Author :Iflowpower – Portable Power Station Supplier
Studies have shown that one of the important causes of ordinary lithium-ion batteries, is because the electrolyte of traditional lithium-ion batteries is organic liquid, and the side react can occur in less than high temperatures, and the gas is caused by explosion. In response to this situation, researchers from Stanford University, the use of artificial intelligence and machine learning, found approximately 21 solid electrolyte materials, there is hope to completely solve this problem. Recently, the team published their research results in the "Energy and Environmental Science" magazine.
The article shows that scientists do not use traditional random testing methods to find new solid electrolyte materials, but use artificial intelligence and machine learning, through experimental data constructing prediction model. They trained a computer algorithm, based on existing data, learn how to identify the quality of the compound. This process and face recognition algorithm, after observing several examples, the process of identifying faces is similar.
The first author of the paper, the leader of the study, Ph.D. in physics said: "The existing amount of a compound containing lithium elements is tens of thousands, most of which are not tested.
Some of them may be a conductor excellent in performance. We have developed a computational model, and our existing limited data is studied, from a large-scale database, screening the right materials. This speed of screening is a million times of existing screening methods.
"In order to design this model, AustInsendek spends almost two years, collects almost all scientific data on solid compounds containing lithium elements, and then treats the stability, cost, richness, and lithium ions of solid compounds through these data. Many physicochemical properties have been evaluated, and 21 of the most suitable solid electrolyte materials were finally screened. AustInsendek said: "We scree more than 12,000 compounds containing lithium elements, and finally found 21 ideal materials as solid electrodes.
Screening as long as it takes a few minutes. Most of my time, actually used to collect and manage all data, develop more reliable metrics related to predictive models. "Researchers plan to test these 21 materials in the laboratory environment, further confirmation of whether they are the best choice.
The meaning and future of this study, another author, Material Science and Engineering Assistant Professor Evanreed, said: "Our approach can handle many of the problems of materials, which help new research investment effects in this field. With the increase in data volume in the real world and the improvement of computer performance, our innovation capabilities will increase in an index method. Whether it is a lithium-ion battery, fuel power lithium battery is still other batteries, and research in this field has significant significance.
".