Centrifugal pump self-adaptive regulation and control system and method based on virtual sensor

Abstract

The invention relates to the technical field of centrifugal pumps, in particular to a centrifugal pump self-adaptive regulation and control system and method based on a virtual sensor, and is characterized in that the system comprises a data acquisition module for acquiring current, voltage, frequency and rotating speed signals of a centrifugal pump in real time; a sensor model is virtualized, a hydraulic characteristic database of the centrifugal pump is established, and the current lift, flow and pipe network resistance coefficient are calculated according to motor parameters, collected in real time, of the centrifugal pump; the self-adaptive adjustment module introduces a feedforward-feedback composite control model, and when pipe network resistance sudden change is detected, the system adjusts the output frequency of the frequency converter in advance through model prediction control; and the tail end constant-pressure module automatically recognizes the working condition of the most unfavorable point, and the output characteristic curve of the pump is adjusted so that the output characteristic curve and the pipe network characteristic curve can be always intersected at a set efficient working point. According to the invention, a mathematical model of motor parameters and pump working conditions is established to replace a physical sensor, so that accurate and rapid constant pressure control is realized.

Description

Technical Field

[0001] This invention relates to the technical field of centrifugal pumps, and specifically to an adaptive control system and method for centrifugal pumps based on virtual sensors. Background Technology

[0002] In existing variable frequency speed control centrifugal pump systems, physical pressure sensors or flow meters installed on the inlet and outlet pipelines typically rely on feedback signals to adjust the motor speed. However, physical sensors suffer from problems such as susceptibility to media corrosion, signal transmission delays, complex installation and maintenance, and long-term operational drift, leading to decreased control accuracy and sluggish system response. Furthermore, when pipeline resistance changes due to valve adjustments or equipment aging, traditional PID control struggles to adapt quickly, easily causing over-adjustment or under-adjustment, resulting in increased energy consumption. Summary of the Invention

[0003] To address the technical problems and shortcomings of existing technologies, this invention provides a centrifugal pump adaptive control method based on virtual sensing. By establishing a mathematical model of motor parameters and pump operating conditions, physical sensors are replaced to achieve precise and rapid constant pressure control.

[0004] To achieve the above and other related objectives, the present invention adopts the following technical solution: An adaptive control system for centrifugal pumps based on virtual sensors, comprising: The data acquisition module collects the current, voltage, frequency, and speed signals of the centrifugal pump in real time. A virtual sensor model is used to establish a database of the hydraulic characteristics of centrifugal pumps. Based on the real-time collected centrifugal pump motor parameters, the current head, flow rate, and pipeline resistance coefficient are calculated. The adaptive adjustment module introduces a feedforward-feedback composite control model. When a sudden change in pipeline resistance is detected, the system uses model predictive control to adjust the inverter output frequency in advance. The terminal constant pressure module automatically identifies the most unfavorable operating conditions and adjusts the pump's output characteristic curve to ensure that it always intersects with the pipeline characteristic curve at the set high-efficiency operating point.

[0005] Preferably, the data acquisition module uses a Kalman filter-based data fusion algorithm to perform weighted fusion of multi-source sensor data and principal component analysis to provide acquired data.

[0006] Preferably, it also includes a soft measurement model. First, a large amount of historical operating data is obtained through experiments to establish a mapping relationship between input and output. Second, machine learning algorithms are used to train the data to generate a prediction model. Finally, cross-validation and parameter optimization are used to ensure the generalization ability and prediction accuracy of the model.

[0007] Preferably, the adaptive adjustment module employs an adaptive algorithm based on a second-order linear active disturbance rejection controller. First, in the parameter initialization phase, initial values ​​are set according to the nominal operating parameters of the centrifugal pump, including rated flow and head, and the main parameters of the controller are determined. Second, in the model update phase, the system state is estimated by real-time acquisition of sensor data and using extended Kalman filtering, and the model parameters are updated based on the estimation results. Finally, in the control quantity calculation phase, the optimal control quantity is solved using an optimization algorithm based on the updated model parameters and setpoints.

