基于实时变量的重介质分选过程参数预测研究

Dense Medium Coal Separation acts as the main coal preparation method in China and the level of separation efficiency has a crucial effect on coal preparation production. At present, the control lag problem is prominent because the control of medium density cannot change synchronously with the feed raw coal quality, leading to a low separation accuracy and efficiency. The project researches on principles and algorithms of real-time prediction of process parameter such as separation density according to changes of feed raw coal quality, building homologous mathematical models, and improving controlling level of dense medium separation process and separation efficiency. On condition that real-time variables be introduced into traditional prediction methods, the ash data that have been tested by online ash testing instrument can repalce the data be obtained through testing in laboratory after being corrected, then act as the basic data to pridict density compositions of raw coal and products and to find the optimal mathematical models and algorithms that can fit washability curves and distribution curve. As a result, process parameters such as separation density, product yield and quantity efficiency can be predicted and can meet the requirements. Besides, prediction models can be built, verified and optimized experimentally, which provides reference basis for accurate real-time control during process of dense medium coal preparation.

重介选煤是我国的主要选煤方法，其分选效率的高低对选煤生产影响巨大，目前由于介质密度的控制不能随入洗原煤煤质同步变化，控制滞后问题较突出，导致分选精度和效率较低。本项目研究根据入洗原煤煤质变化实时预测分选密度等工艺参数的工作机理和算法，并建立相应数学模型，提高重介分选过程的控制水平和分选效率。 通过在传统预测方法中引入实时变量，将在线测灰仪测量的灰分通过算法校准，作为可以代替化验灰分的指标，以此为基准预测原煤密度组成和产品密度组成，寻找可实时拟合可选性曲线和分配曲线的拟合模型和最优拟合算法，根据要求的产品指标预测分选密度、产品产率、数量效率等工艺参数，建立预测模型并通过实验对其进行实验验证和优化，为重介选煤的实时精确控制提供参考依据。

针对在当前重介质选煤过程中，介质密度控制是依赖于通过精煤灰分反馈的，不能随入洗原煤煤质同步变化，控制滞后导致分选精度和效率较低，且密度的调整主要依赖于密控员的经验，洗选效果受人为影响较大的问题，本项目对重介质选煤生产过程中工艺参数的的在线预测机理进行了研究。主要内容包括：研究通过box-plot分析法，消除原煤测灰仪产生的极端异常数据；采用最近10min测量灰分值平均法，消除原煤灰分波动过大的不利影响；建立了原煤测灰仪灰分与快灰数据的线性关系模型，利用快灰数据对测量仪数据进行校准，作为预测工艺参数的灰分基准；研究了原煤灰分与各密度级产率的关系模型，根据原煤灰分预测其密度组成数据；通过月综合数据或系统检查数据，可建立生产系统的分配曲线模型，输入当前原煤浮沉数据后，预测得到精煤产率、灰分、分选密度、数量效率等工艺指标；通过实验分析得出了实际分选密度和循环介质密度之间的线性关系模型，利用预测的实际分选密度计算得到循环介质密度，用于对重介分选系统进行控制调节；根据在现场实验的数据分析，预测的生产工艺指标基本符合实际，个别指标稍有偏差。本项目的研究为提高重介分选过程的控制水平提供了理论和方法支持，对提高重介分选效率，减少宝贵煤炭资源的损失，有非常积极的作用。