Industrial process closed-loop feedback adjustment method based on instrument operation data deviation

Abstract

The invention relates to the technical field of industrial automation control, in particular to an industrial process closed-loop feedback adjustment method based on instrument operation data deviation, and the method comprises the steps: calculating an inherent lag parameter through cross-correlation analysis, and constructing a phase alignment reference instruction sequence of time sequence alignment; calculating an excitation intensity index of the instruction by using a sliding window, and calculating a response intensity index of the measured value based on a second-order difference algorithm; constructing a nonlinear coupling model to quantify an excitation response mismatch index; and dynamically adjusting a feed-forward gain based on the excitation response mismatch index, and introducing an instruction trend correction term to generate a phase-advanced reconstruction feedback signal. On the premise that hardware cost is not increased, fault features are accurately recognized, data phases are repaired in real time, and closed-loop control stability is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology. More specifically, this invention relates to a closed-loop feedback regulation method for industrial processes based on deviations in instrument operating data. Background Technology

[0002] In thermal power plants or chemical plants, control systems primarily rely on pressure and flow transmitters to sense production conditions and adjust valve openings accordingly. To prevent control failure due to the malfunction of a single instrument, multiple instruments are typically installed at critical measuring points, connected to the main pipeline via pressure-sensing lines. However, in actual operation, the measured fluid often contains dust or viscous impurities, making the pressure-sensing lines prone to blockage. From a physical perspective, blockage means increased resistance to the fluid in the pipeline, which slows down the transmission time of the pressure signal to the sensor, resulting in phase lag.

[0003] Currently, the common method for determining instrument malfunction is numerical deviation alarm. This involves comparing readings from multiple instruments, and triggering an alarm if the difference exceeds a set range. However, when the pressure tapping line is partially blocked, the signal transmission is only slower, but the displayed pressure value remains accurate. Under stable operating conditions, the readings of the faulty instrument are almost identical to those of the normal instrument, rendering the existing numerical deviation alarm ineffective in detecting such faults. Furthermore, relying on manual periodic purging of the pipeline is not only labor-intensive but also fails to provide real-time monitoring of whether the pipeline is actually blocked.

[0004] This signal lag is highly detrimental to control systems. Because the feedback signal lags behind the actual process, when the control system issues a control command, the instruments may not respond for a long time. The controller may then mistakenly believe that the control is insufficient and continue to increase the valve opening. By the time the instruments finally reflect the true value, the valve has often already been over-opened. This causes system parameters to fluctuate wildly, forcing the control valve to repeatedly reciprocate, accelerating wear on mechanical components, and shortening the equipment's lifespan. Therefore, a technical solution is needed that can identify and compensate for this lag. Summary of the Invention

[0005] To address the problem that existing technologies cannot detect blockages in pressure-tapping pipelines through numerical deviations, leading to delayed signals participating in control and causing system oscillations, this invention proposes a closed-loop feedback control method for industrial processes based on instrument operating data deviations. This method includes the following steps: Acquire historical steady-state data of the industrial control system, calculate the maximum cross-correlation delay time between control commands and instrument measurements, use this as a parameter to perform time-series backtracking of real-time control commands, and generate a phase-aligned reference command sequence. Based on the phase alignment reference command sequence, the excitation intensity index characterizing the input action amplitude is calculated using a sliding window, and the response intensity index characterizing the high-frequency capture capability of the sensor is calculated based on the real-time instrument measurement value. A nonlinear coupling model incorporating exponential decay characteristics is constructed. Using the excitation intensity index as the amplitude and the reciprocal characteristic of the response intensity index as the exponent, the excitation-response mismatch index reflecting the degree of kinetic energy mismatch is calculated. The phase alignment reference command sequence is mapped to theoretical physical values ​​using a pre-calibrated static curve. Dynamic feedforward weights are calculated by combining the excitation response mismatch index. The trend deviation between the theoretical physical values ​​and the current measured values ​​is weighted and superimposed onto the instrument measured values ​​through the dynamic feedforward weights to generate a reconstructed feedback signal, thereby realizing real-time adjustment of the closed-loop control system.

