Ultrahigh-precision load data implementation method, circuit and measurement and control instrument

Abstract

The invention relates to the technical field of industrial automation field data acquisition, in particular to an ultrahigh-precision load data implementation method and circuit and a measurement and control instrument. The method comprises the steps that an excitation signal is provided for the load sensor through the precise low-drift direct-current power supply, and the excitation signal is applied to the power input end of the sensor after high-frequency noise is eliminated through the pi-type filter network. According to the method, a temperature-load coupled multi-broken-line nonlinear regression curve cluster is constructed, and a temperature variable is introduced into dynamic interpolation calculation of segmented endpoints, so that an error nonlinear cumulative effect caused by endpoint drift under a wide-temperature working condition is effectively inhibited. In the hardware level, the background noise of a sensor signal is suppressed to a microvolt level through the collaborative design of a precise low-drift power supply and multi-stage anti-aliasing filtering, and original data with a high signal-to-noise ratio is provided for a software compensation algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial automation field data acquisition, and particularly to a method, a circuit and a measuring and controlling instrument for realizing ultra-high-precision load data. Background Art

[0002] In high-precision load detection systems in the field of industrial automation, the comprehensive compensation of sensor nonlinearity and temperature drift is the core challenge for improving system accuracy. The traditional multi-segment line nonlinear regression method performs linear fitting within a fixed segmented interval. Although it can improve the sensor nonlinear error at a single temperature, under wide temperature conditions, temperature drift will cause dynamic offsets of the theoretical load values at the endpoints of each segment, thereby triggering the nonlinear cumulative effect of cross-temperature interval errors. This error accumulation exhibits non-uniform distribution characteristics within the entire working temperature range of the sensor. Especially during the temperature transient process, the fixed segmented model cannot track the dynamic changes of the temperature-load coupling characteristics in real time, resulting in systematic deviations in the compensated load data. In the prior art, although methods such as increasing the number of segments or introducing static temperature compensation coefficients can alleviate the errors in a single temperature interval, it is difficult to fundamentally solve the problem of segment endpoint drift caused by temperature gradient changes, and the relationship between the consumption of computing resources and the improvement of accuracy shows non-linear growth, restricting the technical bottleneck of the system accuracy breaking through the one-ten-thousandth level. Summary of the Invention

[0003] The present disclosure provides a method, a circuit and a measuring and controlling instrument for realizing ultra-high-precision load data, aiming to overcome at least one defect existing in the prior art.

[0004] To achieve the above object, the technical solutions disclosed by the present invention are as follows:

[0005] According to one aspect of the present disclosure, a method for realizing ultra-high-precision load data is provided. The steps of the method include:

[0006] Providing an excitation signal to a load sensor through a precision low-drift DC power supply, and applying the excitation signal to the power input end of the sensor after eliminating high-frequency noise through a π-type filter network;

[0007] Performing two-stage anti-aliasing filtering on the differential analog signal output by the sensor, including passive low-pass filtering composed of an RC network and second-order active filtering composed of an operational amplifier;

[0008] Inputting the filtered differential signal into a 32-bit Σ-Δ analog-to-digital converter of an integrated programmable gain amplifier, completing signal amplification and digital sampling under the control of an embedded MCU, and real-time collecting the temperature data of the sensor working environment;

[0009] Dynamically select a preset multi-line nonlinear regression curve cluster based on temperature data, and use the endpoint slope interpolation method to calculate the compensated load value according to the segmented interval to which the current load value belongs;

[0010] The compensated load data is transmitted to a 16-bit high-precision digital-to-analog converter via an isolated SPI bus, and a 4-20mA transmission signal linearly proportional to the load is generated by a voltage-current conversion circuit. At the same time, electrical isolation between the analog output and the digital control circuit is achieved through a capacitive gate isolation module.

[0011] Furthermore, the segmented compensation method of the multi-broken-line nonlinear regression curve cluster includes:

[0012] The full range of the sensor is divided into N continuous sub-intervals. The theoretical load value corresponding to the endpoint of each sub-interval is obtained through calibration experiments and stored as a three-dimensional lookup table related to temperature.

[0013] When the current temperature is detected, the segment endpoint values corresponding to the two adjacent temperature curves are extracted from the lookup table, and the endpoint slope after temperature interpolation is calculated using the following formula:

[0014] Among them, W T1 、W T2 are the theoretical load values at adjacent temperatures T1 and T2, K T1 is the original endpoint slope at temperature T1, K comp is the endpoint slope after compensation, and T is the current temperature.

