| Elevator group control technology is widely used in large public buildings to optimize elevator scheduling,and it has achieved remarkable results in shortening user waiting time,reducing elevator operation energy consumption and mechanical loss.However,in general residential buildings,there are often few elevators in the unit.Space constraints and cost constraints determine the need for simple and cost-effective technologies and solutions.Residential building codes and conventions formed over the years are generally not designed for group control of elevators.system.In recent years,a large number of high-rise residential buildings have emerged in cities.Among a large number of residential buildings that share multiple elevators through fire-fighting corridors,the actual needs such as energy saving and consumption reduction,and improvement of owner’s elevator experience urgently need corresponding elevator group control technology.Combining the travel characteristics of high-rise residential users,this article uses a high-rise residential elevator group in Suzhou as the research background to study the passenger flow pattern recognition of high-rise residential elevator groups and the optimization of elevator group control scheduling algorithms,and conduct semi-physical physical simulation research in order to strengthen the elevator Analyze the potential laws of passenger flow data and improve the rationality of the group control scheduling algorithm decision,thereby effectively improving the service quality of high-rise residential elevator group operation.First,compare the concept of elevator group control system,analyze in detail the demand characteristics of residential elevator group control system,and carry out the overall architecture design of the entire high-rise residential elevator group control system under this premise,laying the foundation for follow-up research and semi-physical physical simulation.Second,in view of the shortcomings of traditional pattern recognition methods,the characteristics of passenger flow in high-rise residential buildings are analyzed and researched,and the method of passenger flow pattern recognition based on LSTM is proposed.This method establishes a passenger flow pattern recognition model through LSTM,learns the long-term dependence in the time series,and classifies it through the Softmax activation function;the Adam algorithm is used to optimize the network training parameters.The experimental results show that this method has a better recognition effect in high-rise residential buildings,and provides theoretical support for the following chapters to adopt reasonable strategies for dispatching elevators for different passenger flow patterns in residences.Third,considering that there are few coordinated scheduling schemes and control systems for high-rise residential elevator groups,and the lack of initial pheromone and slow convergence in elevator group control scheduling of traditional ant colony algorithm,a scheduling based on enhanced ant colony algorithm is proposed.Optimization.Taking the owner’s average waiting time,riding time and system operating energy consumption as the optimization goal;adopting the Q-learning algorithm in reinforcement learning to set the initial pheromone,and at the same time introducing the Q value in the probabilistic path selection to effectively guide the ants to explore.So as to find the best elevator dispatching plan.Experiments show that the enhanced ant colony algorithm has a good effect on the efficiency of elevator group dispatching in high-rise residential buildings.Finally,in order to further test the performance of the algorithm,a semi-physical simulation platform for elevator group control based on PLC and Lab VIEW is built.The lower computer selects Siemens S7-1200 series PLC as the control core,and performs data interaction with the upper computer through Ethernet(TCP protocol)to simulate the dispatching operation of the elevator group control system in high-rise residential buildings. |