De-excitation state detection method for diesel generator car based on voltage-power correlation

Through the detection method based on voltage-power correlation, the support vector machine decision function is used to solve the detection accuracy problem of diesel generators in the state of large power fluctuations and low excitation, achieving higher detection accuracy and stability, avoiding the The problem of uncontrollable output voltage caused by demagnetization.

CN113902059BActive Publication Date: 2024-11-26GUANGZHOU POWER SUPPLY BUREAU GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111265923.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-10-28
Publication Date
2024-11-26
Estimated Expiration
2041-10-28

AI Technical Summary

Technical Problem

The existing method for detecting demagnetization state of diesel generators has poor detection accuracy in low-excitation operation states, and there is a problem of malfunction.

Method used

Using a detection method based on voltage-power correlation, different working conditions are simulated through the simulation platform, sample data is collected, voltage-power correlation is calculated, and the support vector machine decision function is used for detection, to construct functional minimization problems, and iteratively solve them. The normal vector and deviation value realize accurate detection of the demagnetization state of diesel generators.

Benefits of technology

Maintain good detection accuracy under large power fluctuations and low excitation operation conditions, avoiding the uncontrollable output voltage caused by demagnetization, and improving the operating stability of diesel generators in emergency power supply scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113902059B_ABST
    Figure CN113902059B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for detecting the demagnetization state of a diesel generator car based on voltage-power correlation, which solves the technical problem of poor detection accuracy in the prior art under large power fluctuation and low excitation operation state. Collect m groups of samples under different external loads and different working conditions within the sampling time; calculate the correlation between the terminal voltage effective value and the instantaneous value of active power of each group of samples, that is, the voltage-power correlation; normalize the voltage-power correlation, and construct a combined sample #imgabs0#Construct a support vector machine decision function f(x)=ω T x+b, take the combined sample #imgabs1# as the input of the support vector machine, that is, #imgabs2#y i =S i , establish the functional minimization problem expression; iteratively solve to obtain the normal vector and deviation value; according to the effective value of the terminal voltage of the diesel generator inside the diesel generator car and the instantaneous value of the active power obtained by actual monitoring, calculate the voltage-power correlation and substitute it into the optimized support vector machine decision function to detect the demagnetization state of the diesel generator car.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of demagnetization state detection of a diesel generator vehicle. Background Art

[0002] With the advancement of science and technology, the electricity consumption of the whole society is constantly increasing, and people's requirements for the power supply guarantee capability of the power system are also constantly increasing. In order to reduce the power outage maintenance time after the power system fails and improve the power supply flexibility and reliability of the power system, it is necessary to configure an emergency power supply that can be flexibly controlled and convenient to move. As a high-quality mobile emergency power supply, the diesel generator truck can be used in distribution network fault maintenance, emergency power supply and other scenarios to improve the power supply guarantee capability of the power system.

[0003] Diesel generators are used as power sources, and their abnormal operation will seriously affect the safety and stability of their emergency power supply. In particular, when a demagnetization fault occurs, the rotor current inside the generator increases, and the excitation system switches from the voltage control mode to the constant excitation mode. At this time, the output voltage of the diesel generator cannot be controlled, which will have a serious impact on the load that requires emergency power supply. Therefore, it is necessary to accurately and efficiently detect the demagnetization state of the diesel generator inside the diesel generator.

[0004] The existing demagnetization detection method for diesel generator cars is mainly based on the impedance principle. This method works well when the power fluctuation is small and the excitation is normal, but the accuracy is poor and there is a problem of false operation when the power fluctuation is large and the excitation is low. Summary of the invention

[0005] In view of the deficiencies of the above-mentioned technologies, the present invention provides a method for detecting the demagnetization state of a diesel generator vehicle based on voltage-power correlation, which solves the technical problem of poor detection accuracy in the prior art under large power fluctuations and low excitation operating conditions.

[0006] In order to solve the above technical problems, the present invention provides a method for detecting the demagnetization state of a diesel generator vehicle based on voltage-power correlation, comprising the following steps:

[0007] The simulation platform is used to simulate the normal operation and demagnetization operation of the diesel generator car. During the sampling time, m groups of samples are collected under different external loads and different working conditions. The same group of samples is obtained by collecting the terminal voltage effective value and active power instantaneous value of the diesel generator inside the diesel generator car under the same working condition and different load levels; among them, the i-th group of samples is expressed as (V i ,P i ,S i );By adjusting the diesel generator state variable S i Different values ​​are assigned to represent the normal state and the demagnetized state;

[0008] Calculate the correlation between the terminal voltage effective value and the instantaneous value of active power of each group of samples, that is, the voltage-power correlation, according to the following formula:

[0009]

[0010] In the formula, C i represents the voltage-power correlation of the i-th group of samples, P i j represents the instantaneous value of the jth active power in the i-th group of samples, V i j represents the effective value of the j-th terminal voltage in the i-th group of samples, and n represents the number of samples in the i-th group of samples;

[0011] Normalize the voltage-power correlation and construct a combined sample ( S i ):

[0012]

[0013] Among them, C max represents the maximum value of voltage-power correlation in m groups of samples, C min represents the minimum value of voltage-power correlation in m groups of samples, represents the average value of the correlation of m groups of samples, represents the normalized value of the voltage-power correlation of the i-th group of samples;

[0014] Construct support vector machine decision function f(x) = ω T x+b, the combined sample ( S i ) is used as the input of the support vector machine, that is, y i =S i , establish the functional minimization problem expression:

[0015]

[0016] st-y i (ω T x i +b)+1≤0i=1,2,…,m

[0017] Among them, ω represents the hyperplane normal vector, b represents the deviation, m represents the number of combined samples, and x i ,y i Represent the input value and output value of the i-th combined sample respectively;

[0018] Randomly select 80% of the combined samples from all the combined samples as training samples, and the remaining 10% of the combined samples as verification samples; use the training samples to iteratively solve the normal vector and deviation value, and then use the verification samples to verify the data; compare the verification results of each training, and select the normal vector and deviation value obtained in the best effect as the final parameters of the support vector machine decision function;

[0019] According to the effective value of the terminal voltage and the instantaneous value of active power inside the diesel generator car obtained by actual monitoring, the voltage-power correlation is calculated and normalized, and then substituted into the support vector machine decision function obtained by training and optimization to calculate the value of the state variable of the diesel generator car, thereby realizing the demagnetization state detection of the diesel generator car.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] 1. Compared with a single electrical indicator, the voltage-power correlation is significantly different in the normal state and the demagnetization state of the diesel generator car, and the short-term fluctuation in a specific state is small, and no false detection will be caused by the short-term fluctuation of the indicator. Therefore, the present invention can maintain good detection accuracy under the condition of large power fluctuation and low excitation operation, thereby improving the detection accuracy of the demagnetization state of the diesel generator car, avoiding the serious consequences caused by the uncontrollable output voltage of the diesel generator car due to demagnetization, and improving the operation stability of the diesel generator car in the emergency power supply scenario. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a flow chart of a method for detecting demagnetization state of a diesel generator vehicle based on voltage-power correlation; DETAILED DESCRIPTION

[0023] refer to Figure 1 As shown, a method for detecting demagnetization state of a diesel generator vehicle based on voltage-power correlation comprises the following steps:

[0024] The simulation platform is used to simulate the normal operation and demagnetization operation of the diesel generator car. Within the sampling time (with 5 kHz as the sampling frequency and 5 cycles as the sampling time), m (m ≥ 400) groups of samples under different external loads and different working conditions are collected. The same group of samples is obtained by collecting the terminal voltage effective value and active power instantaneous value of the diesel generator inside the diesel generator car under the same working condition and different load levels; among them, the i-th group of samples is expressed as (V i ,P i ,S i );By adjusting the diesel generator state variable S i Different values ​​are assigned to represent the normal state and the demagnetized state; for example, the state variable S of a diesel generator car i =1 indicates normal state, Si =-1 indicates demagnetization state.

