Smart Water Application Architecture Design And Key Technology Research

The emergence of "smart city" has attracted great attention worldwide,and various national governments have formulated various smart city construction plans.As an important component of smart city,smart water has also begun to attract the attention of water managers.Water management departments in various regions have started to try to establish their own smart water application systems.However,smart water has been proposed for a short period of time,and people's understanding of smart water is not clear enough,resulting in various problems in the construction of smart water.Against this background,this paper analyzes and studies the concept,application framework and key technologies of smart water.The detailed work is as follows:(1)Through the analysis and summary of the existing research results of Smart Water,this paper finds that scholars' understanding of smart water has a certain difference.Therefore,this paper analyzes the characteristics of smart water and redefines the concept.In view of the lack of scientific and effective application framework guidance in the current smart water application construction process,this paper designs the application framework of intelligent water with reference to the existing research results and practical experience.(2)In view of the problem of abnormal data often appearing in water data,this paper divides water abnormal data recognition into three different situations from the data type of data set,numerical attribute,Category attribute and mixed attribute.Then,according to these three different situations,three anomaly data recognition methods are designed based on K-means method and K-modes method respectively.In view of the parameter optimization problems in both the K-means method and the K-modes method,this paper proposes to use the fruit fly optimization algorithm to optimize the parameters of the two methods.In addition,since the K-means method and the K-modes method belong to the clustering method,the data set can only be classified,and it is impossible to automatically determine which class in the classification is abnormal data.To this end,this paper proposes an automatic identification algorithm for abnormal data.The algorithm is based on the clustering result to determine the final outlier class based on the principle of inter-class difference,and the outlier value is recognized automatically.In order to verify the validity of this method,three methods are experimented with UCI data and water actual data respectively,and the experimental results show that all three methods showed good accuracy.(3)Aiming at the problem of low accuracy and poor economy in fault diagnosis of water supply network,this paper first proposes a fault diagnosis model of water supply network based on Kernel Extreme Learning Machine,and optimizes its parameters by using Yin-Yang pair optimization algorithm,which improves the diagnostic accuracy of the fault diagnosis model.However,Yin-Yang pair also optimization algorithm has slow convergence speed and easy to fall into local optimum.In order to improve the convergence speed and reduce the possibility of Yin-Yang pair optimization algorithm falling into local optimum,the dynamic adjustment strategy of key parameters and candidate solution generation strategy are designed in this paper.Then,taking the minimum number of monitoring points as the optimization objective,taking the fault diagnosis accuracy of the fault diagnosis model as the necessary condition,and taking the improved Yin-Yang pair optimization algorithm as the tool,this paper proposes an optimized arrangement method of hydraulic monitoring points for fault diagnosis.Finally,the effect of the optimal arrangement method of hydraulic monitoring point is tested by experiments,and it is found that the monitoring point arrangement scheme obtained by this method can not only ensure the accuracy of fault diagnosis,but also improve the economy of monitoring point arrangement scheme.(4)In order to solve the problem that the inaccuracy of current forecasting methods of urban hourly water demand,this paper presents a prediction model of urban short-term water demand based on the idea of "decomposition-prediction and fusion".Since the commonly used Empirical Mode Decomposition(EMD)has the problem of mode confusion,this paper improved the EMD by adding noise and two-layer decomposition,so as to reduce the mode confusion phenomenon existing in the decomposition term.In addition,when choosing the forecasting method,this paper changes the research idea of single forecasting method commonly used,and proposes a suitable forecasting method based on permutation entropy and combining the characteristics of decomposition items.After comparing and analyzing the common forecasting methods,the Least Square Method and the Long-Short Term Memory are selected as the prediction method of the decomposition items.Since the two methods both have the problem of parameter affecting the prediction performance,this paper also introduces the Teaching Learning Based Optimization algorithm to optimize the parameters of the two methods,which improves the prediction accuracy of the overall prediction model.The effect of the prediction model was tested through experiments,and it was found that the prediction model had relatively good prediction accuracy compared with the prediction methods such as LSTM,ARIMA,LSM,BPNN and SVR.(5)In view of the lack of research on multi-source dispatching method for urban agglomeration,this paper proposes a multi-source joint dispatching model for urban agglomeration,which is based on the security demand,economic demand and ecological demand of urban water supply.The model contains many variables and multiple objective functions.In order to solve this problem,this paper proposes a two-subgroup multi-objective fruit fly optimization algorithm that integrates teaching and learning optimization ideas.Considering that multi-objective optimization algorithm can find multiple equilibrium solutions,but it is impossible to determine which of them is the optimal solution,this paper proposes a ranking method of equilibrium solutions based on TOPSIS comprehensive evaluation method and expert scoring method.Through this method,the equilibrium solutions can be ranked according to the decision-making needs of users,so that the ranking problem of equilibrium solutions can be solved.The paper validates the proposed method through a set of urban agglomeration water supply data.The experimental results show that the proposed multi-source joint scheduling method for urban agglomeration has the ability to find the optimal dispatching scheme.This paper is devoted to the research and exploration of basic theory and technical methods in smart water.The paper analyses the connotation and characteristics of smart water.Focusing on a number of core functional requirements in the field of smart water,this paper proposes specific solutions and studies some key technical problems,which further enriches the related research content of smart water and has a certain theoretical value and practical significance.