By Francois Gilbert
Managing hydropower reservoir operations is a tricky science. It involves the ability to accurately predict how much water should be saved in reservoirs so that the utility can continuously produce energy, especially when electricity prices are high. Many utilities struggle with this challenge because it is difficult to produce accurate inflow forecasts.
To make this decision, utility Hydro-Quebec analyzes daily rainfall or snow melt forecasts to determine the amount of potential inflow to its reservoirs. The utility – which generates, transmits and distributes electricity for the entire province – faces the challenge head on with the help of geographic information system (GIS) technology.
 |
| Hydro-Quebec’s previous system could only display one layer of data at a time. In this image, a 24-hour rain forecast is shown using isolines. |
With 62 hydroelectric generating plants and 27 large reservoirs, Hydro-Quebec is the largest electric utility in Canada and one of the world’s largest hydroelectric producers. The government-owned utility has been successful in generating large quantities of renewable electricity at competitive prices due in part to its technological innovation. This innovation is evident in how the utility manages its reservoir operations.
It takes a 10-person team to validate the data that will be entered into mathematical models to support decision-making. Every day, Hydro-Quebec’s in-house meteorologist compares weather forecasts from Environment Canada, as well as forecasts from the U.S. and European models, to determine differences and modify the Environment Canada forecast model to be used for analysis. At the same time, other staff members validate other data, including reservoir levels, water flows and power flows.
After the forecast of how much water will fall into the reservoirs is made, operators then schedule how much energy will be produced for the next day.
Recognizing the value of a geographic approach
In 2006, Hydro-Quebec developed homegrown GIS software to assist operators with their decision-making. However, the system lacked a real geographic component and had limited visualization and analysis capabilities. It integrated bitmap images of rain coming into the reservoirs, but the images were not associated with their location. As well, it was difficult to understand the representation because multiple layers of lines were overlaid on top of one another.
To help improve the accuracy of its forecasts, Hydro-Quebec leveraged ArcGIS technology from Esri, which provides a comprehensive GIS platform for the energy and water resources industry. The utility has used the technology for many years and has found ArcGIS to be the most cost-effective solution for advanced desktop and custom application development.
 |
| The new GIS-based application integrates data from weather forecasts to display both rain (green) and snow (blue), as well as other valuable information, on a digital map for effective planning. |
Hydro-Quebec used Esri’s ArcPy tools to script GIS functionality in the Python programming language to build a program that retrieves weather forecasts by Environment Canada from the utility’s file transfer protocol (FTP) site. It then processes the data in raster catalog format and stores the raster layers in a file geodatabase. Once this is done, the GIS layers are then published and made available in an offline format as map packages, which staff can use when working in remote areas without internet connectivity.
The utility used Esri’s ArcGIS Runtime Software Development Kit (SDK) to quickly build and deploy a new desktop application that includes capabilities for more powerful location services, data management, analysis and mapping. The GIS application georeferences and animates the time-enabled raster images and allows users to zoom in on a digital map to view locations with intense rain, as well as see more clearly if it is rain or snow falling into the reservoirs. More importantly, the new application significantly improves two key processes: weather forecasting and data validation.
Increasing the accuracy of forecasts
Forecasts using the previous system (see Figure 1 on page 22) factored in a lot of human judgment because staff could not clearly see the movement of precipitation and temperature. They could also see only one view layer – rain or snowfall data – at a time. This made it difficult to produce accurate forecasts.
With the new GIS-based application (see Figure 2), forecasting has significantly improved. Staff can distinguish whether rain or snow is falling because the two raster layers are displayed in different colors, one over the other, on the same map. Intense color means large values. Also, the new application displays both temporal and non-temporal data layers, making it easy to incorporate all relevant information for better analysis.
