Smart integrated flow rate control-type positive displacement pump system and flow rate control method using same

Abstract

The present invention relates to a smart integrated flow rate control-type positive displacement pump system to which smart control is applied by integrating a flowmeter function with a positive displacement pump, and a flow rate control method using same. The smart integrated flow rate control-type positive displacement pump system comprises: the positive displacement pump driven by a connected motor; a signal collector for collecting at least one motor signal of a PWM duty ratio, a current, and a voltage of the motor; a calculator that calculates a driving torque of the pump by receiving the motor signal from the signal collector, estimates the pressure of the pump through the calculated driving torque and the radius of the positive displacement pump, and calculates a flow rate through the calculated driving torque and the pressure of the pump; and a controller for controlling the motor to correspond to a required flow rate.

Description

Smart integrated flow control type volumetric pump system and flow control method utilizing the same

[0001] The present invention relates to a volumetric pump system with a flow rate control, and more particularly, to a smart integrated volumetric pump system with a flow rate control that applies smart control by integrating a flow meter function into a volumetric pump, and a method for controlling flow rate using the same.

[0002] In general, a positive displacement pump is a pump that transports a fixed volume of fluid per rotation or reciprocating motion. A positive displacement pump has the characteristic of moving a fixed volume of fluid per rotation or stroke. Because of this characteristic, the flow rate (Q) in an ideal positive displacement pump is equal to the displacement volume per unit angle of the pump (D p ) and the rotational or reciprocating speed (N). In formula form, it is as follows:

[0003] (Formula 1)

[0004] In actual pumps, flow rate changes occur due to internal leakage and other factors. Controlling the flow rate of a positive displacement pump by adjusting valves or other methods to change pipe resistance is not effective. Traditionally, controlling the flow rate of a positive displacement pump required changing the pump's operating speed or using a bypass line.

[0005] Accurate flow control requires a closed-loop pump speed control system, which requires a flow sensor or flow meter capable of measuring the pump's output flow rate in real time. However, conventional systems require separate pumps and flow meters, requiring installation at the pump outlet or in piping. External flow meters complicate the system, increase costs, and are inefficient in terms of installation space and maintenance. Therefore, there is a growing need for technology that integrates flow measurement into the pump itself, enabling the flow rate to be calculated and controlled to a target level based on the pump's rotational speed without the need for an additional flow meter.

[0006] [Prior Art Literature]

[0007] [Patent Document]

[0008] (Patent Document 1) Republic of Korea Patent No. 10-1800709 (Optimal design method for improving the accuracy of an oval flow meter)

[0009]

[0010] Accordingly, the present embodiment was devised to solve the above-mentioned problem, and proposes a smart integrated flow control type volumetric pump system that can calculate the flow rate through the rotational amount of the pump and control it to a target flow rate without an additional flow rate meter by integrating a flow rate meter function into the pump itself, and a flow rate control method utilizing the same.

[0011]

[0012] The present invention includes a volumetric pump driven by a connected motor, a signal collection unit that collects at least one motor signal among a PWM duty ratio, current, and voltage of the motor, a calculation unit that receives the motor signal from the signal collection unit and calculates a driving torque of the pump, estimates a pressure of the pump through the calculated driving torque and a radius of the volumetric pump, calculates a flow rate through the calculated driving torque and the pressure of the pump, and a control unit that controls the motor to correspond to a required flow rate.

[0013] In addition, the calculation unit includes a leakage model based on a laminar flow equation between plates to calculate leakage occurring within the volumetric pump, and the leakage model is characterized in that it calculates separately the velocity-dependent term and the pressure-dependent term of each leakage path according to the position of the gear.

[0014] In addition, the volumetric pump is characterized by being composed of at least one of an oval gear, a vane, a gear, a piston, and a helical / screw structure, and the operation unit applies a correction parameter according to a ripple characteristic according to the structure of the volumetric pump.

[0015] In addition, the above operation unit is characterized by filtering out noise of the motor signal in advance using a Kalman filter or an observer-based estimation algorithm.

[0016] In addition, the operation unit compares the accumulated flow rate with the target flow rate, and the control unit controls the PWM duty ratio so that the accumulated flow rate and the target flow rate correspond.

[0017] In addition, the above leakage model is based on the coefficient C of the laminar flow equation between plates. s , C d , C f It is characterized by applying an improved formula that includes the gear tip gap and the velocity and pressure terms of the side poles. Here, C s is the geometric coefficient, and C d is the viscous drag coefficient depending on the pump geometry, and C f is the friction coefficient according to the pump shape.

[0018] In addition, the operation unit uses past driving history data, real-time measurement signals, and flow rate errors as learning data to create an AI-based prediction model through machine learning, and the prediction model is characterized in that it is configured to calculate or update the viscosity / leakage coefficient of the fluid in real time.

[0019] In addition, the above calculation unit calculates the leakage coefficient (C) based on the geometric tolerance measurement values ​​such as the gear tip clearance and side clearance of the pump. s , C d , C f ) is predicted in advance and reflected in the leakage flow correction. Here, C s is the geometric coefficient, and C d is the viscous drag coefficient depending on the pump geometry, and C f is the friction coefficient according to the pump shape.

