Abstract Rougher flotation is a complex process, making it difficult for a single variable to comprehensively and accurately reflects real-time working conditions. This paper proposes a new multivariable recognition method for rougher flotation conditions using pulp flow and froth size as parameters. It utilizes qualitative trend analysis to recognise rougher flotation condition for the first time, where an improved trend extraction method is used to improve extraction accuracy, and fuzzy logic is adopted to calculate the matching degree of trends. Considering the change in knowledge base caused by the increase in variables, this paper presents a scheme of building knowledge base for multivariable rougher flotation conditions. Experimental results of gold–antimony rougher flotation show that the proposed method can accurately recognise rougher flotation conditions.

中文翻译：

---

基于多变量趋势分析的金锑粗浮选工况识别方法

摘要 粗浮选工艺复杂，单一变量难以全面准确反映实时工况。本文提出了一种新的以矿浆流量和泡沫尺寸为参数的较粗糙浮选条件的多变量识别方法。首次利用定性趋势分析识别较粗糙的浮选条件，采用改进的趋势提取方法提高提取精度，并采用模糊逻辑计算趋势的匹配度。考虑到变量增加引起的知识库变化，提出了一种多变量粗浮选条件的知识库构建方案。