b'QUALITY MEASUREMENTOPTIX Applied IntelligenceThe only real-time quality measurement for tissue productionImproving mill operator and production efficiency via artificial intelligence (AI) and machine learningBy Beth Ann Zarko, Product MarketerDigital Solutions, Global Strategic Marketing & Communications, SolenisReal-time finished qualityWith increasing availability ofThese tools simply cannot handle measurements (i.e., dry tensile, wetinstrumentation and use ofthe computations required to tensile, bulk, etc.) can significantlycentralized data historians, mills areevaluate this level of complexity and improve tissue mill operations andcollecting vast amounts of data thatnonlinearity in real time. decision making.However, real-timeprovide them with ever-increasing physical sensors do not exist for thesevisibility into their processes. TissuePredictive analytics supported by quality parameters.The only way tomanufacturing can have as many asmachine learning can provide real-achieve real-time insights is through10,000 data historian tags that aretime quality measures that remain mathematical predictions generatedassociated through highly complex,robust and accurate in the face using artificial intelligence and machinemulti-dimensional relationships.of changing machine conditions. learning. Solenis OPTIX AppliedTraditional data analytic tools areThese adaptive quality soft sensors Intelligence is leading the industry bylargely reflective, time-consuming,allow for more informed, on-the-providing enhanced process visibilityand have specific limitations infly decision making, rapid change with an analytics platform that uncovershighly variable, continuous-processdetection, and process control mill improvement opportunities notmanufacturing like the tissueoptimization without requiring previously possible.industry.periodic model tuning. The only way to achieve real-time insights is through mathematical predictions generated using artificial intelligence Figure 1. (a) Comprehensible relationships between bulk, tensile, and basis weight; (b) Causal network showingand machine multidimensional relationships between bulk, tensile, and basis weight, and numerous other predictors. learning. 84 TISSUE TECHNOLOGY INTERNATIONAL'