© ecoKI

Increas­ing ener­gy effi­cien­cy with eco­KI: tools, knowl­edge and sup­port for SMEs

Using ener­gy more effi­cient­ly and thus reduc­ing resource con­sump­tion and costs — dig­i­tal­iza­tion and arti­fi­cial intel­li­gence (AI) meth­ods offer a wide range of options here. How­ev­er, small and medi­um-sized enter­pris­es are too often unable to make full use of them, because they often lack the exper­tise to inte­grate these tech­nolo­gies in their own oper­a­tions and the bar­ri­ers to entry are high. In future, they will be able to obtain tools, knowl­edge and a sup­port infra­struc­ture via a plat­form. This is the aim of research in the new joint project “eco­KI” led by the BIBA — Bre­men Insti­tute for Pro­duc­tion and Logis­tics at the Uni­ver­si­ty of Bremen.

Sim­pli­fy and accel­er­ate the step from research lab­o­ra­to­ry to oper­a­tional application

The path from pro­to­type from the research lab to the mar­ket is ardu­ous and takes a long time. The step into the oper­a­tional appli­ca­tion espe­cial­ly of small and medi­um-sized enter­pris­es (SMEs) is still too sel­dom tak­en. The “eco­KI” research and tech­nol­o­gy plat­form is intend­ed to remove hur­dles here and accel­er­ate the process­es along the way. The project focus­es on the trans­fer of find­ings and devel­op­ments from pub­licly fund­ed research.

The goal of the part­ners: To gath­er knowl­edge on dig­i­ti­za­tion and AI meth­ods, espe­cial­ly on machine learn­ing, and make it eas­i­ly and clear­ly acces­si­ble, to build up fur­ther knowl­edge, and to net­work experts with users in order to enable a low-thresh­old and rapid intro­duc­tion to the ben­e­fits of the new tech­nolo­gies for increas­ing ener­gy efficiency.

Per­ma­nent­ly avail­able solutions

Per­ma­nent­ly avail­able, expand­able solu­tions are to be cre­at­ed — with the plat­form itself as well as through the thus gen­er­at­ed imple­men­ta­tion of indi­vid­ual projects in the com­pa­nies with the goal of increased ener­gy effi­cien­cy. In addi­tion, the col­lab­o­ra­tive part­ners expect their research to lead to new ques­tions and fur­ther insights into the needs of industry.

A cen­tral task in the eco­KI project is the devel­op­ment and organ­i­sa­tion of the plat­form as the basis for a long-term, func­tion­ing busi­ness mod­el. The sec­ond essen­tial work is the devel­op­ment of stan­dard build­ing blocks for the plat­form. These should serve as a knowl­edge base for the users and can be used for new tasks. The reusable mod­ules imple­ment­ed in the plat­form are intend­ed to offer com­pa­nies sup­port in devel­op­ing their process­es cost-effec­tive­ly and effi­cient­ly through the use of AI meth­ods. Syn­er­gies from dif­fer­ent use cas­es should be able to be used.

The eco­nom­ic exploita­tion per­spec­tives of the project are strate­gi­cal­ly ori­ent­ed in the long term and are pri­mar­i­ly based on the improved coop­er­a­tion between devel­op­ers and users of inno­v­a­tive AI tech­nolo­gies in oper­a­tional practice.

With CRISP-DM, rig­or­ous mod­els, and machine learning.

For the col­lec­tion, pro­cess­ing and use of the data, the project part­ners rely on the CRISP-DM method (Cross Indus­try Stan­dard Process for Data Min­ing). This is a proven, stan­dard­ized process mod­el that helps achieve a con­sis­tent approach to devel­op­ing data min­ing process­es to iden­ti­fy trends and cor­re­la­tions. To devel­op the gener­ic build­ing blocks for the plat­form, the project also deals with so-called rig­or­ous mod­els and — in the field of arti­fi­cial intel­li­gence — with deep learn­ing, a sub­field of machine learning.

Rig­or­ous mod­els map a tech­ni­cal mech­a­nism with exact sci­en­tif­ic method­ol­o­gy. They have the advan­tage of being able to under­stand sim­u­lat­ed process­es more pre­cise­ly with their help. Machine learn­ing, as opposed to for­mal­ized expert knowl­edge, deals with the auto­mat­ed cre­ation of pre­dic­tive mod­els based on data alone. Espe­cial­ly due to the devel­op­ment of Deep Learn­ing approach­es and their suc­cess­ful appli­ca­tions, the use of machine learn­ing has been expe­ri­enc­ing rapid growth for years.

Key data on the “eco­KI” joint project

In the work­shop “Increas­ing Ener­gy Effi­cien­cy in Pro­duc­tion through Dig­i­ti­za­tion and AI” ini­ti­at­ed by the BIBA — Bre­men Insti­tute for Pro­duc­tion and Logis­tics with rep­re­sen­ta­tives of the Fed­er­al Min­istry for Eco­nom­ic Affairs and Ener­gy (BMWi) and the Project Man­age­ment Organ­i­sa­tion Jülich (PtJ), among oth­ers, the idea for the real­i­sa­tion of a research and tech­nol­o­gy plat­form and net­work struc­ture ori­ent­ed towards the issues of sus­tain­able ener­gy effi­cien­cy emerged. It is intend­ed to make inno­v­a­tive R&D results more eas­i­ly acces­si­ble, in par­tic­u­lar to SMEs, and to pro­mote their application.

Part­ners of the result­ing research project “eco­KI” are, besides the BIBA as coor­di­na­tor, the Ger­man Research Cen­ter for Arti­fi­cial Intel­li­gence (DFKI) in Kaiser­slautern, the Insti­tute for Neu­roin­for­mat­ics (INI) of the Ruhr-Uni­ver­si­ty Bochum and the pro­fes­sor­ship Process Con­trol Engineering/Workgroup Sys­tems Process Engi­neer­ing of the Tech­ni­cal Uni­ver­si­ty Dres­den. The four-year project ends on 30.11.2024, is fund­ed by the BMWi with­in the frame­work of the 7th Ener­gy Research Pro­gramme of the Fed­er­al Gov­ern­ment and is super­vised by the Project Man­age­ment Organ­i­sa­tion Jülich (PtJ).