Intelligent control method and system for variable frequency centrifugal unit

Abstract

The present invention relates to the field of automatic control technology, and further to an intelligent control method and system for a variable frequency centrifugal unit. The method comprises: step 1: establishing a state space model of the variable frequency centrifugal unit based on the dynamic characteristics of the variable frequency centrifugal unit; step 2: designing an adaptive fuzzy neural network controller based on the state space model, and the adaptive fuzzy neural network controller outputs an adaptive power matrix; step 3: estimating load changes through a load torque observer, calculating a load compensation control input matrix, and ensuring the stable operation of the variable frequency centrifugal unit; step 4: calculating a robust power matrix through a robust H∞ controller based on the state space model; and calculating the final control input power matrix of the variable frequency centrifugal unit. The present invention effectively improves the control accuracy, dynamic response capability, nonlinear adaptability and robustness of the variable frequency centrifugal unit under complex working conditions.

Description

Technical Field

[0001] The invention belongs to the technical field of automatic control, and in particular relates to an intelligent control method and system for a variable frequency centrifugal unit. Background Art

[0002] As one of the important equipment in modern industry, variable frequency centrifugal units are widely used in many fields such as petrochemical, medicine, food, metallurgy, etc., for processes such as material separation, purification and concentration. With the increasing demand for industrial production, centrifugal units face many challenges during operation, such as frequent load changes, complex system states, and unpredictable external disturbances. Especially under dynamic load and nonlinear operating conditions, the stability and efficiency of centrifugal units are easily affected, which puts higher requirements on the control technology of the system.

[0003] Traditional centrifuge control systems usually adopt a PID (proportional-integral-differential) control method. The PID controller detects the key parameters of the centrifuge (such as speed, temperature, vibration, etc.) and adjusts the control input according to the set target value to maintain the stable operation of the system. This control method has the advantages of simple structure and easy implementation, and has been widely used in the industrial field. However, with the changes in process complexity and operating conditions, traditional PID control methods have gradually exposed many limitations. Especially under dynamic loads and nonlinear conditions, PID control often cannot cope with rapidly changing conditions, resulting in problems such as system response lag, insufficient control accuracy, and high energy consumption. In industrial production, load changes in centrifuge units are inevitable. For example, in the chemical production process, the feed amount and characteristics of raw materials often fluctuate, which directly affects the working state of the centrifuge unit. Load changes may cause drastic fluctuations in parameters such as speed, torque, and temperature of the centrifuge unit, which in turn affects production efficiency and product quality. The existing PID control method has poor adaptability to nonlinear dynamic load changes due to its linear system design. Traditional controllers usually require manual parameter adjustments to cope with load changes, and these adjustments have lags and cannot respond to load fluctuations in real time in actual applications. Summary of the invention

[0004] The main purpose of the present invention is to provide an intelligent control method and system for a variable frequency centrifugal unit, which effectively improves the control accuracy, dynamic response capability, nonlinear adaptability and robustness of the variable frequency centrifugal unit under complex working conditions.

[0005] In order to solve the above problems, the technical solution of the present invention is implemented as follows:

[0006] An intelligent control method for a variable frequency centrifugal unit, the method comprising:

[0007] Step 1: Based on the dynamic characteristics of the variable frequency centrifugal unit, a state space model of the variable frequency centrifugal unit is established;

[0008] Step 2: Based on the state space model, an adaptive fuzzy neural network controller is designed to combine fuzzy logic with neural network to realize the control of nonlinear system. By updating the adaptive weight, the adaptive fuzzy neural network controller can be dynamically adjusted to adapt to different state changes of the variable frequency centrifugal unit; the adaptive fuzzy neural network controller outputs an adaptive power matrix;

[0009] Step 3: Estimate the load change through the load torque observer and calculate the load compensation control input matrix, so as to ensure the stable operation of the variable frequency centrifugal unit under the condition of dynamic load change;

[0010] Step 4: Based on the state space model, the robust power matrix is ​​calculated through the robust H∞ controller; the final control input power matrix of the variable frequency centrifugal unit is calculated by combining the adaptive power matrix, the robust power matrix and the load compensation control input matrix.

[0011] Furthermore, the state space model of the variable frequency centrifugal unit is expressed using the following formula:

[0012] ;

[0013] in, is the angular velocity; is the angular displacement; is the radial force; is the tangential force; is temperature; is the input power; is the moment of inertia; is the damping coefficient; and are radial stiffness coefficient and tangential stiffness coefficient respectively; is the temperature influence coefficient; For quality; and are radial and tangential unbalanced masses respectively; and are radial and tangential damping coefficients, respectively; For efficiency; is the power of the variable frequency centrifugal unit; is the specific heat capacity; is the heat transfer coefficient; is the surface area; For time; is the differential symbol, Indicates time The differential of .

[0014] Furthermore, the adaptive fuzzy neural network controller in step 2 is expressed by the following formula:

[0015] ;

[0016] in, It is The adaptive weights of the fuzzy rules, is the fuzzy basis function, is the number of fuzzy rules; , is the state vector of the variable frequency centrifugal unit; is the state feedback gain matrix; is the subscript index; is the adaptive power matrix.

[0017] Furthermore, the fuzzy basis function is a Gaussian membership function, which is expressed by the following formula:

[0018] ;

[0019] in, For the The central value of the fuzzy basis function; For the The width parameter of the fuzzy basis function indicates the extension degree of the membership function; the adaptive fuzzy neural network controller continuously adjusts the adaptive weight through the following formula :

[0020] ;

[0021] in, Represents adaptive weight The rate of change of is an adaptive parameter that controls the speed of updating the adaptive weights. The larger it is, the faster the adaptive weights are adjusted; , represents the state vector of the current variable frequency centrifugal unit The state vector of the desired variable frequency centrifugal unit The error between Amplification, used to adjust the adaptive weights to reduce the error; is a positive definite matrix designed according to Lyapunov stability theory.

[0022] Furthermore, the load torque observer in step 3 is expressed using the following formula:

[0023] ;

[0024] in, Input matrix for load compensation control; is the proportional gain, which controls the speed and amplitude of compensation; is the estimated load torque; is the expected load torque; is the integral gain, the control system's ability to compensate for the accumulated error; is the differential gain, which controls the response to the rate of change of the load torque; is the integration variable.

[0025] Furthermore, the estimated load torque Calculate using the following formula:

[0026] ;

[0027] in, is the estimated load torque; is the time constant of load torque observation, which determines the response speed of load estimation; is the angular velocity error gain, which controls the sensitivity of the load estimation to the angular velocity error; is the actual angular velocity; is the estimated angular velocity; It is the load torque nonlinear correction gain, which controls the amplitude of nonlinear compensation; is the integral gain, which controls the ability of the integral term to compensate for the accumulated error.

[0028] Furthermore, the estimated angular velocity Calculate using the following formula:

[0029] ;

[0030] in, is the estimated angular velocity; is the angular velocity correction gain, which controls the linear response to the angular velocity error; It is the angular velocity nonlinear correction gain, which controls the intensity of nonlinear compensation for angular velocity error.

[0031] Furthermore, the robust H∞ controller is expressed using the following formula:

[0032] ;

[0033] in, is the H∞ control gain matrix; is the robust power matrix; H∞ control gain matrix By solving the following Riccati equation:

[0034] ;

[0035] in, is the H∞ control gain matrix; is the solution of the Riccati equation; is the performance index of the H∞ controller, indicating the ability to suppress disturbances. The smaller the value, the stronger the ability to suppress disturbances. The output matrix ; External interference matrix System state matrix and the system input matrix It is obtained through the following process: Based on the state space model of the variable frequency centrifugal unit, the following equation is obtained:

[0036] ;

[0037] From this we get:

[0038] ;

[0039] .

[0040] Furthermore, the final control input power matrix is ​​calculated by the following formula:

[0041] ;

[0042] in, Input power matrix for final control.

