A Dynamic Coordinated Control System for the Electrical Drive of a Roller Press Based on Material Flow Fluctuations
The electric drive dynamic collaborative control system for material flow fluctuations solves the problems of insufficient perception and rigid control of the roller press under complex working conditions, realizes stable operation and production continuity of the roller press, and improves equipment safety and grinding efficiency.
Patent Information
- Application Number
- CN202611354185.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-02
- Publication Date
- 2026-11-17
AI Technical Summary
Existing roller press control systems suffer from limited sensing capabilities, rigid control logic, and a lack of deep electromechanical-hydraulic coordination when facing complex material flow fluctuations. This results in poor system reliability, susceptibility to oscillations, and impacts on production safety and efficiency.
An electric drive dynamic collaborative control system based on material flow fluctuation is adopted. By acquiring frequency converter signals and roller gap displacement signals, a material flow fluctuation intensity index is generated to achieve dynamic torque balance of the dual roller motors and mechanical-hydraulic collaborative buffering, avoiding the reliance on traditional threshold judgment and external sensors.
It achieves stable control under harsh working conditions, avoids system oscillation, ensures equipment safety and production continuity, and improves grinding efficiency.
Smart Images

Figure CN123102483A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of industrial process control and electrical drive technology, and specifically to a dynamic collaborative control system for the electrical drive of a roller press based on material flow fluctuations. Background Technology
[0002] As a core piece of equipment in material crushing and grinding processes, roller presses typically operate in harsh environments with high dust levels and strong vibrations during actual production. During operation, roller presses frequently face significant impacts from random and drastic fluctuations in material flow. Existing roller press control systems suffer from the following technical shortcomings when dealing with complex material flow fluctuations: First, the sensing methods are limited, resulting in poor system reliability. Because roller presses operate under extreme conditions of high dust and strong vibration, there is a lack of suitable high-precision sensors on the market capable of long-term stable operation in such environments. Existing control systems that rely on externally installed sensors to monitor material flow fluctuations and equipment operating status are not only highly susceptible to contamination, interference, and even damage, but also face a very high risk of failure. This leads to the predicament of lacking a reliable means to accurately obtain physical fluctuation information during actual operation.
[0003] Second, the rigid control logic is prone to causing system oscillations. Existing control schemes generally use traditional discrete multi-level manual threshold judgment logic when dealing with material flow fluctuations. Faced with continuous and random material flow impacts, this discontinuous threshold comparison method cannot achieve a smooth transition, which can easily lead to frequent changes in control commands, thereby causing electrical and mechanical oscillations in the entire system.
[0004] Third, the lack of deep coordination between electromechanical and hydraulic systems can easily damage equipment and severely impact production. Existing equipment lacks a dynamic coordination mechanism between the electrical drive system and the mechanical hydraulic system when facing sudden, large material flow impacts. On the one hand, sudden impacts can cause instantaneous inconsistencies in the forces applied to the moving and stationary rollers, resulting in a severe torque imbalance in the dual-roller drive motor. On the other hand, because the system cannot actively release rotor kinetic energy and simultaneously and flexibly reduce hydraulic holding pressure to achieve buffering and yielding, it is not only prone to material spurring, leading to frequent loading of the hydraulic system, but the huge transient load can also cause the frequency converter to trip due to overcurrent, and even cause a sudden drop in the voltage of the power grid in the plant area. Over time, such severe mechanical and electrical impacts can easily damage the torsion plates or bearings of the equipment, causing serious damage to the mechanical structure of the equipment, significantly reducing the grinding efficiency of the system, and seriously affecting the operational safety and production continuity of the equipment.
[0005] Based on the above-mentioned shortcomings, there is an urgent need for a new control scheme that does not rely on additional precision sensors, eliminates multi-level manual threshold judgment, and can achieve adaptive torque balance of dual-roller motors and seamless electromechanical-hydraulic collaborative buffering, so as to solve the technical problems faced by existing roller presses under harsh working conditions. Summary of the Invention
[0006] In view of this, in order to solve the problems mentioned in the background technology, a dynamic collaborative control system for the electrical drive of a roller press based on material flow fluctuation is proposed.
[0007] The objective of this invention can be achieved through the following technical solution: This invention provides a dynamic collaborative control system for the electrical drive of a roller press based on material flow fluctuations, including: a power transient change rate generation module, used to acquire the stator voltage signal and stator current signal of the frequency converter to calculate the real-time active power, filter the real-time active power and calculate the transient change rate to obtain the active power transient change rate.
[0008] The differential signal generation module is used to acquire the physical roll gap displacement signals on both sides and perform differential processing to obtain the roll gap displacement differential signal.
[0009] The material flow fluctuation intensity index generation module is used to continuously map the transient change rate of active power with the differential signal input state mapping matrix of the roll gap displacement to generate the material flow fluctuation intensity index.
[0010] The coupling compensation gain generation module is used to construct a load torque estimation model. It estimates the actual torque of the moving roller and the actual torque of the stationary roller in real time based on the inverter's operating data, calculates the torque deviation, and multiplies the torque deviation by the material flow fluctuation intensity index to obtain the dynamic cross-coupling compensation gain.
[0011] The torque dynamic balancing module is used to continuously and inversely superimpose the dynamic cross-coupling compensation gain onto the speed loop setpoint of the dual-roller frequency converter to achieve dynamic torque balancing of the dual-roller motor.
[0012] The kinetic energy buffer proportional factor generation module is used to extract the current output current of the frequency converter and the hardware limit safety current set as a multiple of the rated current to calculate the current margin approximation rate, and generate the kinetic energy buffer proportional factor based on the current margin approximation rate.
[0013] The mechanical-hydraulic coordinated buffering module is used to synchronously adjust the speed setpoint on the electrical side and the pressure fine-tuning bias command of the proportional relief valve on the hydraulic side according to the kinetic energy buffering ratio factor, so as to realize mechanical-hydraulic coordinated buffering.
[0014] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention generates the material flow fluctuation intensity index by extracting the electrical signals inside the frequency converter and the basic physical roll gap displacement signals to construct a state mapping matrix. This soft measurement scheme does not rely on external additional precision sensors, effectively avoiding the risk of external sensors being easily contaminated or failing in harsh environments such as high dust and strong vibration at the roller press site, and ensuring the long-term stable operation of the control system.
[0015] (2) This invention breaks away from the traditional step-type manual multi-level threshold comparison logic, and transforms the actual physical fluctuations of the material flow into continuous, threshold-free control commands in real time. By combining this index with the torque deviation, a continuous dynamic cross-coupling compensation gain is generated and superimposed on the speed loop of the dual-roller frequency converter. This mechanism avoids abrupt changes in control commands, realizes adaptive, millisecond-level smooth torque balance of the dual-roller motor, and effectively prevents electrical and mechanical system oscillations caused by uneven force on the dual rolls.
[0016] (3) This invention extracts the current margin approximation rate and generates a "kinetic energy buffer ratio factor". When faced with extreme load impact, the system can actively and synchronously perform two coordinated actions based on this factor: first, actively reduce the motor speed on the electrical side to release the rotor's mechanical rotational kinetic energy to resist the load; second, flexibly reduce the pressure holding pressure on the hydraulic side to allow the moving roller to smoothly retract. This seamless dynamic coordination between electrical control and mechanical hydraulics greatly enhances the equipment's ability to absorb sudden impacts, effectively prevents inverter overcurrent tripping faults, and ensures the stability of the power grid voltage and the continuity of production in the plant area. Attached Figure Description
[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a schematic diagram of the system module structure connection of the present invention.
