【作者】 张军朝；

【导师】 陈俊杰；

【作者基本信息】 太原理工大学 ， 计算机应用技术， 2012， 博士

【摘要】 照明控制系统是一项复杂的系统工程,其控制方式有多种,如手动、光控、钟控、集中回路控制、GPRS远程控制等。近年来各地路灯、景观灯的大量应用,耗费大量的电能,因此,其节能技术成为工程中的研究热点之一现阶段路灯、景观灯照明耗能大的主要原因体现在：控制系统监控不集中、实时监控能力不足；控制器路由算法网络性能低；太阳能路灯系统寿命短；光伏控制系统效率低等。针对上述问题,本文在前人工作的基础上进行研究,并取得了以下的创新性成果：1.提出了基于Zigbee与GPRS的市区照明单灯调光节能技术的完整方案。采用Zigbee与GPRS进行数据传输,实现了单灯智能化远程实时监控。基于有源PFC和变频功率控制技术,综合能耗降低40%以上,节能效果显著。2.提出了一种基于蚁群算法和路由算法相结合的单灯调光节能控制器动态路由算法,以使网络性能得到最佳的发挥。并建立了相应的数学模型,理论上验证了改进算法的有效性。该算法已被应用于TDFY3400路灯控制器的设计、生产中,在太原市滨河东路美化亮化景观照明工程中得到推广应用。3.提出一种改进的预约分帧时隙Aloha算法,采用在时间帧头部的微时隙上发送一个较短的随机数来对将要传输序列号的时隙进行预约,通过选择随机数的长度,降低阅读器发生碰撞的概率。4.提出了一种结合RED和HTH算法的逐段控制层次化组播拥寨控制机制。该机制采用逐跳拥塞控制的方法,每个路由器根据当前网络状态判断各自的拥塞状况,若发生拥塞,则按照某种概率进行丢弃,同时向上游路由器通报自己的拥塞情况。路由器收到其它路由器的拥塞状态消息后,采取相应的措施,增加或减少发往下游路由器的发送速率。5.提出一种基于超级电容光伏技术的市区路灯照明控制方案。通过将超级电容连接在光伏系统的前端,在光照不足的情况下,采取超级电容为蓄电池充电的控制策略,从而减少光照变化对充电条件的影响,保证了光伏电池获得最大功率跟踪,蓄电池的合理充放电要求得到满足。最后建立simulink/matlab仿真模型对该设计方法有效性进行了实验,结果表明该方法解决了传统方法的不足,有效地延长了太阳能路灯系统寿命,提高了其使用效率6.提出一种改进的RBF神经网络光伏电池建模方法。以日照、温度和负载电压作为RBF神经网络模型的输入,光伏电池输出功率作为输出,采用具有逼近任意复杂非线性函数能力的RBF神经网络对光伏电池进行建模,并采用粒子群算法对RBF神经网络进行优化,最后建立了光伏电池的动态响应模型。更多还原

【Abstract】 Lighting control system is a complex system engineering, and there are many means in its control such as hand, electric, clock control, centralized circuit control, GPRS remote control, etc. In recent years, with wide application of streetlamps and landscape lights, a large amount of power is used up, therefore, the energy-saving technology has become one of the hot research project.At this stage, the main easons of landscape lamp lighting energy consumption reflect that monitoring of control systems is not focused, and real-time monitoring ability is insufficient; Controller routing algorithm network performance is low; Solar street lamps system has short life; photovoltaic control systems are ineffective. Aimming at the above problems, on the basis of the previous work, this paper has made the following innovative achievements:1. Based on Zigbee and GPRS, the urban lighting single lamp light tone of energy-saving technology complete solutions is put forward. The Zigbee and GPRS for data transmission, realize the single lamp intelligent remote real-time monitor. Based on active PFC technology, comprehensive energy consumption is reduced by40%or more, remarkable energy saving.2. Putting forward SINGLE-NODE routing algorithm, based on a combination of ant colony algorithm and routing algorithm, so that the performance of the network get the best play. And the corresponding mathematical model was established, and in theory the validity of the algorithm is verified. The algorithm has been used in the design of the TDFY3400street light controller. In the River East Road of Taiyuan city, landscape lightings are widely used.3. Put forward an improved algorithm of reserved frame-slotted Aloha, the RFSA algorithm reserves slot for serial to be sent by sending a short random number in the micro slot of time frame header. Using special encoding, the reader can determine data reading status from the received random number to reserve corresponding slot.4. A kind of hop-to-hop controlled hierarchical multicast congestion control mechanism combining RED and HTH was presented. With this mechanism, each router determines its congestion status based on current network state. If congestion occurs, it discards with certain probability and report congestion status to upstream router. After router received congestion state from other routers, it will take appropriate measures to increase or decrease transmission rate to some downstream router. At the same time, packet discard number of each link was considered in congestion control to ensure fairness.5. Put forward a kind of the urban street lamp lighting control scheme based on super capacitor photovoltaic technology. Through the super capacitor in connection to the front of the photovoltaic system, insufficient in the illumination, take super capacitors for battery charging control strategy, and reduce the illumination change to charge the condition the influence, to ensure the photovoltaic cells to obtain the most power tracing, storage battery charging and discharging the reasonable demands are met. Finally, the Simulink/Matlab simulation model is established to prove validity of the method. The results show that the method can solve the shortage of the traditional methods, prolong the life of solar street lamps system, and so improve the efficiency.6. Put forward an improved modeling method of RBF neural network photovoltaic battery. With sunshine, temperature and load voltage as input values of RBF neural network model, output power of photovoltaic cells as output values, by adopting RBF neural network, which possess the approximate complex nonlinear function, the model of photovoltaic cells is built, and by using particle swarm algorithm, the RBF neural network is optimized, finally, the dynamic response model of the photovoltaic cells is built.更多还原