[0043] The invention discloses an intelligent control system for a variable frequency centrifugal unit, the system comprising: a state space model unit, which is used to establish a state space model of the variable frequency centrifugal unit based on the dynamic characteristics of the variable frequency centrifugal unit; an adaptive fuzzy neural network control unit, which is used to design an adaptive fuzzy neural network controller based on the state space model, combine fuzzy logic with neural network, and realize control of nonlinear system. Through updating of adaptive weights, the adaptive fuzzy neural network controller can be dynamically adjusted to adapt to different state changes of the variable frequency centrifugal unit; the adaptive fuzzy neural network controller outputs an adaptive power matrix; a load compensation change unit, which is used to estimate load changes through a load torque observer and calculate a load compensation control input matrix, so as to ensure stable operation of the variable frequency centrifugal unit under dynamic load changes; a control output unit, which is used to calculate a robust power matrix through a robust H∞ controller based on the state space model; and a final control input power matrix of the variable frequency centrifugal unit is calculated by combining the adaptive power matrix, the robust power matrix and the load compensation control input matrix.

[0044] The intelligent control method and system of the variable frequency centrifugal unit of the present invention have the following beneficial effects: the present invention realizes adaptive control of dynamic load changes of the variable frequency centrifugal unit by introducing an adaptive fuzzy neural network controller. The adaptive fuzzy neural network combines the advantages of fuzzy logic and neural network, can dynamically adjust the weight of fuzzy rules according to the changes in system state, and enhances the system's adaptability to complex nonlinear characteristics. When the load of the variable frequency centrifugal unit changes drastically, the adaptive fuzzy neural network controller can respond quickly and adjust the control strategy according to the real-time operating state to ensure that the system remains in a stable and efficient operating state. Compared with the traditional PID controller, the present invention can handle the load change problem more accurately, significantly improve the control accuracy and response speed, especially under complex dynamic conditions. The variable frequency centrifugal unit often has significant nonlinear characteristics during actual operation, especially in the case of high speed or large load fluctuations, the nonlinear behavior of the system is more obvious. Traditional linear control methods (such as PID control) are difficult to effectively deal with these nonlinear problems, which may lead to control failure or insufficient accuracy. The adaptive fuzzy neural network controller of the present invention processes the uncertainty and nonlinearity of the system through fuzzy logic, and dynamically adjusts the control parameters through the self-learning function of the neural network, which can effectively deal with nonlinear problems in the system. This enables the control system to flexibly adjust output in the face of complex situations such as load fluctuations, equipment aging or environmental changes, ensuring system operation stability and control accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A schematic flow chart of an intelligent control method for a variable frequency centrifugal unit provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that shown or described herein.

[0047] Embodiment 1: An intelligent control method for a variable frequency centrifugal unit, the method comprising:

[0048] Step 1: Based on the dynamic characteristics of the variable frequency centrifugal unit, a state space model of the variable frequency centrifugal unit is established;

[0049] In the control process of variable frequency centrifugal units, the role of the state space model is to provide a description of the dynamic characteristics of the unit and provide an accurate mathematical basis for the subsequent control algorithm. It characterizes the behavior of the system by taking the system state (such as speed, torque, etc.) as the state variable and combining it with the external input (such as the motor control signal). Specifically, the state space model of the variable frequency centrifugal unit can reflect how the state of the system evolves when it is subjected to external disturbances or changes in the control signal. For example, when the input signal of the system changes, the system's parameters such as speed, vibration and temperature will change accordingly, and the state space model can describe the change process of these state variables through a set of linear or nonlinear differential equations, so that the controller can foresee and adjust the operation of the system. The state space model is not just a simple description of the system state, it can effectively handle the multi-input and multi-output characteristics of the system. As a complex multi-input and multi-output system, the variable frequency centrifugal unit is not only affected by multiple input variables (such as frequency and torque), but also may have multiple output variables (such as system power, vibration state, etc.) that need to be adjusted simultaneously. Traditional control methods can often only handle systems with a single input and output, while the state-space model can consider multiple state variables at the same time, allowing the controller to comprehensively consider the interaction between these variables when designing, ensuring that the entire system can work in a coordinated manner. For example, there may be a complex coupling relationship between the speed and vibration of a variable frequency centrifugal unit, and the state-space model can capture this coupling behavior and use it in the subsequent control algorithm design.

[0050] In addition, the state space model also has the ability to handle system uncertainty and external disturbances. In actual operation, the variable frequency centrifugal unit may be disturbed by the external environment or changes in system parameters, resulting in the system state deviating from the expected situation. The state space model can help the controller identify and compensate for these uncertainties by accurately describing the system state and input. For example, when the external load changes suddenly, the state space model can quickly reflect the impact of this change on the system state, and the controller can adjust the control strategy in real time based on the state prediction provided by the model, thereby ensuring the stable operation of the system. At the same time, the state space model also provides an important basis for the design of the adaptive fuzzy neural network controller and the robust H∞ controller in the present invention. The adaptive fuzzy neural network controller relies on the description of the system state by the state space model to perform adaptive adjustment of the weights to ensure that it can adapt to the dynamic changes of the system state. The robust H∞ controller uses the state space model to handle the uncertainty in the system to ensure that the system can still maintain robustness and stability in the presence of disturbances. Therefore, the state space model not only provides an accurate dynamic description for the entire intelligent control system, but also directly affects the design and implementation of the controller, and is the key link for the efficient operation of the entire intelligent control method.

[0051] Step 2: Based on the state space model, an adaptive fuzzy neural network controller is designed to combine fuzzy logic with neural network to realize the control of nonlinear system. By updating the adaptive weight, the adaptive fuzzy neural network controller can be dynamically adjusted to adapt to different state changes of the variable frequency centrifugal unit; the adaptive fuzzy neural network controller outputs an adaptive power matrix;

[0052] The adaptive fuzzy neural network controller first uses the characteristics of fuzzy logic to fuzzify the input variables of the variable frequency centrifugal unit, so as to handle the uncertainty and fuzziness of the variables. The state change of the variable frequency centrifugal unit is often not precisely controllable, but has a certain fuzziness. For example, due to factors such as load changes, temperature fluctuations or external interference, the system state fluctuates within a certain range. Through fuzzy logic, these input variables can be converted into fuzzy sets, and the degree to which the input variables belong to a certain fuzzy set can be described according to different membership functions, thereby realizing the fuzzy processing of the input variables. The fuzzy logic controller can map the input variables to the control output according to the pre-set fuzzy rules, thereby providing a preliminary control scheme for the system. However, a simple fuzzy controller may be insufficient when facing a complex nonlinear system, and it is difficult to adapt to the dynamic changes of the system state. Therefore, the present invention enhances the adaptive ability of the system by introducing a neural network. The neural network has a strong self-learning and adaptive ability, and can dynamically adjust its internal weights through training, so that the controller can adapt to the changes in the system state. Specifically, the neural network updates the weight of the network according to the error between the control output and the actual response of the system through the error back propagation algorithm, and finally realizes the precise control of the system. In the present invention, a neural network is combined with a fuzzy controller to form an adaptive fuzzy neural network controller. This combination enables the control system to not only handle the uncertainty of input variables, but also dynamically adjust the control strategy according to changes in system states, thereby enhancing the robustness and adaptability of the system.

