近年来人类保护环境意识逐渐增强，能源危机也深深地引发人类担忧，在此背景下世界各国都在积极发展绿色能源，尤其国内近几年雾霾肆虐，引发了民众对重污染行业，包括发电行业排放的关注。相比较其他类型的绿色能源，风力发电技术更为成熟，基础设施覆盖广，成本低廉。许多研究资料[1-2]表明，风能将会是未来世界能源结构中重要的一部分。为了高效利用风力机海量的运行数据，并适应数据量不断扩大的需求，本文选择采用云平台来搭建风力机状态监测和故障诊断系统。首先研究了市面上常见的云平台方案，然后选择技术开源、适用性强的Cloudfoundry进行云平台的搭建，随后研究并制定了针对私有云服务器的搭建方案。针对风电场维修的实际需求，对风力机状态监测和故障诊断系统进行功能模块设计和工作流程设计。系统包括样本数据库，三大功能模块和UI界面脚本。样本数据库来自风力机实验台，包含58个参数，共12种状态特征。状态监测模块图形化地显示了所有参数。状态监测模块包含了风力机运行参数的列表，并可以通过动态图直观地追踪各运行参数的发展趋势。故障预警模块在传统的固定阈值报警方案上，创新性地提出了建立以各风速下各参数正常运行范围为阈值的动态越限报警方法，大大提高了故障预警功能的可用性和准确性。故障诊断模块则是采用了SVM算法，在对风力机运行参数进行预警的情况下，进一步对故障进行准确分类，并指出异常信号，对工作人员进行维修判断提供依据。准确率高达99.08%。最终，本文在自建的云平台上对该系统进行了测试，使用者可以通过网页浏览器在局域网内登录网址进入到系统中，查看风力机运行状态及故障诊断结果。

Accompanied by the intensification of the energy crisis and the enhancement of human environmental awareness, the world is actively developing renewable green energy. Wind power has an advantage in terms of technology maturity, infrastructure coverage, and cost control. Wind energy structure occupies an extremely important position in the future world energy.In order to use the wind turbine running data effectively, and to adapt to the demand for ever-expanding data , choose to use the cloud platform to deploy wind turbine condition monitoring and fault diagnosis system. First of all, study the common cloud platform on the market, then select an open source, strong applicability one called Cloudfoundry, followed by research and development for private cloud server deployment plan.According to the demand of the actual wind farm maintenance, the function module design and working process design of the wind turbine condition monitoring and fault diagnosis system are carried out. The system includes a sample database, three function modules and a UI interface script.The sample database is from the wind turbine test bed and contains 58 parameters for a total of 12 state features. The status monitoring module graphically shows all the parameters.In the traditional fixed threshold alarm scheme, the fault early warning module is proposed to establish the dynamic overrun alarm method with the normal operating range of each parameter under the wind speed, which greatly improves the usability and accuracy of the fault early warning function.Fault diagnosis module is the use of SVM algorithm, the wind turbine operating parameters in the case of early warning, and further accurate classification of the fault, and pointed out that the abnormal signal, the staff to provide a basis for maintenance judgments. The accuracy rate of up to 99.08%.