unit commitment may lead to cases, in which slow-start thermal units that are initially off- line cannot be scheduled efficiently, due to their long start-up and minimum-up times as well as their large start-up costs. Thus, a new method is proposed, in which the day-ahead scheduling is coordinated with the solution of a weekly unit commitment. The latter is formulated and solved as a two- stage stochastic mixed-integer linear program, under various system and unit operating constraints, according to the provisions of the Greek Grid and Exchange Code. The stochastic parameter of the weekly unit commitment is the unit availability; thus, possible unit outages during the optimization period are taken into account. Test results from the implementation of the proposed method on the medium- scale Greek electricity market are presented. Keywords Day-ahead scheduling, generation scheduling, mixed-integer programming, stochastic weekly unit commitment Nomenclature f ( i F ) index (set) of steps of the energy offer function of generating unit i i ( I ) index (set) of generating units (thermal, hydro) n A ( n L ) index (set) of unit start-up types during stage n { } { } , , hwc un 1 2 , L= L= , where h: hot, w: warm, c: cold start-up and un: unique type of start-up m ( M ) index (set) of reserve types { } 1 ,2 ,2 ,3 + + - M= , where m=1+: primary-up, m=2+: secondary-up, m=2-: secondary-down, m=3: tertiary (spinning - 3S and non-spinning - 3NS) ( ) n N index (set) of stages { } 1, 2 N= ( ) s S index (set) of scenarios 1 T set of hours of the weekly planning horizon referring to the first stage ( 1 ⊆ T T ) 2 T set of hours of the planning horizon referring to the second stage ( 2 ⊆ T T ) DAS T set of hours of the planning horizon for the day- ahead scheduling ( 1 DAS ⊆ T T ) t ( T ) index (set) of hours of the weekly planning horizon ( 1 2 ∪ T=T T ) st B unique scenario bundle for scenario s at hour t ift B size of step f of the i-th unit energy offer function in hour t, in MWh sift b portion of step f of the i-th unit energy offer function dispatched in hour t in scenario s, in MWh ift C marginal cost of step f of unit i energy offer function in hour t, in €/MWh ( ) sit sit c p total production cost of unit i in hour t at level sit p in scenario s, in € t D system load demand in hour t, in MW i DT minimum down time of unit i, in h i NLC no-load cost of unit i (for one hour operation), in € max si P maximum power output of unit i in scenario s, in MW max, AGC si P maximum power output of unit i in scenario s while operating under AGC, in MW min si P minimum power output of unit i in scenario s, in MW min, AGC si P minimum power output of unit i in scenario s while operating under AGC, in MW soak i P fixed power output of unit i while in soak phase, in MW fix it P non-priced component of the energy offer function of generating unit i during hour t, in MWh