[0019] Figure 2 This is a three-dimensional surface diagram of the continuous mapping of material flow fluctuation state and intensity index of the present invention.
[0020] Figure 3 This is an analysis diagram of the dynamic cross-coupling balancing process of the dual-roller torque in this invention. Detailed Implementation
[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0022] Please see Figure 1The present invention provides a dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuation, comprising: a power transient change rate generation module, a differential signal generation module, a material flow fluctuation intensity index generation module, a coupling compensation gain generation module, a torque dynamic balance realization module, a kinetic energy buffer proportional factor generation module, and a mechanical-hydraulic coordinated buffer realization module.
[0023] The power transient rate of change generation module and the differential signal generation module are both connected to the material flow fluctuation intensity index generation module. The material flow fluctuation intensity index generation module is connected to the coupling compensation gain generation module. The coupling compensation gain generation module is connected to the torque dynamic balance realization module. The torque dynamic balance realization module is connected to the kinetic energy buffer proportional factor generation module. The kinetic energy buffer proportional factor generation module is connected to the mechanical-hydraulic synergistic buffer realization module.
[0024] The power transient change rate generation module is used to acquire the stator voltage signal and stator current signal of the frequency converter, calculate the real-time active power, filter the real-time active power, and calculate the transient change rate to obtain the active power transient change rate.
[0025] In a specific embodiment of the present invention, the process of obtaining the stator voltage signal and stator current signal of the frequency converter to calculate the real-time active power, filtering the real-time active power and calculating the transient rate of change to obtain the transient rate of change of active power includes: obtaining the stator voltage signal and stator current signal of the frequency converter to calculate the real-time active power.
[0026] The system receives real-time active power and inputs it into a low-pass filter for high-frequency electromagnetic noise filtering to obtain smoothed active power.
[0027] The power change gradient is obtained by taking the time-dimension derivative of the smoothed active power.
[0028] The absolute value of the power change gradient is extracted as the transient rate of change of active power.
[0029] Specifically, this step aims to extract characteristic quantities that characterize the severity of load impact from the inverter's basic electrical signals. First, the stator voltage and stator current signals of the inverter driving the motor are acquired. These two signals are digital representations of real-time three-phase electrical signals that can be directly read from within the inverter's control system. The real-time active power is calculated using these two signals. Real-time active power refers to the instantaneous power by which the motor converts electrical energy into mechanical energy and transfers it to the load, reflecting the actual energy consumption of the load. Since the acquired electrical signals inevitably contain high-frequency electromagnetic noise caused by factors such as the operation of switching devices, directly using the real-time active power would introduce interference; therefore, filtering of the real-time active power is necessary.
[0030] The calculated real-time active power is input into a low-pass filter. A low-pass filter is a signal processing unit that allows signal components below a certain cutoff frequency to pass through while attenuating or filtering out signal components above that cutoff frequency. Through this operation, high-frequency electromagnetic noise can be effectively filtered out, resulting in smooth active power free from high-frequency disturbances.
[0031] Subsequently, to quantify the severity of load changes, the smoothed active power is differentiated over time. In discrete digital control systems, differentiation calculates the rate of change of the signal between adjacent sampling times. The result of this operation is defined as the power change gradient, which represents how quickly the smoothed active power changes over time.
[0032] Finally, extract the absolute value of the power change gradient. Since load surges, whether increasing or decreasing, manifest as drastic power changes, we focus on the magnitude of the change rather than its direction. The absolute value of the power change gradient is the transient rate of change of active power; it is a scalar value that directly reflects the intensity of the load surge at the current moment.
[0033] Real-time active power can be calculated using the following formula after transforming from a three-phase stationary coordinate system to a two-phase rotating coordinate system (dq coordinate system):
[0034] In the formula, for Real-time active power at any given moment; and They are respectively The stator voltage signal at any given moment in the d-axis and q-axis components in the dq coordinate system; and They are respectively The stator current signal at any given time is represented by its d-axis and q-axis components in the dq coordinate system. The coefficient 1.5 is introduced to ensure constant power when using coordinate transformations that do not conserve power.
[0035] The low-pass filtering process is implemented using a first-order digital low-pass filter, and its mathematical model is as follows:
[0036] In the formula, The smoothed active power output at the current moment; The real-time active power input at the current moment; This represents the smoothed active power output at the previous sampling time. The sampling period of the control system is set to 0.001 seconds based on the system hardware processing capabilities and control accuracy requirements. For the filter coefficients, which determine the filter's cutoff frequency, the calculation formula is as follows:
[0037] In the formula, This is the cutoff frequency of the low-pass filter. Based on spectral analysis of 200 sets of measured data from industrial sensors, the power changes caused by load fluctuations are mainly concentrated below 10 Hz, while electromagnetic noise frequencies are generally higher than 50 Hz. Set to 20 Hz to retain effective load fluctuation information while filtering out noise.
[0038] The power change gradient was calculated using the backward difference method for numerical differentiation:
[0039] In the formula, for The power change gradient at time step. The dimensions of this formula are power / time, consistent with the physical meaning of the rate of change. The calculation of the transient rate of change of active power is as follows:
[0040] In the formula, for The transient rate of change of active power at any given time; This indicates the operation of taking the absolute value.
[0041] For example, suppose that under a certain operating condition, the sampling period of the frequency converter control system is... Set to 0.001 seconds. This is the cutoff frequency of the low-pass filter. The frequency is set to 20 Hz. First, calculate the filter coefficients. : .
[0042] exist At any given time, the system is in a relatively stable state, and the real-time active power is collected and calculated. 30,000 watts, smooth active power It is 30,000 watts. In the current... At a certain moment, due to the load impact caused by the entry of material, the stator voltage and stator current signals of the frequency converter are collected, and after coordinate transformation and calculation, the real-time active power is calculated. It is 35,000 watts.
[0043] Will Input a low-pass filter and calculate the smoothed active power at the current moment. : Next, the time-dimension derivative of the smoothed active power is calculated to obtain the power change gradient. : Watts / second. Finally, the absolute value of the power change gradient is extracted to obtain the transient rate of change of active power. : Watts / second. Ultimately, the transient rate of change of active power is obtained as 558,000 watts / second, which will be used in subsequent calculation steps.
[0044] The differential signal generation module is used to acquire the physical roll gap displacement signals on both sides and perform differential processing to obtain the roll gap displacement differential signal.
[0045] In a specific embodiment of the present invention, the process of acquiring physical roll gap displacement signals on both sides and performing differential processing to obtain differential roll gap displacement signals includes: acquiring physical roll gap displacement signals on both sides of the moving roll.
[0046] The arithmetic mean of the physical roll gap displacement signals on both sides is taken to obtain the average roll gap displacement.
[0047] The average roll gap displacement is subjected to numerical difference calculation based on adjacent sampling times to obtain the differential signal of the roll gap displacement.
[0048] Specifically, after calculating the "transient rate of change of active power" characterizing the electrical impact of the load, this step aims to obtain a characteristic quantity from a mechanical perspective that can synchronously reflect the physical impact of the load. This process begins with acquiring the physical roll gap displacement signals on both sides. In a twin-roll mill, material passes through the gap between the two rolls; this gap is called the roll gap. The physical roll gap displacement signals on both sides are physical quantities measured in real time by displacement sensors installed on the bearing seats or frame on both sides of the moving roll. They accurately represent the instantaneous position of the moving roll relative to the stationary roll on both sides, thus reflecting the actual width of the roll gap. Due to uneven material distribution or slight deviations in equipment installation, the width of the roll gap on both sides may not be completely consistent, therefore, separate acquisition is required.