[0015] Furthermore, the step of dynamically selecting a preset multi-fold nonlinear regression curve cluster further includes:

[0016] When the temperature change rate is detected to be greater than 2°C / min, the dynamic segmentation adjustment mechanism is activated. According to the gradient change direction of the current load value, the original segmentation interval is subdivided into two sub-intervals, and a transition compensation curve generated by the cubic spline interpolation algorithm is inserted at the new endpoints.

[0017] Furthermore, the endpoint slope interpolation method includes:

[0018] After determining the segment interval where the current load value is located in the selected temperature curve cluster, the deviation rate between the actual sampling value and the lower endpoint theoretical value in the interval is calculated;

[0019] The deviation rate is input into the error compensation model trained based on the least squares method, and the corrected endpoint slope weight coefficient is output to adjust the linearity of the multi-fold line fitting.

[0020] Furthermore, the step of collecting sensor working environment temperature data includes:

[0021] By I2 The C bus periodically reads the measurement value of the high-precision digital temperature sensor and switches to high-speed sampling mode when it detects that the standard deviation of three consecutive sampling values exceeds 0.3°C;

[0022] Perform time alignment processing on the temperature data and the load sampling value to ensure that the time stamp deviation between the two is less than 1ms;

[0023] The control method of the voltage-current conversion circuit includes:

[0024] According to the step change of the DAC output value, the integral time constant of the PID controller is dynamically adjusted. When the output current overshoot exceeds 0.05% of the full scale, a reverse compensation pulse is started to suppress oscillation.

[0025] In the steady-state stage, a sliding average filtering algorithm is introduced to eliminate the influence of current ripple on the accuracy of the transmitted signal.

[0026] According to another aspect of the present disclosure, an ultra-high precision load data realization circuit is provided, for realizing the ultra-high precision load data realization method as described above, the circuit comprising:

[0027] A 32-bit embedded MCU based on the ARM Cortex-M4F core, which is connected to a 32-bit Σ-Δ analog-to-digital converter via an SPI bus;

[0028] The sensor excitation module includes a 5V DC power supply with a precision low-noise LDO, a current-expanding transistor, and a current-limiting resistor, which is output to the load sensor through a π-type filter network;

[0029] A signal conditioning module, including a second-order active filter circuit composed of an operational amplifier and an anti-aliasing passive filter network;

[0030] The temperature acquisition module uses a digital temperature sensor mounted on the sensor body. 2 C bus communicates with MCU;

[0031] Isolation communication module, including high-speed capacitive gate isolators, isolating the SPI bus, UART interface and relay control signals respectively;

[0032] The transmitter output module includes a 16-bit digital-to-analog converter, a differential amplifier circuit and a MOSFET current source, and an overcurrent protection transistor and a transient suppression diode are connected in series at the output end.

[0033] Furthermore, the circuit structure of the sensor excitation module includes:

[0034] The secondary winding of the power frequency transformer is connected to a full-wave rectifier bridge, and the output is filtered by an electrolytic capacitor and then input to a low-drift voltage reference source;

[0035] The output end of the reference source drives the sensor through the PNP transistor current expansion circuit, and the sampling resistor connected in series with the emitter forms a current negative feedback loop;

[0036] The current limiting protection circuit includes an N-channel MOSFET and a precision resistor in parallel, and the gate voltage is controlled by a reference source voltage divider network;

[0037] The operational amplifier of the signal conditioning module adopts a zero-drift instrumentation amplifier, the non-inverting input of which is connected to the output of the anti-aliasing filter network, and the inverting input is grounded via an adjustable resistor network;

[0038] The reference voltage input terminal of the analog-to-digital converter is connected to an independent voltage reference source, and its power supply pin is isolated from the digital power supply by a magnetic bead.

[0039] Furthermore, the capacitive gate isolator of the isolated communication module includes a four-channel isolation unit, two of which are used for isolating the SCK and MOSI signals of the SPI bus, and the other two channels are used for isolating the TX / RX signals of the UART;

[0040] The optocoupler output end of the relay drive circuit is connected to the Darlington transistor array, and the collector is connected in series with a self-recovery fuse and then connected to the relay coil.

[0041] Furthermore, the differential amplifier circuit of the transmission output module includes a precision operational amplifier, the non-inverting input terminal of which is connected to the output terminal of the digital-to-analog converter, and the inverting input terminal is connected to the current sampling resistor via a precision resistor network;

[0042] The gate drive circuit of the MOSFET current source includes an accelerating diode and a gate pull-down resistor. The temperature compensation resistor connected in series with the source forms a proportional relationship with the sampling resistor.