[0025] Calculate the correlation between the terminal voltage effective value and the instantaneous value of active power of each group of samples, that is, the voltage-power correlation, according to the following formula:

[0026]

[0027] In the formula, C i represents the voltage-power correlation of the i-th group of samples, P i j represents the instantaneous value of the jth active power in the i-th group of samples, V i j represents the effective value of the j-th terminal voltage in the i-th group of samples, and n represents the number of samples in the i-th group of samples;

[0028] Normalize the voltage-power correlation and construct a combined sample ( S i ):

[0029]

[0030] Among them, C max represents the maximum value of voltage-power correlation in m groups of samples, C min represents the minimum value of voltage-power correlation in m groups of samples, represents the average value of the correlation of m groups of samples, represents the normalized value of the voltage-power correlation of the i-th group of samples;

[0031] Construct support vector machine decision function f(x) = ω T x+b, the combined sample ( S i ) S i As the input and output of the support vector machine, respectively, y i =S i , establish the functional minimization problem expression:

[0032]

[0033] st-y i (ω T x i +b)+1≤0i=1,2,…,m

[0034] Among them, ω represents the hyperplane normal vector, b represents the deviation, m represents the number of combined samples, and x i ,y iRespectively represent the input value and output value corresponding to the i-th combined sample;

[0035] Randomly select some combination samples from all the combination samples as training samples, and the remaining combination samples as verification samples; use the training samples to iteratively solve the normal vector and deviation value, and then use the verification samples to verify the data; compare the verification results of each training, and select the normal vector and deviation value obtained in the best effect as the final parameters of the support vector machine decision function;

[0036] According to the effective value of the terminal voltage and the instantaneous value of the active power of the diesel generator inside the diesel generator car obtained by actual monitoring (according to the sampling frequency, multiple effective values ​​of the terminal voltage and the instantaneous value of the active power of the diesel generator are obtained within a unit time), the voltage-power correlation is calculated and normalized (the voltage-power correlation calculation formula and the normalization formula in the present invention are used), and then the support vector machine decision function obtained through training and optimization is substituted to calculate the value of the state variable of the diesel generator car, thereby realizing the demagnetization state detection of the diesel generator car.

[0037] For example, the parameter of the decision function of the support vector machine after training and optimization is ω b 、b p , at this time the optimized support vector machine decision function is f p (x) = ω p T x+b p , the actual normalized voltage-power correlation x = C A Substitute f p (x) = ω p T x+b p Function, the demagnetization state S can be obtained A (S A =f p (C A )), according to S A You can make a judgment

[0038] In order to facilitate the solution, the functional minimization problem is converted into a Lagrangian dual problem, and the normal vector and deviation value obtained by iteratively solving the Lagrangian dual problem are used with training samples, and then the verification samples are used for data verification. By comparing the verification results of each training, the normal vector and deviation value obtained in the best time are selected as the final parameters of the support vector machine decision function.

[0039] The functional minimization problem is transformed into a Lagrangian dual problem as follows:

[0040] Construct the Lagrangian function:

[0041]

[0042] Where Λ represents the Lagrangian parameter set, Λ={λ k ,k=1,2,…,m},λ k represents the kth Lagrangian operator.

[0043] Using the KKT condition, we get the Lagrangian dual problem of the functional minimization problem:

[0044]

[0045]

[0046] Among them, λ p , k represent the pth and kth Lagrangian operators respectively; y p ,y k Respectively represent the output values ​​corresponding to the pth and kth combination samples, x p 、x k Respectively represent the input values ​​corresponding to the p-th and k-th combined samples.

