• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Neural Networks and Learning Systems, IEEE Transactions on >Minimum Data Requirement for Neural Networks Based on Power Spectral Density Analysis
【24h】

Minimum Data Requirement for Neural Networks Based on Power Spectral Density Analysis

机译：基于功率谱密度分析的神经网络最小数据需求

获取原文
获取原文并翻译 | 示例
       
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

One of the most critical challenges ahead for diesel engines is to identify new techniques for fuel economy improvement without compromising emissions regulations. One technique is the precise control of air/fuel ratio, which requires the measurement of instantaneous fuel consumption. Measurement accuracy and repeatability for fuel rate is the key to successfully controlling the air/fuel ratio and real-time measurement of fuel consumption. The volumetric and gravimetric measurement principles are well-known methods for measurement of fuel consumption in internal combustion engines. However, the fuel flow rate measured by these methods is not suitable for either real-time control or real-time measurement purposes because of the intermittent nature of the measurements. This paper describes a technique that can be used to find the minimum data [consisting of data from just 2.5% of the non-road transient cycle (NRTC)] to solve the problem concerning discontinuous data of fuel flow rate measured using an AVL 733S fuel meter for a medium or heavy-duty diesel engine using neural networks. Only torque and speed are used as the input parameters for the fuel flow rate prediction. Power density analysis is used to find the minimum amount of the data. The results show that the nonlinear autoregressive model with exogenous inputs could predict the particulate matter successfully with $R^{2}$ above 0.96 using 2.5% NRTC data with only torque and speed as inputs.
机译：柴油发动机面临的最严峻挑战之一是，在不损害排放法规的前提下，寻找提高燃油经济性的新技术。一种技术是精确控制空燃比，这需要测量瞬时燃油消耗。燃油费率的测量准确性和可重复性是成功控制空燃比和实时测量油耗的关键。体积和重量测量原理是用于测量内燃机中的燃料消耗的公知方法。但是，由于这些方法的测量是间歇性的，因此用这些方法测量的燃油流速既不适合实时控制，也不适合实时测量。本文介绍了一种技术，该技术可用于查找最小数据[仅由非道路瞬变周期（NRTC）的2.5％的数据组成]，以解决有关使用AVL 733S燃料测量的燃料流量不连续数据的问题使用神经网络的中型或重型柴油发动机仪表。仅扭矩和速度用作燃料流量预测的输入参数。功率密度分析用于查找最小数据量。结果表明，使用2.5％的NRTC数据（仅输入转矩和速度作为输入），具有外源输入的非线性自回归模型可以成功地预测$ R ^ {2} $高于0.96的颗粒物。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号