Parallel control method and device of generator set

Abstract

The invention discloses a parallel control method and device for a generator set. The method comprises the following steps: acquiring a historical working data set and a historical control data set of the generator set; performing modeling processing on the historical working data set and the historical control data set to obtain a working state prediction model; acquiring a current data set of the generator set; the current data set comprises a current working data set and a current control data set; performing control prediction processing on the current data set by using a working state prediction model to obtain a control prediction data set of the generator set; and inputting the control prediction data set as control data of the generator set into the generator set to complete parallel control of the generator set.

Description

Technical Field

[0001] The present invention relates to the fields of industrial data processing, intelligent control and strategy optimization technology, and in particular to a parallel control method and device for generator sets. Background Art

[0002] Generator sets play a vital role in modern power systems, and the performance of their parallel control directly affects the stability and reliability of power supply. Traditional generator set parallel control methods mostly rely on fixed control strategies and manual experience, and are difficult to adapt to complex and changeable operating conditions. For example, when the grid load fluctuates greatly, the generator set performance ages, or a sudden failure occurs, traditional methods are often unable to adjust the operating status of the generator set in a timely and accurate manner, resulting in poor parallel control effects, and may even cause system oscillations, uneven power distribution, and other problems, thereby affecting the safe and stable operation of the power system. In addition, traditional methods lack in-depth mining and utilization of the historical operating data of generator sets, and cannot fully utilize the potential laws contained in the data to optimize the control strategy, making it difficult to achieve refined and intelligent control of generator sets. Summary of the Invention

[0003] The present invention mainly solves the problem of how to accurately and stably control a generator set to improve the stability and reliability of power supply. The present invention discloses a parallel control method and device for a generator set.

[0004] According to a first aspect of an embodiment of the present invention, a parallel control method for a generator set is disclosed, comprising:

[0005] S1, obtaining a historical operating data set and a historical control data set of a generator set;

[0006] S2, performing modeling processing on the historical working data set and the historical control data set to obtain a working status prediction model;

[0007] S3, collecting and obtaining a current data set of the generator set; the current data set includes a current working data set and a current control data set;

[0008] S4, using the working state prediction model, performing control prediction processing on the current data set to obtain a control prediction data set of the generator set;

[0009] S5, inputting the control prediction data set into the generator set as the control data of the generator set to complete the parallel control of the generator set.

[0010] The modeling process of the historical working data set and the historical control data set to obtain a working state prediction model includes:

[0011] S21, combining the historical control data set as a system measurement vector, combining the historical working data set as a target state vector, and using a linear system state equation modeling method to construct a working state prediction model to be solved;

[0012] The expression of the working state prediction model is:

[0013] x(k+1)=Φ(k+1,k)x(k)+G(k)w(k),

[0014] z(k)=H(k)x(k)+v(k),

[0015] Where Φ(k+1,k) represents the state transfer matrix at time k, H(k) represents the measurement matrix at time k, G(k) represents the input matrix at time k, w(k) and v(k) represent the state noise and measurement noise at time k, respectively, x(k) represents the target state vector at time k, z(k) represents the system measurement vector at time k, and x(k+1) represents the predicted target state vector at time k+1.

[0016] S22 , using the historical working data set and the historical control data set, solving the working state prediction model to be solved to obtain the working state prediction model.

[0017] The method of using the working state prediction model to perform control prediction processing on the current data set to obtain a control prediction data set of the generator set includes:

[0018] S41, preprocessing the current data set to obtain a preprocessed data set;

[0019] S42, using the working state prediction model, processing the current control data set in the preprocessed data set to obtain a working data prediction set;

[0020] S43, performing feedback calculation processing on the working data prediction set and the current working data set in the pre-processed data set to obtain a control prediction data set of the generator set.

[0021] The preprocessing of the current data set to obtain a preprocessed data set includes:

[0022] S411, performing data cleaning processing on the current data set to obtain a first data set;

[0023] S412, performing time alignment processing on the first data set to obtain a second data set;

[0024] S413: Perform feature filtering on the second data set to obtain a preprocessed data set.

[0025] The expression of the feature filtering process is:

[0026]

[0027] Among them, s(k) is the kth element of a type of data sequence in the second data set, NS is the length of the data sequence, y1(z,x) is the element of the zth row and xth column of the matrix corresponding to the transformed data information obtained by the spatial feature filtering process, p1() is the transformation function corresponding to the spatial feature filtering process, p1(kT1-zT1) is the value of p1() at kT1-zT1, T1 and F1 are the time domain transformation length and frequency domain transformation length of the spatial feature filtering process, respectively.