[0053] The design principle of the adaptive fuzzy neural network controller also includes a weight adaptive adjustment mechanism. The dynamic characteristics of the variable frequency centrifugal unit will change significantly under different operating conditions, and the control system needs to be able to dynamically adjust the control parameters according to the real-time state. Through the adaptive weight update mechanism, the neural network can automatically adjust the parameters of the fuzzy rules according to the system state feedback, including the shape of the membership function and the range of the fuzzy set, so that the fuzzy controller can adapt to the state changes of the system at any time. For example, when the load suddenly increases or decreases, the adaptive fuzzy neural network controller can quickly adapt to the new load state by adjusting the parameters of the fuzzy rules, thereby ensuring the stable operation of the system. This adaptive adjustment process is based on real-time feedback, so there is no need for human intervention, and the controller can optimize and adjust autonomously, which greatly improves the response speed and stability of the system. In addition, the output of the adaptive fuzzy neural network controller is an adaptive power matrix, which generates corresponding control signals according to the state changes of the system. Specifically, the power matrix reflects the power distribution required by the variable frequency centrifugal unit under different operating conditions, thereby achieving precise energy management and control. In practical applications, the power demand of variable frequency centrifugal units will change with the load. Traditional fixed value control is difficult to adapt to such changes, while the adaptive fuzzy neural network controller can automatically generate a suitable power allocation strategy according to the real-time system status, thereby maximizing the efficiency of the system and reducing energy consumption. The generation process of this power matrix relies on the adjustment of fuzzy rules by the neural network, so that the controller can achieve efficient energy management while ensuring system stability.

[0054] Step 3: Estimate the load change through the load torque observer and calculate the load compensation control input matrix, so as to ensure the stable operation of the variable frequency centrifugal unit under the condition of dynamic load change;

[0055] The core function of the load torque observer is to estimate the actual load torque of the current system in real time. This estimation is based on the system's state feedback information and the existing load model. By observing the dynamic behavior of the system, the load torque observer can continuously predict and adjust the current load. Usually, the change of load is difficult to accurately measure directly through sensors, while the load torque observer uses the system input and output data and the state space model, combined with the observation algorithm, to infer the current load state. The observer processes the state change data of the system in real time and outputs the estimated value of the load. This estimation result provides a basis for the subsequent load compensation control, allowing the system to adjust the operating parameters in a timely manner to adapt to the load changes. On this basis, the calculation of the load compensation control input matrix becomes a key link in the stability of the system. The matrix is ​​used to compensate for the system deviation caused by load changes and ensure that the centrifugal unit can maintain the predetermined operating performance under load fluctuations. Specifically, when the load changes, the system's torque, power, and speed parameters will be affected, which may cause the output power of the centrifugal unit to deviate from the set value. If timely compensation is not performed, the system may be unstable or even fail. The role of the load compensation control input matrix is ​​to compensate for the impact of load changes on system operation by adjusting the input parameters. Through the function of this matrix, the system can automatically adjust the control signal to maintain the stability of output power, torque and speed, thereby achieving effective response to load changes.

[0056] The combination of load torque observer and load compensation control makes the variable frequency centrifugal unit highly adaptable and stable under dynamic load conditions. Unlike the traditional fixed value control method, the load torque observer can dynamically adjust the control strategy by observing the changes in load in real time, so that the system can cope with changes in external environment and working conditions. Traditional control systems often react slowly to load changes or require manual intervention for adjustment. The present invention introduces a load torque observer to enable the system to autonomously predict load changes and automatically compensate according to the dynamic changes of the load, greatly improving the response speed and stability of the system. In addition, the introduction of the load torque observer can also reduce the energy consumption and resource waste of the system. In traditional control systems, it is often necessary to reserve too much power margin for possible load changes, resulting in energy waste. By observing the load state in real time, the load torque observer can ensure that the system operates at the most appropriate power level, thereby avoiding unnecessary energy consumption. This not only improves the energy utilization efficiency of the system, but also extends the service life of the equipment and reduces maintenance costs. The combination of load torque observer and load compensation control shows great technical advantages and innovation in the intelligent control method of variable frequency centrifugal units. First, by observing and compensating for load changes, the system can achieve adaptive control of dynamic load conditions, ensuring smooth and efficient operation. Second, the load torque observer enables the system to respond quickly to changes in external conditions, avoiding the lag phenomenon in traditional control systems. Most importantly, the calculation of the load compensation control input matrix enables the system to accurately adjust the control strategy, avoiding resource waste and improving the energy efficiency and operational stability of the entire system.

[0057] Step 4: Based on the state space model, the robust power matrix is ​​calculated through the robust H∞ controller; the final control input power matrix of the variable frequency centrifugal unit is calculated by combining the adaptive power matrix, the robust power matrix and the load compensation control input matrix.

[0058] The design of the H∞ controller is based on an optimized control strategy that can minimize the impact of external disturbances on system performance. Traditional controllers often fail to achieve precise control when dealing with system uncertainties, which may cause the system to operate under unstable conditions, thereby affecting the efficiency and life of the entire variable frequency centrifugal unit. The H∞ controller accurately models the system state and designs a robust power matrix, so that the system can maintain the set performance indicators even when the external environment changes. The core of this control process is to calculate the impact of disturbances on the system and minimize this impact through control algorithms. In this way, the H∞ controller can significantly improve the anti-disturbance ability of the variable frequency centrifugal unit in a dynamic environment and enhance its stability in adapting to different load conditions. In specific applications, the robustness of the H∞ controller is reflected in its ability to handle uncertainties and external disturbances in the system. This uncertainty may come from changes in the parameters of the centrifugal unit itself, such as mechanical wear, increased friction caused by insufficient lubrication, or changes in material properties caused by temperature changes. These factors may cause the dynamic response of the system to deviate from expectations, thereby affecting the control effect of the system. The H∞ controller accurately describes the dynamic behavior of the system through the state space model and incorporates these uncertainties into the control design, so that the system can cope with these uncertainties by optimizing the control strategy. In addition, external disturbances are also the main threat to the stability of the system. For example, sudden increase in load or power grid fluctuations may affect the normal operation of the centrifugal unit. The H∞ controller can effectively reduce the impact of these external disturbances on the system operation and ensure the stability of the system by estimating and compensating for them.

[0059] In the present invention, the H∞ controller is not just a simple robust controller. It combines the state space model of the system to achieve more sophisticated control. This controller can dynamically adjust the control strategy to cope with different operating conditions by analyzing the state and input-output relationship of the system. For example, when the system is running at high speed, the load change may be more drastic. The H∞ controller can calculate the robust power matrix of the system in real time and automatically adjust the input power to maintain the smooth operation of the system. At the same time, the H∞ controller can take into account the parameter changes of the system when it is running at low speed, and ensure that the system can maintain efficient and stable operation at different speeds through calculation and adjustment. This dynamic adjustment capability enables the H∞ controller to show extremely high control accuracy and stability when dealing with complex and changeable working conditions. The difference between the robust H∞ controller and other control methods is that it pays special attention to the uncertainty and disturbance effects of the system. Traditional controllers usually assume that the parameters of the system are known and fixed when designing. However, in actual applications, the system parameters often change, and there may even be a certain degree of error. These changes and errors will directly affect the performance of the controller, while the H∞ controller optimizes the disturbance transfer function to ensure that the performance of the system can be maintained within the expected range even under uncertain conditions. Therefore, the H∞ controller has significant advantages in anti-disturbance and stability, and is especially suitable for systems with complex dynamic characteristics and changeable external environment such as variable frequency centrifugal units.

[0060] Embodiment 2: The state space model of the variable frequency centrifugal unit is expressed using the following formula:

[0061] ;

[0062] in, is the angular velocity; is the angular displacement; is the radial force; is the tangential force; is temperature; is the input power; is the moment of inertia; is the damping coefficient; and are radial stiffness coefficient and tangential stiffness coefficient respectively; is the temperature influence coefficient; For quality; and are radial and tangential unbalanced masses respectively; and are radial and tangential damping coefficients, respectively; For efficiency; is the power of the variable frequency centrifugal unit; is the specific heat capacity; is the heat transfer coefficient; is the surface area; For time; is the differential symbol, Indicates time The differential of .