[0043] The programmable gain amplifier of the analog-to-digital converter is configured in auto-ranging mode, and its gain control pin is connected to the GPIO port of the MCU through a buffer;

[0044] A tantalum capacitor and a ceramic capacitor are connected in parallel at the output of the reference voltage source to suppress the influence of high-frequency noise on the conversion accuracy.

[0045] According to another aspect of the present disclosure, a measurement and control instrument is provided, comprising the ultra-high-precision load data realization circuit as described above, and:

[0046] The human-computer interaction module includes an 8-digit digital tube display unit and a matrix button, which is connected to the MCU through a latch and a driver chip;

[0047] RS-485 communication module, the bus end of which is connected in parallel with a TVS diode and an RC absorption network, and the signal input end is connected to the MCU via a capacitive gate isolator;

[0048] The digital output module includes two groups of normally open relay contacts, and the drive circuit adopts a composite structure of optocoupler isolation and Darlington tube;

[0049] The power supply monitoring module detects the voltage fluctuations of each power supply branch in real time and triggers the watchdog reset circuit under undervoltage.

[0050] The beneficial effects of the present invention are:

[0051] By constructing a multi-segment non-linear regression curve cluster of temperature-load coupling and introducing the temperature variable into the dynamic interpolation calculation of the segment endpoints, the present invention effectively suppresses the non-linear cumulative effect of errors caused by endpoint drift under wide temperature conditions. At the hardware level, the collaborative design of a precision low-drift power supply and multi-stage anti-aliasing filtering suppresses the background noise of the sensor signal to the microvolt level, providing high-signal-to-noise ratio raw data for the software compensation algorithm; at the algorithm level, based on the dynamic segmentation mechanism of real-time temperature acquisition, the compensation curve is reconstructed by the endpoint slope interpolation method, so that the segment endpoints in each temperature range are always synchronized with the current sensor characteristics, breaking through the static compensation limitation of the traditional fixed segmentation model. In addition, the combination of capacitive grid isolation technology and timing alignment processing eliminates the time-domain interference of digital noise on the analog signal link, ensuring the integrity of high-resolution analog-to-digital conversion data. In the working temperature range of 0-60°C, the system accuracy of the load data is improved to 0.008% FS, and at the same time, the dynamic tracking error under temperature transient conditions is reduced to less than 1 / 5 of the conventional method, achieving a performance breakthrough of an industrial-grade ultra-high-precision load detection system.

[0052] The above description is only an overview of the technical solution of the present invention. In order to understand the technical means of the present invention more clearly and implement it according to the content of the specification, the following uses the preferred embodiments of the present invention and combines the drawings for detailed description. Brief Description of the Drawings

[0053] Figure 1 is the flowchart of the method for realizing ultra-high-precision load data of the present invention;

[0054] Figure 2 is the structural block diagram of the data acquisition instrument of the present invention;

[0055] Figure 3 is the system power supply regulation circuit of the present invention;

[0056] Figure 4 is the high-precision sensor signal conditioning circuit of the present invention;

[0057] Figure 5 is the MCU minimum system and sensor working temperature acquisition circuit of the present invention;

[0058] Figure 6 is the keyboard display circuit of the present invention;

[0059] Figure 7 It is the RS485 and relay output control circuit of the present invention;

[0060] Figure 8 It is a high-precision 4-20mA analog transmission signal output circuit of the present invention;

[0061] Figure 9 It is a linear diagram of the conventional sensor linear fitting method;

[0062] Figure 10 It is a linear schematic diagram of the multi-broken-line nonlinear fitting method for a single temperature sensor;

[0063] Figure 11 It is a linear schematic diagram of the multi-broken-line nonlinear fitting method for the wide temperature sensor of the present invention. DETAILED DESCRIPTION

[0064] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0065] The term "comprise" and any variations thereof in the specification and claims of this application are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or are inherent to these processes, methods, products, or apparatuses. In addition, the use of "and / or" in the specification and claims to indicate at least one of the connected objects, such as A and / or B, means that A alone, B alone, and both A and B are included.

[0066] In the embodiments of the present invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present invention should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0067] The present invention provides the following preferred embodiments:

[0068] Example 1

[0069] In order to solve the problem of reduced load measurement accuracy caused by sensor nonlinear errors and dynamic environmental interference in a wide temperature range, this embodiment integrates multi-level signal conditioning, temperature dynamic compensation and closed-loop feedback control to build an integrated high-precision load data processing system. Figure 1 As shown, the steps of the method for realizing ultra-high precision load data are as follows:

[0070] S100: Provides an excitation signal to the load cell through a precision low-drift DC power supply. The excitation signal is applied to the power input of the sensor after eliminating high-frequency noise through a π-type filter network.