[0028] After obtaining the matrix y1, all row vectors of the matrix are concatenated to obtain a preprocessed data sequence of the class data sequence;

[0029] Using all preprocessed data sequences, a preprocessed data set is constructed.

[0030] The performing feedback calculation processing on the working data prediction set and the current working data set in the pre-processed data set to obtain the control prediction data set of the generator set includes:

[0031] S431, performing a difference quantization calculation on the working data prediction set and the current working data set in the pre-processed data set to obtain a difference value sequence;

[0032] S432 , performing feedback quantization calculation on the difference value sequence and the current control data set in the pre-processed data set to obtain a control prediction data set of the generator set.

[0033] The expression for the difference quantification calculation is:

[0034]

[0035] Among them, h i is the i-th element in the difference value sequence, is the mean value of the i-th type of working parameter data sequence of the working data prediction set, A ij is the jth element of the i-th type of working parameter data sequence of the working data prediction set, B ij is the jth element of the i-th type working parameter data sequence of the current working data set in the preprocessing data set, and N is the length of the working parameter data sequence.

[0036] According to a second aspect of the present invention, a parallel control device for a generator set is disclosed, the device comprising:

[0037] a memory storing executable program code;

[0038] a processor coupled to the memory;

[0039] The processor calls the executable program code stored in the memory to execute the parallel control method of the generator sets.

[0040] According to a third aspect of the present invention, a computer storable medium is disclosed. The computer storable medium stores computer instructions. When the computer instructions are called by a computer, they are used to execute the parallel control method of the generator sets.

[0041] According to a fourth aspect of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the parallel control method of the generator sets.

[0042] The beneficial effects of the present invention are:

[0043] 1. Improve parallel control accuracy: The present invention establishes a working status prediction model and uses the model to perform control prediction processing on the current data set of the generator set. It can dynamically adjust the control strategy according to the actual operating status of the generator set, thereby effectively improving the accuracy of parallel control of the generator set and reducing power fluctuations and system instability caused by control deviations.

[0044] 2. Enhanced System Adaptability: Modeling based on historical operating and control data sets allows the present invention to fully account for the operating characteristics of generator sets under varying operating conditions, including grid load variations, equipment aging, and environmental factors. Consequently, even in complex operating environments, the present invention maintains excellent control effectiveness, significantly enhancing the adaptability of the generator set parallel control system.

[0045] 3. Optimize power distribution: By performing feedback calculations on the predicted working data set and the current working data set, the present invention can monitor the power output of the generator sets in real time and adjust the control parameters of each unit in a timely manner based on the predicted results, thereby achieving optimized power distribution between the generator sets and improving the operating efficiency of the entire power generation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 4 is an implementation flow chart of the method of the present invention. DETAILED DESCRIPTION

[0047] In order to better understand the content of the present invention, an embodiment is given here.

[0048] Figure 1 4 is an implementation flow chart of the method of the present invention.

[0049] According to a first aspect of an embodiment of the present invention, a parallel control method for a generator set is disclosed, comprising:

[0050] S1, obtaining a historical operating data set and a historical control data set of a generator set;

[0051] S2, performing modeling processing on the historical working data set and the historical control data set to obtain a working status prediction model;

[0052] S3, collecting and obtaining a current data set of the generator set; the current data set includes a current working data set and a current control data set;

[0053] S4, using the working state prediction model, performing control prediction processing on the current data set to obtain a control prediction data set of the generator set;

[0054] S5, inputting the control prediction data set into the generator set as the control data of the generator set to complete the parallel control of the generator set.

[0055] The modeling process of the historical working data set and the historical control data set to obtain a working state prediction model includes:

[0056] S21, combining the historical control data set as a system measurement vector, combining the historical working data set as a target state vector, and using a linear system state equation modeling method to construct a working state prediction model to be solved;

[0057] The expression of the working state prediction model is:

[0058] x(k+1)=Φ(k+1,k)x(k)+G(k)w(k),

[0059] z(k)=H(k)x(k)+v(k),

[0060] Where Φ(k+1,k) represents the state transfer matrix at time k, H(k) represents the measurement matrix at time k, G(k) represents the input matrix at time k, w(k) and v(k) represent the state noise and measurement noise at time k, respectively, x(k) represents the target state vector at time k, z(k) represents the system measurement vector at time k, and x(k+1) represents the predicted target state vector at time k+1.

