Study On Intelligent Control Of The Transformer Room Ventilation

In summer，electricity load is the highest year period, most of the transformers areapproaching or exceeding full load running. For the transformer of indoor type, itproduced a large amount of heat in high load operation, the cooling ventilation oftraditional transformer room could not meet its cool live. For this situation, this subjectmade a detailed study for the intelligent control of large power transformer roomventilation, and then it is achieved through engineering.First of all, the transformer top oil temperature prediction model is established,considering the influence of environment temperature on the transformer top oiltemperature, on the basis of the thermoelectricity analogy, add the environmentaltemperature in the form of a variable to the transformer top oil temperature predictionmodel. Will the model calculation value and the transformer top oil temperaturecomparative analysis of the actual value shows that the top oil temperature predictionmodel has high accuracy.Secondly, a detailed analysis of the heat transfer principle of transformer indoorventilation cooling, established a Mathematical model according to the actualparameters of transformer indoor ventilation cooling. Due to this mode is a system ofmulti-variableand delay, Used the fuzzy control to research for the controlled object, inorder to reduced the complexity of the design of fuzzy control strategy, theimplementation structure of control system adopted hierarchical multivariable fuzzycontroller. And through Matlab simulation software to emulated research for theventilation of transformer room fuzzy control system.Then, because the characteristic of time-varying and nonlinear of transformerroom ventilation system models, used the neural network predictive control algorithmfor transformer room ventilation system was studied. For predictive control rollingoptimization excessive-calculation problem, designing a compound control strategy method based on neural network inverse control and BP neural network predictivecontrol. The control system adjusts to the BP neural network predictive control whenthe controlled object is in a dynamic process; into a neural network inverse modelwhen the stable for output of controlled object, did not need to roll optimizationcalculation and reduced the amount of calculation. This method not only reduces theamount of rolling optimization calculations and can guarantee the stability of thecontrol system.In the theoretical study, for110KV transformer of one city’s transformer room,designed the intelligent control system of the transformer room. Designed the dualclosed-loop system of the transformer top oil temperature and indoor temperaturetransformer, used the neural network predictive control to achieve intelligent control ofroom ventilation rate. In the case of low-cost operation, through the double-loopsystem design, control transformer indoor temperature and power transformer top oiltemperature to ensure safe and stable operation of power transformers. In the case ofhigh temperature and high load, that the system actually works to reduce the indoortemperature, the transformer and improved efficiency of heat exchange of indoor air,such that the power transformer in the safe operating range of stability. Avoid hightemperature incited transformer failure.