[0006] This invention overcomes the limitations of traditional fault detection techniques that rely solely on static amplitude deviation, and constructs a kinetic energy coupling basis based on an excitation-response mechanism. By using control commands as a feedforward excitation source and comparing the dynamic energy changes at the input end with the transient response capability at the output end in real time, it can accurately identify micro-blockage faults in the pressure tapping pipeline in the early stage, before the steady-state amplitude of the measured signal shows abnormalities but before the high-frequency dynamic characteristics have decayed. This effectively fills the blind spot of existing technologies in dynamic hysteresis detection.

[0007] Preferably, the generation of the phase alignment reference command sequence involves reading a pre-calibrated intrinsic pure hysteresis parameter, subtracting the intrinsic pure hysteresis parameter from each time point to obtain a backtracking time, and directly retrieving the control command of the backtracking time as the phase alignment reference command for each time point, thereby obtaining the phase alignment reference command sequence.

[0008] This invention takes into account that any fluid transfer requires time, and therefore shifts the control command data backward on the time axis to logically synchronize it with the feedback signal. This eliminates interference caused by physical transmission time and provides an accurate data foundation for subsequent processing.

[0009] Preferably, the response intensity index satisfies the following relationship:

[0010] In the formula, For the first Each meter at time The aforementioned response intensity index; For the instrument's index number; For time series indexing; This is the backtracking index within the sliding window; The length of the sliding window; , and They represent the first Each meter at time , and The instrument measurement value; To prevent extremely small positive constants with a value of zero; The excitation intensity index is obtained by calculating the sum of the absolute values ​​of the first-order differences of the phase alignment reference command sequence within the sliding window and then performing a power operation.

[0011] This invention fully utilizes the physical meaning of differential operations: first-order differential represents the rate of change of a signal, while second-order differential represents the acceleration and curvature characteristics of the signal. When a micro-blockage occurs in the pressure tapping pipeline, its low-pass filtering effect first causes the high-frequency inflection point of the signal waveform to become smooth, that is, the second-order curvature decreases significantly. Compared with monitoring only the rate of change, monitoring the acceleration characteristics can more sensitively perceive subtle distortions in the waveform topology, thereby achieving earlier fault identification before severe blockage occurs.

[0012] Preferably, the excitation-response mismatch index satisfies the following relationship:

[0013] In the formula, For the first Each meter at time The excitation-response mismatch index; For a moment The aforementioned incentive intensity index; For the first Each meter at time The aforementioned response intensity index; The damping characteristic constant is set based on the reference noise level; This is the upper limit of saturation. This is the operator for finding the minimum value. It is an exponential function with the natural constant as its base.

[0014] This invention utilizes a nonlinear coupling model to construct a rigorous logical interlocking mechanism. This model leverages the nonlinear amplification characteristics of the exponential function to establish a clear fault criterion: the mismatch exponent only exhibits explosive growth when a specific physical contradiction—a strong excitation input and a weak response feedback—occurs concurrently. This design effectively shields the system from steady-state noise and signal interference during normal adjustment, thereby significantly reducing the false alarm rate while ensuring high fault detection sensitivity.

[0015] Preferably, before generating the reconstructed feedback signal, a dynamic correction term needs to be calculated. The calculation process of the dynamic correction term includes: calculating the deviation between the theoretical physical value corresponding to the phase alignment reference command sequence and the instrument measurement value, and weighting and summing the deviation with the measurement micro-component at the current moment to obtain the dynamic correction term.

[0016] This invention effectively solves the technical problem of inconsistency in dimensions between control commands and instrument measurements. By introducing a pre-calibrated static mapping function, control commands are transformed into theoretical expected values ​​with clear physical meaning. This establishes a unified physical calculation reference system, enabling the algorithm to accurately quantify the dynamic deviation between theoretical trends and actual measurements under the same dimensions, providing a data foundation for subsequent signal correction.