[0063] Specifically, the angular velocity It is one of the core state variables of the system, which directly reflects the rotation speed of the centrifuge unit. In the state space model, the change of angular velocity is affected by many factors, including the moment of inertia. , damping coefficient , radial stiffness and tangential stiffness For example, in the formula The term represents the effect of the friction damping of the system on the angular velocity. It shows that when the unit is running, the angular velocity will gradually decrease due to the existence of friction. This damping effect is one of the inevitable energy losses in the system, and the state-space model effectively describes the effect of energy loss on the system dynamics through this term. and tangential force They represent the forces in the two main directions that the system is subject to during operation. These forces are not only related to the rotation state of the system, but are also affected by the unbalanced mass. In the state space model, and The term describes the unbalanced mass and Contributions to the radial and tangential forces. These forces are multiplied as the angular velocity increases due to the centrifugal force, indicating that the system is subject to significant vibration and force variations at high speeds. This nonlinear term reflects the complex behavior of the system at high speeds and is key to achieving precise control. Temperature The change of is another important state variable in the system, which directly affects the operating efficiency and safety of the system. In the state space model, the change of temperature is determined by multiple factors. The impact of This means that when the system consumes power, part of the energy will be converted into heat, causing the temperature to rise. At the same time, the system and surface area To dissipate heat, the heat loss is This part of the formula effectively captures the thermal effect of the unit during long-term operation, helping the controller to adjust the operation strategy under high temperature conditions to prevent equipment from overheating and damage or loss of efficiency.

[0064] Input matrix The term describes the input power It shows that the input power of the system affects the moment of inertia by to control the change in angular velocity. This term works together with other coupling terms in the state matrix to enable the system to respond quickly to external inputs. As the input power increases, the angular velocity increases accordingly, and the torque and temperature of the system also change. This dynamic process is fully mathematically described in the state-space model, which the controller can use to predict the response of the system for more precise control. From a physical point of view, the operating state of a variable frequency centrifugal unit is affected by many factors, and the state-space model combines these factors and couples them together to form a complete system description. For example, the angular velocity is not only directly affected by the input power, but also by the inertia. , damping coefficient The effect of physical parameters such as , further affects the radial force and tangential force. This coupling relationship is expressed in the state-space model in the form of a matrix. The controller can adjust the input in real time according to these relationships to ensure that the system can still maintain stable operation under complex working conditions. Another important feature of the state-space model is its robustness. During the operation of the variable frequency centrifugal unit, the system parameters may change over time, such as increased friction coefficient, material aging, etc. By introducing these parameters in the state matrix, the model can better adapt to changes in the system, thereby ensuring the effectiveness of the controller in an uncertain environment. Especially in terms of temperature control, the state-space model can help the controller take compensation measures in time when the temperature rises abnormally by introducing the coupling relationship between heat transfer and power input, so as to avoid overheating of the system or a decrease in operating efficiency. In addition, nonlinear terms in the state-space model such as etc., fully reflecting the dynamic complexity of the system. When rotating at high speed, the vibration caused by unbalanced mass will significantly affect the operating state of the unit. This nonlinear relationship is difficult to handle in traditional linear control methods. Through the state space model, the controller can model these nonlinear effects to better predict the dynamic response of the system. This feature makes the model perform well in dealing with complex and nonlinear systems.

[0065] Embodiment 3: The adaptive fuzzy neural network controller in step 2 is represented by the following formula:

[0066] ;

[0067] in, It is The adaptive weights of the fuzzy rules, is the fuzzy basis function, is the number of fuzzy rules; , is the state vector of the variable frequency centrifugal unit; is the state feedback gain matrix; is the subscript index; is the adaptive power matrix.

[0068] Specifically, the core idea of ​​the controller is to process the system state through fuzzy rules. ,in The state vector representing the variable frequency centrifugal unit covers multiple key variables such as angular velocity, angular displacement, radial force, tangential force, and temperature. For such a complex multivariable system, traditional control methods often cannot take into account nonlinearity and uncertainty, while fuzzy logic can fuzzify these variables, thereby maintaining effective control in the presence of uncertainty. The fuzzy basis function in the formula It is a fuzzy expression of the system state, which defines how fuzzy rules in different states act on the control process. These fuzzy rules provide a flexible control mechanism for the controller based on different combinations of system states, which can handle the complex nonlinear relationships commonly seen in the system. The adaptability of the controller is reflected in the weight adjustment of the fuzzy rules. Specifically, Represents the adaptive weight of each fuzzy rule. These weights are not fixed, but dynamically adjusted according to the operating status of the system. Through adaptive weight adjustment, the controller can optimize the control strategy according to the real-time status, so as to adapt to uncertain factors such as load changes and environmental changes that may occur during the operation of the unit. The weight adjustment process is realized through the self-learning mechanism of the neural network. The controller can automatically update these weights according to the feedback information output by the system, so that it can always be adjusted according to the optimal control strategy. This adaptive adjustment enables the controller to cope with complex dynamic environments and ensure the stability and efficiency of the system.

[0069] In addition, the state feedback gain matrix in the controller and reference gain Further enhances the control accuracy of the system. State feedback gain The current state of the system Feedback adjustment is performed to ensure that the control signal can be accurately adjusted according to the actual state. This feedback mechanism can effectively reduce the dynamic deviation of the system, so that the centrifuge unit can correct the state deviation in time during operation and avoid instability caused by external disturbances or internal uncertainties. Reference gain It is used to convert the desired reference signal Introducing the control process helps the controller maintain the desired operating state of the system. Through these two gain matrices, the controller can accurately adjust the system state in real-time operation and ensure that it runs according to the predetermined reference trajectory. Compared with traditional control methods, the adaptive fuzzy neural network controller of Example 3 shows obvious technical advantages. First, through the fuzzy basis function, the controller can handle the fuzziness and uncertainty in the nonlinear system, so that the system can still maintain stable operation in a complex environment. Secondly, the adaptability of the neural network enables the controller to update the fuzzy rule weights in real time according to the dynamic changes of the system, avoiding the lag of traditional controllers when facing uncertain factors. In addition, the introduction of the state feedback mechanism and the reference gain further improves the control accuracy, so that the system can still maintain an efficient and stable working state when facing disturbances. Especially in the specific application of variable frequency centrifugal units, this controller can cope with dynamic changes in system loads. For example, when the external load suddenly increases, the angular velocity There may be a significant drop, and through state feedback and adaptive adjustment, the controller can improve the adaptive power in time , so that the system returns to normal operation. At the same time, the radial force in the system and tangential force It will also change with the change of load, and the fuzzy basis function can perform real-time fuzzy processing according to the changes of these forces to ensure that the control signal matches the actual operating state. Changes in temperature are also taken into account, ensuring that the controller can effectively manage the temperature during long-term operation or high load conditions to avoid overheating of the system.

[0070] Embodiment 4: The fuzzy basis function is a Gaussian membership function, which is expressed by the following formula:

[0071] ;

[0072] in, For the The central value of the fuzzy basis function; For the The width parameter of the fuzzy basis function indicates the extension degree of the membership function; the adaptive fuzzy neural network controller continuously adjusts the adaptive weight through the following formula :

[0073] ;

[0074] in, Represents adaptive weight The rate of change of is an adaptive parameter that controls the speed of updating the adaptive weights. The larger it is, the faster the adaptive weights are adjusted; , represents the state vector of the current variable frequency centrifugal unit The state vector of the desired variable frequency centrifugal unit The error between Amplification, used to adjust the adaptive weights to reduce the error; is a positive definite matrix designed according to Lyapunov stability theory.

