CN-122018334-A - Electromechanical equipment energy-saving management and control system and method based on intelligent monitoring

CN122018334ACN 122018334 ACN122018334 ACN 122018334ACN-122018334-A

Abstract

The application discloses an electromechanical equipment energy-saving control system and method based on intelligent monitoring, and relates to the field of intelligent energy-saving control. On the basis, a regression feature vector representing the current physical state of the equipment is established through physical feature extraction and linearization conversion. And then dynamically adjusting the algorithm gain by using a forgetting factor based on posterior error, performing time-varying tracking on model parameters, introducing a projection correction mechanism under physical constraint, and forcing the updated parameters to be always in a reasonable physical feasible region. Finally, the self-adaptive energy efficiency model based on real-time correction performs global optimization under the constraint of working conditions, so that the control model is ensured to accurately follow the aging process of the equipment, and the accurate energy-saving control of the full life cycle is realized.

Inventors

XU BAOWEI
XU BAOPENG
LI SHUJU
LI CHAODE
SONG JIALI
Qi Longyi

Assignees

浙江九烁光电工程技术有限公司

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. An electromechanical device energy-saving control method based on intelligent monitoring is characterized by comprising the following steps: s1, performing low-pass filtering processing and excitation effectiveness pre-judging on an original sensor data stream of acquired electromechanical equipment to obtain an excitation effective zone bit and a filtered data set; S2, carrying out physical feature extraction and linearization conversion on the filtered data set to obtain a regression feature vector and an observation scalar for representing the physical state at the current moment; S3, when the excitation valid bit indicates valid, determining a forgetting factor based on the posterior error at the last moment, and determining a gain vector capable of suppressing noise interference by combining the regression feature vector and the covariance matrix at the last moment; S4, carrying out parameter updating and projection correction under physical constraint on the model parameters at the previous moment based on the gain vector, the forgetting factor and the observation scalar to obtain corrected model parameters and a covariance matrix at the current moment; s5, constructing an adaptive energy efficiency prediction model based on corrected model parameters, and performing global optimization on the adaptive energy efficiency prediction model under the condition that the working condition constraint corresponding to the filtered data set is met to obtain an optimal set point; And S6, issuing an optimal set point to the electromechanical equipment for control execution.
2. The intelligent monitoring-based energy-saving control method for electromechanical equipment according to claim 1, wherein the raw sensor data stream comprises three-phase voltage, three-phase current, active power, power factor, working medium inlet temperature, working medium outlet temperature, working medium flow, system pressure, operating frequency, valve opening, start-stop state, environment dry-bulb temperature and relative humidity, wherein step S1 comprises: Carrying out multichannel signal synchronization and digital low-pass filtering on the original sensor data stream to obtain a filtered data set; performing Euclidean distance-based instantaneous change rate estimation on the filtered data set at the previous moment and the filtered data set at the current moment to obtain a data change rate index reflecting the disturbance degree of the system; And carrying out logic comparison judgment on the data change rate index and a preset dead zone threshold value to obtain an excitation effective zone bit.
3. The method for controlling energy conservation of electromechanical equipment based on intelligent monitoring according to claim 1, wherein step S2 comprises: Analyzing and extracting independent state variables and dependent variables required by constructing an energy consumption equation from the filtered data set to obtain a structured extraction variable group; performing mechanism-based nonlinear feature transformation on independent state variables in the extracted variable group to obtain a transformed feature item list; the transformed feature term list map is assembled into a linearized version of the regression feature vector and the dependent variables in the set of extracted variables are mapped into the observed scalar.
4. The method for controlling energy conservation of electromechanical equipment based on intelligent monitoring according to claim 1, wherein step S3 comprises: when the excitation valid bit is true, determining a forgetting factor based on the posterior error at the last moment; Determining a gain molecular vector based on the covariance matrix and the regression feature vector at the previous moment; Determining a gain denominator scalar based on the quadratic form of the regression feature vector with respect to the covariance matrix at the previous moment and combining the forgetting factor; kalman gain vector synthesis is performed on the gain numerator vector and the gain denominator scalar to obtain a gain vector for determining the current observation weight.
5. The method for controlling energy conservation of electromechanical equipment based on intelligent monitoring according to claim 1, wherein step S3 comprises: When the excitation effective zone bit is true, calculating random gradient self-adaptive forgetting factors of the gradient flow based on posterior error at the last moment to obtain variable forgetting factors; Determining a gain molecular vector based on the covariance matrix and the regression feature vector at the previous moment; Determining a gain denominator scalar based on the quadratic form of the regression feature vector with respect to the covariance matrix at the previous moment and combining the forgetting factor; kalman gain vector synthesis is performed on the gain numerator vector and the gain denominator scalar to obtain a gain vector for determining the current observation weight.
6. The intelligent monitoring-based energy-saving control method for electromechanical equipment according to claim 5, wherein when the excitation valid flag bit is true, performing random gradient adaptive forgetting factor calculation of the gradient flow based on the posterior error at the previous moment to obtain a forgetting factor, comprising: calculating the gradient of the cost function on the forgetting factor based on the posterior error of the last moment and the sensitivity vector of the last moment, and updating the forgetting factor along the negative gradient direction; And performing sensitivity vector recursion evolution based on the gain vector at the last moment and the covariance matrix at the last moment to obtain the sensitivity vector at the last moment.
7. The method for controlling energy conservation of electromechanical equipment based on intelligent monitoring according to claim 1, wherein step S4 comprises: Carrying out pre-estimation calculation on the regression feature vector by using the model parameter at the previous moment to obtain a calculation result, and carrying out difference between the calculation result and an observation scalar to obtain a priori error reflecting the deviation of the old model; correcting the model parameters at the previous moment based on the gain vector and the prior error to obtain preliminary updated parameters, and updating the covariance matrix at the previous moment by using the forgetting factor to obtain a covariance matrix at the current moment; and carrying out physical feasible domain projection correction on each element of the preliminary updated parameters to obtain corrected model parameters.
8. The method for controlling energy conservation of electromechanical equipment based on intelligent monitoring according to claim 1, wherein step S5 comprises: constructing a self-adaptive objective function capable of sensing the current aging degree of the equipment based on the corrected model parameters; Constructing a feasible region of the operating variables based on the device security specification and the load requirements in the filtered data set to obtain a feasible region constraint set comprising inequality and equality constraints; and searching minimum values of the adaptive objective function in the feasible region constraint set to obtain an optimal set point.
9. The method for power saving management and control of an electromechanical device based on intelligent monitoring as set forth in claim 1, wherein step S6 comprises mapping the optimal set point to an industrial bus register address and sending the industrial bus register address to the device controller, and triggering a control execution completion signal after waiting for a preset response settling time.
10. Electromechanical device energy-saving management and control system based on intelligent monitoring, characterized by comprising: The electromechanical device data processing module is used for carrying out low-pass filtering processing and excitation effectiveness pre-judging on the acquired original sensor data stream of the electromechanical device so as to obtain an excitation effective zone bit and a filtered data set; the feature extraction and linearization conversion module is used for carrying out physical feature extraction and linearization conversion on the filtered data set to obtain a regression feature vector and an observation scalar which are used for representing the physical state at the current moment; The gain vector generation module is used for determining a forgetting factor based on posterior error at the last moment when the excitation valid bit indicates valid, and determining a gain vector capable of suppressing noise interference by combining the regression feature vector and the covariance matrix at the last moment; The parameter updating and projection correcting module is used for carrying out parameter updating and projection correction under physical constraint on the model parameter at the previous moment based on the gain vector, the forgetting factor and the observation scalar so as to obtain a corrected model parameter and a covariance matrix at the current moment; the optimal set point determining module is used for constructing an adaptive energy efficiency prediction model based on the corrected model parameters, and carrying out global optimization on the adaptive energy efficiency prediction model under the condition that the working condition constraint corresponding to the filtered data set is met so as to obtain an optimal set point; and the control execution module is used for issuing an optimal set point to the electromechanical equipment for control execution.