[0049] To obtain a single signal representing the overall roll gap change, the signals from both sides need to be processed comprehensively. Typically, the arithmetic mean is taken to obtain the average roll gap displacement. Then, this average roll gap displacement is differentiated. The purpose of differentiation is to calculate the rate of change of the roll gap displacement over time, which directly reflects the speed at which the moving roll is forced to retreat or move due to sudden changes in material hardness or thickness. A high rate of change implies a severe mechanical impact. This differentiation is implemented in the digital control system through numerical difference operations, and the result is the differential roll gap displacement signal. This signal, together with the obtained transient rate of change of active power, constitutes a multi-dimensional description of material flow fluctuations.
[0050] First, calculate the average roll gap displacement:
[0051] In the formula, for Average roll gap displacement at any given time; and They are respectively The values of the physical roll gap displacement signals collected at any time on the left and right sides.
[0052] Next, the average roll gap displacement is differentiated to obtain the differential roll gap displacement signal:
[0053] In the formula, for The differential signal of the roll gap displacement at any given moment; It is the average roll gap displacement at the current sampling moment; It is the average roll gap displacement at the previous sampling time; This refers to the sampling period of the data acquisition and control system. To ensure the time alignment of electrical and mechanical signals, the sampling period here... Consistent with the sampling period used in the steps above, it is set to 0.001 seconds. The dimensions of this formula are length / time, i.e., velocity, which is completely consistent with the physical meaning of the rate of change of roll gap displacement.
[0054] For example, in a specific implementation scenario, the system's sampling period Set to 0.001 seconds. At that moment, the physical roll gap displacement signals collected by the system on both sides were as follows: Left side displacement The displacement is 5.000 mm on the right side. It is 5.002 mm. First, calculate... Average roll gap displacement at time 1 : Millimeters.
[0055] In the present At a certain moment, due to a piece of hard material passing through the roll gap, causing the roll gap to expand rapidly, the physical roll gap displacement signals collected by the system on both sides are updated to: left displacement. The displacement was 5.015 mm on the right side. It is 5.019 mm. Calculate the current... Average roll gap displacement at time 1 : Millimeters. Finally, differentiation is performed to calculate the differential signal of the roll gap displacement. : mm / s.
[0056] The above calculations yield a differential signal of 16 mm / s for the current roll gap displacement. This value quantifies the roll gap expansion rate caused by material impact and will be used as a characteristic quantity describing the intensity of mechanical impact, inputting it along with the transient rate of change of active power into subsequent steps.
[0057] See Figure 2 The material flow fluctuation intensity index generation module is used to continuously map the transient change rate of active power with the differential signal input state mapping matrix of the roll gap displacement to generate the material flow fluctuation intensity index.
[0058] In a specific embodiment of the present invention, the transient rate of change of active power and the differential signal of roll gap displacement are continuously mapped to the input state mapping matrix to generate a material flow fluctuation intensity index, including: obtaining the transient rate of change of active power and the differential signal of roll gap displacement to construct a two-dimensional state vector.
[0059] By inputting the two-dimensional state vector into the state mapping matrix whose weights are the reciprocals of the expected maximum values of each state variable, a linear algebraic transformation is performed to obtain the comprehensive fluctuation characteristic value.
[0060] The comprehensive fluctuation characteristic value is normalized to generate a material flow fluctuation intensity index ranging from zero to one.
[0061] Specifically, the core task of this step is to fuse the fluctuation characteristics extracted from the electrical and mechanical dimensions in the previous steps to generate a single index that can comprehensively and comprehensively evaluate the severity of the current material flow fluctuation. This process acquires the calculated transient rate of change of active power and the differential signal of the roll gap displacement, and constructs these two independent scalar values into a two-dimensional state vector. The two-dimensional state vector is a mathematical object containing two ordered elements; its position in the two-dimensional coordinate space can simultaneously represent the magnitude of both the electrical and mechanical impacts.
[0062] Next, the constructed two-dimensional state vector is input into a threshold-free state mapping matrix for linear algebraic transformation. The state mapping matrix is essentially a set of predefined weighting coefficients used to weight the two components of the two-dimensional state vector. The linear algebraic transformation here is specifically a matrix-vector multiplication operation, mapping the two-dimensional input information into a one-dimensional scalar output. This process is threshold-free, meaning there is a continuous functional relationship between the output and input; any small change in input will produce a corresponding continuous change in the output, thus ensuring the smoothness of the evaluation. The result of this transformation is a comprehensive fluctuation characteristic value, a dimensionless value that integrates electrical and mechanical impact information.
[0063] Finally, to give this characteristic value a clear physical meaning and a unified comparison benchmark, the comprehensive fluctuation characteristic value is normalized. Normalization maps a value within an arbitrary range to a specific, finite interval using a mathematical function. In this embodiment, this target interval is zero to one. The final result generated after processing is the material flow fluctuation intensity index, a value between zero and one, where zero indicates that the system is completely stable with no fluctuations, and one indicates that the system has withstood the maximum fluctuation impact within its design range.
[0064] First, construct a two-dimensional state vector:
[0065] In the formula, for A two-dimensional column vector of states at time t; for The transient rate of change of active power at time t, with the dimension of power / time; for The differential signal of the roll gap displacement at time t is expressed in units of length / time.
[0066] By performing a linear algebraic transformation, the comprehensive fluctuation eigenvalues are obtained:
[0067] In the formula, for The comprehensive fluctuation characteristic value at any given time. It is a state mapping matrix, which is a row vector. Because... and Since their dimensions are different, direct addition has no physical meaning; therefore, the weighting coefficient... and It is also used here as a coefficient to eliminate dimensions. and The settings are based on the normal operating limit of their respective physical quantities, and are specifically set as follows: ,
[0068] In the formula, It is the expected maximum value of the transient rate of change of active power. These are the expected maximum values of the differential signal of the roll gap displacement. The setting of these two maximum values is based on statistical analysis of measured data from over 300 sets of similar equipment under extreme operating conditions, ensuring coverage of the vast majority of operating scenarios. In this way, and All are converted to dimensionless values, and their sum is... It is also dimensionless.
[0069] In one specific embodiment of the present invention, and The specific method for obtaining the data is as follows: During the full-load commissioning phase of the roller press, the maximum permissible instantaneous feeding impact condition is artificially created, and the absolute value of the power change gradient at the instant the frequency converter reaches 150% of its rated power (maximum permissible overload capacity) is recorded as the data. Simultaneously, record the displacement rate when the physical roll gap is widened to the mechanical limit protection point as... .
[0070] Finally, normalization and amplitude limiting are performed to generate the material flow fluctuation intensity index:
[0071] In the formula, for The intensity index of material flow fluctuation at any given time. This is the normalization coefficient, whose value is equal to the theoretical maximum sum of the maximum values of the physical quantities corresponding to the weight coefficients in the state mapping matrix. (The outer layer of the formula...) The function is a saturation limiting function, used to force the calculation result to be clamped to 1 when the system encounters an extreme material impact that exceeds design expectations, thereby ensuring... The value of is strictly limited and always falls between zero and one to prevent subsequent compensation instructions from exceeding the limit and causing system oscillations. In this embodiment, when and hour, Reaching the theoretical maximum value, that is Therefore, setting Thus ensuring The value of is strictly between zero and one.
[0072] For example, the example calculation result following the above steps, in the current... At time t, the transient rate of change of active power The differential signal of the roll gap displacement is 558,000 watts per second. The value is 16 mm / s. Based on the analysis of a large amount of historical data, the expected maximum value of the transient rate of change of active power is set. The expected maximum value of the differential signal of the roll gap displacement is 2,000,000 watts per second. It is 50 mm / s.
[0073] Construct the two-dimensional state vector at the current moment : Then, determine the state mapping matrix. coefficient: Seconds per watt Seconds per millimeter. Therefore, Next, the two-dimensional state vector is input into the state mapping matrix and subjected to linear algebraic transformation to calculate the comprehensive fluctuation eigenvalues. : This value is a dimensionless numerical value that comprehensively reflects the fluctuation situation at the current moment.
[0074] Finally, the comprehensive fluctuation characteristic value is normalized and limited to generate the material flow fluctuation intensity index. Set the normalization coefficient. The final material flow fluctuation intensity index was 0.2995. This value, between zero and one, intuitively indicates that the current material flow fluctuation is approximately 29.95% of the maximum fluctuation that the equipment can withstand, providing a precise quantitative basis for subsequent dynamic compensation and buffer control.
[0075] The coupling compensation gain generation module is used to construct a load torque estimation model. It estimates the actual torque of the moving roller and the actual torque of the stationary roller in real time based on the inverter's operating data, calculates the torque deviation, and multiplies the torque deviation by the material flow fluctuation intensity index to obtain the dynamic cross-coupling compensation gain.
[0076] In a specific embodiment of the present invention, a load torque estimation model is constructed, and the actual torque of the moving roller and the actual torque of the stationary roller are estimated in real time based on the inverter operating data. The torque deviation is calculated, and the dynamic cross-coupling compensation gain is obtained by multiplying the torque deviation by the material flow fluctuation intensity index. This includes: collecting the stator current and rotor speed of the inverter as inverter operating data.
[0077] The inverter operating data is input into the load torque estimation model based on the second-order sliding mode observer. The sliding mode surface and state equation are constructed using the rotor speed error and stator current to perform state observation, and the actual torque of the moving roller and the actual torque of the stationary roller are output.
[0078] The torque deviation is obtained by calculating the absolute value of the difference between the actual torque of the moving roller and the actual torque of the stationary roller.
[0079] Multiplying the torque deviation by the material flow fluctuation intensity index yields the dynamic cross-coupling compensation gain.
[0080] Specifically, this step aims to generate a compensation signal for dynamically adjusting the load on the dual-roller motor. This process first requires accurately estimating the actual load on each of the two rolls and, combined with the material flow fluctuation severity obtained from previous steps, calculating a dynamic compensation amount. This is achieved by collecting operating data from the respective frequency converters driving the moving and stationary rolls. Here, frequency converter operating data specifically refers to key electrical quantities that reflect the motor's operating status: the stator current output by the frequency converter to the motor and the rotor speed fed back by the motor encoder. These two data points are standard measurable parameters in the frequency converter control system, providing necessary input for subsequent torque estimation.
[0081] After acquiring the inverter's operating data, it is input into a pre-built load torque estimation model. This model is a technique that calculates load torque without relying on physical torque sensors, using only the mathematical model of the motor and measurable parameters. In this embodiment, a load torque estimation model based on a second-order sliding mode observer is used. This is a robust state observation technique that constructs a mathematical observation model parallel to the actual motor model and uses sliding mode control theory to force the output of the observation model to track the output of the actual system, thereby accurately deriving unknown disturbances in the system, i.e., load torque. This process is performed independently for the moving roller motor and the stationary roller motor, thus outputting the actual torque of the moving roller and the actual torque of the stationary roller.
[0082] In one specific embodiment of the present invention, the observer is established based on the mechanical motion equation of the motor, and its observer state equation is as follows: When the system reaches the sliding surface At that time, the unknown load torque is characterized by a sliding mode control law, i.e. = (If the superspiral algorithm is used, it is characterized by its integral term). In the formula, The observed rotor speed; Let the system's rotational inertia be denoted by . The electromagnetic torque of the motor can be calculated from the stator current of the frequency converter; The load torque that needs to be estimated (i.e., the actual torque of the moving or stationary roller). The coefficient of friction; This is a second-order sliding mode control law. The sliding surface is defined. ,in Given the actual rotor speed acquired by the frequency converter, the second-order sliding mode control law... Designed as follows: . In the formula, , For sliding mode observer gain, The sign function. When the system reaches the sliding surface, the sliding control law... This is equivalent to an unknown load torque disturbance, thereby achieving control over the actual torque. The precise back-calculation estimate.
[0083] The mathematical model of an asynchronous motor can be represented in a two-phase rotating coordinate system (dq coordinate system) as follows:
[0084] The relationship between electromagnetic torque and load torque is as follows:
[0085]
[0086] In the formula, This represents the number of pole pairs of the motor. For mutual inductance between stator and rotor; For rotor self-inductance; and These represent the components of the rotor flux linkage along the d-axis and q-axis, respectively. The second-order sliding mode observer estimates the load torque by observing the error between the estimated and actual values of the rotor current. Its output is the actual torque of the moving roller. Actual torque of the fixed roller .
[0087] In the formula, The stator current has d and q axis components; For the d and q axis components of the rotor flux linkage; The d and q-axis components of the stator voltage; Electromagnetic torque; This is the load torque; This refers to the mechanical angular velocity of the motor. It is the moment of inertia; This represents the number of pole pairs of the motor. For mutual inductance and rotor inductance; This is a coefficient that includes the motor's resistance, inductance, and speed. These parameters are all known parameters provided by the motor manufacturer.
[0088] After obtaining the actual torques of the two motors, the degree of imbalance between them is calculated. The torque deviation is obtained by calculating the absolute value of the difference between the actual torque of the moving roller and the actual torque of the stationary roller. This value directly quantifies the load difference between the two drive systems at the current moment.
[0089]
[0090] In the formula, for Torque deviation at any given time; This represents the actual torque of the moving roller; This represents the actual torque of the fixed roller.
[0091] Finally, this torque deviation is multiplied by the generated material flow fluctuation intensity index. The material flow fluctuation intensity index, acting as a weighting factor between zero and one, adjusts the intensity of the compensation. When the material flow fluctuation is severe, the index approaches one, allowing for significant compensation; when the material flow is stable, the index approaches zero, suppressing the compensation effect and avoiding unnecessary adjustments. The result of this multiplication is the dynamic cross-coupling compensation gain, a value with torque dimensions. Its magnitude reflects not only the torque imbalance but also whether the current system is in a state requiring strong intervention.
[0092]
[0093] In the formula, for Dynamic cross-coupling compensation gain at any given time; for The intensity index of material flow fluctuation at any given time. Due to... It is dimensionless, therefore Dimensions and The value is consistent with that of Newton-meters (N·m), which aligns with the physical meaning of compensation gain as an additional quantity in the velocity loop.
[0094] For example, taking the calculation results from the previous steps, in the current... At any given time, the intensity index of material flow fluctuation It is 0.2995.
[0095] The system collects inverter operating data for both the moving roller motor and the stationary roller motor. Assuming the moving roller side experiences a greater impact, its inverter operating data is: stator current (q-axis component) 250 A, rotor speed 1480 rpm. The stationary roller side has a lighter load, and its inverter operating data is: stator current (q-axis component) 220 A, rotor speed 1485 rpm.
