110KV voltage regulation type high-capacity dynamic reactive power compensation method for transformer substation

Abstract

The invention discloses a 110KV voltage regulation type high-capacity dynamic reactive compensation method for a transformer substation, and relates to the technical field of dynamic reactive compensation. Compared with previous grouping fixed switching, the dynamic reactive compensation device provided by the invention has the advantages that the capacity of hierarchical switching is small, the maximum of each gear is 4664Kvar, the inrush current and the voltage rise are small, and the dynamic reactive compensation effect is good. The impact on a power grid caused by inrush current and overvoltage due to high-capacity capacitor switching is effectively reduced; and no operation overvoltage and switching inrush current exist. Due to the fact that the capacitors are fixedly connected and are not switched in a grouping mode, the output capacity of the capacitors can be finely adjusted according to system requirements, no overvoltage exists in the adjusting process, the switching-on and switching-off operation processes are stable, and the problems of overvoltage, switching inrush current and the like caused by switching of the capacitors are solved. A proper input voltage can be selected, the closing inrush current of the input capacitor is effectively reduced, the impact on a power grid and the capacitor is reduced, and the device does not generate harmonic waves and does not amplify the harmonic waves.

Description

Technical Field

[0001] The present invention belongs to the technical field of dynamic reactive power compensation, and specifically relates to a 110KV voltage regulating type large-capacity dynamic reactive power compensation method for a substation. Background Art

[0002] In a 220KV substation with a 110KV calcium carbide furnace as the main power supply load, usually, low-voltage short network compensation for the calcium carbide furnace at 110 - 200V is configured to ensure that the active power factor of the power grid is above 0.95, and only a small-capacity 35KV is used in the substation to compensate for the reactive power of the dynamic load.

[0003] However, there is generally a problem with low-voltage short network compensation, that is, when the load of the calcium carbide furnace is low or the load is below 50%, the short network compensation cannot be put into operation, or overvoltage caused by the input of short network compensation leads to system protection tripping.

[0004] Under this background, a new type of 110KV voltage regulating type large-capacity 85Mvar dynamic reactive power compensation is adopted in the 220KV substation to make up for the deficiency of short network reactive power compensation.

[0005] When the load is below 50%, the 110KV dynamic voltage regulating reactive power compensation device in the substation is responsible for reactive power compensation. When the load of the calcium carbide furnace is above 50%, the substation reactive power compensation is gradually withdrawn, and the short network compensation is put into operation. By cooperating with high and low voltage compensation, the active power factor of the power grid is ensured to reach above 0.95.

[0006] At the same time, it is also considered that in an extreme environment, that is, when all the short network compensations of the calcium carbide furnace cannot be put into operation, the 110KV voltage regulating type large-capacity dynamic reactive power compensation of the 220KV substation is used to overall meet the reactive power compensation requirements of the calcium carbide furnace. After calculation, the installed capacity of the 110KV voltage regulating type large-capacity dynamic reactive power compensation is 85Mvar. Summary of the Invention

[0007] The present invention aims to solve at least one of the technical problems existing in the prior art; for this purpose, the present invention proposes a 110KV voltage regulating type large-capacity dynamic reactive power compensation method for a substation, which specifically includes the following steps:

[0008] Obtain the load of the calcium carbide furnace, and make different treatments according to the load situation. The specific method is as follows:

[0009] When the load is below 50%, the 110KV dynamic voltage regulating reactive power compensation device in the substation is responsible for reactive power compensation; when the load of the calcium carbide furnace is above 50%, the substation reactive power compensation is withdrawn;

[0010] The dynamic reactive power compensation device includes a disconnector QS. One side of the disconnector QS is connected to a voltage regulator T. One side of the voltage regulator is connected to a reactor L and a lightning arrester FV. The lightning arrester FV is successively connected in series with a capacitor C, a bridge differential current transformer ΔI, and an earthing switch QE2.

[0011] Furthermore, an earthing switch QE1 is also arranged between the disconnector QS and the voltage regulator T.

[0012] Furthermore, the load of the calcium carbide furnace is obtained by real-time acquisition to determine the real-time load.

[0013] Furthermore, the load of the calcium carbide furnace is predicted through an intelligent model. If it is predicted that the load of the calcium carbide furnace will drop below 50%, the dynamic reactive power compensation device is started to adjust the reactive power output capacity of the calcium carbide furnace. At the same time, the real-time load of the calcium carbide furnace is monitored in real time. If no mutation behavior occurs in the calcium carbide furnace within the set time period T2, the dynamic reactive power compensation device is stopped from adjusting the reactive power output capacity of the calcium carbide furnace.

