Multi-Objective Optimization of Multistage Pump Balance Drum System Based on BP Neural Network and Genetic Algorithm

As the centrifugal pump is running, the fluid usually flows into the impeller along pump shaft, and the fluid flows out radially by the force of the impeller. The force is mutual, so the impeller is also subjected to the reaction force of the fluid, but the distribution of this force on the blades is uneven. In addition, the front and rear shrouds of the impeller are asymmetric, which are the main causes of axial force. This paper adopts numerical calculation method studying the mechanism of axial force of impeller at all stages of multistage pump at various working conditions, and exploring the formation mechanism of shroud pressure differential force and blade twisting axial force and its variation laws of similarities and differences, analyzing the steady state and transient characteristics between axial force and hydraulic property of double-casing multistage pump. The results show that the rotational angular velocity of the fluid in the front and rear pump chamber at each stage impeller is distributed along the axial direction in three regions, the regions are pump body boundary layer, core region, and impeller boundary layer. The working surface and back surface of the blade twist have the high and low axial force area, and its distribution is staggered, at the same number of stages, the greater the flow rate, the smaller the blade twisting axial force. The shroud pressure differential force with the increase of impeller stages presents a linear increasing trend, conforms to the principle of linear superposition of cover pressure differential force. The total axial force pulsation of multi-stage pump is related to the number of secondary impeller blades, its primary frequency coincides with the secondary impeller blade frequency, increasing the flow rate can reduce the multi-stage pump axial force pulsation amplitude. The pulsation period of single-stage impeller head and efficiency are related to the number of impeller blades, the smaller the number of impeller stages, the stronger the pressure dynamic, and static interference effect of the impeller inlet and outlet. Rotation of the secondary impeller causes dynamic and static interference, which is the main reason for the pulsation of the axial force coefficient in double-casing multistage pumps, the pulsation intensity is related to the periodic generation and shedding of the blade vortex. The results of the study can be used as a reference for optimizing the axial force of double-casing multistage pumps.

In centrifugal pumps, axial thrust bearings are required to compensate for the axial loading on the pump impeller(s). These forces are mainly the result of unequal distributions of pressure on projected areas at both sides of the impeller(s). Axial loading tends to be highest in multistage pumps in which all impellers face the same direction. Balancing devices are often used to reduce the loading of the axial bearings. Common types of balancing devices are the balance drum and the balance disc. Both of them have their advantages and disadvantages. A stepped balance drum is a new concept for a balancing device that can be used in any type of turbomachinery. It combines the benefits of the balance drum and the balance disk without the drawbacks of both. In this paper, the new concept is presented along with a semi-empirical model of its performance. A CFD analysis is also presented which serves to verify the analytical model. An optimization procedure of a stepped balance drum is given using a multistage centrifugal pump as an example.

Based on the three-dimensional Reynolds-Averaged N-S equations and the standard k-ε turbulence model, the inner flow of a multistage vane water centrifugal pump was simulated. The distribution law of inner flow was analyzed by using the SIMPLEC algorithm and multiple reference frame (MRF) model provided by CFD software Fluent. The pressure and velocity distribution law in impeller and vane diffuser are obtained. Results indicated that the lowest pressure area , where cavitation was easy to occur, was in the inlet of the first impeller blade back. There is reverse flow in outlet of impeller. An area of low pressure can be seen in the inlet of return guide vanes. The guide vane could reduce fluid velocity and eliminated fluid rotating component.

Currently the application of traditional centrifugal pump design theory to develop the multi-stage centrifugal pumps is very mature, but the factor of affecting hydraulic performance of centrifugal self-priming pump is still unclear. In this paper, using CFD method to simulate the hydraulic performance of the multistage self-priming pump in different operating conditions, and get the performance prediction curve. The simulation results were compared with experimental data, the result shows that simulation method can accurately predict the performance and internal flow characteristics of multi-stage self-priming pump. The impact on the hydraulic performance by water-keeper chamber and the gas-liquid separation chamber is very limited. The self-priming components can provide a better hydraulic performance. This will be as a reference for the design of self-priming pump.

Due to the advantages of high head and no leakage, multistage canned motor pump is widely used in oil industry, chemical industry, national defense and atomic energy. In order to meet the needs of the market, the multistage canned motor pump is designed. This paper introduced the hydraulic design and structural design. In order to optimizing the performance of the pump, this paper designed and used multistage canned motor pump DBP15–50×8 as the research object. Three-dimensional model of the main flow passage components is built and the mesh is generated respectively by using Pro/E and ICEM software, and we calculated the whole internal flow field of the pump that was selected by using ANSYS CFX14.0 software, achieving the pressure and velocity distribution in the pump and the internal details of flow in impeller and other main flow components. The post-processing showed the fluid in sliding bearing section rotates around the shaft, so the local flow is disorder. The comparison of the performance prediction and the experiment shows that the error is low. The cavitating turbulent flow in the flow field was numerically simulated by using the cavitation model. The cavitation phenomena didn’t occur in the experiments. The condition meets the result of numerical simulation.

