ModiaResult is part of ModiaSim. See also the ModiaResult documentation.
ModiaResult defines an abstract interface for simulation results with a potentially segmented time axis (on different segments of the time axis, different variables might be defined).
A simulation result consists of a set of result signals. A result signal is identified by its name::String
(e.g. "robot.joint1.angle"). It provides an approximation of a piecewise continuous variable v = v(t)
which is a (partial) function of the independent, monotonically increasing variable t. Typically, the
independent variable t is called "time". The approximation consists of the values of variable v
at particular time instants, v_i = v(t_i) together with the information how to interpolate between these
time instants. If a variable is not defined in some phase, it has a value of missing at the corresponding time instants.
A value v_{ji}(t_i) of a variable v_j at time instant t_i is represented as vj[i] and is
typically a sub-type of Real or of AbstractArray with an element type of Real
(e.g. a (2,3) array variable v_a at time instant t_i is represented as va[i,1:2,1:3]).
A simulation result can also hold constants (parameters), that have the same value at all time instants.
Constants are compactly stored as OneValueSignal and can be of any Julia type (e.g. v = "data.txt").
If a variable is not defined in some phase, it has a value of missing at the corresponding time instants.
Optionally, a unit (via Unitful.jl) can be associated with a
variable v (so the same unit for all elements, if the variable is an array).
The ModiaResult package provides an abstract interface to operate on such simulation results, for example,
- to provide the simulation result in a form to allow signal calculations (e.g.
v_{diff} = v_2 - v_1), - to provide a table view of the signals via DataFrames.jl, or
- to produce line plots in multiple diagrams within multiple windows/figures in a convenient way (see example below).
Concrete implementations of the ModiaResult Abstract Result Interface are provided for:
- Modia.jl (a modeling and simulation environment)
- DataFrames.jl (tabular data; first column is independent variable)
- Tables.jl (abstract interface for tabular data, e.g. CSV tables; first column is independent variable),
- Dictionaries with String keys (if OrderedDict, independent variable is first variable, otherwise independent variable is "time").
Concrete implementations of the ModiaResult Abstract Line Plot Interface are provided for:
- PyPlot (plots with Matplotlib from Python),
- GLMakie (interactive plots in an OpenGL window),
- WGLMakie (interactive plots in a browser window),
- CairoMakie (static plots on file with publication quality).
Furthermore, there are two dummy modules included in ModiaResult, that are useful when performing tests with runtests.jl, in order that no plot package needs to be loaded during the tests:
- NoPlot (= all plot calls are ignored and info messages are instead printed), or
- SilentNoPlot (= NoPlot without messages).
More details:
All packages are registered and are installed with:
julia> ]add ModiaResult
add ModiaPlot_PyPlot # if plotting with PyPlot desired
add ModiaPlot_GLMakie # if plotting with GLMakie desired
add ModiaPlot_WGLMakie # if plotting with WGLMakie desired
add ModiaPlot_CairoMakie # if plotting with CairoMakie desiredIf you have trouble installing ModiaPlot_PyPlot, see Installation of PyPlot.jl
Assume that the result data structure is available, then the following commands
using ModiaResult
# Define plotting software globally
usePlotPackage("PyPlot") # or ENV["MODIA_PLOT"] = "PyPlot"
# Execute "using ModiaPlot_<globally defined plot package>"
@usingModiaPlot # = "using ModiaPlot_PyPlot"
# Generate line plots
plot(result, [("sigA", "sigB", "sigC"), "r[2:3]"])generate the following plot:
Martin Otter, DLR - Institute of System Dynamics and Control