【作者】 杨洁明；

【导师】 熊诗波；

【作者基本信息】 太原理工大学 ， 机械电子工程， 2004， 博士

【摘要】 跳汰机作为一种通用、高效、可靠的选煤设备，广泛应用于选煤厂，在我国，大约60％的入洗原煤采用跳汰分选工艺，跳汰选煤在我国洁净煤战略中占有相当重要的地位。 但与此不相适应的是，目前我国跳汰选煤的自动化水平仍处于十分落后的状态：自动排料系统靠人工设定重物料层厚度期望值，控制算法基本停留在简单的逻辑控制或常规PID控制上，达不到床层稳定的控制要求；床层的分层过程控制系统虽然采用了数字风阀，但风水制度的调节完全由跳汰司机根据经验进行，不仅工作量大，而且人为因素影响大，无法做到稳定、准确的调节。这导致了物料不能按密度良好分层，不同密度物料间的错配现象严重和大量精煤流失或精煤污染，严重影响了选煤生产的效率和效益。 跳汰过程是一个典型的机理复杂、影响因素多、变量间相互耦合、时变、严重非线性的动态过程，产品质量及分选效率与床层厚度和风阀参数之间无法建立确切的数学模型，传统控制理论和方法难以对此类系统进行有效的控制。近年来，以专家系统、模糊控制、神经网络为代表的人工智能技术被引入复杂工业过程控制领域，同时也为跳汰过程的自动控制提供了一条有效途径。 本文研究用智能控制方式实现跳汰过程自动化的策略和实现方法。 首先对影响跳汰分选效果的主要因素做了深入分析，得出床层的密度分布与产品质量密切相关，其正态分布的标准离差可反映出床层中的错配物含量的结论，提出跳汰过程的自动控制应以优化分层状态为目标的控制策略。作者分析了脉动水流特性、松散度和排料过程对分层状态的影响，为明确过程控制中的状态变量与控制要素奠定了基础。 某些重要位置上沿床层深度方向的密度分布信息对跳汰过程控制至 摘要关重要，为此本文研究了获取床层状态信息的Y射线密度检测原理与技术，研制了以cs’37为射源的Y射线探测器及其后续处理电路，解决了计数可靠性、抗噪声干扰等关键问题，并应用于现场跳汰机。试验表明，用Y射线密度探测器能够真实地反映床层在不同深度的密度值，该信息与其它状态信息相结合，可以较准确地反映出跳汰机床层按密度分层的状态。 跳汰过程中，可定量信息与不确定(模糊)信息共存，且多种因素间互相关联，一些关键参数无法用数学公式表征。对此作者研究了基于模糊推理的状态识别方法，提出了用密实期和松散期的密度差值表征床层松散度的方法;提出了基于密度均值、标准离差、原煤灰分及床层厚度等信息的跳汰机分层状态模糊评判方法。为了研究各状态参数之间的关系，设计了现场数据采集与记录系统，对床层在不同深度、不同时间段(密实期与松散期)的密度值、床层厚度、原煤灰分、给料速度、风阀操作参数等作了长时间记录，并辅以定时人工筛分浮沉实验。通过对大量现场数据记录的分析表明，跳汰机分层状态模糊评判方法具有较高的识别精度。 在上述研究的基础上，进行了模拟跳汰机的实验室实验，从中得到了风阀操作制度对脉动水流特性的调节关系、脉动水流特性对床层松散状态的调节作用、床层松散度与分层状态的关系，从而建立了风阀操作制度与分层状态之间的关系链，为控制规则的制定提供了可靠依据。 本文根据跳汰过程的特点，提出了专家控制系统的构成框架，研究了分层过程控制专家系统的知识获取、知识的表示与检索方法及推理机制，并根据分层状态、床层松散状态、给料速度信息和现场运行记录及定时人工筛分浮沉结果进行推理，建立了专家系统规则库，给出了具体的风阀操作参数的调节方法。同时从跳汰机综合控制的角度，对排料过程模糊控制方法做了改进性研究，提出了床层厚度期望值的自动修正方法;在此基础上，设计并实现了跳汰机总体协调控制系统。 现场运行情况表明:基于专家系统和模糊控制的跳汰机自动控制策略是可行的。尽管该系统还需不断完善，但对稳定精煤灰分、减少精煤太原理工大学博士研究生学位论文流失、提高产品质量与分选效率起到了重要作用，取得了很好的经济效 、，盆。更多还原

【Abstract】 Jig is utilized widely in coal preparation plant as a universal, efficient and reliable coal preparation equipment. About 60 percent of raw coal is cleaned by jigging process in coal preparation plant. Jigging plays an important role in coal cleaning in our country.But the automation of jigs falls behind in china. In the automatic discharge system of jigging, the expectation value of bed depth depends on the manual work, and the simple logical control or conventional PID algorithm used in discharging can not satisfies the bed stabilizing. Although the digital air valve is adopted in jigs, the adjustment of air-water rule depends on the experiences of jig operators but automatic system. The parameters can not be adjusted stably and accurately due to the excessive human factors. The satisfying stratification according to density distribution can not be obtained, and the serious misplacing of different density material causes large number of cleaning coal lost or contaminated, which affects the efficiency and benefits of coal cleaning seriously.Multi-variable coupling each other, time-variant, heavy nonlinear characteristic are contained in jigging process. The precise mathematical model between product quality or separation efficiency and operational parameters can not be built because the process mechanism description is too complex and a large of uncertain factors exist. The effective control can not be achieved by conventional control theory and methods. Recently, artificial intelligent technology, such as expert system, fuzzy control and neural network, is introduced to complex process control, which provides an effective method of the jigging automation.This article studies the strategy and realized method of jigging process automation with intelligent control theory.The main influencing factors of jigging separation are analyzed deeply. The analysis shows that the distribution characteristic is closely relative toproduct quality. The misplaced material content in bed can be reflected by the standard deviation of normal distribution. The author puts forward that the target of jigging process control should be the optimum stratification state and the bed stratification state is affected by pulsing water characteristic, mobility of the jig bed and discharging process.The bed stratification state is a key factor in jigging process control. Y radial density detecting principle and technology is studied in the article, y radial sensor and subsequent circuit is developed. The key problems, such as the reliability of counting and antinoise are resolved, and this detector is applied in practice. Y radial source is Cs . Experiment shows that the Y radial density detector can measure the coal density in different bed depth precisely. Bed status can be reflected by combining density information with other state information.In jigging process, quantitative information and uncertain information (fuzzy information) appears at the same time and multi-factors associate each other. Some key factors cannot be described by mathematic model. The article studies the status identification based on fuzzy reasoning, and puts forward that bed mobility is expressed by density difference between compact period and mobility period, also the fuzzy judgment method based on density mean value, standard deviation, raw coal ash content and bed depth information. In order to study the relationship of all kinds of parameters, data acquisition and recording system is designed. Operational parameters, such as density, bed depth, raw coal ash content, feed velocity, air valve etc, are recorded for a long time. Simultaneously, the timing artificial float and sink test is carried out. The high identification precision of bed stratification status is acquired through the analysis of recording data on site using fuzzy judgment method.By large numbers of experiments, the adjustment relationship between air valve operational rule and pulsing water characteristic, the pulsing water characteristic and mobility of bed, also the mobility of be更多还原