With small grids to the big ener­gy turnaround

Sta­ble ener­gy sup­ply through arti­fi­cial intel­li­gence: Pader­born researchers and com­pa­nies focus on oper­at­ing and con­trol meth­ods for microgrids

The trans­for­ma­tion toward a sus­tain­able, effi­cient and cost-effec­tive ener­gy sup­ply is one of the key chal­lenges of the 21st cen­tu­ry. Local net­works, so-called micro­grids, offer great poten­tial. In the case of decen­tral­ized and cel­lu­lar ener­gy sys­tems, the bal­ance between ener­gy sup­ply and demand should already be estab­lished at the local lev­el. The chal­lenge here: a con­sis­tent and effi­cient ener­gy sup­ply based on green sources. Data-dri­ven and self-learn­ing meth­ods could rem­e­dy the sit­u­a­tion, but the intel­li­gent solu­tions have numer­ous weak­ness­es to date. This is where a new project comes in, in which sci­en­tists from the Uni­ver­si­ty of Pader­born, the SICP — Soft­ware Inno­va­tion Cam­pus Pader­born and the busi­ness part­ners West­falen­WIND GmbH and West­falen Weser Netz GmbH are work­ing hand in hand. Their goal is to devel­op an open source frame­work that address­es issues that may arise in the oper­a­tion of dis­trib­uted ener­gy net­works. Freely acces­si­ble and stan­dard­ized tools for explor­ing data-dri­ven con­trollers for ener­gy tech­nol­o­gy will help col­lec­tive­ly dri­ve the tran­si­tion of the cur­rent ener­gy sup­ply sys­tem to a sus­tain­able and renew­able ener­gy-dri­ven struc­ture. The Ger­man Fed­er­al Min­istry of Edu­ca­tion and Research (BMBF) has been fund­ing the project “Train­ing, Val­i­da­tion and Bench­mark Tools for the Devel­op­ment of Data-Dri­ven Oper­a­tion and Con­trol Meth­ods for Smart Local Ener­gy Sys­tems” (DARE) since Octo­ber 2021 for two years with approx­i­mate­ly 988,537 euros.

Small nets, big effect

Micro­grids rep­re­sent an impor­tant solu­tion com­po­nent for the ener­gy tran­si­tion: They con­sist of sus­tain­able ener­gy sources such as wind tur­bines, ener­gy stor­age sys­tems such as bat­ter­ies, and ener­gy con­sumers from var­i­ous sec­tors, such as elec­tric­i­ty, heat, or mobil­i­ty. The local grids can sup­ply ener­gy to house­holds and indus­tri­al com­pa­nies both grid-con­nect­ed and autonomous­ly in iso­lat­ed oper­a­tion. “Micro­grids have the advan­tage that through their local inte­gra­tion, regen­er­a­tive ener­gy can be pro­vid­ed close to con­sump­tion and thus used direct­ly by the con­sumer over a short dis­tance. This can relieve the strain on suprare­gion­al ener­gy grids and reduce the need for grid expan­sion. In addi­tion, the share of renew­able ener­gies is increased, since the lossy trans­port over long dis­tances and unnec­es­sary shut­downs of renew­able pow­er plants due to grid bot­tle­necks are avoid­ed,” explains Dr. Gun­nar Schomak­er, “Research and Devel­op­ment Man­ag­er — Smart Sys­tems” at SICP.

Chal­lenges in the oper­a­tion of microgrids

How­ev­er, a key chal­lenge in the oper­a­tion of micro­grids must be over­come: ensur­ing a con­tin­u­ous and effi­cient ener­gy sup­ply through oper­at­ing and con­trol pro­ce­dures. “A sta­ble ener­gy sup­ply is much more dif­fi­cult to main­tain in decen­tral­ized grids — due to the volatil­i­ty, i.e. fluc­tu­a­tions, of regen­er­a­tive pow­er plants and typ­i­cal­ly only low stor­age and reserve capac­i­ties — than in cen­tral­ized grids that are sup­port­ed by con­ven­tion­al large-scale pow­er plants,” explains the pro­jec­t’s sci­en­tif­ic leader Dr.-Ing. Oliv­er Wallscheid from the “Con­trol and Automa­tion Tech­nol­o­gy” depart­ment at the Uni­ver­si­ty of Pader­born. “For the oper­a­tion and con­trol of such sto­chas­tic, i.e. ran­dom­ly depen­dent, het­ero­ge­neous and volatile ener­gy grids, the tra­di­tion­al top-down strate­gies of cen­tral­ized large-scale grids can­not be trans­ferred,” adds Jun.-Prof. Dr. Sebas­t­ian Peitz from the Depart­ment of “Data Sci­ence for Engineering”.

“Instead, data-dri­ven and self-learn­ing meth­ods are emerg­ing as a pos­si­ble solu­tion, e.g. from the field of so-called ‘rein­force­ment learn­ing’. The prob­lem here, how­ev­er, is that these learn­ing and nov­el intel­li­gent con­trol meth­ods can­not be used direct­ly in the field due to safe­ty and avail­abil­i­ty aspects, but must first be improved and eval­u­at­ed on the basis of syn­thet­ic, i.e. arti­fi­cial, data in a closed sim­u­la­tion cycle,” the sci­en­tist con­tin­ues. Although there are already approach­es to solu­tions, they are very het­ero­ge­neous and are often based on high­ly sim­pli­fied mod­el envi­ron­ments, so that no state­ments can be made about future prac­ti­cal trans­fer. In addi­tion, there is no estab­lished stan­dard of com­par­i­son against which data-dri­ven con­trollers can be objec­tive­ly and quan­tifi­ably evaluated.

The project team wants to change that: “The goal of our project is there­fore to build a so-called ‘open source sim­u­la­tion and bench­mark frame­work’ that maps the cur­rent prob­lem frame­work in the oper­a­tion of decen­tral­ized ener­gy net­works. Through eas­i­ly acces­si­ble as well as stan­dard­ized train­ing, val­i­da­tion and com­par­i­son tools, the research of data-dri­ven con­trollers for ener­gy tech­nol­o­gy should be accel­er­at­ed and made com­pa­ra­ble by means of col­lec­tive knowl­edge,” says Wallscheid. By com­bin­ing the­o­ry and prac­tice, the project part­ners want to enable real­is­tic eval­u­a­tion sce­nar­ios and trans­fer data-dri­ven con­trollers from sim­u­la­tion to field use.

Solu­tion for ener­gy sup­ply in emerg­ing and devel­op­ing countries

Micro­grids are a core ele­ment of the ener­gy tran­si­tion, but also a cen­tral build­ing block for estab­lish­ing basic ener­gy sup­plies in emerg­ing and devel­op­ing coun­tries, Wallscheid said. “The fact that micro­grids can oper­ate not only grid-con­nect­ed but also autonomous­ly in island mode is a typ­i­cal case for remote, off-grid areas. In addi­tion to its con­tri­bu­tion to the ener­gy turn­around in Europe, the micro­grid accord­ing­ly rep­re­sents a cen­tral build­ing block for estab­lish­ing the basic ener­gy sup­ply in emerg­ing and devel­op­ing coun­tries, espe­cial­ly in sub-Saha­ran Africa, where the devel­op­ment of a cen­tral ener­gy infra­struc­ture in sparse­ly pop­u­lat­ed, rur­al areas is not in sight, even in the long term,” the sci­en­tist emphasizes.