[0075] Specifically, the fuzzy basis function adopts the form of Gaussian membership function, which describes how the system state is fuzzified so as to facilitate the processing of fuzzy neural network. The main feature of this Gaussian membership function is to use the membership relationship between state variables and fuzzy rules to reflect the degree of belonging of the system state in a specific fuzzy set. Through this function, the state of the variable frequency centrifugal unit is mapped into fuzzy sets and is based on the actual state of the system and the central value The size of this deviation determines the degree to which the state variable belongs to the fuzzy set, and the width parameter This affects the expansion range of the fuzzy basis function, that is, the range of state changes that the system can perceive. In the control of variable frequency centrifugal units, the selection of Gaussian membership function is crucial. Since this function can handle the nonlinearity and uncertainty in the system in a smooth manner, it allows the controller to effectively judge and handle changes in different state variables when faced with complex operating conditions. This smooth characteristic enables fuzzy logic to cope with a variety of interferences and disturbances that may be encountered during the actual operation of the variable frequency centrifugal unit, thereby enhancing the robustness and stability of the system. When the state of the system and the central value of the fuzzy rule The closer the degree, the The higher the value, the higher the matching degree of the state in the current fuzzy rule. As the state deviates from the center value, the membership degree gradually decreases, and the system can gradually adjust to the most matching fuzzy rule. This process ensures the flexibility of the controller in different states.

[0076] The adaptive fuzzy neural network controller adjusts the weights To optimize the control strategy, this is achieved partly through error feedback of the system state. In actual operation, the current state vector of the variable frequency centrifugal unit is and expected state There is often a deviation between The existence of error means that the current system state does not match the ideal state. The controller needs to reduce this error by adjusting the adaptive weights and finally achieve the desired operating state of the system. In this process, the fuzzy basis function The output of will affect the correction of error feedback, ensuring that the adaptive adjustment mechanism can effectively reduce the deviation in system operation. In order to ensure the stability of this adjustment process, the control system uses Lyapunov stability theory to design a positive definite matrix , which ensures the global stability of the system during weight adjustment. Through this matrix, the error feedback can maintain stable asymptotic convergence during the weight adjustment process. Even in the face of a complex dynamic system environment, the controller can gradually approach the ideal control effect through the weight adjustment mechanism. This stability ensures that the variable frequency centrifugal unit can still maintain effective control under high load, dynamic changes and disturbance conditions, without causing system instability due to excessive response or lag during the weight adjustment process. The adjustment speed of the adaptive weight is determined by the parameter Decide, The size of directly affects the controller's response speed to the error. The value enables the controller to adjust the weights faster and adapt to rapid changes in the system state; while smaller It is suitable for situations where the system changes relatively smoothly and can avoid system oscillations that may be caused by too fast adjustment. Therefore, in practical applications, the controller can dynamically adjust according to different working conditions. The value of , thus ensuring that the system achieves the best balance between stability and response speed. In combination with the entire control process, the adaptive fuzzy neural network continuously adjusts the weights , so that the system can automatically optimize the control signal according to the actual state changes. This mechanism ensures that the variable frequency centrifugal unit can always maintain efficient and stable operation when facing complex working conditions. For example, when the system load changes suddenly, the system state May quickly deviate from the desired state , at this time, by adjusting the weight , the controller can quickly correct the control signal and bring the system back to the desired operating track. In addition, the adjustment process based on Lyapunov stability theory ensures that the system as a whole remains stable while the weights are adjusted, and will not cause system instability due to over-adjustment or hysteresis.

[0077] Embodiment 5: The load torque observer in step 3 is expressed using the following formula:

[0078] ;

[0079] in, Input matrix for load compensation control; is the proportional gain, which controls the speed and amplitude of compensation; is the estimated load torque; is the expected load torque; is the integral gain, the control system's ability to compensate for the accumulated error; is the differential gain, which controls the response to the rate of change of the load torque; is the integration variable.

[0080] Specifically, during the operation of the variable frequency centrifugal unit, the load torque is one of the key parameters that determine the system performance. Due to the continuous changes in external working conditions, the load torque will also fluctuate, which poses a challenge to the stability of the system. Therefore, the load torque observer estimates the load torque and compare it with the expected load torque By comparing the actual load torque and the expected value, the current load state of the system can be effectively identified. When the observer detects a deviation between the actual load torque and the expected value, it will generate a corresponding compensation signal to reduce or eliminate the deviation, thereby maintaining the stability of the system. The first term in the formula, the proportional control part of the load torque, reflects the basic mechanism of this compensation. Through direct feedback of the error, the system can respond quickly to load changes and adjust the load torque by proportional gain. to control the magnitude and speed of this response. It will make the system more sensitive to the feedback of load torque changes and can quickly reduce errors, but it may also cause the system to over-respond and increase the risk of oscillation; smaller This helps maintain system stability, but the response speed may be slower. Therefore, the choice of proportional gain is critical to balancing system stability and response speed. In addition to proportional control, the formula also includes an integral part, which provides compensation for long-term system deviations by integrating the accumulated error. The role of integral control is to deal with the steady-state error of the system, that is, when the load change persists, the proportional control may not be able to completely eliminate the error. At this time, the integral term can apply a greater correction force through the accumulated error. The integral gain in the formula Controls the response of the system to the accumulated error. The design of the integral term enables the system to gradually eliminate the persistent error caused by external interference or internal nonlinearity of the system, ensuring that the unit can maintain accurate load control during long-term operation. However, excessive integral gain may cause the system to produce integral saturation, that is, excessive compensation of the accumulated error may cause system instability. Therefore, the adjustment of the integral gain needs to consider the overall performance of the system to ensure that no new unstable factors are caused while compensating for the accumulated error.

[0081] In addition, the differential control part in the load torque observer formula further improves the dynamic response capability of the system by calculating the error change rate. The differential term is used to predict the error change trend and suppress possible error growth by responding in advance. The differential gain in the formula Controls the sensitivity of the system to the rate of error change. By introducing differential control, the system can adjust when the error begins to appear to prevent the error from further expanding. This feedforward control mechanism enables the system to respond to changes in dynamic loads more quickly, especially when the load changes rapidly, differential control can effectively reduce overshoot and improve the stability of the system. Differential control helps to smooth the dynamic response of the system and avoid oscillation or transient instability caused by too fast load changes. However, the design of differential control also needs to be cautious. Excessive differential gain may cause the system to be too sensitive to noise and produce unnecessary oscillations, so it needs to be optimized in combination with the actual system conditions. Overall, the design of the load torque observer realizes real-time compensation for the load torque changes of the variable frequency centrifugal unit through a comprehensive control strategy of proportion, integration and differentiation. This control strategy can not only respond quickly to changes in load, but also handle long-term deviations of the system and improve the dynamic performance of the system through differential control. In this way, the load torque observer ensures that the system can still maintain stable and reliable operation under different working conditions. When the external load increases suddenly, the system can quickly reduce the error through proportional control, while the integral part gradually eliminates the accumulated deviation, and the differential control predicts and responds to the rapid change of load in advance. Through the coordinated effect of these mechanisms, the load torque observer provides strong support for the entire intelligent control system, ensuring the efficient and stable operation of the variable frequency centrifugal unit, especially under complex working conditions and dynamic load conditions.

[0082] Example 6: Estimated load torque Calculate using the following formula:

[0083] ;

[0084] in, is the estimated load torque; is the time constant of load torque observation, which determines the response speed of load estimation; is the angular velocity error gain, which controls the sensitivity of the load estimation to the angular velocity error; is the actual angular velocity; is the estimated angular velocity; It is the load torque nonlinear correction gain, which controls the amplitude of nonlinear compensation; is the integral gain, which controls the ability of the integral term to compensate for the accumulated error.

