Optimal regulation and control method for centripetal turbine with adjustable nozzle
By integrating working fluid selection and turbine design into the system optimization design, and utilizing the NSGA-II algorithm and PSO-BP neural network optimization model, the problem of accuracy in predicting turbine performance under varying operating conditions in organic Rankine cycles was solved, thereby improving system performance and enabling rapid adjustment of turbine parameters.
Patent Information
- Application Number
- CN202411252143.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-09
- Publication Date
- 2024-11-12
AI Technical Summary
Existing technologies struggle to accurately predict turbine performance under varying operating conditions in organic Rankine cycles, and the random assignment of weights and thresholds in traditional BP neural networks reduces the model's prediction accuracy, failing to achieve simultaneous improvement in turbine performance and system performance.
By coupling working fluid selection, turbine design and system design, the NSGA-II algorithm is used for multi-objective optimization. Combined with PSO-BP neural network and three-dimensional CFD method, the optimal control model of adjustable nozzle centripetal turbine is established. Through multi-objective optimization and neural network prediction accuracy optimization, the optimal matching between turbine parameters and system parameters is achieved.
It improves the operating performance of organic Rankine cycle power generation systems under different cold and heat source conditions, enhances the turbine's adaptability to changes in cold and heat sources, improves prediction accuracy and computational efficiency, and enables rapid adjustment of turbine parameters.
Smart Images

Figure CN118934102A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of centripetal turbine optimization and regulation, in particular to an optimal regulation method for a centripetal turbine with an adjustable nozzle. Background Art
[0002] The organic Rankine cycle has the advantages of simple and compact structure and low operation and maintenance costs. It is a medium- and low-temperature waste heat power generation technology with great development prospects. It has been applied in the fields of waste heat utilization such as industrial waste heat, geothermal energy, and solar energy. The expander is the core equipment of the organic Rankine cycle power generation system. Among different types of expanders, the centripetal turbine has the advantages of large single-stage enthalpy drop, small flow rate and high efficiency. It can adapt to the application conditions of the organic Rankine cycle and is an ideal expansion device.
[0003] However, the driving heat source of the organic Rankine cycle is often affected by process conditions or ambient temperature, and has significant fluctuations, which significantly reduces the thermal economic performance of the system. Therefore, how to improve the operating performance of the system under different heat source conditions is a key issue that needs to be solved in the practical application of the organic Rankine cycle.
[0004] The adjustable nozzle centrifugal turbine can effectively improve the performance of the centrifugal turbine under different operating conditions by adjusting the nozzle outlet angle, thereby improving the thermal economic performance of the organic Rankine cycle power generation system. It is an important variable operating condition control method. However, at present, the centrifugal turbine mainly predicts the turbine performance through a one-dimensional method, which is difficult to accurately predict the actual performance of the turbine under variable operating conditions. Although the three-dimensional CFD simulation method can accurately predict the variable operating condition performance of the turbine, it consumes a lot of computer resources and simulation time.
[0005] Although BP neural network has the ability of self-learning and approximating any nonlinear function, and the use of BP neural network model has the advantages of fast and efficient, the traditional BP neural network determines its weights and thresholds by random assignment, which significantly reduces the prediction accuracy of the model.
[0006] In addition, the traditional adjustable nozzle centrifugal turbine control method often ignores the optimal design of the turbine, and adjusts the turbine parameters according to the turbine performance without considering the system performance of the organic Rankine cycle, making it difficult to achieve the simultaneous improvement of turbine performance and system performance. Therefore, if the turbine design is coupled with the system design, and the three-dimensional CFD method and the optimized BP neural network model are used to predict the variable operating performance of the centrifugal turbine, and the system integrated operation optimization is carried out, the optimal control of the adjustable nozzle centrifugal turbine can be achieved. Summary of the invention
[0007] The purpose of the present invention is to overcome the shortcomings of the above-mentioned background technology, propose an optimal control method for an adjustable nozzle centrifugal turbine, achieve optimal control of the adjustable nozzle centrifugal turbine under different cold and heat source conditions, and maximize the operating performance of the organic Rankine cycle power generation system.
