A maximum static pressure intelligent prediction and surge early warning method for axial flow fan based on coupling of maximum efficiency point static pressure and impeller linear speed

Abstract

The application discloses a kind of based on maximum efficiency point static pressure and impeller linear velocity coupling axial flow fan maximum static pressure intelligent prediction and surge early warning method, to solve the problems of traditional maximum static pressure acquisition high cost, low prediction accuracy, lack of integrated surge protection.This method first collects fan maximum efficiency point static pressure, impeller diameter, rated speed basic parameters, calculates impeller tip linear velocity;Build the introduction static pressure quadratic term and static pressure-linear velocity interactive coupling term multivariate regression prediction model, quickly calculate maximum static pressure;Through pressure reserve coefficient, complete the effectiveness of the prediction result verification, then combined with real-time operation static pressure quantitative safety margin, realize hierarchical surge early warning.The model determination coefficient R of the application reaches 0.9718, and the average absolute percentage error is only 7.04%, without wind tunnel test and CFD simulation, just rely on conventional parameters to realize dynamic prediction, with prediction, verification, margin evaluation and early warning function, greatly improve the safety and operation efficiency of fan operation.

Description

Technical Field

[0001] This invention belongs to the field of fluid machinery technology, specifically relating to a method for predicting the maximum static pressure of axial flow fans, assessing operational safety margins, and providing intelligent early warning of surge. It is particularly applicable to the design, selection, operation monitoring, and preventive maintenance of axial flow fans in the fields of ventilation, air conditioning, and industrial cooling. Background Technology

[0002] Axial flow fans are the core power components of ventilation, industrial heat dissipation, and building ventilation systems. The maximum static pressure, as the ultimate static pressure parameter under low flow conditions, directly determines the safe operating range, surge boundary, and aerodynamic design margin of the fan. It is a key indicator for fan selection, stability control, and fault early warning.

[0003] In the current project, the acquisition of maximum static pressure mainly relies on whole-machine aerodynamic performance tests and CFD numerical simulations: whole-machine performance tests require the construction of a professional wind tunnel test rig, which has a long test cycle, high manpower and material costs, and cannot achieve dynamic prediction during operation; CFD simulations are highly dependent on mesh generation, turbulence models, and boundary conditions, which takes a long time to calculate, and the calculation results of different modeling methods have large deviations, resulting in poor real-time performance in engineering; the maximum static pressure point is close to the surge zone, and CFD simulations often fail to converge or have large calculation deviations.

[0004] Existing empirical prediction formulas mostly use rotational speed and impeller diameter as single input variables, failing to utilize the static pressure at the highest efficiency point—a core operating parameter directly representing the aerodynamic work capacity of the fan—and neglecting the coupled interaction between impeller linear velocity and static pressure level. These formulas can only achieve coarse linear estimations with low prediction accuracy, failing to meet the refined requirements of surge warning and EC (Electronic Control System) fan intelligent control. Furthermore, current technologies can only perform numerical predictions of static pressure, lacking an integrated solution for combining prediction results with quantitative assessment of operational safety margins and graded surge warning and linkage control. Summary of the Invention

[0005] To address the shortcomings of existing technologies, such as low accuracy in maximum static pressure prediction, high acquisition cost, inability to dynamically predict, and lack of safety verification and surge warning linkage, this invention provides an intelligent prediction and surge warning method for the maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity. Using the static pressure at the highest efficiency point, impeller diameter, and rotational speed as input parameters, a high-precision prediction model is constructed by introducing a quadratic term for static pressure and an interactive coupling term between static pressure and linear velocity. This enables rapid prediction of maximum static pressure, verification of the validity of prediction results, assessment of operational margin, and graded surge warning, solving the technical problems of insufficient accuracy, poor adaptability to operating conditions, and inability to link protection in traditional methods.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A method for intelligent prediction of maximum static pressure and surge warning of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity includes the following steps: S1: Parameter acquisition, acquiring the basic parameters of the target axial flow fan, including the static pressure P at the highest efficiency operating point. fs Impeller diameter D, rated speed n; S2: Impeller linear velocity calculation, based on the impeller diameter and rated speed, calculates the impeller tip linear velocity: u tip = π × D × n / 60; In the formula: u tip π is the impeller tip linear velocity in m / s; D is the impeller diameter in m; and n is the fan speed in r / min. S3: Coupled Prediction Model Construction and Maximum Static Pressure Prediction. A multivariate regression prediction model is constructed, including a quadratic term for static pressure, a linear velocity term, and a static pressure-linear velocity interactive coupling term. P... fs u tip Substitute into the model to calculate the maximum static pressure P of the fan. max ; The prediction model expression is: P max = α1 P fs ² + α² P fs + β1u tip + β2P fs u tip + γ In the formula: α1, α2, β1, β2, and γ are the regression coefficients of the model; P fs ² represents the static pressure quadratic term at the point of highest efficiency; P fs u tip This is an interactive coupling term between static pressure and linear velocity; S4: Verify the effectiveness of the prediction, calculate the pressure reserve coefficient K=P max / P fs The validity of the prediction result is determined. If K∈[0.95,1.35], the prediction result is valid. If it exceeds the range, it is determined that there is a measurement error or abnormal working condition. S5: Operational safety margin assessment and surge classification early warning. It collects the real-time operating static pressure of the wind turbine, combines it with the predicted maximum static pressure to quantify the safety margin, sets classification early warning thresholds, and realizes surge protection and EC wind turbine preventive maintenance.

[0007] Preferably, in step S3, the regression coefficients of the model are: α1 = 8.47 × 10⁻ 4Pa⁻¹, α2 = 1.0799, β1 = 0.7242 Pa·s / m, β2 = -5.94×10⁻³, γ = -8.38 Pa. .

[0008] Preferably, the physical mechanism of the prediction model is as follows: (1) Quadratic term α1P fs ²: Characterizes the nonlinear accelerated growth of maximum static pressure with respect to static pressure at the highest efficiency point, reflecting the increasing pressure reserve caused by the increase in aerodynamic load of the fan under high-pressure conditions. (2) Linear term α2 P fs The basic linear proportional relationship between the maximum static pressure and the static pressure at the highest efficiency point represents the stable characteristics of the pressure reserve coefficient under low-pressure conditions. (3) Linear velocity term β1u tip : Characterizes the positive correction effect of impeller rotation speed on static pressure reserve; the higher the linear velocity, the greater the theoretical pressure limit of the fan. (4) Interaction coupling term β2P fs u tip It characterizes the secondary flow loss and the static pressure reserve suppression effect caused by the tip clearance effect under high static pressure and high linear velocity conditions, and achieves accurate correction under different operating conditions.

[0009] Preferably, the method for constructing the prediction model includes: S301: Data acquisition, collecting experimental datasets of axial flow fans of various specifications, with impeller diameters of 250~1250mm, rotation speeds of 450~3320r / min, and static pressures at the highest efficiency point of 37~537Pa. S302: Data preprocessing, calculate the pressure reserve coefficient K for each sample group, remove outliers with K∉[0.95,1.35], and retain valid samples; S303: Multiple regression modeling, using P fs ²、P fs u tip P fs u tip The maximum static pressure P is the independent variable. max For the dependent variable, multiple linear regression was used to fit the model coefficients; S304: Model accuracy verification, model determination coefficient R0 2 =0.9718, mean absolute percentage error (MAPE) = 7.04%, 73.1% of data deviation is within ±10%, and 94.1% of data deviation is within ±15%.

[0010] Preferably, the prediction model is applicable to the following operating conditions: static pressure P at the highest efficiency point. fs =37~537Pa, impeller linear velocity u tip=20~70m / s.

[0011] As a preferred option, the specific strategy for the graded early warning in step S5 is as follows: S501: Margin calculation, calculating the ratio η=P between real-time static pressure and predicted maximum static pressure. real / P max ; S502: Level 1 warning. When η≥95%, the controller issues a surge warning, indicating that the system resistance is approaching the safety boundary. S503: Level 2 warning. When η≥105%, the controller triggers a surge alarm.

[0012] Preferably, the method is applied to the safety range estimation during the wind turbine design phase, and the aerodynamic design margin is determined by the pressure reserve coefficient K to determine whether the safety requirements are met.

[0013] Preferably, the method is applied to the intelligent preventive maintenance of EC wind turbines, where the EC controller dynamically collects operating parameters and updates the maximum static pressure prediction value and safety margin in real time to achieve closed-loop surge protection.