[0020] In addition, the above prediction model is characterized by performing a predictive control function that learns and predicts cumulative flow rate change trends and proactively adjusts the PWM duty ratio of the motor to achieve a target flow rate.

[0021] In addition, the above prediction model is characterized by being configured to detect pump gap change or wear at an early stage and calculate a predictive maintenance time.

[0022] In addition, the present invention relates to a method for calculating a discharge flow rate of a positive displacement pump driven by a motor, comprising: a motor signal collection step in which a sensor and an encoder are coupled to the motor, and current, voltage, and PWM duty cycle (Pulse Width Modulation Duty Cycle) generated by the motor are collected; a pressure and torque estimation step in which torque and rotation speed of the pump are estimated through the collected motor signal, and an operating pressure of the pump is estimated based on the estimated torque; a leakage model application step in which an actual flow rate is calculated based on a leakage model based on laminar flow between flat plates in the theoretical volumetric flow rate of the pump, and the leakage model is applied to correct the leakage flow rate to calculate an actual instantaneous flow rate; a digital integration correction step in which a digital error according to the resolution (PPR) and phase offset of the encoder is modeled and the digital error is corrected for the instantaneous flow rate; and an accumulated flow rate integration step in which the corrected instantaneous flow rate is accumulated over time to calculate a digital accumulated flow rate.

[0023] In addition, in the motor signal collection step, the control unit collects the temperature of the fluid supplied to the pump through a temperature sensor, and in the pressure and torque estimation step, the control unit includes a viscosity correction step for correcting the viscosity parameter in real time based on the measured temperature of the fluid.

[0024] Additionally, it includes a feedback step for comparing the actual flow rate with the calculated flow rate and, if a difference occurs, adjusting the output of the motor to compensate.

[0025] In addition, in a non-transitory computer-readable storage medium having recorded thereon a program for commanding a flow control method to be performed, the program is characterized in that it includes a geometric tolerance-based leakage coefficient prediction and an artificial intelligence-based correction / control algorithm.

[0026]

[0027] The present invention eliminates the need for a separate flow sensor, as the pump itself calculates and controls the flow rate. This reduces system costs, simplifies design and installation, and reduces leak points and improves maintenance convenience due to the lack of sensor installation.

[0028] Furthermore, in addition to basic flow rate calculations using encoder signals, the actual pump flow rate is accurately estimated by applying leakage compensation and viscosity compensation algorithms. As a result, flow rate fluctuations due to fluid temperature changes and pump wear are compensated for in real time, significantly reducing errors compared to the target flow rate. This particularly reduces flow rate errors in low-speed areas and during start / stop operations, which were issues in existing systems.

[0029] Furthermore, by performing motor drive control via a real-time feedback loop, the system responds quickly to load and flow environment changes, stabilizing flow rate. Compared to conventional open control, this system offers increased system stability and a shorter time to reach the target flow rate.

[0030] Furthermore, it can be applied to various types of volumetric pumps (gear pumps, vane pumps, piston pumps, etc.), and its algorithm can be adjusted based on encoder resolution and sensor presence, providing high flexibility. It can be applied to various fields requiring precise flow control, such as hydraulic systems such as accumulators, fuel injection systems, and metering pumping systems, thereby improving quality and performance.

[0031] Additionally, real-time data on the internal conditions of the pump (e.g., leakage level, viscosity change) can be obtained and used for predictive maintenance or system optimization.

[0032] In addition, by predicting the leakage coefficient in advance based on the geometric tolerance information measured during the manufacturing stage, the dependence on experimental correction can be reduced and the predictability of manufacturing quality can be improved.

[0033] Additionally, geometric tolerance-based predictions can be combined with AI-based viscosity correction logic to achieve high flow rate estimation accuracy even in high-temperature, high-viscosity environments.

[0034] Additionally, it provides a cost-effective and intelligent solution compared to existing technologies, and can deliver reliable performance in various industrial fields that require precise flow control.

[0035]

[0036] Figure 1 is a diagram of the overall structure of the present invention.

[0037] Figure 2 is a block diagram of a smart integrated flow control type volumetric pump system.

[0038] Figure 3 is a control and operation flowchart of the present invention.

[0039] Figure 4 is a graph of the idealized flow waveform of an oval gear.

[0040] Figure 5 is a graph of the idealized flow waveform of a helical / screw structure.

[0041] Figure 6 is a graph of the idealized flow waveform of the spa gear structure.

[0042] Figure 7 is a graph of a 4-vein idealized flow waveform.

[0043] Figure 8 is a graph of the idealized piston structure flow waveform.

[0044] Figures 9 and 10 are graphs showing the maximum flow rate error according to PPR.

[0045] Figure 11 is an example of cumulative error change according to phase offset.

[0046] Figure 12 is an example of leakage flow in an oval flow meter.

[0047] Figures 13 and 14 are exemplary diagrams showing the leakage flow between the gear tip and the housing.

[0048] Figure 15 is an example diagram showing the leakage flow between the side and the housing.

[0049] Figures 16 to 18 are examples showing that the area where leakage flow actually occurs changes depending on the rotation angle of each gear, depending on the velocity of the fluid.

