Introduction
We extend our best wishes for safety and a rapid recovery to all those affected by Hurricanes Harvey and Irma, but we have a responsibility as system engineers to do more. The impact of these catastrophic weather events can be reduced by intelligent preparation and Model-Based Systems Engineering (MBSE) offers a better way to anticipate the often-unexpected cascades of events that catch us by surprise.
Back in 2009, Intercax started investigating modeling for watershed management with a state agency responsible for a very flat geographic area. Without altitude differences to encourage natural drainage, such agencies often need to take an active role in moving water around the district using pumps, locks and other regulators, especially during emergency situations. Good simulations support these active interventions. New MBSE tools bridge the gap between architecture modeling and simulation.
In this blog post, we want to show how a SysML model, captured in IBM Rational Rhapsody or MagicDraw, can be transformed into a Simulink model that provides a foundation for modern simulation techniques in a very important domain. For those interested in looking more closely, the SysML models can be downloaded using the links at the end.
Architecture of a Watershed Management District
For this simple example, we will start with two simple physical elements
- A Reservoir is a body of water, which might represent a lake or pond, a river or canal, or water flow channel. It is characterized by volume, water level and flow characteristics. It has internal sources and sinks of water (e.g. springs, seepage, evaporation, etc.) and it may be connected to one or more other Reservoirs through Regulators.
- A Regulator has the ability to move water in either direction between the connected Reservoirs through pumps, locks, spillways and other components.
Both of these elements can be modeled in detail in SysML using value, part and constraint properties, but in this example, we are simply concerned with the interfaces between them, modeled with proxy ports.
Figure 1 Schematic diagram of Watershed_System_01 example
Figure 1 shows a network of nine Reservoirs (labeled A) connected in a specific fashion by eight Regulators (labeled QS). Unlike the arrows in the diagram, the Regulators can move water in either direction, uphill and downhill being relatively unimportant factors in flat geographies. The watershed district operators often need to shift water around to avoid local flooding during storms, depending on current capacity and water level, as well as ultimate drainage into rivers and oceans.
Figure 2 SysML Block definition diagram (IBM Rational Rhapsody) of Watershed_System_01 example
Our SysML model begins with a system composition captured in the block definition diagram in Figure 2.
Figure 3 Internal block diagram of Watershed_System_01 (MagicDraw), with color coding blue for Reservoir and tan for Regulator
We use an internal block diagram (from the MagicDraw version) in Figure 3 to show the connections between the parts of the Watershed_System_01 block. In this diagram, all ports are proxy ports typed by the interface block Pump, which has a single flow property of direction out and type Real. Port type and stereotypes have been hidden to reduce clutter.
Model Transform into Simulink
Figure 4 MATLAB Simulink model generated from SysML model
Using Syndeia, our MBSE platform, we can take the models shown above and transform them into a Simulink model using a simple drag and drop action. The result is shown in Figure 4. The model transformation process carries over blocks, ports and connectors, with the Simulink ports typed by a single real parameter. The MATLAB code within these blocks for the simulation will be provided by the Simulink expert. However, Syndeia manages the connections between the SysML and Simulink elements so that changes in the structure of one or the other model can be recognized and reconciled as parallel model development proceeds.
Next Steps
Ultimately, our goal as model-based systems engineers is to build a single model that extends over all the software tools and data repositories needed to adequately describe the combined natural/artificial system in operation. This includes requirements, electrical and mechanical design, control logic, physical and hydrological models and much more. Only such an approach can give us both timely and accurate guidance during periods of system stress when the preservation of people and property is at serious risk.
Download SysML Models here: