论文部分内容阅读
旋涡泵内部流动理论不完善,其叶轮和流道水力尺寸计算通常采用相似计算法和经验统计计算法。文中针对这两种方法在应用中存在的诸多局限性,提出了基于人工神经网络的智能算法。该方法既不以寻求相似的模型泵并进行换算为基础,也不以统计方法获取经验数据,而是以现有旋涡泵产品数据为基础,应用神经网络构建旋涡泵性能参数与水力尺寸之间的关系模型,从而实现泵的性能参数与水力尺寸的非线性映射,并利用这一关系模型进行新产品设计。文中介绍了这一方法的基本原理和关键技术,其中包括神经网络的结构设计、模型选择、样本对设计和网络训练等。解决了由于旋涡泵叶轮和流道结构型式繁多、设计中缺乏足够的经验数据支持及缺乏统一的经验统计方法的问题,提供了使不同企业能够有效地利用已有产品开发新产品的统一方法,该方法不仅支持知识重用,而且具有通用性,对提高旋涡泵的设计质量和缩短设计周期具有重要意义。
The theory of internal flow of the vortex pump is imperfect, and the calculation of the hydraulic dimensions of the impeller and the runner generally adopts the similar calculation method and the empirical statistical calculation method. In view of the limitations of these two methods in application, this paper proposes an intelligent algorithm based on artificial neural network. Based on the existing data of the vortex pump, the method is based on the existing model data of the vortex pump, and the neural network is used to construct the relationship between the performance parameters of the vortex pump and the hydraulic dimensions The relationship between the model to achieve pump performance parameters and hydraulic dimensions of non-linear mapping, and the use of this relationship model for new product design. This paper introduces the basic principles and key technologies of this method, including the structural design of neural networks, model selection, sample design and network training. Solves the problem of lack of sufficient empirical data support and lack of a unified empirical statistical approach due to the large variety of vortex pump impellers and runner structures in the design, a uniform approach that enables different enterprises to effectively utilize existing products to develop new products, The method not only supports knowledge reuse but also has universality, which is of great significance to improve the design quality of the vortex pump and shorten the design cycle.