[0050] Figures 19 to 21 are examples showing that the area where leakage flow occurs due to the pressure gradient actually changes depending on the rotation angle of each gear.

[0051] Figures 22 and 23 show the leakage flow, Q, by modifying the pitch curve and geometric parameters of the oval gear. sv and Q sp This is an example of calculating .

[0052] Figure 24 is an artificial intelligence-based correction and control flowchart.

[0053]

[0054] The present invention is susceptible to various modifications and embodiments. Specific embodiments are illustrated in the drawings and described in detail. However, this is not intended to limit the invention to specific embodiments, but rather to encompass all modifications falling within the spirit and technical scope of the present invention.

[0055] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs.

[0056] Terms defined in commonly used dictionaries should be interpreted to have a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense unless expressly defined in this application.

[0057] Referring to the attached drawings below, a smart integrated flow control type volumetric pump system and a flow control method using the same, which are embodiments of the present invention, will be described in detail.

[0058] Figure 1 is a schematic diagram of the overall structure of the present invention. The purpose of the present invention is to calculate the flow rate discharged from the pump through a motor signal generated from the motor when the pump is in operation, and to discharge an accurate flow rate by controlling the output of the motor.

[0059] Referring to Figure 1, the volumetric pump system consists of a motor, a pump, an encoder, a temperature sensor, a display, and a board. The motor is coupled to the pump and transmits power to operate it. The pump discharges a constant flow rate of fluid supplied from the suction side to the discharge side. The encoder collects the PWM duty ratio generated by the motor. The temperature sensor measures the temperature of the fluid supplied to the pump. The display outputs data collected and generated from the motor. The board receives data collected from the motor, encoder, temperature sensor, etc., and outputs data according to a pre-entered algorithm. At this time, the data includes generated torque, pressure, flow rate calculation, control feedback, etc.

[0060] The board includes a signal collection unit that collects motor signals from the motor, and a computation unit that estimates pressure, torque, and flow rate using the collected motor signals. Furthermore, the computation unit calculates the amount of leakage occurring within the pump through a mathematical analysis model, reduces errors due to encoder resolution, and integrates the accumulated flow to calculate the result. The flow rate is then adjusted by considering the viscosity of the fluid using collected temperature data.

[0061] Afterwards, the actual flow rate is compared with the calculated flow rate, and the difference is fed back through motor output control. The collected and calculated data can be used as AI-based correction data to build an optimization system.

[0062] Fig. 2 is a block diagram of a smart integrated flow control type positive displacement pump system. Referring to Fig. 2, the positive displacement pump (100) system includes a positive displacement pump (100) driven by a connected motor (110), a signal collection unit (120) that collects at least one motor signal among a PWM duty ratio, current, and voltage of the motor (110), a calculation unit (130) that receives the motor signal from the signal collection unit and calculates the driving torque of the pump (100), estimates the pressure of the pump through the calculated driving torque and the geometric characteristics of the positive displacement pump, and calculates the flow rate through the calculated driving torque and the pressure of the pump, and a control unit (140) that controls the motor (110) to correspond to a required flow rate.

[0063] The pump consists of a positive displacement pump such as an oval gear, vane, screw, or piston. The motor consists of a BLDC or servo motor. The signal collection unit collects PWM duty ratio (Pulse Width Modulation Duty Cycle), current, and voltage in real time through a sensor and an encoder (200).

[0064] The flow rate can be calculated using pressure and torque by referring to Equations 2, 3, and 4 below.

[0065] (Formula 2)

[0066] (Formula 3)

[0067] (Formula 4)

[0068] Here, is the flow rate, is the displacement per unit angle of the pump.

[0069] Equation 2 shows that the flow rate is proportional to the pump's displacement volume (D) and rotational speed (n). Equation 3 shows that the power is expressed as the product of the pressure difference and the flow rate. Equation 4 shows that the power (the product of torque and rotational speed) is expressed as the pressure difference and the pump's displacement per unit angle. Torque can be calculated from the motor signal of the motor connected to the pump, and the pump pressure can be estimated from the mechanical torque and the pump's geometric characteristics.

[0070] The current and voltage of the BLDC or servo motor driving the pump, PWM duty ratio, etc. are closely related to the pump driving torque and rotation speed, and through this, the pressure and torque required to calculate the flow rate can be estimated.

[0071] (Formula 5)

[0072] Here, T is the torque, I is the current measured in real time, and k t is the torque constant of the motor. Next, the relationship between the PWM duty ratio (d) and the rotation speed (n) is as follows.

[0073] (Formula 6)

[0074] Here, k n is the speed constant, d is the PWM duty cycle, k b is the counter electromotive force coefficient, V bemf is the counter electromotive force. Finally, the pressure of the pump can be estimated from the torque due to the pressure and the displacement volume per unit angle of the pump as follows.

[0075] (Formula 7)

[0076] Additionally, the pump temperature (T f ) is measured by a temperature sensor (121), and the viscosity μ(T) is measured through this. f ) can improve the accuracy of flow rate calculation by reflecting changes in the flow rate.

[0077] The calculation unit includes a leakage model based on a laminar flow equation between plates to calculate leakage occurring within the volumetric pump, and the leakage model is characterized in that it calculates separately the velocity-dependent term and the pressure-dependent term of each leakage path according to the position of the gear.