[0014] Preferably, the regression coefficients of the model can be optimized by partitioning according to the fan size, static pressure range, and speed range to achieve precise adaptation to fans of different specifications.

[0015] The beneficial effects of this invention are: (1) Significantly improved prediction accuracy: This invention introduces a static pressure quadratic term and a static pressure-linear velocity interactive coupling term to accurately capture the nonlinear aerodynamic characteristics of the wind turbine and the coupling effect of the operating conditions. The model determination coefficient R^2=0.9718, and the average absolute percentage error is only 7.04%, which is far superior to the traditional linear prediction model. The data within ±10% deviation accounts for 73.1%, which meets the high-precision prediction requirements of engineering; (2) Easy to obtain parameters and high prediction efficiency: Relying only on the three basic parameters of the wind turbine's conventional test, there is no need to build a test bench or carry out complex CFD simulation. The prediction can be completed in a few seconds, and dynamic real-time estimation can be realized during operation; (3) Wide range of applicable operating conditions :Covering low-pressure to high-pressure, small to large axial flow fans, applicable to static pressure 37~537Pa and linear velocity 20~70m / s, with strong versatility, suitable for ventilation, cooling and air conditioning fan scenarios in multiple fields; (4) Complete integrated functions: integrates prediction, effectiveness verification, margin assessment and graded surge warning functions, can be directly connected to EC fan controller to achieve closed-loop control, no need to install additional monitoring sensors, reduce operation and maintenance costs, and improve the safety and service life of fan operation; (5) Physical mechanism fits reality: the parameters of the model have clear aerodynamic physical meaning, can explain the static pressure reserve change law under different working conditions, which is different from the pure black box fitting model, and has stronger reliability and robustness. Attached Figure Description

[0016] Figure 1 Relationship between static pressure at the highest efficiency point and maximum static pressure, and prediction curves for different linear velocities; Figure 2 A comparison plot of predicted and actual values; Figure 3 Prediction deviation distribution and deviation variation with static pressure; Figure 4 Relationship between pressure reserve coefficient and linear velocity, and curve of predicted pressure reserve coefficient; Figure 5 A bar chart comparing the actual and predicted values ​​of stratified validation samples; Figure 6 Flowchart of the overall method implementation of this invention. Detailed Implementation

[0017] The technical solution of the present invention will be further described in detail below through specific embodiments and with reference to the accompanying drawings: Example 1: Model Building and Data Filtering Combined with appendix Figure 1 and attached Figure 4 As shown, 677 sets of axial flow fan performance test data were collected, with impeller diameters ranging from 250 to 1250 mm, rotational speeds from 450 to 3320 r / min, static pressures at the highest efficiency point ranging from 37 to 537 Pa, and maximum static pressures ranging from 40 to 750 Pa.

[0018] 1. Data screening: Calculate the pressure reserve coefficient K=P for each sample group. max / P fs Outliers with K < 0.95 or K > 1.35 were removed, and 580 valid samples were retained. 2. Static pressure P at the point of highest efficiency fs Impeller linear velocity utip and its interaction term P fs u tip The maximum static pressure P is the independent variable. max Using [variable name] as the dependent variable, a predictive model was established using multiple linear regression. The regression analysis results show that: Table 1 shows the comparison results of different prediction models: As shown in Table 1, the model proposed in this invention, which includes the quadratic term of Pfs, linear velocity, and their interaction correction term, has the highest coefficient of determination (R² = 0.9718) and the lowest mean absolute percentage error (MAPE = 7.04%), with 73.1% of data points falling within ±10%, indicating that this model has the best predictive effect on the maximum static pressure of axial flow fans. In particular, the introduction of the interaction term β²×Pfs×u_tip enables the model to capture the coupling effect between static pressure level and linear velocity. Compared with the model without the interaction term, the MAPE is reduced by 0.1 percentage points, and the proportion of data points within ±10% is increased by 1.1 percentage points.