[0008] Preferably, the establishment of the prediction model includes the following steps: First, by analyzing the working principle and operating characteristics of the centrifugal pump, its mathematical description equation is established, using a nonlinear state-space model or a transfer function model; Second, the model parameters are identified and optimized using actual operating data to ensure that the model can accurately reflect the dynamic behavior of the system.

[0009] Preferably, the prediction model includes a rolling optimization step: First, the system output for a future period is calculated based on the prediction model at the current time; second, an optimization objective function is defined, which typically includes a weighted combination of indicators such as control deviation, energy consumption, and control quantity change rate; finally, the optimal control sequence is obtained by solving the optimization problem, and the first control quantity is applied to the actual system; the above process is repeated in the next control cycle.

[0010] Preferably, the process includes the following steps for adjusting the inverter output frequency: First, the required speed setpoint is determined based on the optimal control sequence calculated by the predictive model; second, the setpoint is converted into the inverter output frequency, and a proportional-integral (PI) controller is used for closed-loop regulation to ensure that the actual output frequency can quickly track the setpoint; finally, the output frequency is finely adjusted by monitoring the actual operating status of the centrifugal pump in real time to compensate for load changes.

[0011] On the other hand, a centrifugal pump adaptive control method based on virtual sensors is also provided, which is operated using the aforementioned centrifugal pump adaptive control system.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Operation without physical hardware sensors: Eliminating vulnerable physical sensors reduces hardware costs and maintenance difficulty, and improves system reliability.

[0013] 2. Fast response speed: The response speed of the algorithm model is much faster than the signal transmission and sampling cycle of physical sensors, and the dynamic response time is greatly shortened.

[0014] 3. Energy-saving and efficient: By accurately matching the needs of the pipeline network, unnecessary high-frequency operation is avoided, and the overall energy saving rate is significantly improved.

[0015] 4. High versatility: This system and method are applicable to centrifugal pump units of different diameters and heads, requiring only the loading of the corresponding hydraulic model parameters.

[0016] Other additional advantages and benefits of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the system module architecture of an embodiment of this application.

[0018] Explanation of reference numerals for major components: 100. Data acquisition module; 200. Central controller; 201. Virtual sensor model; 202. Adaptive adjustment module; 203. Terminal constant pressure module; 204. Soft measurement model; 205. Predictive model; 300. Frequency converter. Detailed Implementation

[0019] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. The following specific examples illustrate the embodiments of the present invention, and those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0020] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. The illustrations only show the components related to the present invention and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be changed at will, and the layout of the components may also be more complex.

[0021] It should be noted that in the description of this application, the terms "upper," "lower," "left," "right," "inner," and "outer," etc., indicating directional or positional relationships, are based on the directional or positional relationships shown in the accompanying drawings. These are merely for ease of description and do not indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the invention. Furthermore, it should be noted that in the description of this application, unless otherwise explicitly specified and limited, the terms "installed," "connected," "linked," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two elements. Those skilled in the art can understand the specific meaning of the above terms in the invention based on the specific circumstances.

[0022] Example 1: This invention discloses an adaptive control system for a centrifugal pump based on virtual sensors, comprising: The data acquisition module 100 collects the current, voltage, frequency, and speed signals of the centrifugal pump in real time; the virtual sensor model 201 establishes a hydraulic characteristic database of the centrifugal pump and calculates the current head, flow rate, and pipeline resistance coefficient based on the real-time collected centrifugal pump motor parameters; the adaptive adjustment module 202 introduces a feedforward-feedback composite control model, and when a sudden change in pipeline resistance is detected, the system uses model predictive control to adjust the output frequency of the frequency converter 300 in advance; the terminal constant pressure module 203 automatically identifies the most unfavorable operating condition and adjusts the pump's output characteristic curve to ensure that it always intersects with the pipeline characteristic curve at the set high-efficiency operating point.

[0023] In this solution, the virtual sensor model 201, adaptive adjustment module 202, terminal constant voltage module 203, soft measurement model 204, and predictive model 205 are hardware circuit modules or software program modules installed within the central controller 200. The central controller 200 is a PLC with data processing capabilities, or is typically integrated into a high-end frequency converter 300 (such as Siemens G120, ABB ACS880) or a dedicated industrial control computer (IPC).