[0078] Leakage models for fluid machinery have evolved from the Wilson model, which is based on laminar flow between flat plates (Newton's law of viscosity), to detailed models that reflect the geometric characteristics of each pump. The present invention is based on various leakage models that reflect the geometric characteristics of various fluid machines, and is explained based on the Wilson model, the most representative leakage model. The leakage equation in the gap between the flat plates is as follows: Equation 8.

[0079] (Formula 8)

[0080] In the above formula, Q l Silver leakage flow rate, C s is the leakage coefficient (dimensionless coefficient), D is the exclusion volume, Δp is the pressure difference, and μ is the viscosity.

[0081] Equation 8 can be obtained from Equations 1 and 9 below. In the Wilson model, the preceding term in Equation 9 (the velocity term in Equation 20) is ignored assuming that the rotor is symmetrical and that the leakage toward the suction side and the leakage toward the discharge side are the same.

[0082] (Formula 1)

[0083] (Formula 9)

[0084] Coefficient C based on the laminar flow equation between plates s , C d , C f It is characterized by applying an improved formula that includes the gear tip gap and the velocity and pressure terms of the side poles.

[0085] By distinguishing the pressure gradient and viscosity terms for each leakage path, the velocity term and pressure term can be separated as in Equations 10 and 11 below.

[0086] (Formula 10)

[0087] (Formula 11)

[0088] Here, Cs is a geometric coefficient, Cd is a viscous drag coefficient depending on the pump geometry, and Cf is a friction coefficient depending on the pump geometry.

[0089] The computational unit of the present invention is configured to estimate a leakage coefficient (C) by utilizing geometric tolerance data obtained before and after pump manufacturing. This method predicts leakage flow rate in advance due to effective area or flow resistance calculated based on gaps (such as side gaps and tip gaps) and shape errors between each component.

[0090] In particular, by utilizing precision processing technology (CMM, etc.) to measure key gaps (geometric tolerances) at the component manufacturing stage and reflecting the values ​​in the leakage model, it is possible to calculate the leakage flow rate according to viscosity or operating pressure conditions without experimental correction, which is effective for design verification and quality assurance.

[0091] Because geometric tolerances are not the same for all products during processing, these coefficients were traditionally determined experimentally when evaluating the performance of volumetric pumps. However, with advancements in measurement technology, geometric tolerances can now be predicted by precisely measuring the dimensions of each component during the manufacturing process. This allows the present invention to predict leakage coefficients during the manufacturing process, even without experimental determination.

[0092] This has the effect of reducing the burden of repetitive experimental calibration, reducing quality variation within the manufacturing process, and simultaneously improving the reliability of pump-based flow estimation and manufacturing efficiency.

[0093] The computational unit is characterized by pre-filtering noise from the motor signal using a Kalman filter or an observer-based estimation algorithm. The Kalman filter is an algorithm used to estimate the state of a dynamic system. It statistically processes noise contained in sensor measurements and the uncertainty of the system itself to accurately predict the current state. The Kalman filter can predict the next state using a mathematical model (state equation) of the motor. Furthermore, it corrects the predicted value using actual sensor (voltage, current) measurements. During this process, the state estimate with the smallest error is calculated by considering both prediction and measurement errors.

[0094] The Observer-Based Estimation Algorithm (OBEA) is an algorithm that estimates the unmeasurable internal state of a system. It can estimate the motor's state without attaching sensors to the motor. It constructs a mathematical model representing the motor's state, receives the voltage and current applied to the motor, and continuously corrects the error between the model's predicted values ​​and the actual system output to restore the internal state (e.g., back electromotive force).

[0095] It has the advantage of being able to collect accurate information through prior noise filtering.

[0096] In addition, the operation unit compares the accumulated flow rate with the target flow rate, and the control unit controls the PWM duty ratio so that the accumulated flow rate and the target flow rate correspond. By controlling the output of the motor, the actual required flow rate is satisfied.

[0097] Fig. 3 is a control and operation flowchart of the present invention. It shows a control and operation flowchart for calculating a flow rate from a motor signal. Referring to Fig. 3, the flow rate calculation method is a flow rate calculation method for calculating a discharge flow rate of a positive displacement pump driven by a motor, comprising: a motor signal collection step in which a sensor and an encoder are coupled to the motor, and current, voltage, and PWM duty cycle (Pulse Width Modulation Duty Cycle) generated from the motor are collected; a pressure and torque estimation step in which the torque and rotation speed of the pump are estimated through the collected motor signal, and the operating pressure of the pump is estimated based on the estimated torque; a leakage model application step in which the leakage flow rate is corrected based on a leakage model based on laminar flow between flat plates in the theoretical volumetric flow rate of the pump to calculate an actual instantaneous flow rate; a digital integration correction step in which a digital integration error according to the resolution (PPR) and phase offset of the encoder is modeled and the digital error is corrected for the instantaneous flow rate; and an accumulated flow rate integration step in which the corrected instantaneous flow rate is accumulated over time to calculate a digital accumulated flow rate.

[0098] The motor signal collection stage collects the current, voltage, and PWM duty ratio generated by the motor when the pump is running through one or more sensors and encoders coupled to the motor.

