Generator outlet circuit breaker temperature rise simulation method under model order adaptive strategy

Abstract

The invention discloses a generator outlet circuit breaker temperature rise simulation method under a model order self-adaptive strategy. A three-dimensional model for finite element simulation of a generator outlet circuit breaker is established; performing temperature field finite element simulation based on the three-dimensional model to obtain temperature data of the generator outlet circuit breaker under different working conditions; sampling the temperature data to establish a sample point data matrix; performing order reduction processing on the sample point data matrix to generate an order-reduced data matrix; using the reduced-order data matrix to train the proxy model to generate a reduced-order model; when an error between a temperature simulation value output by using the reduced-order model and a temperature measured value meets a constraint, simulating temperature rise of the generator outlet circuit breaker by using the reduced-order model, and if the error does not meet the constraint, performing order expansion on the reduced-order model; and when the error meets the constraint, simulating the temperature rise of the generator outlet circuit breaker by using the order-expanded reduced-order model. The dimension of the reduced-order model is automatically determined and dynamically adjusted, and the real-time performance and the accuracy of the GCB temperature rise simulation reduced-order model are effectively considered.

Description

Technical Field

[0001] This invention relates to the field of equipment parameter simulation technology, and in particular to a method for simulating the temperature rise of a generator outlet circuit breaker under a model order adaptive strategy. Background Technology

[0002] Generator outlet circuit breakers (GCBs) possess superior characteristics such as rapid interruption of large currents, extremely high reliability, and long-term maintenance-free operation. GCBs are crucial for the rapid disconnection of power generation units when they are connected to the grid or when equipment in the generator set fails; their importance to the safe operation of power generation equipment is self-evident.

[0003] With the continuous development and upgrading of power systems, circuit breaker (GCB) plays a crucial role in maintaining grid stability and security. However, with increasing service life, factors such as wear and tear, and loosening of preload can lead to poor contact, causing localized overheating. Contact resistance increases with temperature, further exacerbating contact deterioration. In recent years, overheating faults in circuit breaker contacts have occurred frequently and are on the rise, sometimes resulting in burnt-out joints and short-circuit accidents caused by discharge to the casing or adjacent conductors. Therefore, researching rapid simulation of the GCB temperature field to achieve online monitoring of its temperature field distribution is of great significance in addressing overheating faults and ensuring the safe and stable operation of GCB equipment.

[0004] In existing technologies, the main approach is to use the offline-online separation POD (Proper Orthogonal Decomposition) + ROM (Reduced Order Model) method. That is, the offline stage generates a reduced order model based on the POD principle and then uses it for online monitoring. The reduction dimension usually relies on a single criterion such as manual experience or fixed energy ratio setting, which has problems such as difficulty in quantifying error, difficulty in balancing real-time performance and accuracy, and lack of online order expansion mechanism. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a generator outlet circuit breaker temperature rise simulation method under a model order adaptive strategy. This method automatically determines and dynamically adjusts the dimension of the reduced-order model, effectively balancing the real-time performance and accuracy of the reduced-order model in GCB temperature rise simulation.

[0006] The technical solution adopted in this invention is as follows.

[0007] This invention proposes a simulation method for temperature rise of generator outlet circuit breaker under an adaptive model order strategy, including:

[0008] A three-dimensional model of the generator outlet circuit breaker for finite element simulation is established; a temperature field finite element simulation is performed based on the three-dimensional model to obtain temperature data of the generator outlet circuit breaker under different operating conditions; the temperature data is sampled to establish a sample point data matrix; the sample point data matrix is ​​reduced in order to generate a reduced-order data matrix.

[0009] A surrogate model is trained using a reduced-order data matrix to generate a reduced-order model. If the error between the simulated temperature value and the measured temperature value output by the reduced-order model meets the constraints, the temperature rise of the generator outlet circuit breaker is simulated using the reduced-order model. If the error does not meet the constraints, the reduced-order model is expanded until the error meets the constraints. Then, the expanded-order reduced-order model is used to simulate the temperature rise of the generator outlet circuit breaker.

[0010] Collect the actual structural dimensions of the generator outlet circuit breaker and construct a three-dimensional model for finite element simulation of the generator outlet circuit breaker; set material parameters and boundary conditions for the three-dimensional model according to the working conditions.

[0011] The sample point data matrix is ​​reduced in order to generate a reduced-order data matrix, including:

[0012] The original singular value matrix is ​​obtained by performing eigenorthogonal decomposition on the sample point data matrix.

