350MW级多炉型能源站的配置和运行优化

根据350MW级用户热负荷持续时间图,结合链条炉、煤粉炉、循环流化床锅炉的成本与容量、效率与负荷率的关系,建立以投资成本和运行成本为目标的数学模型,依据能量平衡、燃烧稳定性和锅炉调节特性形成约束条件,对该数学模型采用等微增率算法和序列二次规划算法分别进行最优化求解,对于配置优化,等微增率算法给出一台容量为51MW的链条炉,一台容量为187MW煤粉炉,一台容量为130MW的循环流化床锅炉;序列二次规划算法给出一台容量为38MW的链条炉,一台容量为214MW煤粉炉,一台容量为108MW的循环流化床锅炉,对比能源站规范设计三台180MW煤粉炉的配置,等微增率算法优化配置的投资成本减小17.6%,序列二次规划算法优化配置的投资成本减小21.2%。针对该用户侧连续14天共336个时间点的热负荷数据进行运行优化,相比于规范设计方法,等微增率法的运行成本降低10.4%,序列二次规划算法的运行成本降低15.7%。

Configuration and Operation Optimization of 350MW Multi-furnace Energy Station

According to the heat load duration chart of 350MW users,and the relationship between the cost,capacity,efficiency and load rate of chain boiler,pulverized coal boiler and circulating fluidized bed boiler,a mathematical model is established aiming at the investment cost and operation cost.According to the constraints of energy balance,combustion stability and boiler regulation characteristics,the mathematical model is optimized by using the equal incremental rate algorithm and the sequence quadratic programming algorithm,respectively.For configuration optimization,the equal incremental rate algorithm gives a chain boiler with a capacity of 51MW,a pulverized coal boiler with a capacity of 187MW and a circulating fluidized bed boiler with a capacity of 130MW;The sequential quadratic programming algorithm gives a chain boiler with a capacity of 38 MW,a pulverized coal boiler with a capacity of 214MW and a circulating fluidized bed boiler with a capacity of 108MW.Comparing the configuration of three 180MW pulverized coal furnaces in the energy station design specification,the investment cost of the optimal configuration of equal incremental rate algorithm is reduced by 17.6%,and the investment cost of the optimal configuration of sequential quadratic programming algorithm is reduced by 21.2%.Compared with the standard design method,the operation cost of equal incremental rate algorithm is reduced by 10.4%,and the operation cost of sequential quadratic programming algorithm is reduced by 15.7%according to the heat load data of 336 time points on the user side for 14 consecutive days.