Grid Generation Process

The functionalities of the grid generation process are divided into three classes:

The higher level functions of the GridGenerator are explained in more detail since they contain the assumptions and logic of grid generation. For a visual representation refer to Short Summary of the Grid Generation Steps.

Step 1

The selected zip code (PLZ) is searched in the table postcode and stored in the postcode_result table. The zip code defines the geographical area for which the network generation takes place. The buildings, which are located in the area of the zip code, are selected from res and oth and stored on buildings_tem. The tables ending with tem are tables that temporarily store the data for grid generation. For the buildings of the zip code the house distance is calculated and the settlement type is is derived from it. The settlement type decides which transformer types are installed in the zip code. Each building is assigned a maximum load. This depends on the building type. For residential buildings, the load is scaled to households, for other buildings (commercial, public, industrial buildings), the building floor area is taken into account in the calculation of the power. of the power. Buildings without load or with load over 100kW are not part of the low voltage network and are therefore deleted. Finally, the transformers from transformers are also transferred to buildings_tem are transferred.

Step 2

The ways from ways, which are located in the zip code area, will be stored on ways_tem. In way sections that overlap, connection nodes are created. Then the buildings are connected to the ways. For this purpose, a path section is created that leads perpendicularly from the existing ways to the center of the building. Finally, the buildings from buildings_tem are assigned a node from ways_tem.

Step 3

Since the number of buildings in a postal code is too large for a coherent network, the buildings are divided into subgroups. The kmeans cluster algorithm divides the buildings into subgroups based on the geographic distance. The number of kmeans clusters for a postal code is usually single-digit. Each kmeans cluster is assigned an ID (kcid, kmeans cluster ID).

Step 4

For the positioning of the transformers, existing transformers from OSM are considered first. Buildings at a certain distance from the transformer are connected to the transformer. The shape of the supply area transforms from a circle (allowed linear distance), to a polygon, because the distance to the consumers is measured along the streets.

Step 5

The kmeans clusters are further divided into so-called building clusters. A buildings cluster becomes a network in which all buildings are connected to a low-voltage transformer. The buildings are grouped by means of an agglomerative hierarchical average linkage clustering. The results of the hierarchical clustering can be displayed as a Dendogram. The distance between two clusters is calculated as average of all distances of buildings from cluster A to cluster B. For this purpose, a distance matrix of the buildings is set up. After each clustering step, there is a loop which verifies that the transformer power suffices to supply the consumers taking into consideration coincidence factor. The multilevel coincidence factor for each cable section is evaluated by summing the classified consumers (residential, public, commercial) of that cable section

Step 6

Branches are created to connect the consumers to the transformer, resulting in a radial network. First, the consumers are connected to the road by cables. A consumer is created as a consumer node bus and every point where several cables intersect as a connection node bus. Next, the connections between the connection nodes are drawn. Finally, the connection nodes are connected to the transformers LVbus. The cables are run along the streets from ways_tem. When the branches are created, the Minimal Spanning Tree Algorithm determines a configuration of the network, whose edge lengths are as short as possible and thus inexpensive. From a repertoire of cable types, a suitable cable is selected. The process of cable installation is based on the realistic decision making of a technician and avoids the use of additional costly network components.

Step 7

The data from the tables with the extension tem are deleted and transferred to the result tables bulidings_result, ways_result.