[0024] Regarding the data acquisition module 100, it can specifically include temperature, pressure, and vibration sensors, and uses a Kalman filter-based data fusion algorithm to weightedly fuse multi-source sensor data and perform principal component analysis to provide acquired data. Specifically, sensor data fusion can improve the accuracy and reliability of centrifugal pump operating status perception. In practical applications, the operating status of centrifugal pumps is usually monitored by multiple sensors, including temperature, pressure, and vibration sensors. The data provided by these sensors have different spatiotemporal characteristics and measurement accuracies, thus requiring specific data fusion algorithms for processing. A typical method is a Kalman filter-based data fusion algorithm, which effectively reduces the impact of measurement noise on the results and improves the accuracy of state estimation by weighted fusion of multi-source sensor data. Furthermore, principal component analysis (PCA) is also widely used in sensor data fusion, reducing data redundancy by extracting key feature variables while retaining crucial information. Experimental results show that the data fusion method combining Kalman filtering and principal component analysis can significantly improve the reliability of centrifugal pump operating status monitoring, providing high-quality data support for subsequent adaptive control.

[0025] This solution includes a soft sensor model 204, whose construction is a crucial step in achieving accurate estimation of the centrifugal pump's operating status based on virtual sensing technology. The core technology of the soft sensor model 204 is to utilize easily measurable parameters (such as temperature and pressure) to estimate difficult-to-measurable parameters (such as flow rate and efficiency), thereby providing data support for virtual sensing. Specifically, the construction of the soft sensor model 204 typically involves the following steps: First, acquiring a large amount of historical operating data through experiments to establish a mapping relationship between input and output; second, training the data using machine learning algorithms (such as support vector machines or neural networks) to generate a prediction model 205; finally, ensuring the model's generalization ability and prediction accuracy through cross-validation and parameter optimization. For example, in centrifugal pump applications, by constructing a soft sensor model 204 based on a second-order linear active disturbance rejection controller, accurate estimation of the centrifugal pump's flow rate was successfully achieved. Experimental results show that the prediction error of this model is smaller than that of traditional methods, significantly improving the practicality of virtual sensing technology.

[0026] The adaptive adjustment module 202 employs an adaptive algorithm based on a second-order linear active disturbance rejection controller (ADRC). In the adaptive control of centrifugal pumps, selecting a suitable algorithm is a key factor in ensuring system performance. This scheme chooses an adaptive algorithm based on a second-order linear ADRC primarily because it possesses strong anti-interference capabilities and high control accuracy, enabling it to adapt to the complex dynamic characteristics of centrifugal pumps. The design principle of the second-order linear ADRC is based on frequency domain analysis methods. By introducing observers and compensation components, it effectively suppresses the impact of external disturbances and model uncertainties on the system. Furthermore, this algorithm can adjust control parameters online based on real-time operating data, thereby achieving rapid response and precise control of the centrifugal pump flow rate. Compared to traditional PI control algorithms, the second-order linear ADRC exhibits superior robustness and adaptability under nonlinear systems and time-varying parameter conditions, making it an ideal choice for the adaptive control of centrifugal pumps.

[0027] The adaptive algorithm's calculation process mainly includes three key steps: parameter initialization, model update, and control quantity calculation. First, in the parameter initialization stage, initial values ​​need to be set based on the centrifugal pump's nominal operating parameters (such as rated flow rate and head), and the main parameters of the controller, such as observer bandwidth and control gain, need to be determined. Second, in the model update stage, the algorithm estimates the system state using an extended Kalman filter (EKF) by acquiring sensor data in real time, and updates the model parameters based on the estimation results. This process not only considers the dynamic characteristics of the system but also corrects model errors through a feedback mechanism, thereby improving control accuracy. Finally, in the control quantity calculation stage, the algorithm solves for the optimal control quantity based on the updated model parameters and setpoints using an optimization algorithm. For example, in a centrifugal pump experiment, the above algorithm was implemented using the Matlab / Simulink simulation platform, and its effectiveness under different operating conditions was verified. Experimental results show that the adaptive algorithm based on a second-order linear active disturbance rejection controller can maintain the stability of flow control under load changes, and the control error is significantly reduced compared to traditional methods.