[0085] Specifically, the basic principle of load torque estimation relies on the angular velocity feedback of the system. The angular velocity of the centrifugal unit is a direct manifestation of the system dynamics and can reflect the changing trend of the load torque. , the system compares the actual angular velocity with the estimated angular velocity to calculate the angular velocity error. The change in angular velocity is closely related to the load torque, so the angular velocity error becomes an important basis for estimating the load torque. The angular velocity error feedback term in the formula This term controls the system response to this error. To adjust the sensitivity to angular velocity errors, the system can accurately and timely capture small changes in angular velocity and adjust the load torque estimation accordingly. In actual operation, load changes not only affect the angular velocity, but also introduce complex nonlinear characteristics. Especially when the load changes drastically, the system may exhibit significant nonlinear behavior. To address this challenge, the nonlinear correction term is introduced into the formula. , which is obtained by the hyperbolic tangent function To handle the nonlinear part of the angular velocity error. When the error is small, the impact of the nonlinear correction is small, but as the error increases, the hyperbolic tangent function gradually saturates and the correction strength increases, providing additional load torque compensation. This mechanism allows the system to accurately estimate the load torque when the load fluctuates violently, avoiding estimation errors caused by insufficient linear feedback. Nonlinear Correction Gain The magnitude of this correction is controlled, which determines whether the system requires stronger nonlinear compensation to ensure the accuracy of the estimate when faced with larger errors.

[0086] In addition to real-time adjustment of the dynamic behavior of the system through angular velocity error and nonlinear correction, load torque estimation also needs to deal with the accumulated deviation of the system during long-term operation. To deal with this problem, the integral term is introduced into the formula , which is used to compensate for the long-term deviation of the system. The principle of integral control is to gradually eliminate the estimation error caused by long-term load changes by accumulating the angular velocity error. For complex systems such as variable frequency centrifugal units, load changes are usually not instantaneous and may continue for a certain period of time, resulting in gradual accumulation of errors. The role of the integral term is to apply a stronger compensation force by accumulating these errors to ensure that the system maintains a high-precision load torque estimate in long-term operation. Integral gain Controlling this process, a higher gain value enables the system to respond faster to cumulative errors, but too large a gain may also cause the system to over-respond to short-term errors and produce unnecessary oscillations. Therefore, the setting of the integral gain needs to be optimized according to the operating characteristics of the system to ensure that the system can eliminate long-term deviations without overreacting to short-term errors. In addition, the time constant in the formula As the time scale of load torque observation, it determines the response speed of the estimation. The smaller the time constant, the faster the load torque estimation responds to the system, which is suitable for use in cases where the load changes rapidly; while a larger time constant can effectively smooth the estimation and prevent over-adjustment caused by short-term fluctuations. By adjusting the time constant, the system can strike a balance between dynamic response and estimation accuracy, ensuring a quick response when the load fluctuates violently, and avoiding instability caused by over-adjustment under stable loads.

[0087] Example 7: Estimated angular velocity Calculate using the following formula:

[0088] ;

[0089] in, is the estimated angular velocity; is the angular velocity correction gain, which controls the linear response to the angular velocity error; It is the angular velocity nonlinear correction gain, which controls the intensity of nonlinear compensation for angular velocity error.

[0090] Specifically, the first term in the formula is the linear feedback term of the angular velocity error. Here, As the angular velocity correction gain, it determines the linear response strength of the system to the current angular velocity error. The actual angular velocity When there is a difference between the angular velocity and the actual angular velocity, the system directly corrects the error through this linear term. The working mechanism of this term is to adjust the estimated angular velocity in real time by proportionally feeding back the angular velocity error so that it gradually approaches the actual angular velocity. The magnitude of this adjustment is controlled. A larger gain value means that the system responds more quickly to angular velocity errors, thus being able to quickly eliminate errors under dynamic conditions. This may cause the system to oscillate, so a balance needs to be struck between the system's response speed and stability. is the nonlinear correction term for the angular velocity error, which uses the inverse hyperbolic cosine function to handle larger angular velocity errors. Unlike the linear feedback term, the nonlinear correction term is specifically designed to deal with situations where the angular velocity error is large. When the error is small, linear feedback can quickly adjust the estimated angular velocity, but when the error is large, relying solely on linear feedback may not be enough to achieve effective correction. Therefore, by introducing nonlinear compensation, the system can provide additional correction strength when the error is large. The inverse hyperbolic cosine function has a strong nonlinear characteristic. As the angular velocity error increases, the correction amplitude also increases significantly. This nonlinear mechanism helps the system quickly return to a stable state when the angular velocity changes greatly.

[0091] Nonlinear correction gain Controls the strength of this nonlinear compensation. The value means that the nonlinear correction of the angular velocity error by the system will be more significant, and it can provide strong compensation under large error conditions, ensuring that the system maintains accurate angular velocity estimation under dynamic loads and complex working conditions. However, the introduction of nonlinear compensation also needs to be cautious. Excessive nonlinear correction gain may cause the system to overreact to the error and affect the smooth operation of the system. Therefore, The choice of needs to be optimized according to the system's working environment and load changes to ensure that the system achieves the best balance between dynamic response and stability. The design of this angular velocity estimation formula fully takes into account the various complex situations that the system may encounter in actual operation. By combining linear and nonlinear correction mechanisms, the formula can handle a variety of situations from small errors to large errors. Under the linear feedback mechanism, the system can quickly adjust the angular velocity estimate through proportional control when the error is small to ensure precise control of the system under stable conditions. When the load changes suddenly or external disturbances cause large angular velocity errors, the nonlinear correction provides additional compensation to ensure that the system can quickly return to normal. This control strategy that combines linear and nonlinear methods enables the angular velocity estimation to maintain accuracy under a wide range of operating conditions. In actual applications, when the variable frequency centrifugal unit encounters drastic fluctuations in external loads, the angular velocity of the system may deviate from the normal value, resulting in the estimated angular velocity. The actual angular velocity At this time, the linear correction term can immediately feedback the error and gradually adjust the estimated value according to the size of the error. However, if the load changes drastically, simple linear feedback is not enough to quickly eliminate the error, so the nonlinear compensation term plays an important role at this time. Through nonlinear correction, the system can provide stronger compensation when the error is large, and quickly bring the estimated angular velocity back to normal levels. This correction mechanism effectively improves the dynamic response capability of the system under complex working conditions.

[0092] Example 8: The robust H∞ controller is expressed using the following formula:

[0093] ;

[0094] in, is the H∞ control gain matrix; is the robust power matrix; H∞ control gain matrix By solving the following Riccati equation:

[0095] ;

[0096] in, is the H∞ control gain matrix; is the solution of the Riccati equation; is the performance index of the H∞ controller, indicating the ability to suppress disturbances. The smaller the value, the stronger the ability to suppress disturbances. The output matrix ; External interference matrix System state matrix and the system input matrix It is obtained through the following process: Based on the state space model of the variable frequency centrifugal unit, the following equation is obtained:

[0097] ;

[0098] From this we get:

[0099] ;

[0100] .

[0101] Specifically, in the variable frequency centrifugal unit intelligent control method and system of the present invention, Example 8 introduces a robust The controller is used to deal with system disturbances and uncertainties under complex working conditions. The controller suppresses the impact of external disturbances on system performance by optimizing the control signal, and ensures stable and efficient operation of the variable frequency centrifugal unit based on the system dynamic model. In the design of the controller, special attention is paid to the robustness of the system, that is, the stability of the system performance can be maintained in the presence of external interference or changes in internal parameters. Control Gain Matrix To generate the robust power matrix , thereby achieving precise control of the system.

[0102] The working principle of the controller can be briefly summarized as follows: by calculating the current system state vector in real time and the control gain matrix Combined to obtain the control signal This signal adjusts the input power of the variable frequency centrifugal unit to suppress deviations caused by external disturbances or system uncertainties. Formula The basic mechanism of this control signal generation is described. The system state vector Contains the key operating parameters of the variable frequency centrifugal unit, including angular velocity, angular displacement, radial force, tangential force and temperature. These variables jointly determine the current state of the system, and the control gain matrix By feeding back these status information, corresponding adjustment signals are generated so that the system can respond to dynamic changes in a timely manner. , we need to solve the Riccati equation first. This is The core step of controller design. The Riccati equation is an important tool in control theory, especially in the design of robust control. It provides a method to optimize system performance so that the system can maintain stable and efficient operation in the presence of disturbances. The Riccati equation is calculated by the system state matrix and the input matrix Describes the dynamic behavior of the system. These matrices are derived from the state-space model of the variable frequency centrifugal unit. The state-space model describes the change of the system state over time in the form of differential equations, such as how the angular velocity is affected by the input power, and the dynamic response of temperature and force. The state matrix reflects the interaction between system states, while the input matrix It describes how the control input affects the system state.