[0017] Preferably, the calculation of the dynamic feedforward weights is performed using an S-shaped growth curve function, by setting a hysteresis determination threshold and a steepness coefficient, to map the excitation-response mismatch index into the normalized dynamic feedforward weights.

[0018] Preferably, the generation of the reconstruction feedback signal includes:

[0019] In the formula, Indicates time The reconstructed feedback signal; The total number of instruments involved in the calculation; For the instrument's index number; For the first Each meter at time The instrument measurement value; For the first Each meter at time Dynamic feedforward weights; For the first Each meter at time Dynamic correction items.

[0020] This invention fuses the low-frequency component of the original measurement with the high-frequency component predicted by the command. The signal output to the controller retains accuracy while being advanced in time, thus offsetting the lag effect.

[0021] Preferably, the value range of the damping characteristic constant in the nonlinear coupling model is 3 to 5 times the average value of the response intensity index under steady-state noise conditions.

[0022] Preferably, the calculation of the maximum cross-correlation delay time is only triggered when the unit load change rate is detected to exceed the threshold, and the inherent pure time delay parameter remains unchanged under steady-state conditions.

[0023] Preferably, the reconstructed feedback signal directly replaces the original measurement signal in the deviation calculation of the PID closed-loop control loop.

[0024] The present invention has the following beneficial effects: This invention changes the traditional detection logic that only considers numerical values, instead diagnosing faults by calculating the rate of change and its corresponding energy. When a pressure-sensing line is just beginning to become blocked, the instrument readings are often still accurate, just reacting slowly, which traditional methods cannot detect. This invention utilizes a second-order differential algorithm to identify minute blunting features in the waveform, thus issuing an alarm before the signal becomes severely distorted. This means maintenance personnel no longer need to blindly purge the lines weekly; they only need to address the alarm precisely, significantly reducing unnecessary workload.

[0025] Furthermore, this invention completely solves the problem of control system oscillation caused by signal lag. Traditional PID controllers, unaware of signal lag, mistakenly assume the target hasn't been reached and continue adjusting, often resulting in over-adjustment. This invention, through an algorithm, compensates for the lag at the software level, generating a reconstructed signal that is ahead in time. When the controller receives this signal, it can make timely and correct adjustments, avoiding repeated valve movements that fluctuate wildly and ensuring stable parameters in the production process. Attached Figure Description

[0026] Figure 1 This is a flowchart of the steps of the closed-loop feedback regulation method for industrial processes based on instrument operating data deviation provided in this embodiment of the invention; Figure 2 This is a schematic diagram of phase alignment provided in an embodiment of the present invention; Figure 3 This is a trend chart of the excitation-response mismatch index provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the reconstruction feedback signal provided in an embodiment of the present invention. Detailed Implementation

[0027] Please see Figure 1 The diagram illustrates the steps of the closed-loop feedback control method for industrial processes based on instrument operating data deviation provided in Embodiment 1. The method includes the following steps: S1: Obtain historical steady-state data of the industrial control system, calculate the maximum cross-correlation delay time between control commands and instrument measurements, use this as a parameter to perform time-series backtracking of real-time control commands, and generate a phase-aligned reference command sequence.

[0028] It's important to note that the transfer of matter or energy objectively requires time. From the issuance of a control command to the action of the actuator, and then to the change in fluid state and its transmission to the sensor, there is an inherent time lag. This means that the instrument readings collected at the current moment actually reflect the results of control commands issued some time ago. Directly comparing the current command with the current measurement will lead to calculation errors. Therefore, to achieve accurate data comparison, this time lag must first be eliminated at the data processing level.

[0029] Preferably, as an example, historical steady-state data of the industrial control system is acquired, the maximum cross-correlation delay time between control commands and instrument measurements is calculated, and this is used as a parameter to perform time-series backtracking of real-time control commands, generating a phase-aligned reference command sequence, including: From the historical steady-state data of the DCS database, a sequence of control commands containing significant load variations or valve opening step changes is extracted, along with the corresponding instrument measurement sequence. Using a cross-correlation analysis algorithm, the correlation coefficient between this control command sequence and the instrument measurement sequence at different time displacements is calculated. The time displacement corresponding to the maximum correlation coefficient is identified and locked as the intrinsic pure time delay parameter.