[0028] The core of model predictive control lies in establishing a predictive model 205 that can accurately describe the future operating state of a centrifugal pump. The construction of the predictive model 205 is typically based on the system's dynamic characteristics and historical operating data, achieved through mathematical modeling and system identification methods. In this scheme, the establishment of the predictive model 205 involves two main steps: First, by analyzing the working principle and operating characteristics of the centrifugal pump, its mathematical description equations are established, usually using a nonlinear state-space model or a transfer function model; second, the model parameters are identified and optimized using actual operating data to ensure that the model accurately reflects the dynamic behavior of the system. For example, in a centrifugal pump application case, by collecting flow rate, pressure, and speed data of the centrifugal pump under different operating conditions, a predictive model 205 based on least squares identification was established. Experimental results show that the prediction error of this model is less than 5% over a future period, providing a reliable basis for subsequent control decisions.

[0029] Predictive model 205 includes a rolling optimization step. The rolling optimization strategy is a crucial component of model predictive control, its core being the continuous optimization of the control sequence based on prediction results to achieve dynamic adjustment of the centrifugal pump's operating state. Specifically, the rolling optimization strategy is implemented through the following steps: First, the system output for a future period is calculated based on predictive model 205 at the current moment; second, an optimization objective function is defined, typically including a weighted combination of indicators such as control deviation, energy consumption, and the rate of change of control variables; finally, the optimal control sequence is obtained by solving the optimization problem, and the first control variable is applied to the actual system. In the next control cycle, the above process is repeated, thus achieving rolling optimization of the control. This strategy not only effectively addresses system uncertainties and external disturbances but also improves the system's response speed and control accuracy by adjusting the control sequence in real time. For example, in centrifugal pump applications, after adopting the rolling optimization strategy, the system's flow control error decreased by approximately 15%, while energy saving was improved by 10%.

[0030] Adjusting the output frequency of the inverter 300 is a crucial step in achieving precise flow control of the centrifugal pump based on model predictive control results. In this scheme, the adjustment process of the inverter 300's output frequency is as follows: First, the required speed setpoint is determined based on the optimal control sequence calculated by the predictive model 205. Second, the setpoint is converted into the output frequency of the inverter 300, typically using a proportional-integral (PI) controller for closed-loop regulation to ensure that the actual output frequency can quickly track the setpoint. Finally, the output frequency is fine-tuned by monitoring the actual operating status of the centrifugal pump in real time to compensate for the effects of load changes and other uncertainties. For example, in centrifugal pump experiments, comparing the output frequency adjustment effects of the inverter 300 under traditional PI control and model predictive control revealed that the latter exhibits higher control accuracy and stability under conditions of large flow fluctuations. Experimental data shows that after adopting model predictive control, the output frequency adjustment time of the inverter 300 is shortened by approximately 20%, the flow control error is reduced by approximately 12%, and the energy-saving effect is significantly improved.

[0031] Specifically, regarding sensor selection and placement, this implementation plan utilizes multiple types of sensors, including pressure sensors, flow sensors, and temperature sensors, to meet the needs of multi-dimensional monitoring of the centrifugal pump's operating status. The pressure sensor is a high-precision strain gauge sensor with a measurement range of 0-2.5 MPa and an accuracy class of 0.1%FS (full scale), capable of real-time acquisition of pressure data at the pump's inlet and outlet. The flow sensor is an electromagnetic flow meter with a measurement accuracy of ±0.5%, adaptable to flow variations under different operating conditions. Furthermore, the temperature sensor uses a PT100 resistance temperature detector (RTD), featuring high sensitivity and long-term stability, used to monitor temperature changes in the pump body and the medium. In terms of placement, the pressure sensors are installed on the inlet and outlet pipes of the centrifugal pump to ensure global perception of pressure distribution; the flow sensor is located downstream of the outlet pipe to avoid measurement errors caused by uneven flow velocity; and the temperature sensor is embedded inside the pump body, close to the bearing area, to capture temperature changes at key hot spots. Through this rational arrangement, these sensors collectively form an efficient hardware sensing network, providing reliable data support for the subsequent application of virtual sensing technology.