[0008] The object of the present invention is achieved by the following technical solution: An optimal control method for an adjustable nozzle centripetal turbine comprises the following steps:
[0009] (1) Input the design conditions of the organic Rankine cycle power generation system, including the heat source temperature and flow rate, the cold source temperature and flow rate, and the working fluid pump characteristic curve;
[0010] (2) Establish a one-dimensional design mathematical model for the centripetal turbine, couple the candidate cycle working fluid and turbine mathematical model into the system model, and use the system thermal efficiency (η E ), average generation cost (LEC), and turbine size factor (DF) as the goals, the working fluid screening and turbine design are integrated into the system optimization design, and a one-dimensional integrated optimization design mathematical model of the system is established;
[0011]
[0012] (3) First, the NSGA-II algorithm is used to iteratively calculate the one-dimensional integrated optimization model to obtain the Pareto frontier solution set of multi-objective optimization, and then multi-objective decision-making is performed to obtain the optimal solution;
[0013] (4) Using the optimal solution obtained through decision making, the evaporator, condenser, centrifugal turbine, and working fluid pump in the system are designed to obtain the evaporator and condenser areas, working fluid pump head and flow rate, and centrifugal turbine structural parameters;
[0014] (5) Taking the main aerodynamic parameters of the centrifugal turbine as variables, the central composite design method was used to design the test scheme, and a three-dimensional simulation model of each test scheme was established. The centrifugal turbine performance of each scheme was obtained using the CFD method;
[0015] (6) The simulation results of each scheme are used as the training set of the PSO-BP neural network to obtain the prediction model of the variable operating performance of the centrifugal turbine;
[0016] (7) Establish the variable operating condition mathematical model of the evaporator, condenser and working fluid pump, and integrate the centripetal turbine PSO-BP neural network model with the variable operating condition models of other components to establish the variable operating condition mathematical model of the organic Rankine cycle power generation system;
[0017] (8) Based on the system thermal efficiency (η E ), average electricity generation cost (LEC) and net output power (W net ) as the target, a multi-objective optimization model of system variable working conditions is established:
[0018]
[0019] (9) The NSGA-II algorithm is used to optimize the variable operating condition model, obtain the Pareto solution set and make decisions on it, determine the optimal operating parameters of the centrifugal turbine under different cold and heat source conditions, and obtain the correlation between the optimal operating parameters of the centrifugal turbine and the cold and heat source parameters.
[0020] According to claim 1, the optimal control method of an adjustable nozzle centrifugal turbine is characterized in that: the optimal control method first couples the candidate circulating working fluid and the centrifugal turbine model to the organic Rankine cycle power generation system, performs multi-objective optimization design on the system, and then performs system variable operating condition optimization based on the component structural parameters obtained by the optimization design, so as to achieve the optimal matching of the centrifugal turbine operating parameters and the cold and heat source conditions, and improve the variable operating condition performance of the system.