[0030] Next, the control commands acquired in real time are processed using a timing backtracking algorithm to shift them, thereby obtaining the phase alignment reference command sequence. The specific implementation satisfies the following relation:

[0031] In the formula, Indicates at time The generated phase alignment reference command; Indicating a historical moment The control commands issued; Represents a time series index; This represents the inherent pure time delay parameter.

[0032] The phase alignment reference commands at all times are arranged in chronological order to obtain the phase alignment reference command sequence.

[0033] Understandably, the above processing shifts the control command sequence backward on the timeline. This ensures that the translated phase alignment reference command sequence and the measurement value sequence are physically and causally aligned. This eliminates time interference and provides a data foundation for subsequent processing.

[0034] In addition, the unit load change rate is used as the monitoring object. The above cross-correlation analysis is only triggered again to update the inherent pure time delay parameter when the unit load change rate exceeds the preset threshold. Under steady state, the inherent pure time delay parameter is kept unchanged to save computing resources.

[0035] Figure 2 This diagram illustrates phase alignment. The gray dotted lines represent the original control commands issued by the DCS system; the blue dashed lines represent the sensor signals actually acquired by the field sensors; and the black solid lines represent the phase alignment reference command sequence. As can be seen from the image, the transition points of the black solid lines and the blue dashed lines are precisely aligned on the time axis, thus proving that this invention successfully eliminates interference caused by transmission lag.

[0036] S2: Based on the phase alignment reference command sequence, the excitation intensity index characterizing the input action amplitude is calculated using a sliding window, and the response intensity index characterizing the high-frequency capture capability of the sensor is calculated based on the real-time instrument measurement value.

[0037] It should be noted that when the pressure tapping line becomes clogged, it increases fluid resistance, which is equivalent to adding a low-pass filter to the signal transmission path. This physical characteristic causes attenuation of the high-frequency components of the measurement signal, resulting in a smoother, more rounded waveform. Since the blockage has the most significant impact on the waveform curvature, it is necessary to calculate the amplitude of the input movement and the curvature characteristics of the output separately.

[0038] Construct a length of A sliding window, using the phase alignment reference command sequence as input, calculates the excitation intensity index using a first-order differential accumulation algorithm, specifically satisfying the following relationship:

[0039] Simultaneously, using the real-time acquired instrument measurement sequence as input, the response intensity index is calculated using the second-order difference curvature algorithm, specifically satisfying the following relationship:

[0040] In the formula, , They are time points and time Phase alignment reference command, For a moment Incentive intensity index; For the first Each meter at time The response intensity index; For the instrument's index number; For time series indexing; This is the backtracking index within the sliding window; The length of the sliding window, for example, Take 15; For example, an energy-sensitive factor. Take 2; Indicates the first Each meter at time The instrument measurement value; To prevent the denominator from being zero, a very small positive constant is set, for example, Take 0.001.

[0041] Understandable, The larger the value, the greater the cumulative change in the control command within the window period; The smaller the value, the closer the waveform of the measured signal is to a straight line within the window period, lacking inflection points of high-frequency changes. Through the above processing, the abstract physical phenomenon of waveform blunting can be transformed into two numerical indicators.

[0042] S3: Construct a nonlinear coupling model that includes exponential decay characteristics, and calculate the excitation-response mismatch index, which reflects the degree of kinetic energy mismatch, using the excitation intensity index as the amplitude and the reciprocal characteristic of the response intensity index as the exponent.

[0043] It should be noted that a single value of the response intensity index or excitation intensity index is insufficient to accurately diagnose a fault, as both indices are very small under steady-state conditions, which is normal. Only when both conditions are met—a drastic change in the input command and an extremely weak change in the output feedback—does the physical characteristics of a pipe blockage hold. Therefore, a mathematical model is needed to reflect these physical characteristics.