[0071] S200: Perform two-stage anti-aliasing filtering on the differential analog signal output by the sensor, including passive low-pass filtering composed of an RC network and second-order active filtering composed of an operational amplifier.

[0072] S300: The filtered differential signal is input into a 32-bit Σ-Δ analog-to-digital converter with an integrated programmable gain amplifier. Under the control of the embedded MCU, signal amplification and digital sampling are completed, and the sensor working environment temperature data is collected in real time.

[0073] S400: Dynamically select a preset multi-line nonlinear regression curve cluster based on the temperature data, and calculate the compensated load value using the endpoint slope interpolation method according to the segment interval to which the current load value belongs.

[0074] S500: The compensated load data is transmitted to a 16-bit high-precision digital-to-analog converter via an isolated SPI bus. A voltage-current conversion circuit generates a 4-20mA transmission signal that is linearly proportional to the load. At the same time, a capacitive gate isolation module is used to achieve electrical isolation between the analog output and the digital control circuit.

[0075] Specifically, a precision low-drift DC power supply provides an excitation signal to the load sensor. Its output is filtered through a π-type filter network to remove high-frequency noise before being connected to the sensor power supply. The sensor's differential analog signal then passes through a passive RC low-pass filter and a second-order active filter consisting of an operational amplifier. The cutoff frequencies of the two filters are set to suppress high-frequency interference and provide anti-aliasing, respectively, to ensure signal purity in the frequency domain before analog-to-digital conversion. The filtered differential signal is then fed into a 32-bit Σ-Δ analog-to-digital converter with an integrated programmable gain amplifier. The embedded MCU dynamically adjusts the gain to align the signal amplitude with the ADC's range. A synchronous clock trigger is used to achieve time alignment of the load and temperature data.

[0076] Furthermore, the embedded MCU pre-stores a three-dimensional lookup table obtained from calibration experiments. The table contains continuous sub-intervals divided by the full range of the sensor and their theoretical load endpoint values at different temperatures. When the current ambient temperature is detected, the segmented endpoint parameters of adjacent temperature calibration points are extracted from the lookup table. The endpoint slope at the current temperature is calculated through linear interpolation to form a dynamic temperature compensation coefficient. The slope is calculated as follows:

[0077] Among them, W T1 、W T2 are the theoretical load values at adjacent temperatures T1 and T2, K T1 is the original endpoint slope at temperature T1, K comp is the endpoint slope after compensation, and T is the current temperature.

[0078] It's important to understand that this interpolation process ensures the continuity of the compensation model during temperature gradients by weighting the slope trends of adjacent temperature points. Furthermore, if the temperature change rate exceeds a preset threshold, the MCU subdivides the current subinterval into two smaller intervals based on the load gradient direction and inserts a transition curve generated by cubic spline interpolation at the new endpoints to mitigate model mismatch caused by sudden temperature changes.

[0079] Furthermore, after determining the segmented interval to which the current load value belongs, the deviation rate between the actual sampled value and the theoretical lower limit of the interval is calculated and input into an error compensation model trained offline using the least squares method. This model outputs modified endpoint slope weight coefficients, which are used to dynamically adjust the linearity of the multi-polygonal line fit. For example, when the deviation rate indicates that the current sampled value is close to the upper limit of the interval, the model automatically increases the weight coefficient of the upper endpoint slope to improve the fitting accuracy of the local interval. This closed-loop feedback mechanism effectively suppresses the accumulation of nonlinear residuals within the interval.

[0080] Furthermore, the high-precision digital temperature sensor is 2 The C bus transmits data periodically. When the standard deviation of consecutive temperature samples exceeds a specified limit, the system switches to high-speed sampling mode. A timestamp alignment algorithm ensures that the time deviation between temperature and load data is within a set threshold. The compensated load data is transmitted to the digital-to-analog converter via an isolated SPI bus. Its output signal is converted into a 4-20mA transmission signal through a voltage-to-current conversion circuit. During this process, the PID controller dynamically adjusts the integral time constant based on the step change in the digital-to-analog conversion value. When an output current overshoot is detected, a reverse compensation pulse is injected to suppress oscillations. A sliding average filter algorithm is introduced during the steady-state phase to smooth the output current and eliminate the impact of high-frequency ripple on signal stability.