[0021] The optimal control method for an adjustable nozzle centrifugal turbine according to claim 1 is characterized in that: in step (2), the circulating medium (N f ), system thermal parameters (relative evaporator temperature R Te , Evaporator superheat ΔT e,s , condensation temperature T c , evaporator pinch point temperature ΔT e,pp , Condenser pinch point temperature difference ΔT c,pp ) and turbine aerodynamic parameters (reaction degree Ω, wheel diameter ratio σ, speed ratio Φ, impeller inlet absolute flow angle α4, outlet relative flow angle β5, nozzle outlet absolute flow angle α3, turbine speed N) are used as decision variables for multi-objective optimization design, so that working fluid screening and turbine design are integrated into the system optimization design. Through optimization calculation, the optimal working fluid and the optimal structural parameters of system components can be obtained at the same time;
[0022] The relative evaporation temperature is defined as follows:
[0023]
[0024] Where, T e is the evaporation temperature, T c,f is the critical temperature of the working fluid;
[0025] The definition of the objective function is as follows:
[0026]
[0027] Where W net is the net work of the system, Q e is the heat absorbed by the system, r n is the turbine nozzle inlet radius, W tis the output power of the centrifugal turbine, CRF is the cost recovery factor, C s is the system investment cost, OP s is the annual operating time of the system;
[0028] The constraints considered in the multi-objective optimization model are as follows:
[0029]
[0030] The optimal control method of an adjustable nozzle centrifugal turbine according to claim 1 is characterized in that: in the step (3), the entropy method is used to determine the weights of different objective functions, and then the Topsis method is used to determine the optimal solution in the Pareto frontier solution set, so that the selection of the optimal solution is more reasonable. The basic idea of the method is as follows:
[0031] First, determine the weight coefficients of each objective function:
[0032]
[0033]
[0034]
[0035] In the formula, x ij is the dimensionless processing value of the jth indicator in the ith sample, m is the number of individuals in the optimal solution set, n is the number of objective functions, and w j is the weight coefficient of the jth indicator, E j is the entropy value of the j-th indicator;
[0036] Then, calculate the distances of each solution individual to the ideal point and the negative ideal point:
[0037]
[0038]
[0039] In the formula, f i ideal is the positive ideal point function value of the jth objective, f i nadir is the negative ideal point function value of the jth objective;
[0040] Finally, the TOPSIS decision method is used to calculate the relative distance Y from each point on the Pareto frontier solution set to the negative ideal point i , and Y i The point with the largest value is taken as the optimal solution, and its expression is:
[0041]
[0042] The optimal control method for a centripetal turbine with an adjustable nozzle according to claim 1 is characterized in that: in the step (5), the system adjusts the centripetal turbine inlet pressure (p t,in )、Outlet pressure (p t,out ), nozzle angle (α3) and inlet temperature (T t,in ) and other four parameters as adjustment variables, and these four parameters are used as design variables of the central composite design method. This method is used to determine more than 70 experimental schemes, so that the design variables of each scheme are evenly distributed in the variable range.
[0043] According to claim 1, the optimal control method of an adjustable nozzle centrifugal turbine is characterized in that: in the step (6), in order to improve the prediction accuracy of the neural network, the PSO optimization algorithm is used to determine the weight and threshold of the BP neural network; the PSO-BP neural network takes the four design variables in the central composite design as input, and the output power, turbine efficiency and working fluid flow of the centrifugal turbine as output, and establishes a PSO-BP neural network prediction model for the variable operating performance of the centrifugal turbine, wherein the number of layers of neurons in the intermediate hidden layer of the neural network is determined by the following formula:
[0044]
[0045] Where n1 is the number of input layer nodes, n2 is the number of output layer nodes, and a is a constant between 1 and 10.
[0046] The optimal control method of an adjustable nozzle centrifugal turbine according to claim 1 is characterized in that: in the step (7), a variable operating condition mathematical model of the working fluid pump is established by similarity theory:
[0047]
[0048]
[0049]
[0050] In the formula, N represents the speed, q is the volume flow rate, H is the head, m is the mass flow rate, η is the efficiency, the subscript off is the variable working condition, d is the design working condition, and p is the pump;
[0051] When the evaporator and condenser operate under variable operating conditions, the outlet temperature of the heat exchanger is assumed first, and then the required area of the heat exchanger is calculated based on the heat source and working fluid parameters. If the calculated area is inconsistent with the actual area of the heat exchanger, the outlet temperature of the heat exchanger is adjusted to continue the iterative calculation until the requirements are met; the PSO-BP neural network prediction model of the adjustable nozzle centripetal turbine is integrated with the variable operating condition mathematics of the system evaporator, condenser and working fluid pump to establish a variable operating condition mathematical model of the system.