[0044] Preferably, as an example, a nonlinear coupling model incorporating exponential decay characteristics is constructed. Using the excitation intensity index as the amplitude and the reciprocal characteristic of the response intensity index as the exponent, an excitation-response mismatch index reflecting the degree of kinetic energy mismatch is calculated, including: By substituting the excitation intensity index as the amplitude term and the response intensity index as the denominator term of the exponential function into the exponential decay coupling model for calculation, the excitation-response mismatch index is obtained. Specifically, this implementation satisfies the following relationship:

[0045] In the formula, Indicates the first Each meter at time The incentive-response mismatch index; For a moment Incentive intensity index; For the first Each meter at time The response intensity index; This is the damping characteristic constant, and its value is usually set to 3 to 5 times the average response intensity index under steady-state noise conditions. In this embodiment, it is set to 8. This is the upper limit of saturation, for example. Take 100; This is the operator for finding the minimum value. It is an exponential function with the natural constant as its base. It is a nonlinear coupled model that incorporates exponential decay characteristics.

[0046] Understandably, when the system is in a steady state, Approaching zero, as a multiplier term, it makes the overall result... Approaching zero; when the system is adjusting normally, The exponent is relatively large, resulting in a large exponent term. Smaller The function value is close to a constant, and the overall result... Maintain a low level; when a blockage occurs, Approaching zero, causing the exponential term to... Approaching infinity The function value increased explosively, greatly amplifying The numerical value. This model can accurately identify and amplify specific fault characteristics of no response to stimulation.

[0047] Figure 3 The graph shows the trend of the excitation-response mismatch index. The solid red line represents the time series image of the excitation-response mismatch index; the light yellow background area represents the fault period when the system is in a state of high hysteresis or pipeline congestion. The graph shows that under normal operating conditions, the red line remains consistently near zero, indicating that this solution has strong noise immunity and no false alarms. However, after entering the fault period, with drastic changes in commands but a slow response in measured values, the red line value spikes dramatically, indicating that this solution has high sensitivity.

[0048] S4: The phase alignment reference command sequence is mapped to theoretical physical values ​​using a pre-calibrated static curve. Dynamic feedforward weights are calculated by combining the excitation response mismatch index. The trend deviation between the theoretical physical values ​​and the current measured values ​​is weighted and superimposed onto the instrument measured values ​​through the dynamic feedforward weights to generate a reconstructed feedback signal, thereby realizing real-time adjustment of the closed-loop control system.

[0049] It should be noted that in order to eliminate oscillations in the closed-loop control system, a feedback signal without phase lag needs to be provided to the controller. Since the control command is lag-free, if the trend of the control command's change can be converted into a physical quantity and superimposed on the lag measurement value, a signal with phase lead can be mathematically constructed.

[0050] Preferably, as an example, the phase alignment reference command sequence is mapped to theoretical physical values ​​using a pre-calibrated static curve. A dynamic feedforward weight is calculated based on the excitation response mismatch index. The trend deviation between the theoretical physical values ​​and the current measured values ​​is weighted and superimposed onto the instrument measured values ​​using the dynamic feedforward weight to generate a reconstructed feedback signal, thereby achieving real-time adjustment of the closed-loop control system. This includes: First, using the phase alignment reference command sequence as input, the theoretical physical value is calculated using the pre-calibrated static curve, and the deviation between this value and the current actual measurement value is calculated. Combined with differential components, a dynamic correction term is obtained, specifically satisfying the following relation:

[0051] In the formula, Indicates the first Each meter at time Dynamic correction terms; Indicates the first Each meter at time The instrument measurement value; Indicates the first Each meter at time The instrument measurement value; Indicates time-based The theoretical physical values ​​obtained by mapping the phase alignment reference command. It is a static curve; For example, the feedforward weighting coefficients. Take 0.8, where This item reflects the dynamic gap between theoretical and actual values.

[0052] Understandable, This represents the actual amount of change that is happening at the sensor. This represents the gap between the theoretical target value and the current actual value. It utilizes the characteristic of no instruction delay to tell the system what the output should be. When a pipe blockage causes... When lagging, The value will increase.