Description

Electromechanical equipment energy-saving management and control system and method based on intelligent monitoring Technical Field The application relates to the field of intelligent energy-saving control, in particular to an electromechanical equipment energy-saving control system and method based on intelligent monitoring. Background The existing electromechanical equipment energy-saving management and control technology mainly depends on a static physical model or a black box model based on historical data offline training, and the defects of poor adaptability and physical constraint deletion to the physical characteristic change of the whole life cycle of equipment are commonly existed. The prior art is generally based on the static assumption that the key physical parameters of the device, such as heat transfer coefficient, compressor efficiency, coefficient of friction resistance, etc., are constant values or are only a fixed functional relationship with load changes. However, in a practical long-period operating environment, the electromechanical device is inevitably affected by physical factors such as fouling deposition of the heat exchanger, deterioration of lubricating oil, wear of mechanical parts, and the like, and the physical characteristics of the electromechanical device are subjected to irreversible nonlinear time-varying degradation (namely soft failure) along with the operating time. The degradation phenomenon causes the actual physical parameters of the equipment to have significant drift with the initially calibrated model parameters after a period of operation, and serious model mismatch problems are generated. Because the pure physical model is difficult to accurately capture the complex microscopic dynamic attenuation process, the pure data driving model lacking physical mechanism constraint is easy to output control instructions which violate physical common sense (such as negative pressure, ultrahigh efficiency and the like) in an extreme working condition or a sample sparse region, thereby bringing control potential safety hazards. In addition, the existing control modes are mostly applied by an open-loop model, namely, the model is used for a long time once training or calibration is completed, and an online and automatic closed-loop mechanism of residual observation-parameter correction is lacked to resist equipment aging in real time. The energy-saving control strategy can only calculate the optimal working point based on outdated model parameters, so that the calculated control parameters deviate from the actual optimal state, the expected energy-saving effect cannot be realized, and continuous and invisible energy waste is caused. Therefore, there is a need for an energy-saving management and control scheme for electromechanical devices that can integrate physical mechanism constraints and sense and adaptively modify the aging characteristics of the devices in real time according to the monitored data. Disclosure of Invention The application provides an electromechanical device energy-saving management and control method based on intelligent monitoring, aiming at the problems in the prior art, which comprises the following steps of S1, conducting low-pass filtering processing and excitation effectiveness pre-judging on an original sensor data stream of the acquired electromechanical device to obtain an excitation effective zone bit and a filtered data set, S2, conducting physical feature extraction and linearization conversion on the filtered data set to obtain a regression feature vector and an observation scalar used for representing the physical state at the current moment, S3, when the excitation effective zone bit indicates effective, determining a forgetting factor based on a posterior error at the last moment, determining a gain vector capable of restraining noise interference by combining the regression feature vector and a covariance matrix at the last moment, S4, conducting parameter updating and projection correction on the model parameter at the last moment based on the gain vector, the forgetting factor and the observation scalar to obtain a corrected model parameter and a covariance matrix at the current moment, S5, constructing an adaptive energy efficiency prediction model based on the corrected model parameter, conducting global optimal prediction model under the condition constraint corresponding to the filtered data set, and conducting optimal setting point control to the electromechanical device under the condition that the optimal setting point is met. The application further provides an electromechanical device energy-saving management and control system based on intelligent monitoring, which comprises an electromechanical device data processing module, a feature extraction and linearization conversion module, a set point optimal determination module, an adaptive energy efficiency prediction model and an optimal control set point optimal control module,