[0096] The two sets of inverter operating data are input into their respective load torque estimation models based on second-order sliding mode observers. After state observation and calculation within the models, the actual torques are output. Assume the model estimates the actual torque of the moving roller. Newton-Mehr. Actual torque of the fixed roller. 00 Newtons per meter.
[0097] Calculate torque deviation : Newton-meters. This result indicates that at the current moment, the moving roller is bearing 15,000 more Newton-meters of load torque than the stationary roller.
[0098] Finally, the torque deviation is multiplied by the material flow fluctuation intensity index to calculate the dynamic cross-coupling compensation gain. : The final calculated dynamic cross-coupling compensation gain is 4492.5 Newton-meters. This value will serve as the core basis for dynamically adjusting the speed of the dual-roller motor in subsequent steps to achieve load rebalancing.
[0099] See Figure 3 The torque dynamic balancing module is used to continuously and inversely superimpose the dynamic cross-coupling compensation gain onto the speed loop setpoint of the dual-roller frequency converter to achieve dynamic torque balancing of the dual-roller motor.
[0100] In a specific embodiment of the present invention, the dynamic cross-coupling compensation gain is continuously and inversely superimposed onto the speed loop setpoint of the dual-roller frequency converter to achieve dynamic torque balancing of the dual-roller motor, including: obtaining the dynamic cross-coupling compensation gain and determining the load distribution characteristics of the dual-roller motor by judging the load lightness and heaviness state of the dual-roller motor.
[0101] Based on the load distribution characteristics, the dynamic cross-coupling compensation gain is allocated into positive compensation signals and negative compensation signals.
[0102] The negative compensation signal is superimposed on the speed loop setpoint of the motor on the heavily loaded side, and the positive compensation signal is superimposed on the speed loop setpoint of the motor on the lightly loaded side, thus completing the dynamic torque balance.
[0103] Specifically, the purpose of this step is to convert the calculated dynamic cross-coupling compensation gain into an actual adjustment command for the speed of the dual-roller motor, thereby actively and dynamically balancing the load torque borne by the two motors. The entire process is continuous and closed-loop, ensuring a rapid response to load imbalances.
[0104] This process obtains the calculated dynamic cross-coupling compensation gain and needs to determine which motor is currently under heavier load. By comparing the estimated actual torque of the moving roller with that of the stationary roller, the load status of the dual-roller motors can be determined. This comparison result is the load distribution characteristic, which is a clear directional piece of information indicating which of the moving and stationary rollers is the heavily loaded motor and which is the lightly loaded motor.
[0105] Based on this load distribution characteristic, the dynamic cross-coupling compensation gain is allocated into positive and negative compensation signals. The core idea of this "reverse superposition" is: for motors with heavy loads, their speed setpoint needs to be reduced to "yield" some load; for motors with light loads, their speed setpoint needs to be increased to "bear" more load. Therefore, the dynamic cross-coupling compensation gain needs to be converted into a speed regulation quantity. To keep the overall average speed of the system constant, the compensation quantity is usually divided equally, with half used to increase the speed and half used to decrease the speed.
[0106] Finally, these compensation signals are applied to the control loops of their respective frequency converters. Specifically, a positive compensation signal aimed at increasing speed is superimposed on the speed loop setpoint of the lightly loaded motor, while a negative compensation signal aimed at decreasing speed is superimposed on the speed loop setpoint of the heavily loaded motor. The speed loop setpoint is the input point in the frequency converter's internal control logic that sets the target motor speed. Through this reverse, cross-compensation method, the system can quickly reduce torque deviation and achieve dynamic torque balance of the dual-roller motor.
[0107] First, determine the load distribution characteristics by comparing the actual torque of the moving roller. Actual torque of the fixed roller Implementation. Then, the dynamic cross-coupling compensation gain will be applied. (Unit: Newton-meter) Convert to speed adjustment. Since the unit at the speed loop input is speed (e.g., radians / second), a conversion factor is needed to ensure dimensional consistency.
[0108]
[0109] In the formula, It is the total speed adjustment amount; This is a torque-speed conversion coefficient, measured in radians per second (N·m). The coefficient is set based on system identification of the motor and load characteristics, or through online experimental tuning, to ensure the stability and response speed of the compensation process. Based on optimization through over a hundred simulation experiments, this coefficient is set... This is an empirical value that enables the system to recover torque balance within 0.5 seconds after a typical shock.
[0110] The total speed adjustment is distributed into positive compensation signals and negative compensation signals:
[0111]
[0112] In the formula, This is a positive compensation signal; This is a negative compensation signal.
[0113] Finally, update the speed loop setpoint based on the load distribution characteristics: if Then the moving roller is the motor on the heavy-load side, and the stationary roller is the motor on the light-load side.
[0114]
[0115] like Then the moving roller is the motor on the lightly loaded side, and the stationary roller is the motor on the heavily loaded side.
[0116]
[0117] In the formula, and These are the new speed loop setpoints for the moving roller and stationary roller motors, respectively. and It is the original velocity loop setpoint before compensation.
[0118] For example, taking the calculation results from the previous steps, in the current... At any given moment, dynamic cross-coupling compensation gain The actual torque of the moving roller is 4492.5 Newton-meters. The actual torque of the fixed roller is 120,000 Newton-meters. The initial velocity loop setpoint of the system is 105,000 Newton-meters. and Both are 157.08 radians / second (corresponding to 1500 revolutions per minute). Torque-speed conversion factor. Set to 0.01 (radians / second) / (Newton-meter).
[0119] The dynamic cross-coupling compensation gain is obtained as 4492.5 N·m. Comparing the actual torques of the two motors, 120000 > 105000, therefore the load distribution characteristics are: the moving roller is the motor on the heavily loaded side, and the stationary roller is the motor on the lightly loaded side. The compensation signal is allocated according to the load distribution characteristics. First, the total speed regulation is calculated: Radius per second. Then, it is assigned as positive and negative compensation signals: positive compensation signal. Radius per second. Negative compensation signal. Radius per second.
[0120] The compensation signal is superimposed onto the speed loop setpoint. Since the moving roller is the motor on the heavily loaded side and the stationary roller is the motor on the lightly loaded side, the new speed loop setpoint for the moving roller motor is: Radius per second. New speed loop setpoint for the fixed roller motor: Radius per second.
[0121] This adjustment causes the loaded moving roller motor to slow down, thus reducing its torque output; while the lightly loaded stationary roller motor is instructed to speed up, thus increasing its torque output. This process will bring the loads of both motors closer together, ultimately achieving dynamic torque balance.
[0122] The kinetic energy buffer proportional factor generation module is used to extract the current output current of the frequency converter and the hardware limit safety current set as a multiple of the rated current to calculate the current margin approximation rate, and generate the kinetic energy buffer proportional factor based on the current margin approximation rate.
[0123] In a specific embodiment of the present invention, the current margin approximation rate is calculated by extracting the current output current of the frequency converter and the hardware limit safety current set as a multiple of the rated current, and the kinetic energy buffer ratio factor is generated based on the current margin approximation rate, including: obtaining the current output current of the frequency converter and dividing it by the hardware limit safety current to obtain the basic approximation ratio.
[0124] The current margin approximation rate is obtained by performing nonlinear function mapping on the basic approximation ratio.
[0125] The current margin approximation rate is input into the scaling factor generation function for continuous transformation to generate the kinetic energy buffer scaling factor.