[0013] Based on the original singular value matrix, the cutoff order r is determined when the energy accumulation ratio α is higher than the first threshold. * ;

[0014] The order is greater than the truncation order r * The ratios of each singular value to the first-order singular value are used as the relative singular value ratios for each order. These relative singular value ratios are then compared with a second threshold in ascending order to determine the order r of the first relative singular value ratio smaller than the second threshold. ** ;

[0015] Truncate r from the original singular value matrix ** The singular value matrix after the order is used to generate a reduced-order data matrix.

[0016] The formula for calculating the energy accumulation ratio is:

[0017]

[0018] In the formula, α is the energy accumulation ratio, r is the total order, and σ is the total energy accumulation ratio. i Let be the i-th singular value.

[0019] Using a reduced-order data matrix to train a surrogate model to generate a reduced-order model includes:

[0020] Finite element simulation is performed on the reduced-order model obtained from the trained surrogate model. When the error between the simulation data and the data obtained from the finite element simulation of the 3D model meets the accuracy requirements, the accuracy verification of the reduced-order model is completed.

[0021] If the error between the simulated temperature value and the measured temperature value output by the reduced-order model meets the constraints, then the temperature rise of the generator outlet circuit breaker is simulated using the reduced-order model. If the error does not meet the constraints, the reduced-order model is extended until the error meets the constraints. Then, the temperature rise of the generator outlet circuit breaker is simulated using the extended-order reduced-order model, including:

[0022] When the error between the temperature simulation value and the actual temperature value output by the reduced-order model after accuracy verification does not meet the constraints, the reduced-order model is extended online based on the augmented SVD principle.

[0023] After each additional mode, it is determined whether the error between the temperature simulation value and the measured temperature value output by the reduced-order model after the expansion meets the constraint. This process is repeated until the error meets the constraint.

[0024] The absolute value of the difference between the temperature simulation value output by the reduced-order model and the actual temperature value is used as the error, and the error threshold is set at 2℃; the error is considered to meet the constraint when the error is less than or equal to the error threshold.

[0025] The beneficial effects of this invention are as follows: Compared with the prior art, this invention establishes an adaptive model order strategy, and jointly considers the energy accumulation ratio of the singular value matrix, the relative singular value ratio of each order, and the temperature error threshold as order criteria. This allows for the selection of the truncation order of the singular value matrix generated by the model order reduction based on intrinsic orthogonal decomposition. The resulting reduced-order model has the characteristic of fast computation, while solving the problem of difficulty in quantifying errors caused by previous truncation based on manual experience or other single criteria. This balances the real-time performance and accuracy of the reduced-order model simulation.

[0026] This invention proposes an online order expansion mechanism based on the augmented SVD principle to solve the problem that when the order of the reduced model changes due to the change of the singular value matrix, the SVD operation needs to be repeated to update the reduced model, thus affecting the real-time performance of the reduced model simulation. Attached Figure Description

[0027] Figure 1 This is a flowchart of a simulation method for temperature rise of generator outlet circuit breaker under a model order adaptive strategy proposed in this invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0029] This invention proposes a simulation method for temperature rise of generator outlet circuit breaker under an adaptive model order strategy, such as... Figure 1 As shown, it includes:

[0030] Step 1: Establish a 3D model of the generator outlet circuit breaker for finite element simulation; set material parameters and boundary conditions for the 3D model according to the operating conditions.

[0031] Specifically, step 1 includes:

[0032] Step 1.1: Collect the actual structural dimension parameters of the generator outlet circuit breaker and construct a three-dimensional model for finite element simulation of the generator outlet circuit breaker;

[0033] Step 1.2: Set material parameters and boundary conditions for the 3D model according to the working conditions;

[0034] Specifically, step 1.2 includes:

[0035] Step 1.2.1: Mesh the 3D model and optimize the mesh quality. For parts with strong geometric nonlinear structures that result in poor mesh quality, re-mesh the mesh to improve the mesh quality.

[0036] Step 1.2.2: Based on the physical constraints required for finite element simulation, set material parameters and boundary conditions for the 3D model after mesh quality optimization;

[0037] Step 1.2.3: Set the temperature correction characteristics for the added material parameters to reflect the feedback effect of temperature rise on conductor conductivity under actual conditions.

[0038] Step 2: Perform finite element simulation of the temperature field based on the three-dimensional model to obtain temperature data of the generator outlet circuit breaker under different operating conditions; sample the temperature data to establish a sample point data matrix.

[0039] Based on the set material parameters and boundary conditions, a temperature field finite element simulation is performed on the full-order model, and a data space set is formed under multiple working conditions with temperature as the output parameter. Based on the data space set, considering the ambient temperature and hot spot temperature range when the equipment is actually running, the data space set is subjected to Latin hypercube sampling. The density of sampling points is set according to the strength of geometric nonlinear changes in different regions of GCB to generate a sample point data matrix.

