重介质洗煤悬浮液密度和液位的控制精度直接关系到洗煤效果。目前，洗煤厂通常只关注密度的调节；然而，随着生产过程持续进行，液位不断下降，严重时甚至无法实现对密度的稳定控制。因此，对重介质洗煤悬浮液密度和液位协调控制的研究十分必要。本文对密度和液位进行协调控制，主要工作如下： 1）以石圪台和漳村洗煤厂为对象，通过分析清水阀、加介阀和分流阀等阀门的开度变化对重介质洗煤悬浮液密度和液位的影响，建立了密度和液位的带参数传递函数模型；结合现场实验数据拟合曲线进行参数辨识，最终确定密度和液位的传递函数模型。其中，石圪台洗煤厂是受加水和加介影响的两入两出模型；漳村洗煤厂是受加水、加介和分流影响的三入两出模型。 2）对两入两出密度液位模型，采用解耦控制；对三入两出密度液位模型采用前馈补偿和解耦控制，具体地，在解耦控制的基础上，将分流阀的开度变化当作扰动，对其采用前馈补偿。 3） 为改善密度和液位控制的动态性能，采用基于状态空间的预测控制算法。首先，通过子空间辨识中的MOESP方法辨识获得系统状态空间；在此基础上，采用有限时域二次性能指标构造优化问题；然后，使用Kalman滤波预测更新状态；最后，对比解耦控制仿真结果，在调节时间、超调量和鲁棒性等方面预测控制更优。 本论文对重介质洗煤悬浮液密度和液位的协调控制研究，实现了对悬浮液密度和液位的有效调节，解决了实际生产过程中悬浮液密度难以长期保持稳定的问题。

The control precision of the suspension density and liquid level in dense medium coal preparation is directly related to the coal preparation effect. At present, the coal washery usually only pays attention to adjust the suspension density. However, as the production process continues, the liquid level continues to decrease, and the stable control of density cannot be achieved when serious conditions occur. Therefore, it is necessary to study the coordinated control of the suspension density and liquid level in dense media coal preparation. In this dissertation, the coordinated control of the density and liquid level is studied. The main work is as follows: 1) For the Shigetai and Zhangcun Coal Washery, by analyzing the influence of the opening change of valves, such as the clean water valve, the add medium valve and the split-flow valve, on the suspension density and liquid level in the dense medium coal preparation, the transfer function model of the density and liquid level with the parameter is established. The transfer function model for density and liquid level is finally determined based on the parameters are identified by the combination of experimental data fitting curves. In the established models, Shigetai Coal Washery is a two-input and two-output model affected by adding water and medium, Zhangcun Coal Washery is a three-input and two-output model affected by adding water, medium and split flow. 2) Decoupling control is applied to the two-input and two-output model. Based on the decoupling control, the feed-forward is applied to the three-input and two-output model, with the change of opening the split-flow valve, which is treated as a disturbance and feed-forward compensation is adopted. 3) In order to improve the dynamic control performance of suspension density and liquid level, a predictive control algorithm based on the state-space model is used in this dissertation. Firstly, the MOESP in subspace identification is used to identify the state-space model; and then, the optimization problem is formulated with the quadratic performance index in finite time domain, and the updated state is got by Kalman filtering; finally, the simulation results are compared and analyzed to conclude that the predictive control performance, such as regulation time, overshoot, and robustness, is better than that of the decoupling control. The study on the coordinated control of suspension density and liquid level in the dense medium coal preparation, in this dissertation, have realized the effective adjustment of suspension density and liquid level, and solved the problem in the actual production process that the suspension density is difficult to remain stable for a long time.