[0019] Example 2: Accuracy Verification of Stratified Samples Combined with appendix Figure 2 Appendix Figure 3 and attached Figure 5 As shown in Table 2, 10 groups of low, medium, and high-pressure operating condition samples were selected from the valid samples for verification. As shown in Table 2, among the 10 validation samples, 7 samples had prediction deviations within ±10%, and 9 samples had deviations within ±15%. Samples with larger deviations were mainly concentrated in the region of medium static pressure and high linear velocity (such as samples 4 and 9), which is related to the rapid change in the slope of the wind turbine performance curve in this region. Statistical analysis of the entire dataset shows that 94.1% of the data points had deviations within ±15%, and 73.1% were within ±10%, which meets the prediction requirements in engineering design.

[0020] Example 3: Static Pressure Prediction for Typical Axial Flow Fans Combined with appendix Figure 6 As shown, taking a certain type of axial flow fan as an example, its static pressure P at the highest efficiency operating point is... fs =200Pa, impeller diameter D=0.8m, rotational speed n=1000rpm. The prediction steps are as follows: (1) Calculate the linear velocity: u tip = π × 0.8 × 1000 / 60 = 41.89 m / s (2) Calculate the interaction term: P fs × u tip = 200 × 41.89 = 8378 (3) Substitute into the prediction formula: P max = 8.47 × 10⁻ 4 × 200² + 1.0799 × 200 + 0.7242 ×41.89 + (-5.94×10⁻³) × 8378 + (-8.38) = 33.88 + 215.98 + 30.37 - 49.77 -8.38 = 222.08 Pa (4) The predicted maximum static pressure of the fan is 222 Pa, and the pressure reserve coefficient K = 222 / 200 = 1.11, indicating that the fan has a static pressure reserve of about 11% when operating at the highest efficiency point, and the operating point is a certain margin away from the surge boundary.

[0021] Example 4: EC wind turbine surge early warning application Combined with appendix Figure 6 As shown, taking a certain EC axial flow fan as an example, its P is known. fs=150Pa, D=0.63m, n=1200rpm. First, the maximum static pressure is predicted using this method: u tip = π × 0.63 × 1200 / 60 = 39.58 m / s; P fs u tip = 150 × 39.58 = 5937; P max = 8.47 × 10⁻ 4 × 150² + 1.0799 × 150 + 0.7242 × 39.58 +(-5.94×10⁻³) × 5937 + (-8.38) = 19.06 + 161.99 + 28.66 - 35.27 - 8.38 =166.06 Pa.

[0022] The controller of the EC wind turbine can be set with warning and alarm thresholds. For example, a warning is triggered when the real-time static pressure reaches 95% of the predicted maximum static pressure (approximately 158 Pa), indicating that the system resistance is approaching the safety boundary; an alarm is triggered when the real-time static pressure reaches 105% of the predicted maximum static pressure (approximately 175 Pa). This intelligent protection strategy based on maximum static pressure prediction enables preventative maintenance, significantly improving the operational safety and service life of the EC wind turbine.

[0023] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Other variations and modifications are possible without departing from the technical solutions described in the claims.

Claims

1. A method for intelligent prediction of maximum static pressure and surge warning of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity, characterized in that, Includes the following steps: S1: Parameter acquisition, acquiring the basic parameters of the target axial flow fan, including the static pressure P at the highest efficiency operating point. fs Impeller diameter D, rated speed n; S2: Impeller linear velocity calculation, based on the impeller diameter and rated speed, calculates the impeller tip linear velocity: the tip = π × D × n / 60; In the formula: u tip π is the impeller tip linear velocity in m / s; D is the impeller diameter in m; and n is the fan speed in r / min. S3: Coupled Prediction Model Construction and Maximum Static Pressure Prediction. A multivariate regression prediction model is constructed, including a quadratic term for static pressure, a linear velocity term, and a static pressure-linear velocity interactive coupling term. P... fs u tip Substitute into the model to calculate the maximum static pressure P of the fan. max ; The prediction model expression is: P max = α1 P fs ² + α2 P fs + β1u tip + β2P fs you tip + c In the formula: α1, α2, β1, β2, and γ are the regression coefficients of the model; P fs ² represents the static pressure quadratic term at the point of highest efficiency; P fs u tip This is an interactive coupling term between static pressure and linear velocity; S4: Verify the effectiveness of the prediction, calculate the pressure reserve coefficient K = P max / P fs The validity of the prediction result is determined. If K∈[0.95,1.35], the prediction result is valid. If it exceeds the range, it is determined that there is a measurement error or abnormal working condition. S5: Operational safety margin assessment and surge classification early warning. It collects the real-time operating static pressure of the wind turbine, combines it with the predicted maximum static pressure to quantify the safety margin, sets classification early warning thresholds, and realizes surge protection and EC wind turbine preventive maintenance.