[0081] Furthermore, a capacitive barrier isolation module is deployed between the digital control circuit and the analog output circuit, blocking digital noise coupling through electrical isolation and electromagnetic shielding. A differential routing layout is used between the DAC output and the voltage-to-current conversion circuit, combined with a common-mode choke to suppress conducted interference. The signal conditioning circuit's ground plane forms a hybrid grounding architecture through a single-point connection and an RC network, further reducing high-frequency noise interference on the analog signal chain.

[0082] The benefits of this embodiment include: The synergy between a temperature-driven dynamic segmented compensation algorithm and a deviation rate feedback mechanism enables adaptive tracking of sensor nonlinear characteristics and local precision optimization across a wide temperature range. A multi-stage filtering and isolation architecture design ensures low noise in the signal chain from a hardware perspective. The integration of synchronous data acquisition and adaptive transmission control strategies ensures real-time performance and output stability under complex operating conditions. This method significantly improves the overall accuracy of load data processing without requiring additional hardware resources.

[0083] Example 2

[0084] This embodiment provides a reliable hardware platform for ultra-high-precision load data acquisition through precision analog circuit design. On this basis, nonlinear compensation of the sensor is achieved through multi-fold nonlinear regression. By collecting the real-time working environment temperature of the sensor, a set of multi-fold nonlinear regression curve clusters are obtained according to different working temperatures, thereby realizing ultra-high-precision data acquisition under wide temperature working conditions.

[0085] like Figure 2 The hardware block diagram of the system for ultra-high-precision load data acquisition is shown below. A high-performance 32-bit M4 embedded MCU serves as the control core, providing the necessary computing power for the software compensation algorithm and other logic control functions. To minimize the impact of power supply noise on data accuracy, the system uses a linear regulated power supply. A power-frequency transformer regulates the 220V AC voltage, then rectifies and stabilizes it to generate three system power supplies, which respectively power the sensor conditioning circuit, the MCU control circuit, and the RS485 / disconnector and analog module circuits. A 5V low-drift precision DC power supply is used for sensor excitation. A high-precision, high-resolution 32-bit ADC with an integrated PGA amplifies and converts the sensor output differential signal. An NST112 digital temperature sensor collects the sensor's ambient temperature for subsequent sensor temperature compensation. Four independent keyboards and an 8-digit digital display serve as the human-machine interface. An RS485 interface and a switch output control module are also included. A high-precision 16-bit DAC and precision V / I conversion circuitry are used to generate a high-precision 4-20mA analog output signal. In order to prevent the crosstalk between the circuits of each module from affecting the accuracy of load data acquisition, highly integrated capacitive gate isolators are used for electrical isolation between the modules.

[0086] Furthermore, if Figure 3 The power supply regulation circuit for the data acquisition system shown in Figure 1 is shown. Compared to switching power supplies, linear power supplies have lower power ripple noise, which is particularly important for high-precision data acquisition, especially at high sampling speeds. The 220V AC power supply passes through the power line filter T2 and then into the power frequency transformer for voltage regulation. F1 is a resettable fuse that provides overload protection. Four diodes, D1-D4, form a full-wave rectifier, which is filtered by a capacitor before passing through an LDO linear regulator, U1, to generate the 3.3V system power supply for the MCU control circuit. The power supply for the analog circuitry first undergoes a full-wave rectifier, D5-D6, and capacitor filtering to generate a 10V DC power supply. This is then passed through a precision low-drift voltage reference, U11, to generate a 5V precision reference voltage, Vref. This voltage is then passed through a current expansion circuit, comprised of U12, Q4, and Q3, to generate the 5V excitation power supply, E+, for the precision sensor with a specified load capacity. Q3 and R8 form a current limiting circuit, with the current limit value being related to R8's resistance. The power supply of the relay, RS485 and analog output module is rectified by D9~D12 and filtered by capacitor to obtain 24V DC power, which drives the relay and provides power for the 4-20mA transmitter circuit. It is then stepped down by LDO to obtain 5V power supply VCC2, which provides DC regulated power supply for the related circuits of this module.

[0087] Furthermore, if Figure 4 The high-precision signal conditioning circuit for a strain gauge sensor is shown. The sensor differential signals S+ and S- are first filtered through a passive low-pass filter and an anti-aliasing filter composed of C10, C12, and C13 before being fed into U12 for analog-to-digital conversion. U12 is a high-performance, domestically produced 32-bit Σ-Δ ADC with an integrated, extremely low-noise PGA, capable of up to 128x amplification for mV-level differential signals. U21's analog and digital power supplies are independent. The sensor analog excitation power supply E+ is filtered through a π-type filter to generate DVCC, which provides the digital power supply for U21. U21's reference voltage is provided by U11. The ADC conversion results are transmitted to the MCU via the SPI bus. The ADC's SPI interface is electrically isolated from the MCU's SPI bus interface by the highly integrated capacitive gate isolator U10.

