I do engineering, statistics, and data. I think of Mathematics as a language for organizing thought. Mathematics differs from linguistics in that it is a language far more precise, but less capable of expressing human emotion. Mathematics can be used to describe nearly everything around us. Human behavior, biology, physiology, psychology, social events, natural disasters, economics, and politics, and even divinity can be more precisely discussed in mathematics. Our awareness of the surroundings generates data. Data aren't necessarily numbers. In fact, numbers are used only as surrogates for data. Mathematics differ from arithmetic, and statistics essentially is mathematics. The common perception that statistics is all about numbers stems from that data processing in statistics require the data to be represented numerically. Mathematics allow us see patterns in things. Patterns tend to repeat. This is what allows us predict things, set expectations, and create things that behave in an expected manner. This is how we have working televisions, develop our relationships, and maintain our health and well-being. I often hear "Statistics can prove anything". No it cannot, unless one knows too little statistics to examine stories and ask questions. My language of science and engineering is mathematics. What excites me most is seeing things that I had not seen before and going places I haven't been to before. This is why I find joy in discovering patterns that I didn't know exist as well as proving a scientific result that had not been proven before.
Multi-phase flow in thermal-hydraulics is one of the least understood phenomena in nature. Behavior of phases such as liquids, vapor, and solid matter, existing under harsh conditions and rapid motion become extremely hard to predict. Yet, adequate understating of this phenomenon is vital for the prediction of many physical systems around us; in energy generation of all forms - from Nuclear to Solar, air craft and space vehicle design, defense, geology, and cardiology.
Data science sometimes is thought of as manipulation of data with tools. Tools are important and learning the skills of using tools is not all that difficult, where repetitive use of tools affords mastery. However, what can be done with tools depends entirely on the art of data science. A chisel and a hammer was all that Michelangelo used -the same tools for removing a rusted bolt from a modern day bathroom commode. A job important, but not curated at Galleria dell' Accademia.
"We are out of money" Prof. Hochreiter informed me in the middle of a research project. "They want about fifteen grand to do the experiment" "Do we have lab space?" "Of course". "Lets go see it" I said. I followed him and his bear walk to a seldom used second floor lab. We looked around. There was water, an air compressor and an assortment of used tools in a yellow toolbox.
A numerical data-based approach to conflict resolution. Method helped resolve a deep rooted conflict that lasted for over three years -2015
Statistical modeling of energy data shows that the cost of electricity is to rise if grid continues to mix conventional base load generation and renewable generation -2016
A data-based prediction of our energy future. Decentralized renewable generation and storage poised to rise -2016
Data Science
Engineering
Large amounts of CFD model data shows combustion and fragment flow through a doorway such as from a bomb explosion. Rear floor affected only by the shock wave -2015
It took years of education of the human form, imagination and ingenuity for Michelangelo to create David and Pietà. Yet of course, marble cannot be carved with bare hands.
Data prediction through Infinite convolution of surface response modeling, a mathematical discovery that allowed continued safe operation of 14 US commercial nuclear reactors - Awarded "Engineer of the year"
A combination of experimental and Monte-Carlo model data are used to satisfy regulatory requirements of code validation
"What else do you need?" "I' will let you know" I said. An experimental rig for understanding two-phase flow in nuclear reactor cores was put together with what was around that included office supplies. Prof. Hochreiter visited me in the lab several
weeks later in an early afternoon. He clapped his hands and roared into laughter when my experimental rig held together with paper clips got into action. Looking at the array of paperclips he said "Lets call this "CENTIPEDE!" Results from CENTIPEDE lead to discoveries that helped understand two-phase flow past spacer grids in modern nuclear reactors. Prof. Hochreiter would proudly show CENTIPEDE to his visitors and say "Take note! This is the first time you ever saw inside a reactor, this close!" He laugh and points at me: "This guy put it together on a shoestring! He turns towards me "HOW MUCH WAS THE TOTAL?" "Ninety four dollars and thirty two cents... I needed an air regulator". He will then burst into another round of roaring laughter.
Cross flow velocities in accelerating flow through a nuclear fuel assembly. High cross flow is caused by flow diversion indicating to the importance of mixing features in gaining thermal margin. The velocities are well distributed with bifurcations that provide an excellent visual characterization that justifies the use of "mean cross flow velocity".
When a system is sufficiently well characterized, statistical models can generate high fidelity data that can predict its behavior. This allows treating "generated data" when "hard data" are scarce and difficult to obtain. In a methodology developed in 2015, available "hard data" were combined with "generated data" to demonstrate that a computer code is capable of simulating complex phenomena. The combination of data compensate each other for fidelity and availability.
Perhaps nothing can replace human ingenuity and information processing power in decision making. However, organization and visualization of informtion can help this process.
Predictions made base on mathematical theories have guided humans to most unlikely discoveries that later proved the correctness of those predictions. The recent detection of gravitational waves, predicted nearly a century ago perhaps is one of the best examples.
Mathematical transforms reveal patterns in chaotic big data -2016