[0126] Specifically, when the system encounters extreme load shocks, even after torque balancing adjustments, the inverter output current may still approach its hardware protection limit. This step aims to generate a proportional factor for initiating and regulating the mechanical-hydraulic co-buffered operation, proactively preventing the inverter from tripping due to overcurrent.
[0127] First, obtain the current output current of the frequency converter. Since the dual-roller is driven by two frequency converters, the one with the larger current is chosen as the evaluation criterion, as it represents the maximum electrical risk currently faced by the system. The current output current of the frequency converter refers to the effective value of the three-phase current at the output terminal of the frequency converter or its vector magnitude in the dq coordinate system, which is a key parameter for real-time monitoring within the frequency converter. Simultaneously, obtain the hardware limit safe current of the frequency converter, a fixed value determined by the frequency converter's hardware specifications, representing the maximum output current at which the frequency converter can operate safely continuously. Dividing the current output current of the frequency converter by the hardware limit safe current yields a dimensionless ratio, the basic approximation ratio, which linearly reflects the degree of proximity between the current current and the limit current. The hardware limit safe current refers to the critical current value that the frequency converter's power devices (IGBTs) can withstand without triggering overcurrent protection tripping; it is typically set to 1.5 to 2.0 times the rated current of the frequency converter, hence the multiple is set to 1.5 to 2.0 times.
[0128] Subsequently, a nonlinear function mapping is applied to the basic approximation ratio. Directly using a linear ratio may cause the system to be overly sensitive when the current is far from its limit, or under-respond when it is close to its limit. The purpose of the nonlinear function mapping is to suppress buffering action when the current is low, and to rapidly amplify its effect once the current exceeds a certain safety threshold, thereby generating an indicator that more accurately reflects the "degree of danger approach." The processed result is the current margin approximation rate, a value between zero and one, which nonlinearly characterizes the urgency of the current margin being consumed.
[0129] Finally, the current margin approximation rate is input into the proportional factor generation function for continuous conversion. This function directly converts the current margin approximation rate, representing the urgency level, into a proportional factor that can be used for control. In this embodiment, the function is a direct mapping, meaning the output equals the input. The resulting kinetic energy buffer proportional factor is a continuously adjustable value ranging from zero to one, which will serve as the intensity command for subsequent mechanical-hydraulic coordinated buffering actions.
[0130] First, calculate the fundamental approximation ratio:
[0131] In the formula, for The basic approximation ratio at any given time; for The current output current of the inverter at any given moment is the larger of the two inverter output currents. This is the hardware's maximum safe current, which is the rated or peak current specified on the inverter's nameplate; it is a fixed parameter.
[0132] By performing a nonlinear function mapping on the basic approximation ratio, the current margin approximation rate is obtained:
[0133] In the formula, for Current margin approximation rate at time step. This is a preset activation threshold used to define the safe zone and the warning zone. Based on the analysis of inverter safety operating procedures and over 500 hours of equipment operation data, this threshold is set to 0.8 to avoid the control system becoming overly sensitive, while ensuring sufficient safety margin. When the basic approximation ratio is below 0.8, the system is considered to be in the safe zone, and the approximation rate is zero. When the ratio enters the warning zone (0.8 to 1.0), the approximation rate increases rapidly with a quadratic relationship, reflecting the accelerated accumulation of risk. When the ratio exceeds 1, the approximation rate reaches its maximum value of 1.
[0134] Finally, the kinetic energy buffer scaling factor is generated through the scaling factor generation function:
[0135] In the formula, for The kinetic energy buffer scaling factor at time t. The generating function here is an identity mapping, directly using the current margin approximation rate after nonlinear processing as the final control scaling factor.
[0136] For example, suppose a frequency converter has a hardware-limited safe current. The activation threshold for the nonlinear mapping is 400 amps. Set to 0.8. In the current... At that moment, due to the severe load impact, the system monitored that the output currents of the two frequency converters were 375 amps and 380 amps, respectively.
[0137] Get the current output current of the frequency converter Choose the larger of the two, that is Okay. Then calculate the fundamental approximation ratio. : The basic approximation ratio Input a nonlinear function for mapping processing. Because... The second segment of the applicable function. Calculate the current margin approximation rate. : .
[0138] Approximation of current margin Input the scaling factor generation function to obtain the kinetic energy buffer scaling factor. : The final kinetic energy buffer ratio factor is 0.5625. This value, between zero and one, precisely quantifies that the current system needs to buffer at 56.25% of its maximum buffer capacity. This value will be used in the next step to synchronously adjust the buffering action on the electrical and hydraulic sides to protect the equipment.
[0139] The mechanical-hydraulic coordinated buffering module is used to synchronously adjust the speed setpoint on the electrical side and the pressure fine-tuning bias command of the proportional relief valve on the hydraulic side according to the kinetic energy buffering ratio factor, so as to realize mechanical-hydraulic coordinated buffering.
[0140] In a specific embodiment of the present invention, the speed setpoint on the electrical side and the pressure fine-tuning bias command of the proportional relief valve on the hydraulic side are adjusted synchronously according to the kinetic energy buffer ratio factor to achieve mechanical-hydraulic coordinated buffering, including: obtaining the speed attenuation coefficient by subtracting the product of the kinetic energy buffer ratio factor and the maximum allowable speed reduction ratio.
[0141] The speed attenuation coefficient is multiplied by the initial speed setpoint to obtain a new speed setpoint on the electrical side, which guides the motor to actively decelerate in order to release the rotor's mechanical rotational kinetic energy.
[0142] The kinetic energy buffer ratio factor is converted into a proportional relief valve pressure fine-tuning bias command on the hydraulic side, reducing the holding pressure to allow the moving roller to retreat smoothly.
[0143] Specifically, this step is the core of the mechanical-hydraulic coordinated buffering action. Based on the generated kinetic energy buffering ratio factor, it synchronously sends commands to the electric drive system and the hydraulic control system to absorb and mitigate severe load shocks in a coordinated manner. This process consists of two parallel adjustment paths: one on the electrical side and one on the hydraulic side.
[0144] On the electrical side, the obtained kinetic energy buffer scaling factor is first subtracted from the numerical value to obtain an intermediate variable called the velocity attenuation coefficient. The kinetic energy buffer scaling factor characterizes the required buffer strength, and its value is between zero and one; therefore, the value range of the velocity attenuation coefficient is also correspondingly between zero and one. When no buffer is needed, the kinetic energy buffer scaling factor is zero, and the velocity attenuation coefficient is one, indicating that the velocity does not attenuate; when maximum buffer strength is required, the kinetic energy buffer scaling factor is one, and the velocity attenuation coefficient is zero, indicating that the velocity needs to attenuate to the maximum extent.
[0145] Next, the calculated speed attenuation coefficient is multiplied by the initial speed setpoint. The initial speed setpoint is the target speed set during normal, stable system operation. The result of the multiplication is the new electrical-side speed setpoint. This new setpoint will be lower than the initial value. Upon receiving this instruction, the inverter will actively reduce the motor speed. The motor rotor and its connected mechanical components store a large amount of mechanical rotational kinetic energy during high-speed rotation. The process of reducing the speed releases this stored kinetic energy to counteract external impact loads, thereby reducing the pressure on drawing power from the grid. The direct effect is a reduction in the inverter's output current, avoiding overcurrent protection.