[0088] Furthermore, if Figure 5The following figure shows a minimum MCU system and temperature acquisition circuit. U13 is a 32-bit embedded microcontroller based on the ARM Cortex-M4F core. It supports DSP instructions, integrates a floating-point unit, and boasts extensive internal resources, including SPI, IIC, and UART. As the system's control core, it provides the necessary computing power for software compensation algorithms. P1 is the SWD interface for MCU simulation debugging and program downloads. R19 and C7 form the power-on reset circuit. The high-precision digital temperature sensor NST112 acquires the sensor's on-site operating temperature, exchanging data with the MCU via the IIC.

[0089] Furthermore, if Figure 6 The keyboard display circuit shown in Figure 1. K1-K4 are independent keyboards used for parameter settings and other functions. U18 and U19 are four common-cathode diodes, which together form an 8-bit digital tube display for real-time display of load data and other functions. Buffer U20 is used for segment drive, resistor R25 is used for current limiting, and U22 is a Darlington driver chip for driving the 8-bit digital tube.

[0090] Furthermore, if Figure 7 The RS485 and relay output control circuit shown. U17 is the RS485 interface chip, used for TTL to RS485 logic level conversion. It is electrically isolated from the MCU's UART via capacitor-gate isolator U16. Resistors R16 and R20 effectively prevent the impact of this node failure on RS485 bus communication. R17 and R18 absorb bus reflection signals. D16 and D17 are bidirectional transient suppression diodes that provide bus overvoltage protection. F2 and F3 are resettable fuses for bus overload protection. Two relays, JD1 and JD2, are controlled by the MCU via capacitor-gate isolator U14 and a transistor drive circuit.

[0091] Furthermore, if Figure 8 The high-precision 4-20mA analog transmitter signal output circuit shown in the figure. The MCU controls the high-precision 16-bit DAC chip U8 via another SPI bus interface through a capacitive gate isolator U9. U8 is a high-performance voltage-output DAC manufactured by SiRuiPu. Its reference voltage is provided by U7. A precision current source circuit, consisting of differential amplifier U6, Q1, Q2, R3, and R5, performs high-precision V / I conversion on the DAC output voltage. Under the control of the MCU, a 4-20mA transmitter signal output that changes with load data is generated. This transmitter signal has overcurrent protection and reverse-bias function. Q2 and R4 implement overcurrent protection, which depends on the size of the current sampling resistor R4. Diode D13 provides reverse-bias protection.

[0092] Furthermore, if Figures 9 to 11The conventional sensor nonlinear compensation method shown and the nonlinear regression method adopted by the present invention. It can be seen that a main factor affecting the load data acquisition is the nonlinearity of the sensor. For example, Figure 9 As shown, generally speaking, there is an objectively existing input-output characteristic curve for the sensor, that is, the theoretical actual curve. This curve is generally obtained through actual measurement. In actual operation, the method of two-point linear regression such as connecting the end points is often used to fit the theoretical actual curve. Undoubtedly, the residual error brought by simply using the method of connecting two points to fit the theoretical actual curve within the full scale range is relatively large.

[0093] Furthermore, if the full scale of the sensor is divided into several segments, such as Figure 10 As shown, within each segment, the method of fitting by connecting the two end points can well reduce the fitting error, so that a higher-precision fitting can be achieved, the fitting accuracy can be improved, and the nonlinear error can be reduced. The more the number of segments, the higher the fitting accuracy, but of course it will also bring an increase in the calculation amount. Generally speaking, for ordinary sensors such as S-type tension-compression sensors, the full scale is equally divided into several segments (generally less than ten segments), and then two-end linear fitting is carried out in each segment, and a nonlinear fitting error better than one ten-thousandth can be obtained. For piecewise linear fitting, within the full scale range, since the straight line slopes in each segment are different, it is essentially a nonlinear regression, which is usually called multi-segment line nonlinear regression.

