* Add the AHU tutorial, #67

Tutorial/AHU.md

@@ -0,0 +1,38 @@
+# AHU use case
+
+## Overview
+This is a demonstration, in which the Scikit models are used to control an
+air handling unit (AHU).
+
+The task is to control the supply air temperature, while the outdoor
+temperature fluctuates.
+
+### Provide the Scikit models
+The models are Multi-Layer Perceptron (MLP) models created with Scikit-learn.
+You can use the frozen models stored in
+
+pyDMPC\pyDMPC\SpecialStudies\FrozenModels\AHU
+
+and copy them and the corresponding scalers (required to scale the inputs for
+the MLP) to the working directory. If you would like to create the models from Modelica models, you can use this [script](../pyDMPC/Utilities/mlp_train_modelica.py). This script requires an FMU (see settings below) of the
+heater and cooler models in
+
+ModelicaModels.SubsystemModels.DetailedModels.HeaterML and
+ModelicaModels.SubsystemModels.DetailedModels.CoolerML
+
+### Controlled system (AHU)
+The central AHU model is stored in this path in the ModelicaModels package of
+pyDMPC
+
+ModelicaModels.ControlledSystems.ControlledSystemBoundaries
+
+Use Dymola to export the model to a functional mockup unit (FMU) according to
+the following settings:
+
+![FMU](../pyDMPC/Resources/Images/FMU.png)
+
+The FMU filename and the file's location has to be included into the Init.py.
+
+### Settings
+All settings are given in the Init.py according as described in the
+[Settings](Settings.md).