[0052] According to the optimal control method of an adjustable nozzle centrifugal turbine according to claim 1, it is characterized in that: in the step (9), the importance sequence of the cold and hot source conditions to the system performance is determined by the range analysis method, and then the correlation between the optimal operating parameters of the centrifugal turbine and the key cold and hot source parameters is analyzed, and the correlation formula between the optimal operating parameters and the key cold and hot source parameters is determined by multivariate regression.
[0053] The beneficial effects of the present invention are:
[0054] (1) The present invention integrates working fluid screening, turbine design and organic Rankine cycle power generation system design. Through design optimization and operation optimization, the adaptability of the adjustable nozzle centrifugal turbine to changes in cold and hot source conditions is enhanced, thereby maximizing the variable operating performance of the system.
[0055] (2) The PSO algorithm is used to optimize the weights and thresholds of the BP neural network to improve the prediction accuracy of the neural network. The PSO-BP neural network is combined with the three-dimensional CFD method to improve the prediction accuracy and calculation efficiency of the variable operating performance of the adjustable nozzle centripetal turbine.
[0056] (3) Establishing the predictive correlation between the cold and hot source parameters and the optimal parameters of the adjustable nozzle centripetal turbine is helpful for the rapid adjustment of turbine parameters under the fluctuation of cold and hot source parameters. BRIEF DESCRIPTION OF THE DRAWINGS
[0057] Figure 1 Flow chart of the optimal control method for adjustable nozzle centripetal turbine
[0058] Figure 2 Basic Organic Rankine Cycle Schematic
[0059] Figure 3 Pareto frontier for system integration optimization
[0060] Figure 4 Comparison of PSO-BP neural network prediction and experimental values of centripetal turbine power
[0061] Figure 5 Relationship between the optimal nozzle angle and heat source parameters of a centripetal turbine DETAILED DESCRIPTION
[0062] The present invention will be further described in detail below in conjunction with the accompanying drawings.
[0063] See attached Figure 1 The flowchart shown in the figure and the basic organic Rankine cycle as an example are used to introduce the basic process of the optimal control method of the adjustable nozzle centrifugal turbine of the present invention. The basic organic Rankine cycle mainly includes an evaporator, a centrifugal turbine, a condenser and a working fluid pump. The system principle diagram is shown in the attached figure. Figure 2For the adjustable nozzle centrifugal turbine in the basic organic Rankine cycle, the method is carried out in the following steps:
[0064] According to the specific cold and hot source conditions, the heat source temperature T is given h =150℃, heat source flow rate q v =19.5kg / s, cooling water temperature T w =25℃; select system thermal parameters (relative evaporator temperature R Te , Evaporator superheat ΔT e,s , condensation dew point temperature T c , evaporator pinch point temperature ΔT e,pp , Condenser pinch point temperature difference ΔT c,pp ) and turbine aerodynamic parameters (reaction degree Ω, wheel diameter ratio σ, speed ratio Φ, impeller inlet absolute flow angle α4, outlet relative flow angle β5, nozzle outlet absolute flow angle α3, turbine speed N) are used as design variables; at the same time, R245fa, R600, R600a, R1234ze, R1234yf, R1233zd, R601a, R601 and other working fluids are selected as candidate working fluids, and the working fluids are numbered to convert the circulating working fluid (N f ) as the design variable.
[0065] The performance parameters of the basic organic Rankine cycle are calculated according to the following formula:
[0066] System thermal efficiency:
[0067] η E =W net / Q e (1)
[0068] Net power output of the organic Rankine cycle:
[0069] W net =W t -W p (2)
[0070] Output power of centrifugal turbine:
[0071] W t =m f (h t,in -h t,out ) (3)
[0072] Electric energy consumed by working fluid pump:
[0073] W p =m f (h p,out -h p,in ) (4)
[0074] The amount of heat absorbed by the working fluid in the evaporator:
[0075] Q e =m f (h e,out -h e,in ) (5)
[0076] System average power generation cost:
[0077]
[0078]
[0079] Turbine size factor:
[0080]
[0081] In the formula, h is specific enthalpy, Q is heat, W is power, m is f is the mass flow rate of the working fluid, CRF is the cost recovery factor, C s is the system investment cost, OP s is the annual operating time of the system, r n is the turbine nozzle inlet radius, i is the interest rate, subscript t is the turbine, p is the pump, e is the evaporator, in represents the inlet, and out represents the outlet.