When a temporal layer is selected in the table of contents, a time slider automatically appears to animate the time series. In addition, the application delivers the following map functionalities:
- Spatial selection applied on the current selected layer;
- Zoom and center after an item is selected on the map;
- Zoom back to initial settings;
- Search text on all layer components;
- Dram line with pencil;
- Measure distance and area;
- Display longitude and latitude;
- Display scale;
- Layer visualization scale range; and
- Interactive attribute tables to show selected features.
Forecasts are updated automatically twice a day, with each new forecast overwriting the previous one. After the weather forecast is validated and published by Hydro-Quebec’s meteorologist, another program automatically calculates the equivalent watershed precipitation and temperature using a spatial analyst function. This data is then sent as an input to the hydrological model that will calculate runoff – defined as the water to be stored in the reservoirs or used for power generation – for the next seven days.
Improving the efficiency of data validation
The GIS-based application has provided staff with an efficient way to validate data. Hydro-Quebec has numerous data logger devices installed at its plants that transmit information to a central database. This data is then processed by the utility’s validation software that runs every day and sends a report to the GIS application so the data can be mapped and analyzed.
Sometimes, data errors occur because the data logger device has broken down or the site has been vandalized or blocked by ice. Once the errors have been determined, they need to be corrected so as not to impede efficient reservoir operations.
With the new application, the GIS automatically reviews all operational data and flags suspicious data with a red circle on the map (see Figure 3 on page 26). Staff can quickly navigate to the error and click on the power plant with suspicious data. The application then calls up the time-series software to display the related time-series graph. Users can immediately correct the data because the application integrates the GIS with the utility’s time-series software, enabling a seamless workflow. The time-series software and the GIS are separate components, but users see both as one application.
 |
| Data is inspected, validated and modified with time-series software, which is now integrated with the GIS application, allowing users to correct erroneous data instantly. |
Location is important in resolving data errors. Knowing where the wrong values occur can help determine the type of error or source of the problem. For example, if all data associated with a specific plant is showing in the system as incorrect, it is possible that the data transmission has failed. The application provides valuable information that allows staff to prompt field technicians to check the condition of the data logger devices.When the application flags a series of power plants with incorrect values, as seen in Image Figure 4, this means the plants need to be inspected one-by-one.
However, with the new application, staff can display a data layer superimposed on the power plant map to find a correlation between the plant’s data problem and its location. After staff find the problem and correct the data, they can re-run the validation software to ensure all errors have been corrected. This has significantly reduced the number of plant site inspections by Hydro-Quebec staff.
GIS: A platform for intelligent operations
Empowering Hydro-Quebec’s staff with a GIS-based system for reservoir operations planning has resulted in more accurate forecasts and efficient workflows. The technology has allowed the utility to integrate several systems and software, automate various tasks and leverage a single application to support both forecasting and data validation workflows. About 75 staff in the power scheduling business unit now have access to the application through desktop visualizations. With a common view of data and powerful analysis tools, they have gained the ability to make timely, intelligent decisions.
In addition, the GIS functionality programmed into the new application has been packaged as a specialized user control, which can be easily plugged into the utility’s other Microsoft .NET software to extend the same powerful mapping, data visualization and spatial analysis capabilities to applications supporting other operations such as transmission lines congestion studies.
 |
| Incorrect values flagged by the application can be reviewed by Hydro-Quebec staff, significantly reducing the number of plant inspections required. |
A developer simply needs to add the user control with the accompanying configuration file to their new application to make the GIS functionality available. This introduces application development standards across Hydro-Quebec and reduces software development and maintenance costs. It also provides a standardized user interface for presenting maps in all applications.
The new application also allows for efficient data exchange between the utility’s GIS and time-series software. The GIS application can display graphs from the time-series software and the time-series software can now also display GIS views. This maximizes the return on their technology investments.
With their new GIS-based reservoir operations planning application, Hydro-Quebec saves about seven hours of staff time per day in operational data validation. They have also significantly improved their ability to present and interpret weather forecasts, leading to more precise inflow predictions in support of intelligent operations.
Francois Gilbert is a senior water resources engineer for Hydro-Quebec.