[0053] In summary, the above relationship is based on the slow changes of the sensor itself, with an additional correction based on theoretical prediction added to calculate the trend of data change if the pipeline is not blocked.

[0054] It should be added that the method for obtaining the static curve is as follows: Steady-state operating condition data segments are selected, and the criteria for selection are: within a continuous time window, the rate of change of control commands and the rate of change of instrument measurement values ​​are both less than a preset steady-state threshold. For example, the steady-state threshold is 0.1%.

[0055] The control command values ​​in the steady-state operating condition data segment are extracted as independent variables, and the corresponding instrument measurement values ​​are extracted as dependent variables to construct a sample point set.

[0056] The static curve is obtained by performing a polynomial fitting on the sample point set using the least squares method. .

[0057] Next, using the stimulus-response mismatch index as input, the normalized dynamic feedforward weights are calculated using the S-shaped growth curve function, specifically satisfying the following relationship:

[0058] In the formula, Indicates the first Each meter at time Dynamic feedforward weights; The excitation-response mismatch index is the one calculated above. For example, the gain limit value of the weight. Take 1; For example, the threshold for determining hysteresis is... Take 15; For example, the steepness coefficient. Take 1.5. This is an S-shaped growth curve function.

[0059] Understandably, when When the value is very small, it indicates that the pipeline is unobstructed and there is no blockage. At this point, the exponent in the denominator is very large, leading to a significant increase in the overall result. If the value approaches 0, it indicates that the system is normal and does not need to be corrected using theoretical values; it can remain as is. When the value is very high, it indicates a severe blockage. At this point, the exponential term disappears, and the overall result... Rapidly rising and approaching This indicates that the sensor is no longer reliable and that theoretical predictions must be introduced for significant correction.

[0060] Then, using the original measured values, dynamic correction terms, and dynamic feedforward weights as common inputs, a weighted algorithm is used to generate a reconstructed feedback signal, specifically satisfying the following relationship:

[0061] In the formula, Indicates time The reconstructed feedback signal; The total number of instruments involved in the calculation; For the instrument's index number; For the first Each meter at time The instrument measurement value.

[0062] Understandable, This ensures that the signal values ​​are accurate under steady-state conditions; It injects high-frequency change trends based on command predictions into the signal. Because... Originating from lag-free instructions, the final generated Phase lead was achieved on the time axis. This is similar to artificially adding a compensation signal to an originally sluggish sensor signal, enabling it to keep up with changes in control commands in real time.

[0063] Figure 4 The diagram illustrates the reconstructed feedback signal. The blue dashed line represents the unprocessed instrument measurement value; the gray dotted line represents the theoretical physical value obtained through pre-calibrated static curve mapping based on the hysteresis-free control command; and the red solid line represents the reconstructed feedback signal. As can be seen from the image, in the area marked by the double-headed arrows, at the instant of the command step, the blue dashed line is still slowly rising, while the red solid line has already rapidly jumped and closely overlapped with the gray dotted line. This indicates that the red signal achieves a significant phase lead on the time axis.

[0064] Finally, the reconstructed feedback signal is directly fed into the PID module as a process variable, replacing the original measured value in the closed-loop control. After receiving this signal without lag, the PID controller can make timely adjustments, thereby eliminating overshoot and oscillation caused by feedback lag.

[0065] This concludes the embodiment.

[0066] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A closed-loop feedback control method for industrial processes based on instrument operating data deviation, characterized in that, include: Acquire historical steady-state data of the industrial control system, calculate the maximum cross-correlation delay time between control commands and instrument measurements, use this as a parameter to perform time-series backtracking of real-time control commands, and generate a phase-aligned reference command sequence. Based on the phase alignment reference command sequence, the excitation intensity index characterizing the input action amplitude is calculated using a sliding window, and the response intensity index characterizing the high-frequency capture capability of the sensor is calculated based on the real-time instrument measurement value. A nonlinear coupling model incorporating exponential decay characteristics is constructed. Using the excitation intensity index as the amplitude and the reciprocal characteristic of the response intensity index as the exponent, the excitation-response mismatch index reflecting the degree of kinetic energy mismatch is calculated. The phase alignment reference command sequence is mapped to theoretical physical values ​​using a pre-calibrated static curve. Dynamic feedforward weights are calculated by combining the excitation response mismatch index. The trend deviation between the theoretical physical values ​​and the current measured values ​​is weighted and superimposed onto the instrument measured values ​​through the dynamic feedforward weights to generate a reconstructed feedback signal, thereby realizing real-time adjustment of the closed-loop control system.