[0082] Considering the limiting conditions for the normal operation of the organic Rankine cycle, a multi-objective integrated optimization model of the system is established with the system thermal efficiency, average power generation cost and turbine size factor as the objectives:
[0083]
[0084] The constraints considered in the design optimization are as follows:
[0085]
[0086] The multi-objective optimization program of organic Rankine cycle was written by Matlab, and the 13 selected design parameters were optimized using NSGA-II algorithm to obtain the Pareto frontier solution set of multi-objective optimization, as shown in the attached figure. Figure 3 The entropy method and Topsis method are used to make decisions on the Pareto frontier solution set and determine the optimal design parameters of the system. The optimization results are shown in Table 1.
[0087] Table 1 Optimal solutions obtained by multi-objective optimization design decision
[0088]
[0089] The data in Table 1 are used to design the evaporator, condenser, working fluid pump, and centrifugal turbine in the organic Rankine cycle, and the heat exchange area of the evaporator is 146m 2The heat exchange area of the condenser is 370m 2 , the pump efficiency is 0.7, and the pump head is 125m. At the same time, the geometric parameters of the centripetal turbine are obtained.
[0090] The inlet pressure (p t,in )、Outlet pressure (p t,out ), nozzle angle (α3) and inlet temperature (T t,in ) as a variable, and use the central load design method to design 100 test schemes; establish a three-dimensional CFD simulation model for each test scheme, simulate each simulation model, and obtain the turbine performance under each simulation scheme.
[0091] The inlet pressure of the centripetal turbine (p t,in )、Outlet pressure (p t,out ), nozzle angle (α3) and inlet temperature (T t,in ) as input, turbine output power, turbine efficiency and mass flow as output, the number of neurons in the middle hidden layer is 10, and a PSO-BP neural network model is established; the CFD simulation results of 100 test schemes are used as training samples of the PSO-BP neural network, 80 of which are training sets and 20 are test sets to train the neural network. The results are shown in Figure 4 As shown; the results show that the PSO-BP neural network model can accurately predict the performance of the adjustable nozzle centrifugal turbine under different working conditions.
[0092] The variable operating condition mathematical model of the evaporator, condenser and working fluid pump is established, and the variable operating condition mathematical model of the organic Rankine cycle power generation system is established by combining the PSO-BP neural network model of the adjustable nozzle centripetal turbine. The variable operating condition mathematical model of the organic Rankine cycle power generation system is established with the system thermal efficiency, average power generation cost and net output work as the target, and the turbine inlet pressure (p t,in )、Outlet pressure (p t,out ), nozzle angle (α3) and inlet temperature (T t,in ) is the decision variable, and a multi-objective operation optimization model of the system with variable working conditions is established:
[0093]
[0094] The NSGA-II algorithm optimization calculation program was written using Matlab, and the parameters of the adjustable nozzle centripetal turbine under different cold and hot source conditions were optimized to obtain the Pareto frontier solution set under each operating condition. Then the entropy method and Topsis decision method were used to determine the optimal solution under different cold and hot source conditions.
