溶解氧预报模型及误差校正

针对污水处理过程中，由于溶氧仪表面生成生物氧化膜，产生“钝化”现象，造成了溶氧值测量误差，以至影响生产；又因为需要频繁地清理溶氧仪，造成了人力、物力、财力的浪费，降低生产效率等问题，利用人工神经元网络建立了溶解氧预报模型。基于此模型，给出了传感器误差校正的具体方法。当误差超出给定界限时，给出故障报警。现场应用表明，该方法在一定程度上延长了溶氧仪清理的间隔时间，节约了生产成本，提高了生产效率，收到了较好的预报效果。

Dissolved Oxygen Prediction Model and its Error Revision

During the process of sewage disposal, the surface of dissolved oxygen instrument will form a layer of biological oxidation membrane, which makes the dissolved oxygen instrument insensitive and causes measure errors. So the dissolved oxygen instrument must be cleamed in time. This leads to waste a lot of manpower, material resources, financial ability, and to reduce the productivity. In order to solve this problem, the method of intelligent neural network is proposed for dissolved oxygen predictive model.Based on this model, the method of sensor error revising is presented, and fault alarm is applied when error is over the threshold. This method prolongs the clearing time, saves the productive cost and improves the productivity. The practical application indicates that the proposed method is effective.