[0014] Furthermore, the specific method for predicting the load of the calcium carbide furnace through the intelligent model is as follows:

[0015] The historical load data is the relevant parameters of several calcium carbide furnaces and their maximum change values during the previous T1 time period when mutation behavior occurs. Here, T1 is a preset value. The relevant parameters include furnace temperature, furnace pressure, voltage, current, power factor, and start / stop furnace frequency.

[0016] The collected historical load data is divided into a training set and a validation set. The training set is used as input to train the intelligent model with the help of a long short-term memory network LSTM. After training, the validation set is used to verify its accuracy rate. When the accuracy rate exceeds the set ratio R1, it indicates that the intelligent model is available.

[0017] The intelligent model is used to predict the load data of the calcium carbide furnace. When it is predicted that the load of the calcium carbide furnace will drop below 50%, the dynamic reactive power compensation device is started to adjust the reactive power output capacity of the calcium carbide furnace. At the same time, the real-time load of the calcium carbide furnace is monitored in real time. If no mutation behavior occurs in the calcium carbide furnace within the set time period T2, the dynamic reactive power compensation device is stopped from adjusting the reactive power output capacity of the calcium carbide furnace.

[0018] Furthermore, the specific method for predicting the load of the calcium carbide furnace through the intelligent model is as follows:

[0019] ST1: Mark all relevant parameters of the calcium carbide furnace as target parameters.

[0020] ST2: Mark the behavior of reducing the load of the calcium carbide furnace to less than 50% each time as a mutation behavior, and obtain all the target parameters that have changed during the T1 time period before the mutation behavior; here, T1 is a preset value; here, the numerical change of the target parameter is at least greater than X1, and X1 is a preset value;

[0021] ST3: Then obtain the occurrence ratio of all target parameters in each mutation behavior.

[0022] ST4: Obtain the occurrence ratios of all target parameters, perform screening according to the occurrence ratios, determine the screening line for the occurrence ratios based on the standard deviation of the occurrence ratios and the distribution of the values in the occurrence ratios, and mark the target parameters with occurrence ratios exceeding the screening line as object parameters;

[0023] ST5: Perform numerical constancy analysis on the object parameters, determine the mutation value according to the maximum change value of the numerical change of the object parameter during the T1 time period before the mutation behavior appears, and determine the change line according to the distribution of the mutation values; obtain the change line of each object parameter;

[0024] ST6: Monitor the object parameters of the calcium carbide furnace in real time. When any object parameter changes and the change value exceeds the change line, the T1 time period will be automatically monitored at this time. If there are object parameters with change values exceeding the corresponding change line and the number of object parameters exceeds the set ratio R2 within this time period, a pre-start signal will be automatically generated;

[0025] ST7: When the pre-start signal is generated, start the dynamic reactive power compensation device to adjust the reactive power output capacity of the calcium carbide furnace, and at the same time monitor the real-time load of the calcium carbide furnace in real time. If no mutation behavior occurs in the calcium carbide furnace within the set time period T2, stop the dynamic reactive power compensation device from adjusting the reactive power output capacity of the calcium carbide furnace;

[0026] Furthermore, the specific method for screening object parameters in ST4 is as follows:

[0027] First, obtain the occurrence ratios of the target parameters, automatically obtain the mean value of all occurrence ratios, and then calculate the standard deviation of all occurrence ratios using a formula. When the standard deviation is lower than the preset value X2, mark the mean value of the occurrence ratios as the screening line at this time;

[0028] If the standard deviation is not lower than X2, automatically obtain the number of target parameters with occurrence ratios greater than the mean value, mark it as the upper number, and mark the number of target parameters with occurrence ratios less than the mean value as the lower number;

[0029] When the upper number exceeds the lower number, automatically mark the value obtained by multiplying the mean value by 1.15 as the screening line at this time, otherwise mark the mean value as the screening line;

[0030] Mark the target parameters with occurrence ratios exceeding the screening line as object parameters.

[0031] Further, the specific method for analyzing the constant value in ST5 is as follows:

[0032] Optionally select an object parameter, obtain the maximum change value of the corresponding numerical change for each T1 duration before the occurrence of the mutation behavior, mark it as the mutation value, and obtain several mutation values Bj, where j = 1,..., n;

[0033] Then automatically obtain the average value P of Bj, and calculate the stable value W of Bj using the formula. The specific calculation formula is:

[0034]

[0035] When the value of W does not exceed the preset value X3, mark the minimum value in Bj at this time as the change line of the corresponding object parameter;

[0036] When the value of W exceeds the preset value X3, sort Bj in descending order according to |Bj - P|, then separately select the corresponding Bj where Bj - P is less than zero, sort them according to the original data, select one Bj in turn, delete it each time a Bj is selected, and then recalculate the value of W until the value of W does not exceed the preset value X3. At this time, mark the minimum value in the remaining Bj as the change line of the object parameter; if the value of W still exceeds X3 after deleting all the selected data, automatically mark the average value as the change line.