[0095] The range analysis method is used to find that the influence of the cold and hot source temperature on the operating parameters of the adjustable nozzle centrifugal turbine is much greater than the cold and hot source flow rate. Therefore, a curve diagram of the cold and hot source temperature and the optimal operating parameters of the centrifugal turbine is drawn, and a univariate regression analysis is performed using Matlab to obtain the fitting correlation formula of the optimal operating parameters of the centrifugal turbine. The relationship between the nozzle outlet angle of the adjustable nozzle centrifugal turbine and the heat source temperature is shown in the attached figure. Figure 5 As shown, it can be seen that the nozzle outlet angle and the heat source temperature are in a power function relationship, expressed as α3=68667.9T h -1.62 .
Claims
1. An optimal control method for a centripetal turbine with an adjustable nozzle, It is characterized in that The steps include: (1) Input the design conditions of the organic Rankine cycle power generation system, including the heat source temperature and flow rate, the cold source temperature and flow rate, and the working fluid pump characteristic curve; (2) Establish a one-dimensional design mathematical model for the centripetal turbine, couple the candidate cycle working fluid and turbine mathematical model into the system model, and use the system thermal efficiency (η E ), average generation cost (LEC), and turbine size factor (DF) as targets, a mathematical model for one-dimensional integrated optimization design of the system is established; (3) First, the NSGA-II algorithm is used to iteratively calculate the one-dimensional integrated optimization model to obtain the Pareto frontier solution set of multi-objective optimization, and then multi-objective decision-making is performed to obtain the optimal solution; (4) Using the optimal solution obtained through decision making, the evaporator, condenser, centrifugal turbine, and working fluid pump in the system are designed to obtain the evaporator and condenser areas, working fluid pump head and flow rate, and centrifugal turbine structural parameters; (5) Taking the main aerodynamic parameters of the centrifugal turbine as variables, the central composite design method was used to design the test scheme, and a three-dimensional simulation model of each test scheme was established. The centrifugal turbine performance of each scheme was obtained using the CFD method; (6) The simulation results of each scheme are used as the training set of the PSO-BP neural network to obtain the prediction model of the variable operating performance of the centrifugal turbine; (7) Establish the variable operating condition mathematical model of the evaporator, condenser and working fluid pump, and integrate the centripetal turbine PSO-BP neural network model with the variable operating condition models of other components to establish the variable operating condition mathematical model of the organic Rankine cycle power generation system; (8) Based on the system thermal efficiency (η E ), average electricity generation cost (LEC) and net output power (W net ) as the target, a multi-objective optimization model of system variable working conditions is established: (9) The NSGA-II algorithm is used to optimize the variable operating condition model, obtain the Pareto solution set and make decisions on it, determine the optimal operating parameters of the centrifugal turbine under different cold and heat source conditions, and obtain the correlation between the optimal operating parameters of the centrifugal turbine and the cold and heat source parameters.
2. The optimal control method of an adjustable nozzle centrifugal turbine according to claim 1, Features: This optimal control method first couples the candidate circulating working fluid and the centripetal turbine model to the organic Rankine cycle power generation system, carries out multi-objective optimization design of the system, and then carries out system variable operating condition optimization based on the component structural parameters obtained from the optimization design, so as to achieve the optimal matching of the centripetal turbine operating parameters with the cold and heat source conditions, and improve the variable operating condition performance of the system.
