Photo: ZSW

ZSW AI team devel­ops inter­ac­tive plat­form for companies

Arti­fi­cial intel­li­gence is being used more and more, espe­cial­ly in the ener­gy tran­si­tion indus­try. Self-learn­ing meth­ods help, for exam­ple, to bet­ter pre­dict wind and solar ener­gy feed-in or to opti­mize the devel­op­ment and pro­duc­tion of fuel cells, bat­ter­ies and e‑fuels. To ensure that small and medi­um-sized com­pa­nies are not left behind by the big play­ers in the indus­try, they need future-proof tech­nolo­gies. The Cen­tre for Solar Ener­gy and Hydro­gen Research Baden-Würt­tem­berg (ZSW) has there­fore devel­oped an inno­va­tion plat­form where com­pa­nies can test for them­selves how they can use arti­fi­cial intel­li­gence (AI) to make their process­es from prod­uct devel­op­ment to busi­ness oper­a­tions more future-proof and eco­nom­i­cal­ly viable. Ideas and con­cepts for this were devel­oped by the ZSW as part of the “AI Lab for Renew­able Ener­gies”, which was finan­cial­ly sup­port­ed by the Baden-Würt­tem­berg Min­istry of Economics.

“Our AI plat­form cov­ers a broad range of appli­ca­tions: Whether new tech­nolo­gies for the ener­gy tran­si­tion, whether cli­mate-neu­tral pro­duc­tion process­es or appli­ca­tions — this is what our inno­va­tion plat­form stands for. Small and medi­um-sized enter­pris­es — and espe­cial­ly star­tups — along the entire val­ue chain in the field of renew­able ener­gies can ben­e­fit from it,” explains Anton Kaifel, team leader of the AI and Machine Learn­ing depart­ment at ZSW.

First test, then decide 

Using the ZSW ser­vice is sim­ple: com­pa­nies reg­is­ter on the web­site, upload their data to a pro­tect­ed data room and can use soft­ware and infra­struc­ture to inde­pen­dent­ly train AI mod­els with their data. “Thus, com­pa­nies have a very low entry thresh­old and can use our AI Play­Ground tool to test whether and how arti­fi­cial intel­li­gence can be inte­grat­ed into their com­pa­ny and whether it is worth­while for them — in oth­er words, test-before-invest,” explains Dr. Frank Sehnke, Data Sci­en­tist for AI at ZSW. He and the ten-mem­ber AI research team at the ZSW know the require­ments and needs of com­pa­nies very well. They gained expe­ri­ence in the “AI Lab for Renew­able Ener­gies (KILEE)”, where com­pa­nies were able to use AI meth­ods to opti­mize pro­duc­tion process­es and devel­op new prod­ucts and ser­vices. The project start­ed in April 2020 and has now been suc­cess­ful­ly com­plet­ed. “Thanks to the AI Lab and oth­er research projects, the tech­no­log­i­cal matu­ri­ty of our plat­form is very high. Machine learn­ing can be used to improve prod­ucts, process­es and ser­vices and to devel­op new busi­ness mod­els. This is done using exist­ing or con­tin­u­ous­ly col­lect­ed data,” says Frank Sehnke.

Sev­er­al com­pa­nies from dif­fer­ent indus­tries in the South­west par­tic­i­pat­ed in the region­al AI lab, includ­ing weath­er and ener­gy fore­cast­ing ser­vice providers and com­pa­nies from the wind and solar indus­tries. The Freiburg-based high-tech start-up green­ven­to­ry, which spe­cialis­es in soft­ware-sup­port­ed solu­tions for the inven­to­ry and analy­sis of dis­trib­uted ener­gy sys­tems, also ben­e­fits from the tai­lor-made offers from the ZSW dig­i­tal lab­o­ra­to­ry. “For us as a young start-up com­pa­ny in the ener­gy sec­tor, the AI Lab was and is a very great help. The experts at the ZSW have shown us con­crete pos­si­ble uses of AI and exem­plary appli­ca­tions with our data. I think it is very good that an AI plat­form has now emerged from the lab. Every­one should be able to use the enor­mous poten­tial that AI tech­nol­o­gy offers. The plat­form offers low-thresh­old access for the use of AI in one’s own com­pa­ny,” says green­ven­to­ry co-founder Dr. Kai Mainzer.

Analy­sis tool for com­pa­nies nationwide 

While the focus of the KI-Lab was on com­pa­nies from the south­west, the plat­form with its analy­sis tool is now avail­able to com­pa­nies from all over Ger­many. “We want to give com­pa­nies a com­pet­i­tive edge with our dig­i­tal ser­vice by help­ing them achieve greater effec­tive­ness, effi­cien­cy and agili­ty. The dynam­ic devel­op­ment of the mar­ket requires flex­i­bil­i­ty and an AI solu­tion that sup­ports new busi­ness mod­els,” says Anton Kaifel.

Com­pa­nies that sign up on the plat­form go through a process. First, areas of appli­ca­tion for AI can be deter­mined in the respec­tive com­pa­ny. From this, an AI readi­ness lev­el is deter­mined and fur­ther steps are dis­cussed. Then, a proof-of-con­cept (PoC) appli­ca­tion with AI will test the data and devel­op ini­tial AI mod­els using the plat­form. Tuto­ri­als pro­vide guid­ance in the han­dling of the AI plat­form. If AI mod­els based on a PoC are to be used by the com­pa­nies for process opti­miza­tion for mon­i­tor­ing and pre­dic­tive main­te­nance of plants or for new ser­vices and prod­ucts, the ZSW sup­ports the com­pa­nies to devel­op indi­vid­ual appli­ca­tions and soft­ware. This enables com­pa­nies to sig­nif­i­cant­ly short­en the time hori­zons for new devel­op­ments and to intro­duce their prod­ucts and ser­vices to the mar­ket earlier.

With the AI plat­form, the ZSW pri­mar­i­ly wants to effec­tive­ly use the inter­faces between the dig­i­tal and ener­gy indus­tries, between star­tups and estab­lished com­pa­nies along the entire val­ue chain in the field of renew­able ener­gies, as AI has great inno­v­a­tive pow­er in this area. “A study com­mis­sioned by the Fed­er­al Min­istry for Eco­nom­ic Affairs and Ener­gy (BMWi) shows that com­pa­nies that use arti­fi­cial intel­li­gence achieve high­er prof­its for the same turnover and cre­ate addi­tion­al jobs. Through AI, com­pa­nies are more like­ly to pro­duce inno­va­tions and set mile­stones,” con­cludes Anton Kaifel.

The ZSW AI plat­form can be accessed via