[0040] Step 3: Reduce the order of the sample point data matrix to generate a reduced-order data matrix;

[0041] Specifically, step 3 includes:

[0042] Step 3.1: The sample point data matrix is ​​decomposed into the POD basis vector matrix and the original singular value matrix using the eigenorthogonal decomposition method;

[0043] Step 3.2: Based on the original singular value matrix, determine the cutoff order r when the energy accumulation ratio α is higher than the first threshold. * ;

[0044] In this embodiment, the first threshold is set to 0.99, which means that the total modal energy of the reduced-order model accounts for more than 99% at this time. The first threshold is appropriately changed according to the specific working conditions.

[0045] The formula for calculating the energy accumulation ratio is:

[0046]

[0047] In the formula, α is the energy accumulation ratio, r is the total order, and σ is the total energy accumulation ratio. i Let be the i-th singular value.

[0048] Step 3.3, the order is greater than the truncation order r. * The ratios of each singular value to the first-order singular value are used as the relative singular value ratios for each order. These relative singular value ratios are then compared with a second threshold in ascending order to determine the order r of the first relative singular value ratio smaller than the second threshold. ** ;

[0049] In the embodiment, the rth * The ratio of the +1st order singular value to the 1st order singular value is used as the rth order singular value. * The relative singularity ratio β of order +1 is truncated if β is less than the second threshold. * For singular value matrices of order +1, if β is not less than the second threshold, then r * The relative singularity ratio of the +2nd order is compared with the second threshold, and so on until it is less than the second threshold. In the embodiment, the second threshold is 0.001, indicating that at this time the rth order... * The energy of the +1st mode has decayed to 0.1% of the energy of the first mode. The second threshold is adjusted appropriately according to the specific operating conditions.

[0050] The formula for calculating the ratio β is:

[0051]

[0052] In the formula, β is the relative singular value ratio, and σ1 is the first-order singular value. For the rth * +1 order singularity.

[0053] Step 3.4, truncate r from the original singular value matrix ** The singular value matrix after the first order is used to generate a reduced-order data matrix for the lower-order modes.

[0054] Step 4: Use the reduced-order data matrix to train the surrogate model to generate a reduced-order model; if the error between the temperature simulation value and the measured temperature value output by the reduced-order model meets the constraints, then use the reduced-order model to simulate the temperature rise of the generator outlet circuit breaker. If the error does not meet the constraints, then extend the order of the reduced-order model until the error meets the constraints, then use the extended-order reduced-order model to simulate the temperature rise of the generator outlet circuit breaker.

[0055] Specifically, step 4 includes:

[0056] Step 4.1: Perform finite element simulation on the reduced-order model obtained from the trained surrogate model. When the error between the simulation data and the data obtained from the finite element simulation of the 3D model meets the accuracy requirements, the accuracy verification of the reduced-order model is completed.

[0057] Step 4.2: When the error between the temperature simulation value and the measured temperature value output by the reduced-order model after accuracy verification does not meet the constraints, the reduced-order model is expanded online based on the augmented SVD principle. For each additional mode, the expanded-order reduced-order model is obtained. Step 4.2 is repeated to determine whether the error between the temperature simulation value and the measured temperature value meets the constraints. This process is repeated until the error meets the constraints.

[0058] Based on the validated reduced-order model, using current as the input condition, temperature simulation results are obtained according to the input condition. These results are then compared with the measured temperature to verify the error constraint. The error is considered to satisfy the constraint if the error is less than or equal to an error threshold, where the error threshold ε is the criterion. max Take 2℃ as the reference temperature, and adjust it appropriately according to actual working conditions and requirements. The specific formula for calculating the error is as follows:

[0059] ε=|TT r |

[0060] Where ε is the error, and T is the simulated temperature value output by the reduced-order model. r This is the actual measured temperature value;

[0061] The specific principle of augmented SVD is as follows:

[0062] X old =UΣV T

[0063] In the formula, X old The original reduced-order data matrix is ​​an n×n matrix, U, VT is the matrix of singular value basis vectors generated by SVD decomposition and is an n×r matrix respectively * and r * ×n matrix, Σ is the singular value matrix and is an r * ×r * matrix;

[0064] It is necessary to use the known U, Σ, V T to quickly find X new , U new , ∑ new , V T new , so the new snapshot P can be projected onto the old basis U r×p = U T A, P is the projection coefficient matrix, which is used to measure the components of the new snapshot on the old subspace, and A is the newly appended snapshot matrix;

[0065] Calculate the residual matrix R n×p = A - UP, perform column orthogonalization QR decomposition on R to obtain R = QR2, where if the column rank of R is q ≤ p, only q orthogonal vectors are generated to avoid zero vectors;

[0066] Construct a small block matrix where K is a matrix of (r + q)×(r + p), perform SVD decomposition on K to obtain At this time, because (r + q) << n, the time spent on augmented SVD is much less than the time required to directly perform SVD decomposition on the original sample data matrix again to re-truncate and change its mode;

[0067] At this time where I p is the p-order identity matrix.

