Study On Corrosion Behavior Of Commonly Used Metal For Heat Transfer Equipment Of Water Source Heat Pump

There are phenomenons of corrosion and scaling of metal which appear on the heat exchange equipment and pipelines of the water source heat pump system in the actual engineering. This article considers the problem of metal corrosion and heat exchange performance from the perspective of experiment and numerical simulation and the neural network prediction. The purpose is to provide the corrosion data of metal materials which are commonly used by the heat exchangers in the sewage and seawater heat pump system and offer feasible reference to the prediction method of corrosion rate.Firstly, the article explores influence of the heat exchangers on metal corrosion by controlling the factors which include medium flow rate, oxygen content, salinity and galvanic coupling. The experiment measuring the corrosion rate by designing the self-corrosion experiment of carbon steel 10#, copper T2, stainless steel 316 L and aluminum alloy LF21 and the galvanic corrosion experiment of three galvanic couples which include the carbon steel 10# and aluminum alloy LF21, the carbon steel 10# and copper T2, the copper T2 and stainless steel 316 L. The working medium is seawater and sewage, and the four kinds of working conditions include which mainly set according to the oxygen content and flow velocity. The results show that the fluxion of the working medium not only washing the deposited product, but also accelerating the arrival of dissolved oxygen to the surface of the metal. The fluxion enhances the corrosion to some content. In the continuous oxygen condition, the corrosion rate in the flow status of sea water group and sewage group increased 4.8% and 3.4% respectively compared with the corrosion rate in the non-flow status. The dissolved oxygen is the root cause of corrosion. Usually, the corrosion will enhance with the increase of oxygen content. In this experiment, in the flow condition, the corrosion rate in the continuous oxygen status of sea water group and sewage group increased12.5% and 8.5% respectively compared with the corrosion rate in the uncontinuous oxygen status. In the coupling with carbon steel 10#, the susceptibility of pitting of the aluminum alloy LF21 will extremely intensify. However, aluminum alloy LF21 is more corrosion-resistant than carbon steel 10# and copper T2 as long as it is not in a galvanic couple strictly. In the coupling with copper T2, the corrosion rate of carbon steel 10# is remarkablely accelerated.Observe and shoot the metallographic microstructure by SEM, the results show that the most common corrosion type is pitting corrosion. There are many honeycombed pitting holes on the surface of carbon steel 10# and the max corrosion depth is about 0.25 mm. There are also a few pitting holes appear on the surface of copper T2, the max corrosion depth is about 0.15 mm and the density of pitting holes is about 0.5/μm2. However, stainless steel 316 L suffer the slightest corrosion damage.Combined with the experimental data and using Fluent to carry out the numerical simulation of heat transfer and analyze the impacts to the heat exchange performance after corrosion. The results show that the surface heat transfer coefficient of the fluid and the tube wall when there are corrosion product on the surface reduces about 0.6% compared with that of the non-corrosion state. And when the corrosion product is removed, the surface heat transfer coefficient is about 1.7 times of the non-corrosion state.Using the self-organizing feature map network SOM classify the samples according to the similarity. And then apply the Exact RBF network to train the samples which are already classified by SOM network and predict the corrosion rate of metals in seawater respectively. Compare the results of RBF network with the results of BP network. The relative errors of the SOM-RBF composite pattern are all below 10% except the sample NO.30. Compared with the BP network, its relative error can reduce about 50%. This indicates that in the condition of the training sample is few and distribution is inhomogeneous, the prediction accuracy and calculation speed will significantly improved when use the SOM-RBF composite pattern.The research of this paper is in accord with the trend of the international development of material corrosion and offers feasible and basic research achievements on the establishment of standardized corrosion database, the data modeling, simulation of the corrosion process and experimental verification work and so on. It has the value on engineering application.