[0094] Furthermore, in addition to the influence of the nonlinearity of the sensor itself on the load data acquisition accuracy, the temperature drift of the sensor is also an important factor affecting the load data acquisition. The present invention adopts the multi-segment line nonlinear regression and temperature compensation method as shown in Figure 11 As shown, on the basis of a reasonable hardware circuit and in cooperation with this method, ultra-high-precision load data acquisition can be obtained. There is no doubt that the working environment temperature of the sensor will change. Due to the inherent characteristics of the sensor, the temperature drift objectively exists. If only multi-segment line nonlinear regression is carried out for a single temperature, such as room temperature 25°C, then it is still difficult to ensure ultra-high precision within the full temperature range of the sensor. In fact, for a sensor, its input-output characteristic curve has a certain relationship with temperature. Essentially, as the working temperature of the sensor changes, its input-output characteristic curve is a series of curve clusters. Then, by collecting the working environment temperature of the sensor and using the values of each curve cluster at the end points of each segment, and then respectively performing multi-segment line nonlinear regression, the high precision of load data acquisition within the full working temperature range of the sensor can be ensured. As shown in Figure 11 As shown, there are different curves corresponding to different working temperatures of the sensor. The input-output characteristic curve of the sensor within the full working temperature range is a series of curve clusters. Usually, a curve is obtained every 5-10°C, and good temperature compensation effects can be obtained within the full working temperature range of the sensor.

[0095] Although the present invention has been specifically described above with reference to preferred embodiments thereof, it is to be understood that the invention is not limited to the embodiments described above. Rather, various modifications and variations can be made by those skilled in the art without departing from the spirit of the invention, and these modifications and variations shall fall within the scope defined by the appended claims and their equivalents.

Claims

1. A method for realizing ultra-high-precision load data, characterized in that The steps of the method include: The load cell is supplied with an excitation signal through a precision low-drift DC power supply. The excitation signal is then applied to the power input of the sensor after high-frequency noise is eliminated by a π-type filter network. The differential analog signal output by the sensor is subjected to two-stage anti-aliasing filtering, including a passive low-pass filter composed of an RC network and a second-order active filter composed of an operational amplifier; The filtered differential signal is input into a 32-bit Σ-Δ analog-to-digital converter with an integrated programmable gain amplifier. Under the control of an embedded MCU, the signal is amplified and digitally sampled, and the sensor's operating environment temperature data is collected in real time. Dynamically select a preset multi-line nonlinear regression curve cluster based on temperature data, and use the endpoint slope interpolation method to calculate the compensated load value according to the segmented interval to which the current load value belongs; The compensated load data is transmitted to a 16-bit high-precision digital-to-analog converter via an isolated SPI bus, and a 4-20mA transmission signal linearly proportional to the load is generated by a voltage-current conversion circuit. At the same time, electrical isolation between the analog output and the digital control circuit is achieved through a capacitive gate isolation module.

2. The method for implementing ultra-high-precision load data according to claim 1, wherein The segmented compensation method of the multi-broken-line nonlinear regression curve cluster includes: The full range of the sensor is divided into N continuous sub-intervals. The theoretical load value corresponding to the endpoint of each sub-interval is obtained through calibration experiments and stored as a three-dimensional lookup table related to temperature. When the current temperature is detected, the segment endpoint values corresponding to the two adjacent temperature curves are extracted from the lookup table, and the endpoint slope after temperature interpolation is calculated using the following formula: Among them, W T1 and W T2 are the theoretical load values at adjacent temperatures T1 and T2 respectively, K T1 is the original endpoint slope at temperature T1, K comp is the compensated endpoint slope, and T is the current temperature.

3. The method for implementing ultra-high-precision load data according to claim 1, characterized in that The step of dynamically selecting a preset multi-broken-line nonlinear regression curve cluster further includes: When the temperature change rate is detected to be greater than 2°C / min, the dynamic segmentation adjustment mechanism is activated. According to the gradient change direction of the current load value, the original segmentation interval is subdivided into two sub-intervals, and a transition compensation curve generated by the cubic spline interpolation algorithm is inserted at the new endpoints.

4. The method for realizing ultra-high-precision load data according to claim 1, characterized in that, The endpoint slope interpolation method includes: After determining the segment interval where the current load value is located in the selected temperature curve cluster, the deviation rate between the actual sampling value and the lower endpoint theoretical value in the interval is calculated; The deviation rate is input into the error compensation model trained based on the least squares method, and the corrected endpoint slope weight coefficient is output to adjust the linearity of the multi-fold line fitting.