[0099] The pressure and torque estimation step estimates pressure and torque using the collected motor signals. Torque is estimated using Equation 5 below.

[0100] (Formula 5)

[0101] Here, T is the torque, I is the current measured in real time, and k t is the torque constant of the motor.

[0102] The rotation speed of the motor can be calculated using the measured PWM duty ratio (d) using Equation 6 below.

[0103] (Formula 6)

[0104] Here, k n is the speed constant, d is the PWM duty cycle, k b is the counter electromotive force coefficient, V bemf is the counter electromotive force.

[0105] The pressure of the pump can be estimated from the torque due to pressure and the radius of the pump as shown in Equation 7 below.

[0106] (Formula 7)

[0107] The leakage model application step calculates the actual flow rate by correcting the leakage flow rate based on a leakage model based on laminar flow between plates. A detailed explanation of the leakage flow rate correction will be provided later.

[0108] The digital integral compensation stage models and compensates for the digital integral error according to the encoder's resolution and phase offset.

[0109] The encoder's resolution (PPR: Pulse Per Revolution) significantly impacts the accuracy of digital flow measurement. When the encoder pulse rate is low, ±maximum errors occur depending on the sampling point. To compensate for this, the present invention applies a mathematical model to calculate the digital integration error and precisely compensate for the flow rate.

[0110] (Formula 12)

[0111] Also, the phase offset of the encoder If this exists, periodic fluctuations occur in the integration error, and accordingly, the correction coefficient is defined as in Equation 13.

[0112] (Formula 13)

[0113] Corrected digital flow error δQ dig is calculated based on this coefficient as in Equation 14.

[0114] (Formula 14)

[0115] Here, k represents the correction factor of Equation 13, and Q vol refers to the theoretical flow rate of Equation 2. An explanation of the RMS flow rate error according to the increase in encoder PPR is provided later in Fig. 11.

[0116] The viscosity compensation step measures pump temperature using a sensor, reflecting changes in viscosity to improve flow rate calculation accuracy. The computational unit then inputs viscosity parameters for the fluid based on temperature. Furthermore, viscosity parameters can be compensated based on data input in real time.

[0117] The cumulative flow integration step calculates the digital cumulative flow by accumulating the corrected instantaneous flow rate over time.

[0118] It includes a feedback step that compares the actual flow rate with the calculated flow rate and, if a difference occurs, adjusts the output of the motor to compensate.

[0119] Fig. 4 is a graph of an idealized flow waveform of an oval gear. The positive displacement pump of the present invention is configured with at least one of an oval gear, a vane, a spur gear, a piston, and a helical / screw structure, and the calculation unit is characterized in that it applies a correction parameter according to the ripple characteristic according to the structure of the positive displacement pump. Referring to Fig. 4, an idealized flow waveform of an oval gear flow meter is shown. Two oval gears mesh and discharge a constant volumetric flow rate, but ripple may occur depending on the rotation angle. This ripple can be theoretically approximated as in Equation 15.

[0120] (Formula 15)

[0121] Figure 5 is a graph of the idealized flow waveform for a helical / screw configuration. Referring to Figure 5, helical or screw pumps are known to have virtually no flow ripple, as they continuously transport fluid through a helical path. The theoretical model has a nearly constant average flow rate, and ripple components appear at high frequencies.

[0122] (Formula 16)

[0123] Figure 6 is a graph of the idealized flow waveform of a spur gear structure. Referring to Figure 6, a spur gear pump transports fluid through meshing linear gears, resulting in significant periodic flow rate fluctuations due to gear meshing. The flow rate can be approximated by a ripple of triplet harmonics.

[0124] (Formula 17)

[0125] Figure 7 is a graph of a four-vane idealized flow waveform. Referring to Figure 7, it visually represents the flow waveform generated during ideal rotation in a positive displacement pump with four vanes. Four flow ripples occur within one rotational cycle, and each ripple can be ideally approximated as a sinusoidal shape proportional to the change in pump chamber volume.

[0126] Theoretical flow rate Q theory =V d x Although constant at ω, the actual flow rate Q(t) exhibits pulsating characteristics depending on the number of vanes and eccentricity. This flow rate waveform can be expressed as the equation below, which serves as the basis for explaining the cumulative error ΔE and maximum error limit according to digital measurement and encoder resolution (PPR).

[0127] (Formula 18)

[0128] Figure 8 is a graph of the idealized flow waveform of a piston structure. Referring to Figure 8, the piston pump generates a continuous flow rate based on the reciprocating motion per revolution, and has very high ripple. The ideal waveform can be approximated by the following series of continuous functions.

[0129] (Formula 19)

[0130] Figures 9 and 10 are graphs showing the maximum flow rate error according to pulse per revolution (PPR). The encoder's resolution (PPR: Pulse Per Revolution) significantly affects the accuracy of digital flow measurement. When the encoder pulse rate is low, ±maximum errors occur depending on the sampling point. To compensate for this, the present invention applies a mathematical model to calculate the digital integral error and precisely compensate for the flow rate.

