Development of a new simplified algorithm facilitating GIS-based preliminary planning of small hydropower plants

With the increasing promotion of renewable energy sources, hydropower is expected to play a crucial role in energy storage and grid balancing, supplementing intermittent solar and wind power. However, the complex topography and variability of design parameters often lead to underutilization of small hydropower potential or sub-optimal designs. This study presents a new simplified algorithm for GIS-based positioning optimization and preliminary planning of small hydropower plants.

The practical problem addressed with this algorithm is that during the preliminary design phase of a small run-of-river hydropower plant, the designers are required to preselect, along a given river, the water intake point, the location of the power generation station, and the tailrace discharge point back to the river, as well as the conduit route between the intake and the station. The latter is quite complex, as it may consist of a section of open channel and the remaining section of closed conduit (penstock). The open channel is significantly cheaper than the penstock, but it needs to practically follow the contour line of the intake on suitable ground. The penstock does not have problems with steep slopes, but it is generally much more expensive per unit length than the open channel, especially if it is made of steel. At the same time, during the routing process, areas where pipelines are not allowed to pass must be excluded, e.g., natural reserve areas, or residential areas, or areas with intense geological phenomena. Simultaneously, the expected electricity production and the cost for each candidate design should be considered, in order to examine the technical and economic viability of the project.

Developed in Python within the QGIS environment, which is an open but well-established geographical information system software package, the algorithm operates in raster format and uses as input the digital terrain model, the flow direction and flow accumulation grids, the examined river reaches in raster format, characteristic discharge rate for each cell of the examined river, the open channel and the penstock cost per unit length in raster format. Areas where pipelines are not allowed to pass are designated with a very high unit length cost.

After reading the input data and initializing the required output raster data objects, the algorithm creates the lefthand and righthand contour lines for each upstream cell. Then it iterates between all the possible downstream cells for this upstream cell. For each upstream-downstream positions couple it iterates through all the possible combinations of open channel and penstock and creates a list with the combinations having the lower cost. For each optimal solution it stores all the characteristics in a list. After finalizing all the searches the algorithm sorts the list with the optimal solutions considering the produced energy/cost ratio.

While such algorithms typically exhibit O(n³) complexity, meaning that as the size of the area increases the computation time will become prohibiting, a key characteristic of the proposed algorithm is that includes novel search functions minimising the searched cells and required repetitions making the execution time reasonable in larger areas.