2. The closed-loop feedback control method for industrial processes based on instrument operating data deviation as described in claim 1, characterized in that, The generation of the phase alignment reference command sequence involves reading a pre-calibrated intrinsic pure time delay parameter, subtracting the intrinsic pure time delay parameter from each time point to obtain a backtracking time, and directly retrieving the control command of the backtracking time as the phase alignment reference command for each time point, thereby obtaining the phase alignment reference command sequence.

3. The closed-loop feedback control method for industrial processes based on instrument operating data deviation as described in claim 1, characterized in that, The response intensity index satisfies the following relationship: In the formula, For the first Each meter at time The aforementioned response intensity index; For the instrument's index number; For time series indexing; This is the backtracking index within the sliding window; The length of the sliding window; , and They represent the first Each meter at time , and The instrument measurement value; To prevent extremely small positive constants with a value of zero; The excitation intensity index is obtained by calculating the sum of the absolute values ​​of the first-order differences of the phase alignment reference command sequence within the sliding window and then performing a power operation.

4. The closed-loop feedback control method for industrial processes based on instrument operating data deviation as described in claim 1, characterized in that, The excitation-response mismatch index satisfies the following relationship: In the formula, For the first Each meter at time The excitation-response mismatch index; For a moment The aforementioned incentive intensity index; For the first Each meter at time The aforementioned response intensity index; The damping characteristic constant is set based on the reference noise level; This is the upper limit of saturation. This is the operator for finding the minimum value. It is an exponential function with the natural constant as its base.

5. The closed-loop feedback control method for industrial processes based on instrument operating data deviation as described in claim 2, characterized in that, Before generating the reconstructed feedback signal, a dynamic correction term needs to be calculated. The calculation process of the dynamic correction term includes: calculating the deviation between the theoretical physical value corresponding to the phase alignment reference command and the instrument measurement value, and weighting and summing the deviation with the measurement micro-component at the current moment to obtain the dynamic correction term.

6. The closed-loop feedback control method for industrial processes based on instrument operating data deviation as described in claim 1, characterized in that, The calculation of the dynamic feedforward weights is achieved by using an S-shaped growth curve function, which maps the excitation-response mismatch index to the normalized dynamic feedforward weights by setting a hysteresis determination threshold and a steepness coefficient.

7. The closed-loop feedback control method for industrial processes based on instrument operating data deviation according to claim 1, characterized in that, The generation of the reconstruction feedback signal includes: In the formula, Indicates time The reconstructed feedback signal; The total number of instruments involved in the calculation; For the instrument's index number; For the first Each meter at time The instrument measurement value; For the first Each meter at time Dynamic feedforward weights; For the first Each meter at time Dynamic correction items.

8. The closed-loop feedback control method for industrial processes based on instrument operating data deviation as described in claim 4, characterized in that, The range of the damping characteristic constant in the nonlinear coupling model is 3 to 5 times the average value of the response intensity index under steady-state noise conditions.

9. The closed-loop feedback control method for industrial processes based on instrument operating data deviation according to claim 1, characterized in that, The calculation of the maximum cross-correlation delay time is only triggered when the unit load change rate is detected to exceed the threshold, and the inherent pure time delay parameter remains unchanged under steady-state conditions.

10. The closed-loop feedback control method for industrial processes based on instrument operating data deviation according to claim 1, characterized in that, The reconstructed feedback signal directly replaces the original measurement signal in the deviation calculation of the PID closed-loop control loop.

Images