[0037] Compared with the prior art, the beneficial effects of the present invention are:

[0038] Through the dynamic reactive power compensation device provided by the present application, compared with the previous grouped fixed switching, the capacity of the hierarchical switching is small, up to 4664 Kvar per stage, the inrush current and voltage rise are small, effectively reducing the impact on the power grid caused by the inrush current and overvoltage due to the switching of large-capacity capacitors;

[0039] There is no switching overvoltage and closing inrush current. Since the capacitors are fixedly connected and not switched in groups, its output capacity can be finely adjusted according to the system needs. There is no overvoltage during the adjustment process, and the closing and withdrawal operations are stable, solving the problems of overvoltage and closing inrush current caused by capacitor switching; and a suitable input voltage can be selected to effectively reduce the closing inrush current of the input capacitors and reduce the impact on the power grid and the capacitors themselves. This device does not generate harmonics and there is no harmonic amplification;

[0040] At the same time, the corresponding dynamic reactive power compensation device can be started at the appropriate time through the solution provided by the present application, with predictability; the present invention is simple and effective and easy to use. Description of the Drawings

[0041] Figure 1This is the circuit schematic diagram of the dynamic reactive power compensation device of the present invention. Specific embodiments

[0042] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0043] Please refer to Figure 1 ; This application provides a 110KV voltage regulation type large-capacity dynamic reactive power compensation method for a substation;

[0044] As the first embodiment of this application, the method specifically includes the following steps:

[0045] Obtain the load of the calcium carbide furnace and make different treatments according to the load conditions. The specific method is as follows:

[0046] When the load is lower than 50%, the 110KV dynamic voltage regulation reactive power compensation device of the substation is responsible for reactive power compensation; when the load of the calcium carbide furnace is higher than 50%, the short network compensation is gradually put into operation and the substation reactive power compensation is withdrawn;

[0047] During this process, an intelligent model is used to predict the load of the calcium carbide furnace. The intelligent model here is the long short-term memory network LSTM, which is a special type of recurrent neural network specifically used to process and predict time series data; the specific method is as follows:

[0048] Collect and sort out the historical load data of the calcium carbide furnace. The historical load data here is the maximum change value of the relevant parameters of several calcium carbide furnaces within the first T1 time period before each mutation behavior, that is, the real-time values of the relevant parameters under different historical load conditions. The relevant parameters include electrical parameters, furnace temperature, and furnace pressure. The electrical parameters are voltage, current, power factor, etc. Of course, the relevant parameters here can also include the start-stop furnace frequency. How much the maximum change of these parameters is within the first T1 time period before the mutation behavior; here is to determine whether the load will drop below 50% according to the change of the relevant parameters.

[0049] Ensure the quality and integrity of the data; preprocess the data, including cleaning, denoising, normalization, etc., to make it suitable as the input of the LSTM model; here, the collected historical load data is divided into a training set and a validation set according to a preset ratio. Generally, the ratio of the training set to the validation set is 8:2;

[0050] After that, an LSTM model is constructed using a deep learning framework, which can be any one of TensorFlow, Keras, and PyTorch; the structure of the model includes an input layer, multiple LSTM layers, a fully connected layer, and an output layer; the number of LSTM layers is set by the administrator according to needs;

[0051] The prepared dataset is input into the model for training; a suitable loss function and optimizer are selected. In this embodiment, the loss function can be mean squared error (MSE), and the Adam optimizer is used; and appropriate epochs and batch sizes are set; during the training process, monitor the loss value and other performance metrics of the model to ensure that the model is learning correctly;

[0052] The trained intelligent model is verified using a validation set. When the prediction accuracy rate exceeds the set ratio R1, it indicates that the intelligent model is available; otherwise, historical load data is retrieved again for training until the intelligent model is available;

[0053] With the help of the intelligent model, the load data of the calcium carbide furnace is predicted. When it is predicted that the load of the calcium carbide furnace is about to decrease to less than 50%, the dynamic reactive power compensation device is started to adjust the reactive power output capacity of the calcium carbide furnace. At the same time, the real-time load of the calcium carbide furnace is monitored in real time. If no mutation behavior occurs in the calcium carbide furnace within the set time T2, the dynamic reactive power compensation device is stopped from adjusting the reactive power output capacity of the calcium carbide furnace;

[0054] The specific adjustment method is to adjust the voltage across the capacitor according to Q = 2πfCU2 to change the output capacity, thereby changing the reactive power output capacity to adjust the system power factor;

[0055] The dynamic reactive power compensation device here is as Figure 1 shown, and specifically includes a disconnecting switch QS. One side of the disconnecting switch QS is connected to a voltage regulator T. One side of the voltage regulator is connected to a reactor L and a lightning arrester FV. The lightning arrester FV is successively connected in series with a capacitor C, a bridge differential current transformer ΔI, and an earthing switch QE2;