[0061] S22, using the historical working data set and the historical control data set to solve the working state prediction model to obtain the working state prediction model;

[0062] By modeling and analyzing historical generator set operating data, this invention can predict potential equipment failures or performance degradation trends in advance. Based on this, operators can take preventive maintenance measures in advance to avoid downtime caused by sudden equipment failures, thereby reducing generator set maintenance costs and extending equipment life.

[0063] The method of using the working state prediction model to perform control prediction processing on the current data set to obtain a control prediction data set of the generator set includes:

[0064] S41, preprocessing the current data set to obtain a preprocessed data set;

[0065] S42, using the working state prediction model, processing the current control data set in the preprocessed data set to obtain a working data prediction set;

[0066] S43, performing feedback calculation processing on the working data prediction set and the current working data set in the pre-processed data set to obtain a control prediction data set of the generator set.

[0067] The control method of the present invention can effectively reduce power fluctuations and system oscillations when generator sets are operated in parallel, thereby improving the stability of the power system. This is of great significance for ensuring the reliability and security of power supply, and is particularly suitable for applications with high power quality requirements, such as data centers, hospitals, and industrial production.

[0068] The preprocessing of the current data set to obtain a preprocessed data set includes:

[0069] S411, performing data cleaning processing on the current data set to obtain a first data set;

[0070] S412, performing time alignment processing on the first data set to obtain a second data set;

[0071] S413, performing feature filtering on the second data set to obtain a preprocessed data set;

[0072] The data cleaning process includes filling missing values, smoothing noise data, and smoothing or deleting outliers. Smoothing noise data involves first identifying noise data and then smoothing it based on the preceding and following data. Noise data is defined as values that are less than the sensor's sensitivity or greater than the sensor's upper limit. Kalman filtering can be used to identify outliers. The value to fill missing values can be determined by averaging the measured values within a certain sampling interval before and after the missing value.

[0073] The time alignment process is to unify different types of data to the same time base, which can be achieved by using a time registration algorithm; the time registration algorithm can use an interpolation / extrapolation method, a Lagrange three-point interpolation method, etc.

[0074] The expression of the feature filtering process is:

[0075]

[0076] Among them, s(k) is the kth element of a type of data sequence in the second data set, NS is the length of the data sequence, y1(z,x) is the element of the zth row and xth column of the matrix corresponding to the transformed data information obtained by the spatial feature filtering process, p1() is the transformation function corresponding to the spatial feature filtering process, p1(kT1-zT1) is the value of p1() at kT1-zT1, T1 and F1 are the time domain transformation length and frequency domain transformation length of the spatial feature filtering process, respectively.

[0077] Feature filtering effectively removes noise from the current data set. Different noises exhibit distinct characteristics in the frequency and time domains. This expression utilizes the time and frequency domain transform lengths to process the data, separating noise signatures from valid data features and highlighting key features within the data. This helps improve the accuracy of subsequent operating status prediction models because, when processing purer and more representative data, the model can better learn the inherent patterns within the data, thereby enhancing the accuracy of generator set operating status predictions.

[0078] In practical applications, various types of data from generator sets may differ due to factors such as sampling time and sensor location. Feature filtering can be used to uniformly transform these data and map them into a unified space, facilitating subsequent data processing and analysis.

[0079] After obtaining the matrix y1, all row vectors of the matrix are concatenated to obtain a preprocessed data sequence of the class data sequence;

[0080] Using all preprocessed data sequences, a preprocessed data set is constructed.

[0081] p1() can be a Gaussian function;

[0082] The performing feedback calculation processing on the working data prediction set and the current working data set in the pre-processed data set to obtain the control prediction data set of the generator set includes:

[0083] Performing difference quantification calculation on the working data prediction set and the current working data set in the preprocessed data set to obtain a difference value sequence;

[0084] Performing feedback quantization calculation on the difference value sequence and the current control data set in the preprocessed data set to obtain a control prediction data set of the generator set;

[0085] The expression for the difference quantification calculation is:

[0086]

[0087] Among them, h i is the i-th element in the difference value sequence, is the mean value of the i-th type of working parameter data sequence of the working data prediction set, A ij is the jth element of the i-th type of working parameter data sequence of the working data prediction set, B ij is the jth element of the i-th type working parameter data sequence of the current working data set in the preprocessing data set, and N is the length of the working parameter data sequence.