[0032] As the core component of the system, the controller plays a crucial role in data processing and decision-making. This solution selects an industrial-grade controller based on a PLC (Programmable Logic Controller), which boasts powerful computing capabilities, abundant I / O interfaces, and high reliability, enabling it to adapt to complex industrial environments. This controller supports multiple communication protocols, such as Modbus TCP and Profibus DP, facilitating data interaction with other devices. Simultaneously, its built-in high-speed processor can perform real-time analysis of data collected by sensors and generate control commands based on preset algorithms. Regarding actuators, the frequency converter 300 is a key device for regulating the centrifugal pump flow. This solution can also utilize a high-performance vector control frequency converter 300, with an output power range of 5.5-75kW, a frequency regulation accuracy of ±0.01Hz, and excellent dynamic response characteristics. The frequency converter 300 indirectly changes the centrifugal pump speed by adjusting the motor input voltage and frequency, thereby achieving precise control of flow rate and head. Furthermore, the frequency converter 300 also features multiple protection functions, including overcurrent, overvoltage, and undervoltage protection, effectively improving the system's safety and stability. The controller and the frequency converter 300 are connected via an RS-485 communication interface and use the Modbus RTU protocol for data transmission, ensuring rapid response and execution of control commands.

[0033] Example 2: Based on the above system, the centrifugal pump adaptive control system based on virtual sensing consists of multiple functional modules. These modules cooperate to complete the entire process from data acquisition to control output. The operation method is as follows: First, the data acquisition module 100 is responsible for receiving and preprocessing raw data from various sensors. This module supports multi-channel synchronous sampling, enabling high-speed acquisition of signals such as pressure, flow, and temperature, and removes noise interference through digital filtering algorithms to improve data quality. Second, the virtual sensing module uses a soft measurement model 204 to perform in-depth analysis of the acquired data. By fusing multiple easily measurable parameters, it estimates key variables that are difficult to measure directly, such as pump efficiency and vibration amplitude. The core of this module lies in its machine learning-based prediction algorithm, which can continuously optimize model parameters based on historical data to improve estimation accuracy. Third, the adaptive control module is the core functional module of the system. Its main task is to adjust the control strategy in real time based on the data provided by the virtual sensing module and the operating characteristics of the centrifugal pump. This module integrates a model predictive control algorithm, which can predict the pump's operating state over a future period and generate the optimal control sequence. In addition, the system also includes a human-computer interaction module and a fault diagnosis module. The former is used to visually display the operating status and parameter settings, while the latter uses anomaly detection algorithms to promptly detect potential problems and issue alarms, thereby ensuring the reliable operation of the system.

[0034] To ensure efficient collaboration among the software modules, a comprehensive communication mechanism was designed. The data acquisition module 100 and the virtual sensing module use shared memory for data transmission, a method characterized by low latency and high throughput, meeting the needs of applications with high real-time requirements. The virtual sensing module and the adaptive control module communicate via a message queue mechanism, encapsulating data in JSON format for easy parsing and processing between modules. This communication method not only improves the flexibility of data interaction but also enhances the system's scalability. Furthermore, the human-computer interaction module communicates with other modules via the TCP / IP protocol, using a standard HTTP RESTful API interface to support cross-platform access. The fault diagnosis module sends alarm information to other modules using a publish-subscribe pattern, ensuring timely handling of anomalies. The entire communication mechanism design fully considers data consistency, real-time performance, and reliability, providing a solid technical guarantee for the efficient operation of the system.

[0035] Regarding the hydraulic characteristics of centrifugal pumps, the principle for adjusting the output characteristic curve is as follows: The output characteristic curve of a centrifugal pump reflects its performance under different operating conditions, mainly including the flow-head curve, flow-shaft power curve, and flow-efficiency curve. The shape of these curves is closely related to the pump's design parameters, such as impeller diameter and rotational speed. By changing these parameters, the output characteristic curve of the centrifugal pump can be effectively adjusted to meet actual operating requirements. Specifically, when the rotational speed of a centrifugal pump changes, its flow rate and head will also change accordingly, and the change pattern conforms to the law of similarity. For example, under constant impeller diameter conditions, the flow rate is directly proportional to the rotational speed, while the head is directly proportional to the square of the rotational speed. Furthermore, changing the impeller diameter also has a significant impact on the characteristic curve; generally, reducing the impeller diameter leads to a decrease in flow rate and head, but efficiency may improve. This adjustment method provides a basis for optimizing the operation of centrifugal pumps under different operating conditions.