[0146] On the hydraulic side, the same kinetic energy buffering proportional factor is converted into a proportional relief valve pressure fine-tuning bias command. The proportional relief valve is a key component in the hydraulic system used to set and maintain the holding pressure of the moving roller. The conversion process here involves multiplying the dimensionless kinetic energy buffering proportional factor by a preset maximum allowable pressure drop, resulting in a specific pressure adjustment. This adjustment acts as a negative bias command to reduce the holding pressure of the hydraulic system. Reducing the holding pressure means decreasing the clamping force of the hydraulic cylinder on the moving roller, allowing the moving roller to more easily and smoothly retract when impacted by hard materials, buffering impact energy through physical displacement and protecting the mechanical structure. Through the synchronized and coordinated actions of releasing kinetic energy on the electrical side and allowing retraction on the hydraulic side, synergistic mechanical-hydraulic buffering is achieved.
[0147] Using one minus the kinetic energy buffer ratio factor With the maximum permissible reduction ratio The product of these factors yields the velocity decay coefficient. ,Right now ; Compare the velocity decay coefficient with the initial velocity setpoint Multiply to obtain the new target buffer speed on the electrical side ,Right now This reduces the motor speed to release the rotor's mechanical rotational kinetic energy. The kinetic energy buffer scaling factor... Converted into a proportional relief valve pressure fine-tuning bias command on the hydraulic side The calculation formula is: ,in The maximum permissible pressure drop is set; this bias command reduces the holding pressure, allowing the moving roller to smoothly retract.
[0148] For example, in the current At any given moment, the kinetic energy buffer ratio factor The value is 0.5625. Assume the initial velocity of the system is a given value. The value is 157.08 radians / second (corresponding to 1500 revolutions / minute). The initial holding pressure of the hydraulic system is 10 MPa, and the maximum permissible pressure drop is set according to the equipment safety specifications. It is 5 MPa.
[0149] Assuming the maximum permissible reduction ratio Set to 0.25, the process first calculates the velocity decay coefficient. : Subsequently, the speed setpoint on the electrical side was calculated. : Radius per second. This new speed setpoint (approximately 656 rpm) will be sent to the frequency converter of the twin-roller motor, instructing the motor to decelerate.
[0150] At the same time, calculate the proportional relief valve pressure fine-tuning bias command on the hydraulic side. : MPa. This command will be sent to the hydraulic control system, causing the proportional relief valve to reduce the holding pressure from 10 MPa to Megapascals.
[0151] Ultimately, the electrical system releases kinetic energy by reducing its rotational speed, while the hydraulic system allows for yielding by reducing its pressure. Both operate synchronously based on the same kinetic energy buffering ratio factor, achieving a combined mechanical and hydraulic buffering effect against the current impact.
[0152] In a specific embodiment of the present invention, the method further includes: monitoring the numerical change of the current margin approximation rate to obtain the approximation rate decline state.
[0153] The recovery mechanism is triggered based on the approach rate decline state to generate a recovery curve.
[0154] Obtain the smooth transition parameter in the recovery curve that represents the ratio of the current recovery time to the set total recovery time.
[0155] The transition speed command is obtained by nonlinearly increasing the speed setpoint on the electrical side using smooth transition parameters.
[0156] The transition pressure command is obtained by nonlinearly increasing the holding pressure on the hydraulic side using smooth transition parameters, thus restoring the initial operating state.
[0157] Specifically, after implementing mechanical-hydraulic co-buffering to cope with extreme load shocks, this step aims to define a safe and smooth recovery mechanism that allows the system to automatically recover from the buffered state to the initial operating state after the shock subsides.
[0158] The system continuously monitors the changes in the calculated current margin approximation rate. As the load surge subsides, the motor current decreases, causing the current margin approximation rate to drop. When this value falls from a positive zero to below a preset recovery threshold, the system determines that it has entered the approximation rate decline state. This state is a trigger signal indicating that the buffering requirement has been eliminated and the recovery procedure can be initiated.
[0159] Once the approximation rate fallback state is triggered, the recovery mechanism is activated, generating a recovery curve. The recovery curve is a time-varying function that defines the transition trajectory of the system as it recovers from the buffered state to the initial state. To avoid the recovery process itself causing new system oscillations, this curve must be smooth.
[0160] This describes how to perform the recovery action according to the recovery curve. First, a smooth transition parameter representing the ratio of the current recovery time to the set total recovery time is obtained from the recovery curve. In this embodiment, the smooth transition parameter mainly refers to the total duration of the recovery process. This parameter determines the speed of the recovery process, and its setting is based on considerations of system stability to ensure that the recovery process is sufficiently smooth.
[0161] Subsequently, this smooth transition parameter is used to synchronously perform nonlinear incremental processing on both the electrical and hydraulic sides. The buffer speed value calculated from the speed setpoint on the electrical side increases nonlinearly according to the trajectory defined by the recovery curve until it returns to the initial speed setpoint. This time-varying command is called the transition speed command. Similarly, the holding pressure on the hydraulic side also increases nonlinearly from the low pressure value during buffering, following the same recovery curve trajectory, until it returns to the initial holding pressure. This command is called the transition pressure command. Through this synchronous and smooth nonlinear incremental method, the initial operating state is finally restored.
[0162] The trigger condition for the approximation rate decline state is:
[0163] In the formula, The current margin approximation rate at the current moment; This is the current margin approximation rate at the previous sampling time. It is a preset recovery trigger threshold, which is set to a positive number close to zero, such as 0.01, based on a trade-off between system noise level and sensitivity, to confirm that the impact has completely subsided.
[0164] The recovery curve is composed of a smooth transition function that varies with time. It means that among them This is the time calculated from the start of the recovery mechanism. This function... When it is 0, in The time is 1. This is the smooth transition parameter, representing the total recovery time. To ensure smoothness during startup and termination, a cubic polynomial function is used:
[0165] This function ensures that... and When the rate of change (first derivative) is zero, a smooth start-stop is achieved.
[0166] Calculation of transition speed command:
[0167] In the formula, Transition speed instructions during the recovery process; It is the calculated speed setpoint on the electrical side under buffer conditions; It is the initial velocity setpoint of the system.
[0168] Calculation of transition pressure command:
[0169] In the formula, This is a transitional pressure command during the recovery process; It is the hydraulic holding pressure under buffer conditions; It is the initial holding pressure of the system.
[0170] For example, the system is in a buffered state, assuming it encounters a limit load shock, i.e., the buffer scaling factor. Reaching the maximum value of 1 and When set to 0.25, the speed setpoint on the electrical side The hydraulic holding pressure is 117.81 radians / second. The initial operating parameters of the system are: initial velocity setpoint. (The value is 5 MPa.) Radius per second, initial holding pressure Megapascals.
[0171] The system continuously monitors the current margin approximation rate. Suppose that at a certain moment, its value decreases from 0.015 in the previous period to 0.008 in the current period. Since 0.008 is less than the set recovery trigger threshold... The approximation rate decline state is triggered. The recovery mechanism is activated. Smooth transition parameters are obtained, and the total recovery time is set. The time is 2 seconds. The system generates the recovery curve function based on this. The system begins recovery according to the recovery curve. The recovery process is complete when the system is halfway through... Taking seconds as an example, let's perform the calculation: First, calculate the value of the smooth transition function: Calculate the transition speed command at this moment. : Radius per second. Calculate the transition pressure command at this moment. : Megapascals.