[0088] The expression for quantifying the difference comprehensively considers the mean and element values of various operating parameter data sequences in both the predicted working data set and the preprocessed current working data set. By calculating the difference value sequence, the degree of difference between the two data sets can be accurately quantified. Traditional methods may simply calculate differences or ratios, but this expression incorporates exponential functions and absolute value operations, which can more comprehensively and sensitively reflect subtle differences between the data, providing a more accurate basis for subsequent feedback control.

[0089] During the actual operation of a generator set, operating data fluctuates due to various factors. This expression, by weighting the data, can better adapt to such fluctuations. When data fluctuations are large, exponential functions and absolute value operations can amplify the impact of the differences, allowing the system to promptly perceive data changes. When data fluctuations are small, they can also accurately capture subtle differences, ensuring stable and accurate system operation under various operating conditions. By analyzing the difference value sequence, the system can promptly detect deviations between predicted and actual operating data and adjust control parameters based on the size and trend of the deviation, bringing the generator set's operating status closer to the expected target and improving the stability and reliability of the power supply.

[0090] The performing feedback quantization calculation on the difference value sequence and the current control data set in the pre-processed data set to obtain the control prediction data set of the generator set includes:

[0091] kz i =L2(p i )log|p i |,

[0092]

[0093] Among them, kz i is the predicted data of the i-th type of control parameters, p i is the intermediate control value of the i-th type control parameter, L2 is the second-order Legendre function, C ij is the jth element of the i-th type control parameter data sequence of the current control data set in the pre-processed data set, is the mean value of the i-th type control parameter data sequence of the current control data set in the preprocessed data set, and N is the length of the control parameter data sequence.

[0094] The feedback quantization calculation uses a second-order Legendre function and a logarithmic function to process the intermediate control value to obtain predicted data for the i-th control parameter. The calculation of the intermediate control value comprehensively considers the current control data set after preprocessing and the difference value sequence, and can adaptively adjust the control parameters based on real-time changes in the data. When the difference value sequence changes, the intermediate control value also changes accordingly, allowing the control parameters to promptly respond to changes in the operating state of the generator set, achieving dynamic control of the generator set.

[0095] The second-order Legendre function has excellent approximation properties and can effectively transform intermediate control values, while the logarithmic function can perform nonlinear mapping of data, enhancing the system's sensitivity to data changes. This approach enables more precise control of the operating status of generator sets, reducing problems such as uneven power distribution and system oscillations, and improving the safety and stability of the power system. By comprehensively analyzing historical and real-time data, the system can better understand the operating patterns of generator sets, predict future operating conditions, and adjust control parameters based on these predictions, achieving refined and intelligent control of generator sets.

[0096] The control data set includes four types of control parameters: operating frequency, power distribution coefficient, excitation current and synchronous control parameters;

[0097] The working data set includes four types of working parameters: frequency stability, voltage stability, dynamic response time and mean time between failures;

[0098] The working state prediction model to be solved is solved to obtain the working state prediction model, and the Kalman filter method is used to solve Φ(k+1,k), H(k) and G(k).

[0099] The performing feature filtering on the second data set to obtain a preprocessed data set includes:

[0100] For each type of data attribute of the second data set, autoregressive-sliding average modeling is performed with the data collection information of the data as the independent variable and the data value of the data as the dependent variable to obtain a regression model for each type of data attribute;

[0101] Using the regression model, calculating and processing the independent variable to obtain a regression data value; determining whether the absolute value of the difference between the regression data value and the corresponding dependent variable value is greater than a set first regression discrimination threshold; if the absolute value is greater than the first regression discrimination threshold, deleting the data from the second data set; if the absolute value is less than or equal to the first regression discrimination threshold, not processing the data;

[0102] performing fusion processing on all feature-filtered data of the second data set to obtain a preprocessed data set;

[0103] The data collection information is collection time information.

[0104] According to a second aspect of the present invention, a parallel control device for a generator set is disclosed, the device comprising:

[0105] a memory storing executable program code;

[0106] a processor coupled to the memory;

[0107] The processor calls the executable program code stored in the memory to execute the parallel control method of the generator sets.

[0108] According to a third aspect of the present invention, a computer storable medium is disclosed. The computer storable medium stores computer instructions. When the computer instructions are called by a computer, they are used to execute the parallel control method of the generator sets.