Abaqus finite element simulation technology and BP neural network and genetic algorithm are applied to the study on the optimization of the cutting parameters of the thin-walled ring gear. A simulation model of ring gear shaping was established by Abaqus finite element analysis software, a cutting force prediction model was built using BP neural network. With the help of the BP neural network and genetic algorithm function extreme value optimization feature, the gear ring gear shaping parameter optimization model was established, and the Abaqus simulation model was used to test it. The relevant slotting parameters, namely the slotting speed, the amount of back-cutting and the circumferential feed speed have been successfully optimized, and the slotting force has been reduced. The experimental results show that the parameter optimization model of ring gear shaping based on BP neural network and genetic algorithm can effectively realize the optimization of slotting parameters.

In order to study the evolution mechanism of axial force characteristics of centrifugal pump in the transient process of variable working conditions, the IS80-65-160 centrifugal pump is taken as the research object, and four flow operating points of centrifugal pump 1.0Qr, 0.8Qr, 0.6Qr, and 0.4Qr are selected. Non-constant numerical simulation is carried out, and on the basis that the numerical calculation results of the external characteristics are basically consistent with the experimental results, the changes of the axial force of the impeller during the flow reduction process of the variable valve adjustment and the variable speed adjustment of the centrifugal pump are studied, respectively. The results show that as the flow rate decreases, the force of the variable valve regulating impeller cover plate increases, and the force of the variable speed regulating cover plate decreases. Under the two adjustment methods, the axial force on the pressure surface of the blade becomes smaller, the axial force on the suction surface of the blade is basically unchanged, and the pulsation coefficient of the impeller axial force increases first and then decreases. The axial force pulsation coefficient reaches the maximum value in the variable valve adjustment condition of 0.8Qr and the variable speed adjustment condition of 0.6Qr, respectively, and the axial force on the outer wall of impeller cover plate has the greatest influence on the main frequency axial force ripple coefficient; the axial force acting on the outer wall surface of the front cover plate of the impeller and the inner wall surface of the front and rear cover plates increases with the increase in the radius, The axial force acting on the outer wall surface of the rear cover plate of the impeller has axisymmetrical properties. This study revealed the evolution mechanism of the axial force characteristics of the centrifugal pump impeller during the transient process of variable working conditions, and the research results can provide a reference for improving the stable operation of the centrifugal pump.

Network attacks are characterized by concealment in construction, diversity in implementation and complexity in composition. As far as building concealment is concerned, attackers often hide their attack intentions, secretly collect enough intelligence information, design appropriate attack means, and launch attacks in a sudden way before launching attacks. In this paper, a game model of network attack and defense based on BPNN (BP Neural Network) algorithm is proposed, aiming at the characteristics of conflict of interests and strategic dependence in network attack and defense. At the same time, attackers will choose vulnerabilities that are easy to use and destructive to launch attacks, and discuss the distribution of benefits after cooperation before cooperation, so as to ensure the long-term and stability of the partnership. F1m10minally, through the simulation experiment, it is proved that the model constructed in this paper tends to choose the high-cost attack strategy from the first minute to the tenth minute, and the probability of choosing the high-cost attack strategy among the attack groups remains basically unchanged at the 12th minute. After more than 12 min, due to the long-term stalemate between the offensive and defensive sides, the attack cost is too high, and the attacker gradually tends to a negative state, and finally evolves into no attack.

To investigate the effect of the different wrap angles from the hub to the shroud surface in the space diffuser (i.e., the trailing edge stacking style) on the principle of corner separation vortex flow, a numerical simulation method has been conducted in a multistage submersible pump. Building a linear equation on the profile line of the diffuser trailing edge to optimize the wrap angle on every spanwise from the hub to the shroud, and the mapping response relationship between the wrap angle difference and the hydraulic performance in the space diffuser has been analyzed. Under the variable wrap angle difference (Δϕ=ϕhub−ϕshroud), the secondary flows in different directions, non-uniformity, diffuser efficiency, and pressure recovery are compared. The positive wrap angle difference (i.e., the shroud wrap angle is smaller than the hub one) improves the strength of the secondary flow and partly corner separation vortex in the diffuser, so the hydraulic characteristic of positive cases is better than the negative wrap angle difference. Moreover, in scheme A (in which the hub wrap angle is constant and the shroud wrap angle decreasing), the transversal secondary flow has been weakened, the low-energy fluid located in the corner has been suppressed, the extensional secondary flow has been increased, the diffuser hydraulic performance has been improved, and unidirectionally increases with the wrap angle difference increasing. When the shroud wrap angle is constant, the extensional secondary flow has been enhanced by the increasing hub wrap angle. Meanwhile, the increasing extensional secondary flow has been countered by the deteriorating extensional flow at the diffuser inlet and transversal secondary flow, the diffuser hydraulic performance increases and then decreases as the wrap angle difference increases, with an optimal wrap angle difference is about 20°.