[0172] Within the first second of recovery, the system will send a transition speed command of 137.445 radians / second to the electrical side and a transition pressure command of 7.5 MPa to the hydraulic side. This process will continue for 2 seconds until... hour, The speed and pressure are fully restored to their initial values, completing the restoration of the initial operating state.
[0173] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.
Claims
1. A dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations, characterized in that, include: The power transient change rate generation module is used to acquire the stator voltage signal and stator current signal of the frequency converter to calculate the real-time active power, filter the real-time active power and calculate the transient change rate to obtain the active power transient change rate. The differential signal generation module is used to collect the physical roll gap displacement signals on both sides and perform differential processing to obtain the roll gap displacement differential signal. The material flow fluctuation intensity index generation module is used to continuously map the transient change rate of active power with the differential signal of roll gap displacement input state mapping matrix to generate the material flow fluctuation intensity index. The coupling compensation gain generation module is used to build a load torque estimation model. It estimates the actual torque of the moving roller and the actual torque of the stationary roller in real time based on the inverter's operating data, calculates the torque deviation, and multiplies the torque deviation by the material flow fluctuation intensity index to obtain the dynamic cross-coupling compensation gain. The torque dynamic balancing module is used to continuously and inversely superimpose the dynamic cross-coupling compensation gain onto the speed loop setpoint of the dual-roller frequency converter to achieve dynamic torque balancing of the dual-roller motor. The kinetic energy buffer proportional factor generation module is used to extract the current output current of the frequency converter and the hardware limit safety current set as a multiple of the rated current to calculate the current margin approximation rate, and generate the kinetic energy buffer proportional factor based on the current margin approximation rate. The mechanical-hydraulic coordinated buffering module is used to synchronously adjust the speed setpoint on the electrical side and the pressure fine-tuning bias command of the proportional relief valve on the hydraulic side according to the kinetic energy buffering ratio factor, so as to realize mechanical-hydraulic coordinated buffering.
2. The dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations according to claim 1, characterized in that: The process of acquiring the stator voltage signal and stator current signal of the frequency converter to calculate the real-time active power, filtering the real-time active power, and calculating the transient rate of change to obtain the transient rate of change of active power includes: Calculate the real-time active power by acquiring the stator voltage and stator current signals of the frequency converter; The real-time active power is received and input into a low-pass filter for high-frequency electromagnetic noise filtering to obtain smoothed active power. The power change gradient is obtained by taking the time-dimension derivative of the smoothed active power. The absolute value of the power change gradient is extracted as the transient rate of change of active power.
3. The dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations according to claim 1, characterized in that: The process of acquiring the physical roll gap displacement signals on both sides and performing differential processing to obtain the roll gap displacement differential signal includes: Collect the physical roll gap displacement signals on both sides of the moving roll; The arithmetic mean of the physical roll gap displacement signals on both sides is taken to obtain the average roll gap displacement. The average roll gap displacement is subjected to numerical difference calculation based on adjacent sampling times to obtain the differential signal of the roll gap displacement.
4. The dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations according to claim 1, characterized in that: The step of continuously mapping the transient rate of change of active power with the differential signal of roll gap displacement input state mapping matrix to generate the material flow fluctuation intensity index includes: A two-dimensional state vector is constructed by obtaining the transient rate of change of active power and the differential signal of roll gap displacement; The two-dimensional state vector is input into a state mapping matrix whose weights are composed of the reciprocals of the expected maximum values of each state variable. A linear algebraic transformation is then performed to obtain the comprehensive fluctuation characteristic value. The comprehensive fluctuation characteristic value is normalized to generate a material flow fluctuation intensity index ranging from zero to one.
5. The dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations according to claim 1, characterized in that: The aforementioned load torque estimation model estimates the actual torque of the moving roller and the actual torque of the stationary roller in real time based on the inverter's operating data, calculates the torque deviation, and multiplies the torque deviation by the material flow fluctuation intensity index to obtain the dynamic cross-coupling compensation gain, including: The stator current and rotor speed of the frequency converter are collected as the operating data of the frequency converter; Input the inverter operating data into the load torque estimation model based on the second-order sliding mode observer, construct the sliding mode surface and state equation using the rotor speed error and stator current to perform state observation, and output the actual torque of the moving roller and the actual torque of the stator roller. Calculate the absolute value of the difference between the actual torque of the moving roller and the actual torque of the stationary roller to obtain the torque deviation; Multiplying the torque deviation by the material flow fluctuation intensity index yields the dynamic cross-coupling compensation gain.
6. The dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations according to claim 1, characterized in that: The step of continuously and inversely superimposing the dynamic cross-coupling compensation gain onto the speed loop setpoint of the dual-roller frequency converter to achieve dynamic torque balancing of the dual-roller motor includes: The load distribution characteristics are obtained by acquiring the dynamic cross-coupling compensation gain and determining the load status of the dual-roller motor. Based on the load distribution characteristics, the dynamic cross-coupling compensation gain is allocated into positive compensation signals and negative compensation signals; The negative compensation signal is superimposed on the speed loop setpoint of the motor on the heavily loaded side, and the positive compensation signal is superimposed on the speed loop setpoint of the motor on the lightly loaded side, thus completing the dynamic torque balance.
7. The dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations according to claim 1, characterized in that: The process involves extracting the current output current of the frequency converter and calculating the current margin approximation rate using a hardware-limited safe current set as a multiple of the rated current, and generating a kinetic energy buffer scaling factor based on the current margin approximation rate, including: Obtain the current output current of the frequency converter and divide it by the hardware limit safe current to get the basic approximation ratio; The current margin approximation rate is obtained by performing nonlinear function mapping on the basic approximation ratio. The current margin approximation rate is input into the scaling factor generation function for continuous transformation to generate the kinetic energy buffer scaling factor.
8. The dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations according to claim 1, characterized in that: The method of synchronously adjusting the speed setpoint on the electrical side and the pressure fine-tuning bias command of the proportional relief valve on the hydraulic side according to the kinetic energy buffering ratio factor to achieve mechanical-hydraulic coordinated buffering includes: The velocity attenuation coefficient is obtained by subtracting the product of the kinetic energy buffer ratio factor and the maximum allowable deceleration ratio. Multiplying the speed attenuation coefficient by the initial speed setpoint yields a new speed setpoint on the electrical side, guiding the motor to actively decelerate in order to release the rotor's mechanical rotational kinetic energy; The kinetic energy buffer ratio factor is converted into a proportional relief valve pressure fine-tuning bias command on the hydraulic side, reducing the holding pressure to allow the moving roller to retreat smoothly.
9. The dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations according to claim 1, characterized in that, Also includes: The approximation rate decline state is obtained by monitoring the numerical change of the current margin approximation rate. A recovery curve is generated based on the recovery mechanism triggered by the decline in the approximation rate. According to the recovery curve, the speed setpoint on the electrical side and the pressure holding pressure on the hydraulic side are increased synchronously to restore the initial operating state.
10. A dynamic coordinated control system for the electrical drive of a roller press based on material flow fluctuations, as described in claim 9, is characterized in that: The process of synchronously increasing the speed setpoint on the electrical side and the holding pressure on the hydraulic side according to the recovery curve to restore the initial operating state includes: Obtain the smooth transition parameter in the recovery curve that represents the ratio of the current recovery time to the set total recovery time; The transition speed command is obtained by nonlinearly increasing the speed setpoint on the electrical side using smooth transition parameters. The transition pressure command is obtained by nonlinearly increasing the holding pressure on the hydraulic side using smooth transition parameters, thus restoring the initial operating state.