[0109] According to a fourth aspect of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the parallel control method of the generator sets.

[0110] The foregoing is merely an embodiment of the present invention and is not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A parallel control method for generator sets, characterized in that: include: S1, obtaining a historical operating data set and a historical control data set of a generator set; S2, performing modeling processing on the historical working data set and the historical control data set to obtain a working status prediction model; S3, collecting and obtaining a current data set of the generator set; the current data set includes a current working data set and a current control data set; S4, using the working state prediction model, performing control prediction processing on the current data set to obtain a control prediction data set of the generator set; S5, inputting the control prediction data set into the generator set as the control data of the generator set to complete the parallel control of the generator set.

2. The parallel control method for generator sets according to claim 1, characterized in that: The modeling process of the historical working data set and the historical control data set to obtain a working state prediction model includes: S21, combining the historical control data set as a system measurement vector, combining the historical working data set as a target state vector, and using a linear system state equation modeling method to construct a working state prediction model to be solved; The expression of the working state prediction model is: x(k+1)=Φ(k+1,k)x(k)+G(k)w(k), z(k)=H(k)x(k)+v(k), Where Φ(k+1,k) represents the state transfer matrix at time k, H(k) represents the measurement matrix at time k, G(k) represents the input matrix at time k, w(k) and v(k) represent the state noise and measurement noise at time k, respectively, x(k) represents the target state vector at time k, z(k) represents the system measurement vector at time k, and x(k+1) represents the predicted target state vector at time k+1. S22 , using the historical working data set and the historical control data set, solving the working state prediction model to be solved to obtain the working state prediction model.

3. The parallel control method for generator sets according to claim 1, characterized in that: The method of using the working state prediction model to perform control prediction processing on the current data set to obtain a control prediction data set of the generator set includes: S41, preprocessing the current data set to obtain a preprocessed data set; S42, using the working state prediction model, processing the current control data set in the preprocessed data set to obtain a working data prediction set; S43, performing feedback calculation processing on the working data prediction set and the current working data set in the pre-processed data set to obtain a control prediction data set of the generator set.

4. The parallel control method for generator sets according to claim 3, characterized in that: The preprocessing of the current data set to obtain a preprocessed data set includes: S411, performing data cleaning processing on the current data set to obtain a first data set; S412, performing time alignment processing on the first data set to obtain a second data set; S413: Perform feature filtering on the second data set to obtain a preprocessed data set.

5. The parallel control method for generator sets according to claim 4, characterized in that: The expression of the feature filtering process is: Wherein, s(k) is the kth element of a data sequence of a type in the second data set, NS is the length of the data sequence, y1(z,x) is the element of the zth row and xth column of the matrix corresponding to the transformed data information obtained by the spatial feature filtering process, p1() is the transformation function corresponding to the spatial feature filtering process, p1(kT1-zT1) is the value of p1() at kT1-zT1, T1 and F1 are respectively the time domain transformation length and frequency domain transformation length of the spatial feature filtering process; After obtaining the matrix y1, all row vectors of the matrix are concatenated to obtain a preprocessed data sequence of the class data sequence; Using all preprocessed data sequences, a preprocessed data set is constructed.

6. The parallel control method for generator sets according to claim 1, characterized in that: The performing feedback calculation processing on the working data prediction set and the current working data set in the pre-processed data set to obtain the control prediction data set of the generator set includes: S431, performing a difference quantization calculation on the working data prediction set and the current working data set in the pre-processed data set to obtain a difference value sequence; S432 , performing feedback quantization calculation on the difference value sequence and the current control data set in the pre-processed data set to obtain a control prediction data set of the generator set.

7. The parallel control method for generator sets according to claim 6, characterized in that: The expression for the difference quantification calculation is: Among them, h i is the i-th element in the difference value sequence, is the mean value of the i-th type of working parameter data sequence of the working data prediction set, A ij is the jth element of the i-th type of working parameter data sequence of the working data prediction set, B ij is the jth element of the i-th type working parameter data sequence of the current working data set in the preprocessing data set, and N is the length of the working parameter data sequence.

8. A parallel control device for a generator set, characterized in that: The device comprises: a memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the parallel control method for generator sets according to any one of claims 1 to 7.

9. A computer storable medium, characterized in that The computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the parallel control method for generator sets according to any one of claims 1 to 7.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the parallel control method for generator sets according to any one of claims 1 to 7.

Images