Simulation-based interpretation of spacecraft particle sensor measurements

粒子传感器，如朗缪尔探测器或粒子成像仪，被用于许多卫星和实验室等离子体实验，以推断等离子体参数，如密度和温度。在最简单的形式中，朗缪尔探头由一个可以偏置到可变电压的电极组成，可以从这个电极上测量电流。收集的电流作为偏置电压的函数，即所谓的探针特性，然后可以在理论或计算模型的基础上进行解释，以得出局部等离子体参数。近一个世纪以来，人们在这个问题上做了大量的工作。为了描述朗缪尔探针在等离子体中的响应，已经发展了许多理论模型和最近的几个计算模型。然而，在实践中，探头测量几乎被普遍地解释为能够在操作模式下产生实时答案的简化分析模型。不幸的是，这些解释是出了名的不确定，误差条可能是100%的。这些巨大不确定性的原因来自于理想化模型的使用，在这些模型中，只考虑了相关物理过程的一小部分。例如，一些模型只考虑静止的未磁化等离子体中的探针，另一些模型考虑了等离子体流动而忽略了环境磁场，还有一些模型考虑了磁场，但忽略了等离子体流动或其他影响，如光电子或二次电子发射。在基本上所有的情况下，等离子体都被假定为空间均匀的，而忽略了附近卫星或实验物体的存在及其几何形状。解决这一困境的一个有希望的解决方案是，在详细的计算机模拟的基础上解释粒子传感器的测量结果，该模拟能够解释等离子体-材料相互作用的多物理特性，同时也考虑到进行测量的几何形状。遗憾的是，由于进行这种模拟所需的时间和计算资源，这在实时操作模式下是不可能的。在这项拟议的研究中要探索和开发的一个实际解决方案是，为要支持的每项实验或空间飞行任务建立一个库或数据库，从中可以使用适当的回归技术来推断等离子体参数。与目前的做法相比，新方法将显著提高推断的等离子体参数的准确性。拟议的研究将集中于为选定的航天器和实验产生解决方案库，开发和评估不同的回归技术。这项研究的预期结果是改变了对传感器测量的解释范式，然后将依赖于详细的动力学模拟结果，而不是理想化的分析模型。这一建议是基于多年的卫星环境建模经验。

Particle sensors such as Langmuir probes or particle imagers, are used in many satellites and laboratory plasma experiments to infer plasma parameters such as the density and temperature. In its simplest form a Langmuir probe consists of an electrode which can be biased to variable voltages, and from which the electric current can be measured. The collected current as a function of bias voltage, the so-called probe characteristic, can then be interpreted on the basis of theoretical or computational models, to yield local plasma parameters. Much work has been done on this topic over nearly a century. Many theoretical, and more recently, several computational models have been developed to describe the response of Langmuir probes in a plasma. In practice however, probe measurements are almost universally interpreted in terms of simplified analytic models capable of producing real time answers in operation mode. Unfortunately these interpretations are notoriously uncertain, with error bars that can be of order 100%. The reason for these large uncertainties comes from the use of idealised models in which only a fraction of the relevant physical processes are taken into account. For example, some models only consider a probe in a stationary unmagnetised plasma, others account for plasma flow but neglect ambient magnetic fields, yet others account for a magnetic field, but ignore plasma flow or other effects such as photoelectron or secondary electron emission. In essentially all cases, plasma is assumed to be spatially uniform, and the presence of nearby satellite or experimental objects and their geometry is ignored. A promising solution to this predicament is to interpret particle sensor measurements on the basis of detailed computer simulations capable of accounting for the multiphysics which characterises plasma-material interaction, while also accounting for the geometry in which measurements are made. This is unfortunately not possible in real time operation mode owing to the time and computational resources required to do such simulations. One practical solution to be explored and developed in this proposed research, is to construct a library or data base for each experiment or space mission to be supported, from which plasma parameters could be inferred using a suitable regression technique. Compared to current practice, the new approach would lead to significant improvements in the accuracy of inferred plasma parameters. The proposed research would concentrate on producing solution libraries for selected spacecraft and experiments, developing, and assessing different regression techniques. An expected outcome of this research is a change of paradigm in the interpretation of sensor measurements, which would then rely on detailed kinetic simulation results rather than on idealised analytic models. This proposal is based on many years of experience in satellite-environment modelling.