混合选别浓密过程双速率智能切换控制

赤铁矿混合选别浓密过程是以底流矿浆泵频率为输入，以底流矿浆流量为内环输出，以底流矿浆浓度为外环输出的强非线性串级工业过程.由于受到频繁的浮选过程产生的中矿矿浆和污水的随机干扰，底流矿浆浓度外环和流量内环始终处于动态变化之中，控制器积分作用失效，内外环相互影响，使被控系统的动态性能变坏，底流矿浆浓度与流量超出工艺规定的控制目标的范围，甚至产生谐振.本文针对上述问题利用提升技术建立基于内环流量闭环动态模型的浓度外环动态模型，将基于未建模动态补偿驱动的一步最优PI控制和基于模糊推理与规则推理的切换控制相结合，提出了由浓度外环控制和流量内环控制组成的混合选别浓密过程的双速率智能切换控制算法，建立了由机理主模型和神经网络补偿模型组成的混合选别浓密过程动态模型.所提算法通过混合选别浓密过程的半实物仿真实验结果表明本文所提控制方法的有效性.

Dual-rate Intelligent Switching Control for Mixed Separation Thickening Process

The mixed separation thickening process (MSTP) of hematite beneficiation is a strong nonlinear cascade process with the frequency of underflow slurry pump as the input, the slurry flow-rate as the inner loop output and the concentration as the outer loop output. During its operation, some large and frequent random disturbances generated from the flotation middling and sewage will continuously cause dynamic changes of slurry concentration and slurry flow-rate, which will cause failure of controller integration. The influence between the outer and inner loops will deteriorate the dynamic performance of the controlled system and even cause resonance. To deal with the problem, lifting technique to introduce the dynamic characteristics of the inner closed-loop control system into the outer dynamic model of slurry concentration. We put forward a double-rate intelligent switching control algorithm for MSTP combined with one-step optimal PI control, unmodeled dynamics compensation control and fuzzy switching control. An MSTP dynamic model is established using the master model with mechanism and compensation model with neural network. Simulation experiment on the hardware-in-the-loop simulation system of MSTP proves the effectiveness of our method.