[0131] Referring to Figures 9 and 10, when calculating the flow rate using the digital integration method, the maximum error occurring in the flow rate measurement depending on the encoder resolution (PPR: Pulse Per Revolution) ) is expressed. This graph explains the phenomenon that errors occur because digital sampling is performed in pulses at regular intervals, even though the actual flow rate is a continuous analog function. Theoretically, the worst-case error can be expressed by the equation below. The curve in Fig. 10 visualizes the maximum error limit calculated by Equation 12. It can be seen that the integration error decreases rapidly as the PPR increases.

[0132] (Formula 12)

[0133] Therefore, as the PPR increases, the integration interval becomes tighter and the accumulated error decreases rapidly. The flow correction algorithm of the present invention calculates the theoretical limit of digital error based on the above formula and uses it as a correction factor.

[0134] Figure 11 is an example of cumulative error change according to phase offset. The phase offset (θ) that occurs when the pulse generation point of the encoder does not match the phase of the actual flow curve. off ) is a graph analyzing the integration error. This models the effect of angular error on the accumulated flow rate, and can derive a correction coefficient as in Equation 13.

[0135] Phase offset of the encoder (θ off) exists, periodic fluctuations occur in the integration error, and the correction coefficient is defined as follows.

[0136] (Formula 13)

[0137] The digital error calculated based on this correction factor is calculated as in Equation 14, and the graph in Fig. 11 shows the phase offset (θ off ) shows the cumulative error curve according to the change in value.

[0138] (Formula 14)

[0139] Such phase compensation plays a crucial role in reducing the cumulative error of a quantitative dispensing system.

[0140] Referring to Figure 11, it can be seen that the flow rate error decreases as the encoder PPR increases. For example, when the pump's displacement volume is a 'specified volume' and the rated speed is a 'specified speed', the RMS flow rate error decreases by a 'specified %' when the encoder PPR is increased from 4 to 8.

[0141] In another example, when the vane structure is a 4-vane structure, the error, which was up to ±25%, converges to nearly 0% when the PPR is 8 or higher. This example numerically demonstrates that the accuracy of flow rate measurement can be improved.

[0142] Figure 12 is an example of a leakage flow diagram in an oval flow meter. Referring to Figure 12, this group of drawings analyzes the fluid leakage path within an oval gear by decomposing it into cross sections.

[0143] To calculate the leakage flow considering the geometric characteristics of the oval flow meter, the main leakage paths must be analyzed. Figure 12 shows the main leakage paths of the oval flow meter. (Q t )1 and (Q t ) The arrow marked 2 represents the leakage flow between the gear tip and the housing. Another leakage is (Q s )1 and (Q s)2 is the leakage between the side of the gear and the housing as indicated by the arrow. The leakage amount of each path is described later.

[0144] Figures 13 and 14 are exemplary diagrams illustrating the leakage flow between the gear tip and the housing. The leakage flow in the oval flow meter is analyzed using basic leakage theory and oval gear kinematics, as well as the following basic assumptions: 1) All fluids are in a laminar flow state. 2) The gap between the gear tip and the housing, as well as the gap between the gear flank and the housing, are all flat plates. 3) Leakage between the flank of each gear and the housing occurs in a direction perpendicular to the gear's major axis.

[0145] Referring to Figures 13 and 14, (Q t )1 has two paths (Q tv )1 and (Q tp ) can be divided by 1. (Q tv )1 is the leakage that depends on the velocity in the preceding term in Equation 20 below, and (Q tp )1 is the latter term, i.e., the leakage dependent on pressure. (Q tv )1 and (Q tv )2 flows from the discharge port to the suction port when gear 1 rotates counterclockwise (positive direction), and vice versa (Q tp )1 and (Q tp )2 flows from the inlet to the outlet. Because the pressure of the inlet ( in ) because it is greater than the pressure at the outlet.

[0146] (Formula 20)

[0147] Therefore (Q t )1 and (Q t )2 can be expressed by equations 21 and 22 below through equation 20 above.

[0148] (Formula 21)

[0149] (Formula 22)

[0150] Here, (Q tv )1 and (Q tv )2 is the leakage flow between the gear tip and the housing depending on the fluid velocity, (Q tp )1 and (Q tp )2 is the leakage flow between the gear tip and the housing, which varies with the pressure gradient δ. t is the gap between the gear tip and the housing, p in is the inlet pressure, p out is the outlet pressure, l t is the width of the gear tip.

[0151] Figure 15 is an example diagram showing the leakage flow between the side and the housing. Referring to Figure 15, the direction of the leakage depends on the velocity and pressure difference of the fluid ( ) is affected by the direction of the flow. Therefore, leakage can occur from the intake to the intake, from the intake to the exhaust, or from the exhaust to the intake, or from the exhaust to the exhaust, depending on the situation. Referring to Equation 20, the reason why the preceding term, i.e. the velocity term, is ignored in the case of symmetry in the Wilson model is for this reason.

[0152] (Formula 20)

[0153] In cases where the rotor is not perfectly symmetrical, such as in an oval flow meter, there are cases where each section is symmetrical to each other and the velocity term can be ignored, as shown in Fig. 15, and cases where this is not the case. Therefore, it is necessary to calculate each section separately.

