一种多子群协进化的粒子群算法同步综合换热网络

针对同时存在整型变量及连续变量的换热网络优化问题,提出一种多子群协进化的粒子群算法。为了增强粒子群算法的全局搜索能力,将种群按精英个体、一般个体、较差个体划分为3个子群,针对每个子群的粒子进化状态提出不同的学习算子,用于丰富粒子的进化方式,增加种群多样性;同时建立协进化机制,动态地更新子群,以实现粒子之间的良性竞争,更好地引导粒子进化。采用结构优化策略处理整型变量,并与多子群协进化的粒子群算法结合,实现了连续变量与整型变量的同步优化。通过两个优化实例验证算法的性能,优化结果表明了新方法的有效性。     

采用个体进化能力实时评价和强制更新策略的换热网络优化方法

针对启发式方法在优化换热网络时出现个体进化停滞的现象,提出了一种采用个体进化能力实时评价和强制更新策略的进化算法。首先,建立个体进化能力评价机制,实时考察每个个体进化过程中的性能变化;其次,当个体进化能力退化时,建立强制更新策略,通过随机抽取当前结构中的换热单元,并给予其换热量以随机扰动,提高个体的结构变异能力,促进其整型变量的全局优化,改变个体原先的优化路径从而寻得更优的结构。将该算法应用于无分流换热网络实例中,取得了优于现有文献的结果,表明了该算法具有较强的全局搜索能力。     

基于结构性能连续性的模拟退火算法优化换热网络

换热网络优化是典型的混合整数非线性问题,其整型变量的组合情况(换热网络的结构)对于其优化的走向以及局部最优解的质量具有至关重要的作用.指出换热网络的结构与性能存在连续性,并由此提出以连续性指导整型变量最优搜索的换热网络模拟退火算法,对换热网络的结构进行最优搜索,降低换热网络的综合费用.通过对实际算例的优化,将换热网络的综合费用由1599229美元降至1581165美元.结果表明:换热网络结构与性能的连续性能指导整型变量的优化走向,保证模拟退火算法在换热网络优化过程中的可靠性.     

基于动态拓扑结构的分布式Memetic差分进化算法同步综合换热网络

针对标准DE(差分进化)算法在优化换热网络时出现的局部搜索能力弱、易陷入局部最优等问题,本文建立了一种基于动态拓扑结构的分布式Memetic差分进化算法,同步综合换热网络。首先,在子种群内部采用基于欧拉距离的动态拓扑结构,子种群之间采用冯诺依曼拓扑结构,有效地加快了个体之间的信息交流,保持种群多样性,扩大搜索范围。继之,结合Memetic算法思想,将Hooke-Jeeves算法作为局部搜索策略,增强算法局部搜索能力。同时,对于局部搜索获得的新解,提出了一种协作学习机制,平衡算法的全局寻优与快速收敛能力。最后,为处理整型变量,提出了两条简单有效的整型变量优化策略,使算法实现了连续变量与整型变量的同步优化。选取两个经典算例验证了算法的有效性。算法应用于算例一,相对于现有文献的最优值,本文所得结构的费用值下降了1 783$/a,表明算法的性能优于标准DE算法以及其它改进版本的DE算法。算法应用于算例二,相对于现有文献的最优值,本文所得结构的费用值下降了1 209$/a,表明算法能够有效地处理大规模换热网络问题,具有很强的鲁棒性。     

基于连续性的换热网络整型变量优化策略

换热网络优化是典型的混合整数非线性问题,其整型变量(换热网络的结构)的组合情况对于其优化的走向以及局部最优解的质量具有至关重要的作用.分析了换热网络结构与性能之间的连续性特征;利用最速下降法对换热网络整形变量进行初步优化以确定合适的换热器数目,生成一系列连续变化的换热网络结构;以结构序列作为优化变量,借助换热网络性能连续性原理,将多维整型变量转换为一维连续变量,对整型变量进行启发式的连续性搜索.通过具体算例求解,该方法帮助优化过程跳出了局部最优的陷阱,并得到了较好的优化结果.     