2. The intelligent prediction and surge warning method for maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity as described in claim 1, is characterized in that, In step S3, the regression coefficients of the model are taken as follows: α1 = 8.47×10⁻ 4 Pa⁻¹, α2 = 1.0799, β1 = 0.7242 Pa·s / m, β2 = -5.94×10⁻³, γ = -8.38 Pa.

3. The intelligent prediction and surge warning method for maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity as described in claim 1, is characterized in that... The physical mechanism of the prediction model is as follows: (1) Quadratic term α1P fs ²: Characterizes the nonlinear accelerated growth of maximum static pressure with respect to static pressure at the highest efficiency point, reflecting the increasing pressure reserve caused by the increase in aerodynamic load of the fan under high-pressure conditions. (2) Linear term α2 P fs The basic linear proportional relationship between the maximum static pressure and the static pressure at the highest efficiency point represents the stable characteristics of the pressure reserve coefficient under low-pressure conditions. (3) Linear velocity term β1u tip : Characterizes the positive correction effect of impeller rotation speed on static pressure reserve; the higher the linear velocity, the greater the theoretical pressure limit of the fan. (4) Interaction coupling term β2P fs u tip It characterizes the secondary flow loss and the static pressure reserve suppression effect caused by the tip clearance effect under high static pressure and high linear velocity conditions, and achieves accurate correction under different operating conditions.

4. The intelligent prediction and surge warning method for maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity as described in claim 1, is characterized in that, The method for constructing the prediction model includes: S301: Data acquisition, collecting experimental datasets of axial flow fans of various specifications, with impeller diameters of 250~1250mm, rotation speeds of 450~3320r / min, and static pressures at the highest efficiency point of 37~537Pa. S302: Data preprocessing, calculate the pressure reserve coefficient K for each sample group, remove outliers with K∉[0.95,1.35], and retain valid samples; S303: Multiple regression modeling, using P fs ²、P fs u tip P fs u tip The maximum static pressure P is the independent variable. max For the dependent variable, multiple linear regression was used to fit the model coefficients; S304: Model accuracy verification, model determination coefficient R0 2 =0.9718, mean absolute percentage error (MAPE) = 7.04%, 73.1% of data deviation is within ±10%, and 94.1% of data deviation is within ±15%.

5. The intelligent prediction and surge warning method for maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity as described in claim 1, characterized in that, The predictive model is applicable to the following operating conditions: static pressure P at the highest efficiency point. fs =37~537Pa, impeller linear velocity u tip =20~70m / s.

6. The intelligent prediction and surge warning method for maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity as described in claim 1, is characterized in that, The specific strategy for the graded early warning in step S5 is as follows: S501: Margin calculation, calculating the ratio η=P between real-time static pressure and predicted maximum static pressure. real / P max ; S5 02: Level 1 warning: When η≥95%, the controller issues a surge warning, indicating that the system resistance is approaching the safety boundary; S503: Level 2 warning. When η≥105%, the controller triggers a surge alarm.

7. The intelligent prediction and surge warning method for maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity as described in claim 1, characterized in that, The method is applied to the safety range estimation in the wind turbine design phase, and the pressure reserve coefficient K is used to determine whether the aerodynamic design margin meets the safety requirements.

8. The intelligent prediction and surge warning method for maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity as described in claim 1, characterized in that, The method is applied to the intelligent preventive maintenance of EC wind turbines. The EC controller dynamically collects operating parameters and updates the maximum static pressure prediction value and safety margin in real time to achieve closed-loop surge protection.

9. The intelligent prediction and surge warning method for maximum static pressure of axial flow fans based on the coupling of static pressure at the highest efficiency point and impeller linear velocity as described in claim 1, characterized in that, The regression coefficients of the model can be optimized by partitioning according to the fan size, static pressure range, and speed range to achieve precise adaptation to fans of different specifications.

Images