English translation evaluation method based on BP neural network

Innovation education performance management plays a key role in training innovative students for universities, and how to evaluate innovation education performance is one of the difficulties and hot research fields for the researchers related. The paper presents a new model for evaluating innovation education performance based on improved BP neural network. First an evaluation indicator system of innovation education performance is designed through the aspects of university regulation and teachers and students and education effects; Second, aiming at the shortages of the existing BP neural network algorithm of data-mining for evaluating innovation education performance, BP neural network algorithm is improved through adjusting dynamic strategy and the value of momentum factor to speed up the convergence and simplify the structure and to improve evaluating accuracy of the original BP model. Finally the experimental results show that the new evaluation indicator system and improved BP neural network algorithm can be used practically in evaluating innovation education for different universities and guarantee the evaluation effectiveness and validity.

The mathematic principles and numerical algorithm of BP neural network for gas contents were firstly studied, Then, the actual measurement data of gas contents during geological prospecting and mining of PingdingshanNO.10WU9-10 mine were collected, and 12 reliable dots were gained. By selecting 3 factors including depth, coal seam thickness and coal roof lithology as the input element, and the multivariate forecast models of gas contents based on BP neural network were respectively constructed. According to the calculation and evaluation of results, accuracy of the model to meet the requirements of engineering precision, indicated that BP neural network to predict mine e Pingdingshan 9–10 ten gas content of coal seam gas is feasible.

With the Belt and Road strategy deeply rooted in people's hearts, it is an inevitable choice for China and countries (or regions) along the Belt and Road strategy to build a science and technology innovation community. In this study, the high-tech industries of major provinces and cities in China are taken as the research object, and the collaborative innovation process and mechanism of each innovation subject are taken as the research purpose. By introducing the collaborative innovation concept of the innovation community, we construct the innovation performance evaluation index system to optimize the model of the innovation community performance evaluation via using neural networks. The neural network is one of the most widely used algorithms among deep learning algorithms, and this paper mainly analyzes the characteristics of Back Propagation (BP) neural network in the performance evaluation, which has strong local optimization but slow convergence speed. We build a BP neural network based on the genetic algorithm, and this neural network makes the performance evaluation model has the characteristics of strong global optimization and fast convergence speed. However, at the same time, it also has the risk of early convergence and lacks the ability of the global search. On this basis, we propose an improved genetic BP neural network based on simulated annealing in this paper. The improved simulated annealing algorithm not only can reduce error but also has a stronger local convergence and a faster evolution speed, which can obtain a satisfying global optimal solution. When we apply the improved simulated annealing genetic BP neural network in the performance evaluation of the science and technology innovation community, the simulation results show that this algorithm can be a reliable performance evaluation method and easier to obtain the optimal solution than other algorithms. Besides, the feasibility of the evaluation system and method is verified.

The paper focuses on the fault of emulsifier during the production of emulsion explosive as the research object. Aiming at the conventional Back Propagation (BP) neural network has the defects of tardy convergence rate and undesirable optimal ability in vibration fault diagnosis of emulsifier a method of optimizing the BP neural network based on improved particle swarm optimization was presented. It can optimize initial weight and threshold of the BP neural network, and diagnose the fault of emulsifier. Instance simulation results show that the model of the BP neural network based on improved particle swarm optimization fault diagnosis has better classification effect and improves the fault diagnosis accuracy of the emulsifier.

To study drivers’ driving behaviour in dilemma zones at signalized intersections, dilemma zones at signalized intersections are taken as the research objects. Based on analyzing the driving characteristics of motor vehicles. And collect the driving behavior parameters of video observations and road parameters of field observations, take these parameters as the factors which influence the drivers’ decision, conduct the significance analysis with SPSS(Statistical Product Service Solutions) for significant influencing factors, and then take these factors as the input value of BP(Back Propagation) neural network model, BP neural network model is set up and tested based on TensorFlow in Python. The prediction model of drivers’ decision-making in dilemma zone of signalized intersections under different speed limits is obtained, the research shows that the lower the speed limit, the better the accuracy of the intersection prediction. Further verified by comparison with the prediction model of drivers’ decision-making based on binary logistic, the accuracy of different prediction models is analyzed based on actual driver decisions.