National Institute for Fusion Science

Many people enjoy billiards. There are but ten billiard balls, but it is difficult to predict how the balls will collide into one another as they move on the table, and then to decide upon a shot. But what if there were several hundred million or several trillion balls on the table? Further, what if balls of different size and weight were mixed together? It would be difficult to predict the movement of the balls using even computer simulations. Here, we introduce simulation research on predicting the movements of plasma with various species of ions.

Plasma in the future fusion reactor will not only be deuterium and tritium which are the fuel of fusion, and helium generated from the fusion reaction. It also will include small amount of impurities, such as tungsten, iron, carbon, and others that will be mixed together. Because such impurities cool the plasma, those are not desirable for maintaining high-temperature plasma. Thus, the vessel material to be selected will be one which causes the release of few impurities. We are searching for approaches to the device’s design that will make it easier to remove impurities. At the same time, it also is necessary to understand and to predict whether impurities that entered the plasma will remain or will be expelled outside the device (which is called the transport phenomenon). In order to determine this, in addition to experimental research (see Research Update no. 227), computer simulation research is indispensable. However, because calculations under the condition of various species of ions intermingling are extremely complicated, the first issue is establishing the method. 

Important in conducting calculations is treatment of the phenomenon in which various species of ions collide, which is called the Coulomb collisions. As billiard balls repel other balls, move around the table, and fall into a pocket, ions repelled by the Coulomb collision move toward inside the bottle of the magnetic field lines, and some ions escape to outside the bottle. In the case in which a number of species of ions intermingle, because great differences appear in the condition of the collision depending upon the combination of the colliding ions it is necessary to calculate which ions will collide and to calculate accurately what type of transport phenomenon will occur. At the National Institute for Fusion Science, we have developed a simulation method that can accurately calculate the phenomena in which various species of ions collide.

The discussion will become somewhat academic here. We will explain points of improvement based upon comparisons with methods used to date. The calculation of the Coulomb collision must meet the two following physical principles. These are the principle that the total amounts of momentum and energy do not change from before or after the collision, and the principle that the amount of “entropy” increases due to the collision. “Entropy” is the quantity that expresses in which direction change moves. For example, when one drop of ink is placed in a cup of water, the ink will spread. That ink which has spread once will not naturally recombine into one drop. Change advances in the direction that entropy grows. That is, phenomena in this world will move in the direction of chaos (entropy increases). The technique heretofore for calculating collision phenomena could satisfy the principle of increase in entropy only through ideal conditions in which the temperature of all ion species was the same. The method that we have pioneered enables us to satisfy the principle also in cases in which the ion temperature of various species varies. Moreover, also in cases in which simulations of collision phenomena have been conducted over a long period of time, it was confirmed that the law of conservation for momentum and energy is satisfied.

We will soon undertake this research that utilizes high precision calculation methods. Our first goal is to reproduce the movements of impurities observed in the Large Helical Device (LHD) through numerical simulation. The next step is to model the planned deuterium experiment through numerical simulation. Deuterium has twice the mass of light hydrogen, which is used at present in the LHD experiment. But by using deuterium it is expected that plasma performance will improve. Through these experiments and the comparison and verification with the numerical simulations, we aim to heighten the reliability of the calculation methods, and to predict the behaviors of fusion plasma.