3. The optimal control method for an adjustable nozzle centrifugal turbine according to claim 1, Features: In the step (2), the circulating working fluid (N f ), system thermal parameters (relative evaporator temperature R Te , Evaporator superheat ΔT e,s , condensation dew point temperature T c , evaporator pinch point temperature ΔT e,pp , Condenser pinch point temperature difference ΔT c,pp ) and turbine aerodynamic parameters (reaction degree Ω, wheel diameter ratio σ, speed ratio Φ, impeller inlet absolute flow angle α 4 , outlet relative airflow angle β 5 , absolute airflow angle α at nozzle outlet 3 , turbine speed N) as the decision variables of multi-objective optimization design, so that the working fluid screening and turbine design are integrated into the system optimization design. Through optimization calculation, the optimal working fluid and the optimal structural parameters of system components can be obtained at the same time; The relative evaporation temperature is defined as follows: Where, T e is the evaporation temperature, T c,f is the critical temperature of the working fluid; The definition of the objective function is as follows: Where W net is the net work of the system, Q e is the heat absorbed by the system, r n is the turbine nozzle inlet radius, W t is the output power of the centrifugal turbine, CRF is the cost recovery factor, C s is the system investment cost, OP s is the annual operating time of the system; The constraints considered in the multi-objective optimization model are as follows:
4. The optimal control method for an adjustable nozzle centrifugal turbine according to claim 1, Features: In step (3), the entropy method is used to determine the weights of different objective functions, and then the Topsis method is used to determine the optimal solution in the Pareto frontier solution set, so that the selection of the optimal solution is more reasonable. The basic idea of this method is as follows: First, determine the weight coefficients of each objective function: In the formula, x ij is the dimensionless processing value of the jth indicator in the ith sample, m is the number of individuals in the optimal solution set, n is the number of objective functions, and w j is the weight coefficient of the jth indicator, E j is the entropy value of the j-th indicator; Then, calculate the distances of each solution individual to the ideal point and the negative ideal point: In the formula, f i ideal is the positive ideal point function value of the jth objective, f i nadir is the negative ideal point function value of the jth objective; Finally, the TOPSIS decision method is used to calculate the relative distance Y from each point on the Pareto frontier solution set to the negative ideal point i , and Y i The point with the largest value is taken as the optimal solution, and its expression is:
5. The optimal control method of an adjustable nozzle centripetal turbine according to claim 1, Features: In step (5), the system converts the centrifugal turbine inlet pressure (p t,in )、Outlet pressure (p t,out ), nozzle angle (α 3 ) and inlet temperature (T t,in ) and other four parameters as adjustment variables, and these four parameters are used as design variables of the central composite design method. This method is used to determine more than 70 experimental schemes, so that the design variables of each scheme are evenly distributed in the variable range.
6. The optimal control method of an adjustable nozzle centripetal turbine according to claim 1, Features: In the step (6), in order to improve the prediction accuracy of the neural network, the PSO optimization algorithm is used to determine the weight and threshold of the BP neural network; the PSO-BP neural network takes the four design variables in the central composite design as input and the output power, turbine efficiency and working fluid flow of the centripetal turbine as output, and establishes a PSO-BP neural network prediction model for the variable operating performance of the centripetal turbine, wherein the number of layers of neurons in the middle hidden layer of the neural network is determined by the following formula: Where n 1 is the number of nodes in the input layer, n 2 is the number of nodes in the output layer, and a is a constant between 1 and 10.
7. The optimal control method for an adjustable nozzle centripetal turbine according to claim 1, Features: In the step (7), a variable operating condition mathematical model of the working fluid pump is established by similarity theory: In the formula, N represents the speed, q is the volume flow rate, H is the head, m is the mass flow rate, η is the efficiency, the subscript off is the variable working condition, d is the design working condition, and p is the pump; When the evaporator and condenser operate under variable conditions, first assume the outlet temperature of the heat exchanger, and then calculate the required area of the heat exchanger based on the heat source and working fluid parameters. If the calculated area is inconsistent with the actual area of the heat exchanger, adjust the outlet temperature of the heat exchanger and continue the iterative calculation until the requirements are met.
8. The optimal control method for an adjustable nozzle centripetal turbine according to claim 1, Features: In the step (9), the importance sequence of the cold and heat source conditions to the system performance is determined by the range analysis method, and then the correlation between the optimal operating parameters of the centripetal turbine and the key cold and heat source parameters is analyzed, and the correlation formula between the optimal operating parameters and the key cold and heat source parameters is determined by multivariate regression.
Citation Information
Cited By
1D-3D centripetal turbine collaborative optimization method
CN120911217A
1d-3d centripetal turbine collaborative optimization method
CN120911217B
Aerodynamic thermal design and test feedback correction method for centripetal steam turbine
CN122616439A
