Photo: ZSW

“Grid­Sage” makes dis­tri­b­u­tion net­works trans­par­ent and helps to avoid net­work bottlenecks

In the course of the ener­gy turn­around, more and more elec­tric­i­ty from decen­tral­ized pho­to­volta­ic and wind pow­er plants is flow­ing through the pow­er grids. In order for trans­mis­sion sys­tem oper­a­tors to be able to con­trol plants when nec­es­sary, dis­tri­b­u­tion sys­tem oper­a­tors will soon have to cre­ate more trans­paren­cy in their grids. This is pro­vid­ed for in the new reg­u­la­tion of grid man­age­ment, Redis­patch 2.0 for short, which will come into force on 1 Octo­ber 2021. How­ev­er, it will be dif­fi­cult for many dis­tri­b­u­tion sys­tem oper­a­tors to imple­ment the require­ments on their own due to a lack of know-how and too lit­tle time. In order to sup­port com­pa­nies with the required feed-in fore­casts, the Cen­tre for Solar Ener­gy and Hydro­gen Research Baden-Würt­tem­berg (ZSW) has now devel­oped the “Grid­Sage” tool. It uses arti­fi­cial intel­li­gence to accu­rate­ly fore­cast the gen­er­a­tion of green pow­er plants in the dis­tri­b­u­tion grid for the next 36 hours. It is also pos­si­ble to pre­dict the load. Stadtwerke Schwäbisch Hall is already using Grid­Sage successfully.

Adjust­ing the feed-in of green elec­tric­i­ty pow­er plants in the event of an overload

The legal basis is the Grid Expan­sion Accel­er­a­tion Act. The pre­vi­ous­ly applic­a­ble reg­u­la­tions on feed-in man­age­ment for renew­able ener­gy and com­bined heat and pow­er plants will be sup­ple­ment­ed by the new redis­patch sys­tem. As of the fourth quar­ter, green elec­tric­i­ty and CHP plants with an installed gen­er­a­tion capac­i­ty of 100 kilo­watts or more must be includ­ed in the redis­patch. For this pur­pose, dis­tri­b­u­tion sys­tem oper­a­tors must, among oth­er things, pre­pare feed-in fore­casts and iden­ti­fy redis­patch potentials.

Grid­Sage: Fore­casts for Redis­patch 2.0

The fore­cast of gen­er­a­tion out­put is of cen­tral impor­tance in the Redis­patch 2.0 sce­nario. Good fore­casts ensure that redis­patch mea­sures can be car­ried out cheap­ly and effi­cient­ly. How­ev­er, this pos­es major chal­lenges for small dis­tri­b­u­tion sys­tem oper­a­tors in par­tic­u­lar. The ZSW fore­cast­ing tool Grid­Sage sup­ports them in this. The use of Grid­Sage makes dis­tri­b­u­tion net­works trans­par­ent, equips them for the future and helps to avoid net­work bot­tle­necks. “Grid­Sage fore­casts the pow­er gen­er­a­tion in the dis­tri­b­u­tion grid for the next 36 hours with a res­o­lu­tion of 15 min­utes,” explains Dr. Jann Binder from ZSW. “We update the fore­casts for the indi­vid­ual EEG plants and grid nodes every hour and make them avail­able to the grid oper­a­tor in an auto­mat­ed way.” Oth­er time inter­vals and fore­cast hori­zons or flex­i­ble fore­cast deliv­ery are also pos­si­ble on request.

The ZSW researchers cre­ate the fore­casts with the help of arti­fi­cial intel­li­gence meth­ods: neur­al net­works learn from past data which gen­er­a­tion plant will pro­duce how much pow­er under which weath­er con­di­tions. Grid­Sage uses this infor­ma­tion to auto­mat­i­cal­ly gen­er­ate high-res­o­lu­tion fore­casts. At many points, the ZSW sub­jects the mea­sured data to a plau­si­bil­i­ty check, for exam­ple by com­par­ing the nom­i­nal plant out­put with the annu­al gen­er­a­tion or by com­par­ing it with the gen­er­a­tion of mea­sured neigh­bour­ing plants, in order to exclude erro­neous data.

Grid­Sage can also do more than is required by Redis­patch 2.0: “In addi­tion to gen­er­a­tion, the tool fore­casts the load in the dis­tri­b­u­tion grid,” says Binder. “In the future, this will become increas­ing­ly impor­tant. The more elec­tro­mo­bil­i­ty becomes estab­lished and heat pumps are installed in homes, the greater the impor­tance of pre­cise fore­casts of con­sump­tion in dis­tri­b­u­tion grids. In this way, the grids can con­tin­ue to be oper­at­ed effi­cient­ly in the future and the need for grid expan­sion can be reduced.”