基于改进参数的粒子群算法的换热网络优化

对于换热网络综合优化问题,粒子群算法能有效解决其容易陷入局部最优和无法收敛到全局最优的局限性。标准粒子群算法具有较强的随机性,可调节参数较少,不同的参数配置对算法的优化效果有显著影响。在分析粒子群算法中各参数特点的基础上,通过合理调整参数,使该算法在连续变量优化过程中具有更好的全局收敛性能。采用4股流体的小规模换热网络算例进行验证,优化结果表明,改进参数后的粒子群算法对计算换热网络综合问题有效。     

采用结构进化增强策略的RWCE算法优化换热网络

针对于启发式算法应用于换热网络优化时,后期由于种群多样性消失或其他原因导致年综合费用难以进一步下降的问题,提出了一种结构进化增强策略。该策略在一般启发式算法的整型变量优化中,通过将换热单元的生成与消去分开处理,先以一定概率随机地在网络中生成若干换热单元,再在连续变量优化的过程中实现换热单元的消去,提升网络结构变异能力。最后,将该策略与强制进化随机游走算法(random walk algorithm with compulsive evolution,RWCE)相结合形成一种新的混合算法(ESE-RWCE)。算例研究表明,ESE-RWCE算法相比于RWCE算法实现了全局搜索性能的提升。     

基于分级超结构的换热网络改造优化

建立了基于分级超结构模型的换热网络改造同步优化数学模型。该模型不依赖于夹点约束,不需要预先给定最小传热温差,能够有效权衡改造投资费用与运行费用之间的关系。改造投资费用中考虑了现有换热器的重新配置费用、现有换热器新增传热面积费用和新增换热器费用,符合工程实际要求。针对换热网络改造优化数学模型具有不连续和非线性的特点,数学模型的求解采用双层优化策略,其中表示网络结构调整的离散变量优化利用遗传算法,而表示操作参数的连续变量优化利用粒子群算法优化。2个不同规模的换热网络改造算例用于验证所提出方法的有效性。     

一种流股分流动态调节的换热网络优化策略

现有的节点非结构模型(Node-wise non-structural superstructure, NW-NSS)在优化换热网络时需要预设固定的节点分流数量,难以满足结构进化过程对于求解空间和计算效率的需求,易造成换热单元生成空间受限,影响算法全局优化性能。本文提出一种流股分流动态调节策略,该策略基于实时结构的整型变量分布信息,动态增加结构进化所需的必要分流并减少无效结构对优化的阻碍,辅助算法以更高的效率跳出局部极值,提升优化质量。将策略应用于16SP、20SP算例,分别得到年综合费用为6 653 940和1 711 886$/a的最优换热网络结构,较文献最优结果降低了3 140和3 202$/a。     

考虑压降的换热网络优化设计

提出考虑压降的有分流换热网络合成的优化策略,在分级超结构的基础上,建立换热网络同步综合的MINLP数学模型,以年度总费用最小为优化目标,采用改进粒子群算法对模型求解,并通过两个算例验证了模型和方法的有效性。     

A matrix approach for steady-state simulation of heat exchanger networks

A steady-state simulation scheme for heat exchanger networks is presented. This proposal is based on a matrix approach composed by linear equations where network connectivity is addressed in the model explicitly using graph theory. The structure of the algorithm is flexible allowing the inclusion of the evaluation of heat transfer coefficients, physical properties variation in relation to temperature and pseudo-stationary simulations. The application of the proposed scheme is illustrated by the simulation of a heat exchanger network during a certain period considering the fouling rate.     

换热网络中大步长与接受差解的关联分析及改进策略

RWCE算法优化换热网络时,较大的优化步长能够在接受差解的作用下促进结构变异,从而有效地跳出局部最优解。但持续的大步长无法保留结构变异后新生成的、换热量较小的换热单元,影响结构进化效率。基于大步长与接受差解的关联性,提出了一种大步长与接受差解联动策略,将接受差解和采用大步长同频,仅当接受差解时才使用大步长,并必定接受大步长作用下的新结构,保证大步长对结构变异的促进作用,解决大步长作用下新生成单元难以保留的问题,全面提升结构进化效率。将大步长与接受差解联动策略应用于20sp和15sp算例,分别得到结果为1 395 389和1 511 551$/a的最优换热网络结构,性能较基础RWCE算法提升7.06%和1.69%,验证了该策略对提升算法结构进化效率的有效性。     

换热网络全局优化的多维峰谷轮换法

在峰谷轮换法的基础上提出了利用多维峰谷轮换法对以年综合费用最少为目标函数的无分流换热网络进行优化.在换热网络变量寻优过程中,首先将目标函数极小化得到一个局部极小值点,然后沿着多个方向进行变量搜索,找出最先跳出局部极小值点的变量组合,使得跳出局部极小值的搜索效率高于单变量的搜索效率,并可寻找到更好的换热网络结构,将极小化与跳出局部极小值点过程交替进行直至求得全局最优解.通过算例验证了该方法的可行性,且可找到比其它方法的优化结果更优的换热网络结构.     