[0036] From a system perspective, the intersection of the pipeline characteristic curve and the centrifugal pump characteristic curve determines the pump's actual operating point. By adjusting the rotational speed or impeller diameter, the pump's characteristic curve can be altered, thereby adjusting the operating point to fall within its high-efficiency range. Studies show that centrifugal pumps operate most efficiently when the motor frequency is between 26-50Hz, especially when the flow rate is 0.7 to 1.3 times the rated flow. Therefore, properly adjusting the output characteristic curve can not only improve the centrifugal pump's operating efficiency but also reduce energy consumption and extend equipment lifespan.

[0037] Regarding adjustment methods, in practical applications, the main approaches to adjusting the output characteristic curve of a centrifugal pump include changing the rotational speed and adjusting the impeller diameter. Among these, adjusting the motor speed using a frequency converter 300 is a common and efficient method. The frequency converter 300 achieves precise control of the motor speed by changing the frequency of the input power supply, thereby indirectly adjusting the characteristic curve of the centrifugal pump. In a centrifugal pump system, installing a frequency converter 300 and optimizing system friction significantly improves the pump's operating characteristics. Experimental results show that after reducing system friction, the centrifugal pump's characteristic curve is closer to the theoretical design value, and the system stability is significantly improved.

[0038] Taking a specific case as an example, suppose a centrifugal pump operates at its rated speed with a flow rate of 4.0 m³ / s. 3 / h, with a head of 30m. If the flow rate needs to be reduced to 3.0m... 3 The motor speed can be reduced to 75% of its original speed using a frequency converter 300. According to the similarity law, the head will decrease to 22.5m, and the shaft power will also decrease accordingly, thus achieving precise adjustment of the output characteristic curve. Furthermore, by combining this with changes in the pipeline characteristic curve, the operating point can be further optimized, ensuring the centrifugal pump always operates at high efficiency. This method is not only simple to operate but also highly flexible and economical, and can be widely applied to flow regulation scenarios in industrial production.

[0039] The following is an explanation of the hydraulic characteristic database, including its establishment method: Establishing a centrifugal pump hydraulic characteristic database is fundamental to improving its operating efficiency and control accuracy, involving multiple stages such as data acquisition, data preprocessing, and data storage. First, data acquisition is a crucial step in database construction, and its quality directly affects the accuracy of subsequent analysis results. Data acquisition is typically accomplished through a combination of experimental measurements and on-site monitoring. During experimental measurements, centrifugal pump performance testing equipment is used to record operating parameters under different conditions, including flow rate, head, shaft power, and efficiency. On-site monitoring involves installing sensors at key parts of the pump body to acquire real-time data, such as pressure sensors, flow sensors, and temperature sensors, ensuring the comprehensiveness and representativeness of the data. After data acquisition, the raw data needs to be preprocessed to eliminate noise interference and improve data quality. Preprocessing methods include filtering, smoothing, and outlier detection. For example, for abnormal data points caused by sensor malfunctions or environmental interference, statistical methods or machine learning algorithms can be used for identification and correction. Finally, the processed data is stored in the database according to a specific structure for easy subsequent querying and analysis. Data storage typically uses relational database management systems (such as MySQL) or NoSQL databases (such as MongoDB), with the specific choice depending on the data size and application requirements.

[0040] Regarding the database content, the centrifugal pump hydraulic characteristic database covers operating parameters under various working conditions, mainly including core indicators such as flow rate, head, shaft power, efficiency, and rotational speed. These parameters not only reflect the basic performance of the centrifugal pump but also provide important support for the application of virtual sensing technology and model predictive control algorithms. For example, the variation patterns of flow rate and head under different rotational speeds can be used to verify the applicability of similarity laws; while efficiency curves help determine the high-efficiency operating range of the centrifugal pump, providing a basis for optimizing control strategies. From a data structure perspective, the database typically organizes data in a tabular format, with each row recording parameter values ​​under a specific operating condition and each column corresponding to a specific physical quantity. Furthermore, to improve data retrieval efficiency, indexes can be created for key fields. For example, using rotational speed and flow rate as a combined index allows for quick retrieval of performance data under specific operating conditions. In some complex application scenarios, the database can be further expanded to include more dimensions of information, such as fluid properties parameters like temperature and viscosity, as well as related data such as pipeline characteristic curves. In this way, the operating status of the centrifugal pump can be described more comprehensively.