[0103] In the process of solving the Riccati equation, the matrix is a key variable that represents the system's sensitivity to disturbances. By solving this matrix, the system can determine how to adjust the input power in different states to minimize the impact of disturbances. The term reflects the output matrix of the system , which defines which state variables are the control targets of the system. In the application of variable frequency centrifugal units, angular velocity and temperature are two key output variables, so the output matrix The corresponding elements in are set to non-zero values. This indicates that the controller pays special attention to the stability of angular velocity and temperature, because these two variables are directly related to the operating efficiency and safety of the system. The external disturbance matrix Describes how the external environment affects the system. For example, a sudden increase in load may cause significant fluctuations in angular velocity, while changes in ambient temperature may affect the system's thermal management performance. The introduction of the control matrix allows the controller to predict the effects of these external factors and compensate for them by adjusting the input power. The solution is obtained by combining the state matrix of the system , input matrix , perturbation matrix And the output matrix By optimizing these matrices, the controller can obtain an optimal feedback gain matrix to ensure that the system can effectively resist external disturbances under various working conditions.

[0104] An important control design parameter is Performance Indicators , which reflects the system's ability to suppress external disturbances. The value means that the system has a stronger ability to suppress disturbances, because Defines the maximum disturbance that the system can withstand. If the value is too small, the controller can respond more actively to external disturbances, so that it can make adjustments before the disturbances become apparent, thus ensuring the stability of the system. The value may cause the system to be overly sensitive to noise, so in actual design, it needs to be optimized according to the specific operating environment and disturbance characteristics. The controller can find the best balance between disturbance suppression capability and system stability by adjusting the value. In the actual application of variable frequency centrifugal units, external disturbances may come from a variety of factors, including sudden changes in load, fluctuations in ambient temperature, and wear of the equipment itself. These disturbances will have a significant impact on the dynamic behavior of the system, causing fluctuations in angular velocity, temperature, and other key parameters. If these fluctuations are not effectively controlled, they may lead to a decrease in system efficiency and even cause safety problems. Therefore, the use of The robust control strategy of the controller can effectively cope with these complex working conditions. By monitoring the system status in real time and adjusting the input power according to the feedback signal, the system can respond immediately when disturbances occur, thereby maintaining the stability of key variables such as angular velocity and temperature. In addition, the solution of the Riccati equation Not only used to calculate the control gain matrix , and also provides system status feedback information. By optimizing the solution, the controller can determine which states of the system are most sensitive under the current disturbance conditions, and make targeted adjustments. In other words, the solution process of the Riccati equation is equivalent to finding an optimal control strategy for the system, which can minimize the impact of external disturbances on the system and maintain the robustness of the system under various complex working conditions.

[0105] Embodiment 9: The final control input power matrix is ​​calculated by the following formula:

[0106] ;

[0107] in, Input power matrix for final control.

[0108] Specifically, through the formula , the system performs a weighted average of the control signals of the two different mechanisms, robust control and fuzzy neural network control. This weighted design helps to balance the impact of the two control strategies under different working conditions. The controller mainly deals with the uncertainty and external disturbance of the system, and is suitable for working under the conditions of rapid load changes or large fluctuations in the external environment; while the fuzzy neural network controller can better handle the nonlinearity and fuzziness in the system, and is suitable for scenarios where the system state is stable but there are slight changes. Therefore, combining these two control strategies and weighted averaging can ensure that the system can consider disturbances and nonlinear problems in a dynamic environment. On the basis of weighted averaging, the formula also adds a load compensation term , which is the control input estimated and generated by the load torque observer. The introduction of the load compensation term ensures that the system can adapt to changes in the load in real time. Especially under dynamic load conditions, the load torque observer can accurately estimate the current load torque and generate corresponding compensation control signals. The core of load compensation is to ensure that the system can cope with rapid changes or uncertainties in the load, such as load fluctuations in the production process or changes in the operating status of the equipment. By adding , the system can further adjust the power input to ensure that load changes do not have too much impact on the stability of the system.

[0109] This integration of triple control signals ensures that the variable frequency centrifugal unit can achieve precise control under various complex working conditions. The controller provides strong anti-disturbance capabilities, and it can handle sudden changes in the external environment and uncertainties in the system's internal parameters. Secondly, the fuzzy neural network controller handles the nonlinear characteristics of the system through adaptive adjustment of fuzzy logic and neural networks, especially when the load is relatively stable but there are slight changes, fuzzy control can provide more precise adjustments. Finally, load compensation ensures that the system can cope with the challenges brought by load changes by estimating the load torque in real time and generating corresponding compensation signals. This multi-dimensional combination of control strategies greatly enhances the robustness, flexibility and response speed of the system. The final control input power matrix The generation process not only takes into account the current system state, but also ensures that the system can maintain the best operating state under various load conditions through the load compensation mechanism and the adaptive adjustment of the fuzzy neural network. The biggest advantage of this control method is that it can simultaneously cope with a variety of complex situations such as external disturbances, load changes and system nonlinearity. By weighted integration of different control strategies, the system can make corresponding dynamic adjustments when facing different types of working conditions, which can not only respond quickly to external disturbances, but also accurately adjust the nonlinear state of the system, and finally achieve stable and efficient operation. Compared with the traditional single control strategy, the multiple control signal fusion method in the present invention shows significant technical advantages. Traditional control methods usually rely on only one control mechanism, such as robust control or fuzzy control, but each of them has limitations. Robust control performs well in dealing with external interference, but may not be fine enough when dealing with system nonlinearity or small fluctuations; fuzzy control is good at dealing with nonlinear problems, but may not perform well when dealing with external interference. By combining these control strategies, the system can cope with a variety of complex situations at the same time, greatly enhancing the flexibility and robustness of control.

[0110] Embodiment 10: An intelligent control system for a variable frequency centrifugal unit, the system comprising: a state space model unit, for establishing a state space model of the variable frequency centrifugal unit based on the dynamic characteristics of the variable frequency centrifugal unit; an adaptive fuzzy neural network control unit, for designing an adaptive fuzzy neural network controller based on the state space model, combining fuzzy logic with neural network to achieve control of nonlinear systems, and the adaptive fuzzy neural network controller can be dynamically adjusted by updating adaptive weights to adapt to different state changes of the variable frequency centrifugal unit; the adaptive fuzzy neural network controller outputs an adaptive power matrix; a load compensation change unit, for estimating load changes through a load torque observer, and calculating a load compensation control input matrix, thereby ensuring stable operation of the variable frequency centrifugal unit under dynamic load changes; a control output unit, for calculating a robust power matrix through a robust H∞ controller based on the state space model; and calculating the final control input power matrix of the variable frequency centrifugal unit by combining the adaptive power matrix, the robust power matrix and the load compensation control input matrix.

[0111] Specifically, in the production process of the industry, variable frequency centrifugal units are used for material separation. Due to the frequent load changes in the process (for example, the feed amount and material characteristics will fluctuate from time to time), coupled with the fluctuations in external ambient temperature and changes in operating conditions, the centrifugal unit has frequent changes in dynamic characteristics and nonlinear problems during operation. Traditional control methods are difficult to cope with such complex dynamic loads and system nonlinearity. Therefore, it is necessary to design an intelligent control system to ensure the stable and efficient operation of the unit. At system startup, the state space model unit establishes a state space model of the system based on the dynamic characteristics of the centrifugal unit (such as speed, torque, vibration, temperature and other parameters). The model mathematically describes the various operating state variables of the centrifugal unit and constructs a multi-input and multi-output dynamic system model. The system monitors the dynamic state of the centrifugal unit in real time according to the state space model.

