海洋条件下小型堆稳压器液位智能预测研究

Intelligent prediction of pressurizer water levels in small reactors under ocean conditions

浮动核电站的稳压器等设备受海洋条件影响发生空间位置变化并产生液面晃荡,导致现有压差测量方法不能反映容器内真实液位且易造成信号误触发。为研究海洋条件下稳压器的压差波动规律,优化液位测量方案,本文开展摇摆条件下稳压器液位测量实验;基于实验数据使用麻雀搜索算法优化极限学习机模型,构造了一种液位信号与摇摆角度、摇摆角速度、重力方向加速度、以及空间测点位置等海洋条件之间的回归预测模型。结果表明:本文所提出的麻雀搜索算法优化极限学习机预测模型具有速度快和精度高等优点,预测结果与实验数据在多种典型场景下均吻合良好,可为浮动核电站液位遥测及人工智能技术应用提供参考。

Given the influence of ocean conditions,pressure stabilizers and other equipment in floating nuclear power plants change their spatial position and produce liquid level sloshing,which leads to pressure differences such that measurement methods cannot reflect the actual liquid levels in containers and signals are easily falsely triggered.In this work,we aim to study the rule governing the pressure difference fluctuation of pressurizers under ocean conditions and improve the liquid level measurement scheme.We first conducted a liquid level measurement experiment on pressurizers under rolling conditions and used the extreme learning machine(ELM)with the sparrow search algorithm(SSA)to build a prediction model of the regression between water level signals and ocean conditions,such as rolling angle,angular velocity,gravitational acceleration,and measurement point location.Results show that the proposed SSA-ELM prediction model has fast speed and high precision.The predicted value is in good agreement with the experimental data under a variety of typical conditions.Therefore,our study can provide a reference for the remote measurement of liquid levels and the application of artificial intelligence technology in floating nuclear power plants.