The application of field synergy number in shell-and-tube heat exchanger optimization design

In the present work the field synergy principle is applied to the optimization design of the shell-and-tube heat exchanger with segmental baffles. The field synergy number which is defined as the indicator of the synergy between the velocity field and the heat flow is taken as the objective function. The genetic algorithm is employed to solve the heat exchanger optimization problems with multiple design variables. The field synergy number maximization approach for heat exchanger optimization design is thus formulated. In comparison with the initial design, the optimal design leads to a significant cost cut on the one hand and an improvement of the heat exchanger performance on the other hand. The comparison with the traditional heat exchanger optimization design approach with the total cost as the objective function shows that the field synergy number maximization approach is more advantageous.     

Multiobjective thermo‐economic and thermodynamics optimization of a plate–fin heat exchanger

In the present work, a multiobjective heat transfer search (MOHTS) algorithm is proposed and investigated for thermo‐economic and thermodynamic optimization of a plate–fin heat exchanger (PFHX). Heat exchanger effectiveness and total annual cost (TAC) are considered as thermo‐economic objective functions. Similarly, entropy generation rate and heat exchanger effectiveness are considered as thermodynamic objective functions. Six design variables including flow length of cold and hot streams, no flow length, fin height, fin pitch, and fin offset length are considered as decision variables. Effectiveness and accuracy of the proposed algorithm are evaluated by analyzing application examples of a PFHX. The results obtained using the proposed algorithm for thermo‐economic considerations are compared with the available results of NSGA‐II and TLBO in the literature. Results show that 3.56% to 10.29% reductions in TAC with 0.48% to 0.81% higher effectiveness are observed using the proposed approach compared to TLBO and NSGA‐II approaches. Additionally, the distribution of each design variable in its allowable range is also shown for thermo‐economic consideration to identify the level of conflict on objective functions. The sensitivity analyses of design variables on the objective functions value are also performed in detail.     

Three-dimensional numerical optimization of a manifold microchannel heat sink

A three-dimensional analysis procedure for the thermal performance of a manifold microchannel heat sink has been developed and applied to optimize the heat-sink design. The system of fully elliptic equations, that govern the flow and thermal fields, are solved by a SIMPLE-type finite volume method, while the optimal geometric shape is traced by a steepest descent technique. For a given pumping power, the optimal design variables that minimize the thermal resistance are obtained iteratively. The procedure is robust and the optimal state is reached within six global iterations. Comparing with the comparable traditional microchannel heat sink, the thermal resistance is reduced by more than a half while the temperature uniformity on the heated wall is improved by tenfold. The sensitivity of the thermal performance on each design variable is also examined and presented in the paper. Among various design variables, the channel width and depth are more crucial than others to the heat-sink performance. The optimal dimensions and corresponding thermal resistance have a power-law dependence on the pumping power.     

线性拟合换热网络性能与进口温度的关系

换热网络问题属于典型的混合整数非线性问题,有严重的非凸、非线性特性,且影响换热网络性能的因素众多.主要研究了换热网络进口温度与性能之间的关系.给定一换热网络固定结构,同时在固定面积的条件下,变动每股流体的进口温度,分析流体的进口温度与换热网络性能之间的关系,得出线性拟合的适用性.然后设计正交试验,拟合出性能与进口温度的线性模型,并验证该模型的准确性.最后根据拟合模型,从众多流股中优选冷、热流体组合成目标换热网络.     

Optimal design analysis of a tubular heat exchanger network with extended surfaces using multi-objective constructal optimization

The present work aims to investigate the influence of extended surfaces (fins) on the multi-objective optimization of a tubular heat exchanger network (THEN). An increase in the heat transfer area using various extended surfaces (fins) to enhance the performance of the heat exchanger was used while considering the effectiveness and total heat transfer area as two objective functions. In addition to the simulation of simple fins, a new set of fins, called constructal fins, was designed based on the constructal theory. Tubular heat exchanger network design parameters were chosen as optimization variables, and optimization results were achieved in such a way as to enhance the effectiveness and decrease the total heat transfer area. The results show the importance of constructal fins in improving the objective functions of heat exchangers. For instance, the simple fins case enhances the effectiveness by up to 5.3% compared to that without fins (usual heat exchanger) while using constructal fins, in addition to the 7% increment of effectiveness, reduces the total heat transfer area by 9.47%. In order to optimize the heat exchanger, the heat transfer rate and cold fluid temperature must increase, and at the same time, the hot exiting fluid temperature should decrease at the same constant total heat transfer area, which is higher in the constructal fins case. Finally, optimized design variables were studied for different cases, and the effects of various fins were reported.