[0154] The leakage flow path between the side of the gear and the housing can be divided into four sections. Section 1 is the area bounded by a straight line starting from the contact point of gear 1 and perpendicular to the major axis of gear 1. Section 2 is the area symmetrical to section 1 with respect to the minor axis of gear 1. Sections 3 and 4 are symmetrical areas with respect to the minor axis, excluding the areas of sections 1 and 2.

[0155] The area of ​​section 1 of gear 1 is actually the leakage flow (Q) depending on the velocity of the fluid.sv① )1, leakage flow (Q) according to pressure gradient sp① )1) Since the fluid flows from the inlet to the outlet, there is no leakage (dashed arrow). While gear 1 rotates counterclockwise, the area of ​​section 2 moves from the outlet to the inlet, and the area of ​​section 3 moves from the inlet to the outlet. This is because the areas of sections 2 and 3 are symmetrical with respect to the short axis of gear 1, so the velocities of the two sections have opposite directions. That is, the leakage flow (Q) depending on the area of ​​section 2 (the velocity of the fluid) sv② )1 and Section 3 Area (Leakage flow according to fluid velocity (Q sv③ )1) has the same amount and opposite direction. The leakage flow of the area of ​​section 4 (Q sv④ )1 and (Q sp④ )1 flows from the outlet to the inlet and from the inlet to the outlet, respectively. Therefore, these two leakage flows actually occur in this section. This relationship can be expressed as Equation 23.

[0156] (Formula 23)

[0157] Here, (Q sp① )1= (Q sv① )1= 0, (Q sv② )1+ (Q sv③ )1=0. It can be expressed as in Equation 24.

[0158] (Formula 24)

[0159] The leakage flow of gear 2 can be calculated in the same way as gear 1, as follows.

[0160] (Formula 25)

[0161] Here, (Q sp⑤ )2= (Q sv⑤ )2= 0, (Q sv⑥ )2+ (Q sv ⑦)2=0. Therefore, it can be expressed as in Equation 26.

[0162] (Formula 26)

[0163] Figures 16 to 18 are exemplary diagrams showing that the area where leakage flow actually occurs depending on the velocity of the fluid varies according to the rotation angle of each gear, and Figures 19 to 21 are exemplary diagrams showing that the area where leakage flow actually occurs due to the pressure gradient varies according to the rotation angle of each gear. Leakage due to lateral velocity and leakage due to pressure actually occur only in the shaded areas as in Figures 16 to 21.

[0164] Referring to Figures 16 to 18, in the case of leakage due to speed When the maximum leakage (1 / 2 of the total area of ​​the oval gear side face) occurs, When the minimum leakage (instantaneous leakage = 0) occurs, the leakage occurs repeatedly in a 180 degree cycle.

[0165] Referring to Figures 19 to 21, in the case of pressure-induced leakage, gear 1 At least when Maximum leakage occurs when gear 2 is When the maximum, Maximum leakage occurs when:

[0166] Figures 22 and 23 show the leakage flow, Q, by modifying the pitch curve and geometric parameters of the oval gear. sv and Q sp This is an example of calculating. Referring to FIGS. 22 and 23, the pitch curve and geometric parameters of the oval gear can be defined, and are used for calculating the pitch curve considering the geometric characteristics of the oval flow meter, calculating the rotation angle, and calculating the leakage area of ​​FIGS. 16 to 21.

[0167] Leakage from the side face due to speed (Q sv )1 and (Q sv )2 can be expressed as follows using Equation 20, Fig. 15, Fig. 22, and Fig. 23, respectively.

[0168] (Formula 27)

[0169] (Formula 28)

[0170] Here, am.

[0171] Figure 24 is a flowchart of AI-based compensation and control. Referring to Figure 24, the AI ​​prediction model receives past driving history and collected real-time sensor signals as learning data. The learned AI prediction model receives collected fluid temperature and motor signals and predicts viscosity compensation coefficients, leakage coefficients, and ripple parameters in real time. The predicted compensation coefficients compensate for the flow rate calculation results and perform feedback control to achieve the target flow rate. Feedback control is achieved through motor output control.

[0172] The above prediction model is characterized by performing a predictive control function that learns and predicts cumulative flow rate change trends and proactively adjusts the PWM duty ratio of the motor to achieve a target flow rate.

[0173] Furthermore, the predictive model is designed to detect pump gap changes or wear early, thereby determining the timing of predictive maintenance. For example, if the difference between the calculated theoretical flow rate and the actual discharged flow exceeds a certain level, a pump abnormality can be detected. Furthermore, if a pump undergoes gradual changes while continuously collecting data, the timing of maintenance can be predicted when the change deviates from a certain level.

[0174] Such AI-based correction and control can provide the foundation for digital twin control.

[0175] Past driving history DB and real-time sensor signals (current, speed, temperature, pressure, etc.) are collected, and the AI ​​prediction model learns by receiving inputs such as temperature, pressure, PWM, and flow rate errors.

[0176] The AI ​​prediction model builds viscosity correction coefficients, leakage coefficients, and ripple parameters using input data, and compares them with actual results to make corrections.

[0177] The compensation calculation unit uses predicted values, such as the calculated viscosity compensation coefficient, leakage coefficient, and ripple parameter, as data to compensate for the flow rate calculation formula. The motor output is then adjusted to achieve the target flow rate.