[0112] Based on the state space model, an adaptive fuzzy neural network control unit designed an adaptive fuzzy neural network controller. This controller can dynamically adjust the weight of fuzzy rules according to the real-time operating status of the centrifuge unit, combining fuzzy logic and the self-learning characteristics of neural networks, thereby achieving real-time adaptive control of nonlinear problems. When the feed rate suddenly increases, the system detects changes in angular velocity and torque, and the adaptive fuzzy neural network controller immediately adjusts the control parameters to generate a new adaptive power matrix , ensuring that the centrifugal unit can maintain stable operation under new load conditions. Since material properties often change, the load torque observer continuously estimates the load changes of the system. When the load changes, the load compensation change unit calculates the load compensation control input matrix , real-time compensation for system dynamic deviations caused by load changes. For example, when the load torque suddenly increases, causing the system speed to decrease, the load compensation unit immediately calculates the compensation signal and feeds the compensation control signal back to the control system to ensure that the centrifuge unit maintains normal speed and operating stability. Based on the state space model and real-time monitoring data, robust The controller solves the Riccati equation to calculate the robust power matrix This control signal is used to suppress external disturbances and internal uncertainties of the system, ensuring that the system can maintain good performance in a changing environment. The controller can adjust the system input signal to suppress the impact of ambient temperature fluctuations on system temperature and efficiency. Finally, the system controls the output unit and combines the adaptive power matrix , Robust Power Matrix and load compensation control input matrix , through the formula , calculate the final control input power matrix . This control input directly adjusts the operating power of the centrifuge unit to ensure that the system can still operate efficiently and stably under dynamic load and disturbance conditions. Through the real-time adjustment and compensation of the intelligent control system, the variable frequency centrifuge unit can quickly respond to load changes and external disturbances and maintain a stable working state. Specifically, when the feed rate fluctuates, the system can dynamically adjust the speed and power to avoid system overload or efficiency loss. When the external ambient temperature fluctuates violently, the system can effectively suppress the impact of temperature on the equipment through robust control to ensure that the temperature fluctuates within a reasonable range. The load compensation system effectively solves the torque fluctuation problem caused by changes in material properties and maintains the normal operation of the system. In the end, the energy consumption of the centrifuge unit was reduced by about 10% during the entire production process, the failure rate was significantly reduced, and the production efficiency was increased by 15%, realizing the intelligent and efficient control of the system.

[0113] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. Intelligent control method of variable frequency centrifugal unit, It is characterized in that The method comprises: Step 1: Based on the dynamic characteristics of the variable frequency centrifugal unit, a state space model of the variable frequency centrifugal unit is established; Step 2: Based on the state space model, an adaptive fuzzy neural network controller is designed to combine fuzzy logic with neural network to realize the control of nonlinear system. By updating the adaptive weight, the adaptive fuzzy neural network controller can be dynamically adjusted to adapt to different state changes of the variable frequency centrifugal unit; the adaptive fuzzy neural network controller outputs an adaptive power matrix; Step 3: Estimate the load change through the load torque observer and calculate the load compensation control input matrix, so as to ensure the stable operation of the variable frequency centrifugal unit under the condition of dynamic load change; Step 4: Based on the state space model, the robust power matrix is ​​calculated by the robust H∞ controller; the final control input power matrix of the variable frequency centrifugal unit is calculated by combining the adaptive power matrix, the robust power matrix and the load compensation control input matrix; The state space model of the variable frequency centrifugal unit is expressed using the following formula: ； in, is the angular velocity; is the angular displacement; is the radial force; is the tangential force; is temperature; is the input power; is the moment of inertia; is the damping coefficient; and are radial stiffness coefficient and tangential stiffness coefficient respectively; is the temperature influence coefficient; For quality; and are radial and tangential unbalanced masses respectively; and are radial and tangential damping coefficients, respectively; For efficiency; is the power of the variable frequency centrifugal unit; is the specific heat capacity; is the heat transfer coefficient; is the surface area; For time; is the differential symbol, Indicates time The differential of The adaptive fuzzy neural network controller in step 2 is expressed by the following formula: ； in, It is The adaptive weights of the fuzzy rules, is the fuzzy basis function, is the number of fuzzy rules; , is the state vector of the variable frequency centrifugal unit; is the state feedback gain matrix; is the subscript index; is the adaptive power matrix; The fuzzy basis function is a Gaussian membership function, which is expressed by the following formula: ； in, For the The central value of the fuzzy basis function; For the The width parameter of the fuzzy basis function indicates the extension degree of the membership function; the adaptive fuzzy neural network controller continuously adjusts the adaptive weight through the following formula : ； in, Represents adaptive weight The rate of change of is an adaptive parameter that controls the speed of updating the adaptive weights. The larger it is, the faster the adaptive weights are adjusted; , represents the state vector of the current variable frequency centrifugal unit The state vector of the desired variable frequency centrifugal unit By calculating the error Amplification, used to adjust the adaptive weights to reduce the error; It is a positive definite matrix designed according to Lyapunov stability theory; The load torque observer in step 3 is expressed using the following formula: ； in, Input matrix for load compensation control; is the proportional gain, which controls the speed and amplitude of compensation; is the estimated load torque; is the expected load torque; is the integral gain, the control system's ability to compensate for the accumulated error; is the differential gain, which controls the response to the rate of change of the load torque; is the integration variable; Estimated load torque Calculate using the following formula: ； in, is the estimated load torque; is the time constant of load torque observation, which determines the response speed of load estimation; is the angular velocity error gain, which controls the sensitivity of the load estimation to the angular velocity error; is the actual angular velocity; is the estimated angular velocity; It is the load torque nonlinear correction gain, which controls the amplitude of nonlinear compensation; is the integral gain, which controls the ability of the integral term to compensate for the accumulated error; Estimated angular velocity Calculate using the following formula: ； in, is the estimated angular velocity; is the angular velocity correction gain, which controls the linear response to the angular velocity error; is the angular velocity nonlinear correction gain, which controls the nonlinear compensation strength of the angular velocity error; The robust H∞ controller is expressed as follows: ； in, is the H∞ control gain matrix; is the robust power matrix; H∞ control gain matrix By solving the following Riccati equation: ； in, is the H∞ control gain matrix; is the solution of the Riccati equation; is the performance index of the H∞ controller, indicating the ability to suppress disturbances. The smaller the value, the stronger the ability to suppress disturbances. The output matrix ; External interference matrix System state matrix and the system input matrix It is obtained through the following process: Based on the state space model of the variable frequency centrifugal unit, the following equation is obtained: ； From this we get: ； 。 2. The intelligent control method of the variable frequency centrifugal unit according to claim 1, It is characterized in that The final control input power matrix is ​​calculated by the following formula: ； in, Input power matrix for final control.

3. An intelligent control system for a variable frequency centrifugal unit, using the intelligent control method for a variable frequency centrifugal unit as claimed in claim 1 or 2, It is characterized in that The system comprises: a state space model unit, which is used to establish a state space model of the variable frequency centrifugal unit based on the dynamic characteristics of the variable frequency centrifugal unit; an adaptive fuzzy neural network control unit, which is used to design an adaptive fuzzy neural network controller based on the state space model, combine fuzzy logic with neural network, and realize control of nonlinear system. Through updating of adaptive weights, the adaptive fuzzy neural network controller can be dynamically adjusted to adapt to different state changes of the variable frequency centrifugal unit; the adaptive fuzzy neural network controller outputs an adaptive power matrix; a load compensation change unit, which is used to estimate load changes through a load torque observer and calculate a load compensation control input matrix, so as to ensure stable operation of the variable frequency centrifugal unit under dynamic load changes; a control output unit, which is used to calculate a robust power matrix through a robust H∞ controller based on the state space model; and a final control input power matrix of the variable frequency centrifugal unit is calculated by combining the adaptive power matrix, the robust power matrix and the load compensation control input matrix.

Images