Free Courses Fluids Structures Electronics 3D Design Optics Photonics Materials Installation Python Optimization Safety & Cybersecurity Embedded Software MBSE Digital Mission Engineering STEM NVH Student Teams Semiconductors Indo Kursus Partner Courses My Courses

[FAQ]

Unlike some other models where clustering actually implies coarsening, the meaning of the clustering in the MSMD approach is the opposite. In the battery module, the default option offers an enabled cell-clustering with N_x = N_y = N_z =1 i.e. one big cluster for the entire battery cell. The submodels are solved for the entire battery cell. However, the “Cluster Cells” option offers a user an option to divide the battery into a number of segments in each of the directions (N_x, N_y, N_z) and the submodels are solved for each of these segments (clusters) separately, thus solving more equations and providing a more detailed information. This explains why clustering option is more computationally expensive. ===================================================== For example, a battery-pack usually consists of a number of battery cells that are connected in serial and/or parallel. Each constituent battery-cell is meshed and its active zone consists of a certain number of the finite-volume cells, say 1000. The default “Cluster Cells” option with N_x = N_y = N_z = 1 is used for every battery cell in a pack. The potentials and temperatures will be solved in each finite-volume-cell of a particular battery cell. The average values will be calculated also and these averaged values are used as inputs to either ECM or Newman P2D submodels. In turn, all solved equations for these finite-volume-cells of that zone will get their pro-rata share of the source terms from the submodels. However, should a clustering of, say, N_x = 2, N_y = 3, N_z = 4 be used, then the original 1000 finite-volume-cells are separated into 24 clusters, thus providing a more detailed distributions than the N_x = N_y = N_z = 1 clustering, because all the averaging and the source-term grouping is done per cluster. Please note that disabling the “Cluster Cells” option will force a solution of the submodels for each individual cell and this makes the simulation considerably slower, with very little or no benefit. ===================================================== The way the clusters are constructed are as follows: (1) A bounding box for an active zone of each battery is identified (X_min, X_max, Y_min, Y_max, Z_min, Z_max). (2) The bounding box is divided into N_x * N_y * N_z clusters uniformly (the values N_x, N_y & N_z are defined in the MSMD panel). 3) The control volumes with the cell centre in each cluster belong to its own specific cluster and are used to calculate the cluster temperature and voltage.