[0178] The present invention relates to a non-transitory computer-readable storage medium having recorded thereon a program for executing a flow control method, wherein the program comprises a geometric tolerance-based leakage coefficient prediction and an artificial intelligence-based correction and control algorithm. The program has the advantage of rapidly deriving optimal results by forming an artificial intelligence-based algorithm through learned data.

[0179] The present invention is not limited to one embodiment, and the scope of application is diverse, and various modifications can be implemented without departing from the gist of the present invention as claimed in the claims.

[0180] Below are formulas used to explain the present invention.

[0181]

[0182]

[0183]

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192] [Explanation of symbols]

[0193] 100: Volumetric pump

[0194] 110: Motor 120: Signal collector

[0195] 121: Temperature sensor 130: Operation unit

[0196] 140: Control unit

[0197] 200: Encoder

Claims

A volumetric pump driven by a connected motor, A signal collection unit that collects at least one motor signal among the PWM duty ratio, current, and voltage of the above motor, A calculation unit that receives a motor signal from the signal collection unit, calculates the driving torque of the pump, estimates the pressure of the pump through the calculated driving torque and the radius of the volumetric pump, and calculates the flow rate through the calculated driving torque and the pressure of the pump. A volumetric pump system comprising a control unit for controlling the motor to correspond to the required flow rate. In the first paragraph, The above operation unit In order to calculate the leakage occurring in the above volumetric pump, a leakage model based on the laminar flow between plates is included. The above leak model is A volumetric pump system characterized in that the velocity-dependent term and the pressure-dependent term of each leakage path are calculated separately according to the position of the gear. In the first paragraph, The above volumetric pump is composed of at least one of an oval gear, a vane, a gear, a piston, and a helical / screw structure, A volumetric pump system characterized in that the above operation unit applies a correction parameter according to the ripple characteristic according to the structure of the volumetric pump. In the first paragraph, A volumetric pump system characterized in that the above operation unit filters out noise of a motor signal in advance using a Kalman filter or an observer-based estimation algorithm. In the first paragraph, The above operation unit compares the accumulated flow rate and the target flow rate, A volumetric pump system characterized in that the control unit controls the PWM duty ratio so that the accumulated flow rate and the target flow rate correspond. In the second paragraph, The above leakage model is based on the laminar flow equation between plates with coefficient C s , C d , C f A volumetric pump system characterized by applying an improved formula that separates the velocity term and the pressure term of the gear tip gap and the side pole. Here, C s is the geometric coefficient, and C d is the viscous drag coefficient depending on the pump geometry, and C f is the friction coefficient according to the pump shape. In the first paragraph, The above operation unit uses past driving history data, real-time measurement signals, and flow rate errors as learning data to create an AI-based prediction model through machine learning. A volumetric pump system characterized in that the above prediction model is configured to calculate or update the viscosity / leakage coefficient of the fluid in real time. In the second or sixth paragraph, The above calculation unit calculates the leakage coefficient (C) based on the geometric tolerance measurement values ​​such as the gear tip clearance and side clearance of the pump. s , C d , C f ) in advance and reflecting it in the leakage flow correction. Here, C s is the geometric coefficient, and C d is the viscous drag coefficient depending on the pump geometry, and C f is the friction coefficient according to the pump shape. In paragraph 7, A volumetric pump system characterized in that the above prediction model performs a predictive control function that preemptively adjusts the PWM duty ratio of the motor to achieve a target flow rate by learning and predicting the cumulative flow rate change trend. In paragraph 7 or 9, A volumetric pump system characterized in that the above prediction model is configured to detect pump gap changes or wear at an early stage and calculate a predictive maintenance point. In a flow rate calculation method for calculating the discharge flow rate of a volumetric pump driven by a motor, A motor signal collection step in which a sensor and encoder are combined with the motor and the current, voltage, and PWM duty cycle (Pulse Width Modulation Duty Cycle) generated from the motor are collected. A pressure and torque estimation step that estimates the torque and rotation speed of the pump through the collected motor signal and estimates the operating pressure of the pump based on the estimated torque. The leakage model application step calculates the actual flow rate based on the leakage model based on the laminar flow between the plates in the theoretical volume flow rate of the pump, and corrects the leakage flow rate to calculate the actual instantaneous flow rate. A digital integral correction step that models the digital integral error according to the resolution (PPR) and phase offset of the encoder and corrects the digital error to the instantaneous flow rate, and A flow control method comprising an accumulated flow integration step for calculating a digital accumulated flow by accumulating the corrected instantaneous flow rate over time. In Article 11, In the above motor signal collection step, the control unit collects the temperature of the fluid supplied to the pump through a temperature sensor, A flow control method including a viscosity correction step in which the control unit corrects the viscosity parameter in real time based on the temperature of the measured fluid in the above pressure and torque estimation step. In Article 11, A flow control method including a feedback step for comparing the actual flow rate with the calculated flow rate and, if a difference occurs, adjusting the output of the motor to compensate. A non-transitory computer-readable storage medium having recorded thereon a program for commanding the execution of a flow control method according to any one of claims 11 to 13, A storage medium characterized in that the above program includes a geometric tolerance-based leakage coefficient prediction and an artificial intelligence-based correction and control algorithm.

Images