Application Multi-Objective Robust Decision-Making to the Design of Run-ofRiver Hydropower Plants

Hydropower is a comparatively cheap, reliable, sustainable, and renewable
source of energy. Run of River (RoR) hydropower plants are characterised by a
negligible storage capacity and by generation almost completely dependent on the
timing and size of river flows. Their environmental footprint is minimal compared to that
of reservoir-powered plants, and they are much easier to deploy.
This work uses and extends HYPER, a state-of-the-art toolbox that finds the
design parameters that maximise either the RoR plant’s power production or its net
economic profit. Design parameters include turbine type (Kaplan, Francis, Pelton and
Crossflow), configuration (single or two in parallel), and design flow, along with
penstock diameter and thickness, admissible suction head, and specific and rotational
speed.
This work extends HYPER to realise hydropower system design that is robust
to climate variability and change and to changing economic conditions. It uses the many
objective robust decision making (MORDM) approach through the following steps: (1)
an explicit three objective formulation is introduced to explore how design parameter
choices balance investment cost, average annual revenue, and drought year (first
percentile) revenue, (2) coupling of a multi-objective evolutionary algorithm (here,
AMALGAM) with HYPER to solve the problem using 1,000 years of synthetic
streamflow data obtained with the Hirsch-Nowak streamflow generator, (3) sampling
of deeply uncertain factors to analyse robustness to climate change as well as financial
conditions (electricity prices and interest rates), (4) quantification of robustness across
these deeply uncertain states of the world. We also extend HYPER by adding the
possibility to consider three-turbine RoR plants.
The HYPER-MORDM approach is applied to a proposed RoR hydropower plant
to be built on Mukus River in Van province which is located in Eastern Anatolia region
of Turkey. Preliminary results suggest that applying MORDM approach to RoR
hydropower plants provides insights into the trade-offs between installation cost and
hydropower production, while supporting design with a range of viable alternatives to
help them determine which design and RoR plant operation is most robust and reliable
for given site conditions and river stream characteristics. Results confirm earlier
findings that installation of more than one turbine in a hydropower plant enhances
power production significantly by providing operational flexibility in the face of variable
streamflows. When contrasting robustness of a design with its benefit / cost ratio, a
classic measure of performance of hydropower system design which accounts only for
annual revenues and cost, designs with the highest benefit / cost ratios do not
necessarily perform well in terms of dry year revenue. They also show less robustness
to both climate change (and associated drying) and to evolving financial conditions
than the designs that do better balance average annual revenue with dry year revenue