[0068] Based on the principle of augmented SVD, the online order expansion of the reduced-order model requires less computational effort than re-performing SVD decomposition for model reduction.

[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific embodiments of the present invention, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A simulation method for temperature rise of generator outlet circuit breaker under an adaptive model order strategy, characterized in that, include: Establish a three-dimensional model of the generator outlet circuit breaker for finite element simulation; Temperature field finite element simulation was performed based on a three-dimensional model to obtain temperature data of the generator outlet circuit breaker under different operating conditions. Temperature data is sampled to create a sample point data matrix; The sample point data matrix is ​​reduced in order to generate a reduced-order data matrix. A reduced-order model is generated by training a surrogate model using a reduced-order data matrix. If the error between the simulated temperature value and the measured temperature value output by the reduced-order model meets the constraints, the temperature rise of the generator outlet circuit breaker is simulated using the reduced-order model. If the error does not meet the constraints, the reduced-order model is extended until the error meets the constraints. Then, the temperature rise of the generator outlet circuit breaker is simulated using the extended-order reduced-order model.

2. The method for simulating the temperature rise of the generator outlet circuit breaker under the adaptive model order strategy according to claim 1, characterized in that, Collect the actual structural dimensions of the generator outlet circuit breaker and construct a three-dimensional model for finite element simulation of the generator outlet circuit breaker; set material parameters and boundary conditions for the three-dimensional model according to the working conditions.

3. The generator outlet circuit breaker temperature rise simulation method under the model order adaptive strategy according to claim 1, characterized in that, The sample point data matrix is ​​reduced in order to generate a reduced-order data matrix, including: The original singular value matrix is ​​obtained by performing eigenorthogonal decomposition on the sample point data matrix. Based on the original singular value matrix, the cutoff order r is determined when the energy accumulation ratio α is higher than the first threshold. * ; The order is greater than the truncation order r * The ratios of each singular value to the first-order singular value are used as the relative singular value ratios for each order. These relative singular value ratios are then compared with a second threshold in ascending order to determine the order r of the first relative singular value ratio smaller than the second threshold. ** ; Truncate r from the original singular value matrix ** The singular value matrix after the order is used to generate a reduced-order data matrix.

4. The generator outlet circuit breaker temperature rise simulation method under the model order adaptive strategy according to claim 3, characterized in that, The formula for calculating the energy accumulation ratio is: In the formula, α is the energy accumulation ratio, r is the total order, and σ is the total energy accumulation ratio. i Let be the i-th singular value.

5. The generator outlet circuit breaker temperature rise simulation method under the model order adaptive strategy according to claim 1, characterized in that, Using a reduced-order data matrix to train a surrogate model to generate a reduced-order model includes: Finite element simulation is performed on the reduced-order model obtained from the trained surrogate model. When the error between the simulation data and the data obtained from the finite element simulation of the 3D model meets the accuracy requirements, the accuracy verification of the reduced-order model is completed.

6. The generator outlet circuit breaker temperature rise simulation method under the model order adaptive strategy according to claim 5, characterized in that, If the error between the simulated temperature value and the measured temperature value output by the reduced-order model meets the constraints, then the temperature rise of the generator outlet circuit breaker is simulated using the reduced-order model. If the error does not meet the constraints, the reduced-order model is extended until the error meets the constraints. Then, the temperature rise of the generator outlet circuit breaker is simulated using the extended-order reduced-order model, including: When the error between the temperature simulation value and the actual temperature value output by the reduced-order model after accuracy verification does not meet the constraints, the reduced-order model is extended online based on the augmented SVD principle. After each additional mode, it is determined whether the error between the temperature simulation value and the measured temperature value output by the reduced-order model after the expansion meets the constraint. This process is repeated until the error meets the constraint.

7. The generator outlet circuit breaker temperature rise simulation method under the model order adaptive strategy according to claim 6, characterized in that, The absolute value of the difference between the temperature simulation value output by the reduced-order model and the actual temperature value is used as the error, and the error threshold is set at 2℃; the error is considered to meet the constraint when the error is less than or equal to the error threshold.

Images