Claims

1. A method for detecting demagnetization state of a diesel generator vehicle based on voltage-power correlation, It is characterized in that The following steps are involved: The simulation platform is used to simulate the normal operation and demagnetization operation of the diesel generator car. During the sampling time, m groups of samples are collected under different external loads and different working conditions. The same group of samples is obtained by collecting the terminal voltage effective value and active power instantaneous value of the diesel generator inside the diesel generator car under the same working condition and different load levels; among them, the i-th group of samples is expressed as (V i ,P i ,S i );By adjusting the diesel generator state variable S i Different values ​​are assigned to represent the normal state and the demagnetized state; Calculate the correlation between the terminal voltage effective value and the instantaneous value of active power of each group of samples, that is, the voltage-power correlation, according to the following formula: In the formula, C i represents the voltage-power correlation of the i-th group of samples, P i j represents the instantaneous value of the jth active power in the i-th group of samples, V i j represents the effective value of the j-th terminal voltage in the i-th group of samples, and n represents the number of samples in the i-th group of samples; Normalize the voltage-power correlation and construct a combined sample Among them, C max represents the maximum value of voltage-power correlation in m groups of samples, C min represents the minimum value of voltage-power correlation in m groups of samples, Represents the average value of the correlation of m groups of samples, represents the normalized value of the voltage-power correlation of the i-th group of samples; Construct support vector machine decision function f(x) = ω T x+b, will combine the samples In S i As the input and output of the support vector machine, respectively, y i =S i , establish the functional minimization problem expression: s.t.-y i (ω T x i +b)+1≤0 i=1,2,…,m Among them, ω represents the hyperplane normal vector, b represents the deviation, m represents the number of combined samples, and x i ,y i Respectively represent the input value and output value corresponding to the i-th combined sample; Randomly select some combination samples from all the combination samples as training samples, and the remaining combination samples as verification samples; use the training samples to iteratively solve the normal vector and deviation value, and then use the verification samples to verify the data; compare the verification results of each training, and select the normal vector and deviation value obtained in the best effect as the final parameters of the support vector machine decision function; According to the effective value of the terminal voltage and the instantaneous value of active power inside the diesel generator car obtained by actual monitoring, the voltage-power correlation is calculated and normalized, and then substituted into the support vector machine decision function obtained by training and optimization to calculate the value of the state variable of the diesel generator car, thereby realizing the demagnetization state detection of the diesel generator car.

2. The method for detecting demagnetization state of a diesel generator vehicle based on voltage-power correlation according to claim 1, It is characterized in that The sampling frequency is 5 kHz and the sampling time is 5 cycles.

3. The method for detecting demagnetization state of a diesel generator vehicle based on voltage-power correlation according to claim 1, It is characterized in that m≥400。 4. The method for detecting demagnetization state of a diesel generator vehicle based on voltage-power correlation according to claim 1, It is characterized in that After converting the functional minimization problem into a Lagrangian dual problem, the normal vector and deviation value obtained by iteratively solving the Lagrangian dual problem are obtained using training samples, and then the verification samples are used for data verification. By comparing the verification results of each training, the normal vector and deviation value obtained in the best time are selected as the final parameters of the support vector machine decision function.

5. The method for detecting demagnetization state of a diesel generator vehicle based on voltage-power correlation according to claim 4, It is characterized in that The functional minimization problem is transformed into a Lagrangian dual problem as follows: Construct the Lagrangian function: Where Λ represents the Lagrangian parameter set, Λ={λ k ,k=1,2,…,m},λ k represents the kth Lagrangian operator; Using the KKT condition, we get the Lagrangian dual problem of the functional minimization problem: Among them, λ p , k represent the pth and kth Lagrangian operators respectively; y p ,y k Respectively represent the output values ​​corresponding to the pth and kth combination samples, x p 、x k Represent the input values ​​corresponding to the p-th and k-th combined samples respectively.

6. The method for detecting demagnetization state of a diesel generator vehicle based on voltage-power correlation according to claim 1, It is characterized in that Diesel generator car state variable S i =1 indicates normal state, S i =-1 indicates demagnetization state.