5. The method for implementing ultra-high-precision load data according to claim 1, wherein The step of collecting sensor working environment temperature data includes: Read the measured values of the high-precision digital temperature sensor periodically through the I 2 ²C bus, and switch to the high-speed sampling mode when the standard deviation of the continuously sampled values exceeds 0.3°C for three consecutive times; Perform time alignment processing on the temperature data and the load sampling value to ensure that the time stamp deviation between the two is less than 1ms; The control method of the voltage-current conversion circuit includes: According to the step change of the DAC output value, the integral time constant of the PID controller is dynamically adjusted. When the output current overshoot exceeds 0.05% of the full scale, a reverse compensation pulse is started to suppress oscillation. A sliding average filtering algorithm is introduced in the steady-state stage to eliminate the influence of current ripple on the accuracy of the transmitted signal.

6. A circuit for implementing ultra-high-precision load data, which is used to implement the ultra-high-precision load data implementation method according to any one of claims 1 to 5, characterized in that, The circuit comprises: A 32-bit embedded MCU based on the ARM Cortex-M4F core, which is connected to a 32-bit Σ-Δ analog-to-digital converter via an SPI bus; The sensor excitation module includes a 5V DC power supply with a precision low-noise LDO, a current-expanding transistor, and a current-limiting resistor, which is output to the load sensor through a π-type filter network; A signal conditioning module, including a second-order active filter circuit composed of an operational amplifier and an anti-aliasing passive filter network; The temperature acquisition module uses a digital temperature sensor mounted on the sensor body and communicates with the MCU through the I 2 C bus; Isolation communication module, including high-speed capacitive gate isolators, isolating the SPI bus, UART interface and relay control signals respectively; The transmitter output module includes a 16-bit digital-to-analog converter, a differential amplifier circuit and a MOSFET current source, and an overcurrent protection transistor and a transient suppression diode are connected in series at the output end.

7. The ultra-high-precision load data implementation circuit according to claim 6, wherein The circuit structure of the sensor excitation module includes: The secondary winding of the power frequency transformer is connected to a full-wave rectifier bridge, and the output is filtered by an electrolytic capacitor and then input to a low-drift voltage reference source; The output end of the reference source drives the sensor through the PNP transistor current expansion circuit, and the sampling resistor connected in series with the emitter forms a current negative feedback loop; The current limiting protection circuit includes an N-channel MOSFET and a precision resistor in parallel, and the gate voltage is controlled by a reference source voltage divider network; The operational amplifier of the signal conditioning module adopts a zero-drift instrumentation amplifier, the non-inverting input of which is connected to the output of the anti-aliasing filter network, and the inverting input is grounded via an adjustable resistor network; The reference voltage input terminal of the analog-to-digital converter is connected to an independent voltage reference source, and its power supply pin is isolated from the digital power supply by a magnetic bead.

8. The ultra-high-precision load data implementation circuit according to claim 6, characterized in that, The capacitive gate isolator of the isolated communication module includes a four-channel isolation unit, two of which are used for isolating the SCK and MOSI signals of the SPI bus, and the other two channels are used for isolating the TX / RX signals of the UART; The optocoupler output end of the relay drive circuit is connected to the Darlington transistor array, and the collector is connected in series with a self-recovery fuse and then connected to the relay coil.

9. The ultra-high-precision load data implementation circuit according to claim 6, characterized in that The differential amplifier circuit of the transmission output module includes a precision operational amplifier, the non-inverting input end of which is connected to the output end of the digital-to-analog converter, and the inverting input end of which is connected to the current sampling resistor via a precision resistor network; The gate drive circuit of the MOSFET current source includes an accelerating diode and a gate pull-down resistor. The temperature compensation resistor connected in series with the source forms a proportional relationship with the sampling resistor. The programmable gain amplifier of the analog-to-digital converter is configured in auto-ranging mode, and its gain control pin is connected to the GPIO port of the MCU through a buffer; A tantalum capacitor and a ceramic capacitor are connected in parallel at the output of the reference voltage source to suppress the influence of high-frequency noise on the conversion accuracy.

10. A measuring and controlling instrument, characterized in that, The method comprises the ultra-high precision load data realization circuit according to any one of claims 6 to 9, and: The human-computer interaction module includes an 8-digit digital tube display unit and a matrix button, which is connected to the MCU through a latch and a driver chip; RS-485 communication module, the bus end of which is connected in parallel with a TVS diode and an RC absorption network, and the signal input end is connected to the MCU via a capacitive gate isolator; The switch output module includes two sets of normally open relay contacts, and the drive circuit adopts a composite structure of optocoupler isolation and Darlington tube; The power supply monitoring module detects voltage fluctuations of each power supply branch in real time and triggers the watchdog reset circuit when undervoltage occurs.

Images