[0041] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A centrifugal pump adaptive control system based on virtual sensors, characterized in that, include: The data acquisition module (100) acquires the current, voltage, frequency and speed signals of the centrifugal pump in real time; Virtual sensor model (201) is used to establish a hydraulic characteristic database of centrifugal pumps. Based on the real-time collected centrifugal pump motor parameters, the current head, flow rate and pipeline resistance coefficient are calculated. The adaptive adjustment module (202) introduces a feedforward-feedback composite control model. When a sudden change in pipeline resistance is detected, the system adjusts the output frequency of the inverter (300) in advance through model predictive control. The terminal constant pressure module (203) automatically identifies the most unfavorable operating conditions and adjusts the pump's output characteristic curve to ensure that it always intersects with the pipeline characteristic curve at the set high-efficiency operating point.

2. The centrifugal pump adaptive control system based on virtual sensors according to claim 1, characterized in that, The data acquisition module (100) uses a Kalman filter-based data fusion algorithm to perform weighted fusion of multi-source sensor data and principal component analysis to provide acquired data.

3. The centrifugal pump adaptive control system based on virtual sensors according to claim 1, characterized in that, It also includes a soft measurement model (204). First, a large amount of historical operating data is obtained through experiments to establish a mapping relationship between input and output. Second, machine learning algorithms are used to train the data to generate a prediction model (205). Finally, cross-validation and parameter optimization are used to ensure the generalization ability and prediction accuracy of the model.

4. The centrifugal pump adaptive control system based on virtual sensors according to claim 1, characterized in that, The adaptive adjustment module (202) adopts an adaptive algorithm based on a second-order linear active disturbance rejection controller. First, in the parameter initialization stage, the initial values ​​are set according to the nominal operating parameters of the centrifugal pump, including rated flow and head, and the main parameters of the controller are determined. Second, in the model update stage, the system state is estimated by real-time acquisition of sensor data and extended Kalman filtering, and the model parameters are updated according to the estimation results. Finally, in the control quantity calculation stage, the optimal control quantity is solved by optimization algorithm based on the updated model parameters and setpoints.

5. The centrifugal pump adaptive control system based on virtual sensors according to claim 3, characterized in that, The establishment of the prediction model (205) includes the following steps: First, by analyzing the working principle and operating characteristics of the centrifugal pump, its mathematical description equation is established, and a nonlinear state-space model or transfer function model is adopted; Second, the model parameters are identified and optimized using actual operating data to ensure that the model can accurately reflect the dynamic behavior of the system.

6. The centrifugal pump adaptive control system based on virtual sensors according to claim 3, characterized in that, The prediction model (205) includes a rolling optimization step: First, the system output for a future period is calculated based on the prediction model (205) at the current time; second, the optimization objective function is defined, which usually includes a weighted combination of indicators such as control deviation, energy consumption and control quantity change rate; finally, the optimal control sequence is obtained by solving the optimization problem, and the first control quantity is applied to the actual system; the above process is repeated in the next control cycle.

7. The centrifugal pump adaptive control system based on virtual sensors according to claim 6, characterized in that, The process includes the following steps for adjusting the output frequency of the frequency converter (300): First, the required speed setpoint is determined based on the optimal control sequence calculated by the prediction model (205); second, the setpoint is converted into the output frequency of the frequency converter (300), and a proportional-integral (PI) controller is used for closed-loop regulation to ensure that the actual output frequency can quickly track the setpoint; finally, the output frequency is finely adjusted by monitoring the actual operating status of the centrifugal pump in real time to compensate for load changes.

8. A method for adaptive control of a centrifugal pump based on virtual sensors, characterized in that, The centrifugal pump is operated using the adaptive control system as described in any one of claims 1-7.

Images