Study On The Use Of Gray System For The Water Quality Forecast Of Coke-Plant Wastewater

The coke-plant wastewater is recognized as a hard-biodegradable industrialwaste water, sincce it contains a large number of polycyclic aromatichydrocarbons and heterocyclic organic category. And it is quite difficult forthese organics to be degraded by micro-organisms in aerobic conditions.The coke-plant wastewater thus became one of the most difficult problems inthe filed of China's Water Pollution Control. A huge quantity of Coke-planwastwater is disposed into the receiving water without safe disposals,thatbrings tremendous pollution to the receiving water.In order to prevent the disposal of exceed quota coke-plan wastewaterdischarged by treatment plants and not to bad impact on quality of the waterenvironment, the achievement of on-line control and real-time adjustmentsbecomes necessary for coke-plant wastewater treatment plant. And the keypoint to achive this goal is the establishment of the mathematical model ofeffluent quality.Grey system theory focuses on the"small sample", "poor information"uncertainty problems which are hardly resolved by probability/ statistics andfuzzy math. It explores the real rules of things' movements by sequenceoperator. It's characterized by " Modeling with less data", and suitable forwater quality forecasts.Taking Panzhihua Iron and Steel Group coke-plant wastewater treatmentstation as an example, making use of MATLAB as a numerical computingplatforms,this paper, by applying the grey system theory, constructs a GM (1,1)forecast model of the quality of effluent disposed by the sewage treatmentplant.After testing the model, the model predictes an average relative error of 1.85percent, which satisfies the requirements of the water quality forecast.The establishment of this model is conducive to the achievement of computercontrol and cost savings of monitoring for sewage treatment stations. Itprovides a means to support decision-making for real-time control in sewage treatment plants.And also it provides a feasible way for other similar intelligent Control in sewage treatment plants. It offers experiences for relevant researchs.