[0056] An earthing switch QE1 is also provided between the disconnecting switch QS and the voltage regulator T;

[0057] Certainly, as Embodiment 2 of the present application, this embodiment is implemented on the basis of Embodiment 1. The difference from Embodiment 1 is that in this embodiment, the method for predicting the load of the calcium carbide furnace is different. The method provided in this embodiment is as follows:

[0058] ST1: First, mark all relevant parameters of the calcium carbide furnace as target parameters;

[0059] ST2: Mark the behavior of reducing the load of the calcium carbide furnace to less than 50% each time as a mutation behavior, and obtain the target parameters that have changed during the T1 time period before the mutation behavior; here, T1 is a preset value; here, the numerical change of the target parameter is at least greater than X1, and X1 is a preset value.

[0060] ST3: Then obtain all the target parameters and the occurrence ratio in each mutation behavior, that is, the number of times the target parameter changes when any mutation behavior occurs, and divide this number by the number of mutation behaviors generated, and the result is marked as the occurrence ratio.

[0061] ST4: Obtain the occurrence ratios of all the target parameters and perform screening according to the occurrence ratios. The specific screening method is as follows:

[0062] First, obtain the occurrence ratio of the target parameter, automatically obtain the average value of all the occurrence ratios, and then calculate the standard deviation of all the occurrence ratios using a formula. When the standard deviation is lower than the preset value X2, mark the average value of the occurrence ratio as the screening line at this time.

[0063] If the standard deviation is not lower than X2, then automatically obtain the number of target parameters with occurrence ratios greater than the average value, mark it as the upper number, and mark the number of target parameters with occurrence ratios less than the average value as the lower number.

[0064] When the upper number exceeds the lower number, automatically mark the value obtained by multiplying the average value by 1.15 as the screening line at this time, otherwise mark the average value as the screening line.

[0065] Mark the target parameters with occurrence ratios exceeding the screening line as object parameters.

[0066] ST5: Perform numerical constancy analysis on the object parameters. The specific method of numerical constancy analysis is as follows:

[0067] Select any object parameter, obtain the maximum change value of its corresponding numerical change during the T1 time period before each mutation behavior occurs, and mark it as the mutation value, and obtain several mutation values Bj, where j = 1,..., n.

[0068] Then automatically obtain the average value P of Bj and calculate the stability value W of Bj using a formula. The specific calculation formula is:

[0069]

[0070] When the value of W does not exceed the preset value X3, mark the minimum value in Bj at this time as the change line of the corresponding object parameter.

[0071] When the value of W exceeds the preset value X3, sort Bj in descending order according to |Bj - P|, then separately screen out the corresponding Bj with Bj - P less than zero, sort them according to the original data, select one Bj in turn, delete it each time a Bj is selected, and then recalculate the value of W until the value of W does not exceed the preset value X3. At this time, mark the minimum value among the remaining Bj as the change line of the object parameter; if the value of W still exceeds X3 after deleting all the screened data, automatically mark the average value as the change line;

[0072] Obtain the change line of each object parameter;

[0073] ST6: Real-time monitor the object parameters of the calcium carbide furnace. When any object parameter changes and the change value exceeds the change line, automatically monitor the duration T1 at this time. If there are object parameters whose change values exceed the corresponding change lines and the proportion exceeds the set ratio R2 within this duration, automatically generate a pre-start signal;

[0074] ST7: When a pre-start signal is generated, start the dynamic reactive power compensation device to adjust the reactive power output capacity of the calcium carbide furnace, and at the same time, monitor the real-time load of the calcium carbide furnace in real time. If there is no mutation behavior in the calcium carbide furnace within the set duration T2, stop the dynamic reactive power compensation device from adjusting the reactive power output capacity of the calcium carbide furnace;

[0075] During the above process, the real-time load of the calcium carbide furnace will be monitored synchronously in real time. If a mutation behavior occurs, immediately start the dynamic reactive power compensation device to adjust the reactive power output capacity of the calcium carbide furnace.

[0076] With the help of the dynamic reactive power compensation device, a breakthrough has been made in the adjustment means. Instead of using the method of capacitor switching, according to Q = 2πfCU2, the output capacity is changed by adjusting the voltage across the capacitor, so as to change the reactive power output capacity to adjust the system power factor. At present, the DYWT device can achieve at least nine-level output of the capacity from (100 - 25)%, and meet the substation's demand for reactive power compensation through the hierarchical switching of the voltage regulator.

[0077] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A 110KV voltage-regulating type large-capacity dynamic reactive power compensation method for a substation, characterized in that, The method specifically includes the following steps: Obtain the load of the calcium carbide furnace and make different treatments according to the load conditions. The specific method is as follows: When the load is lower than 50%, the 110KV dynamic voltage regulating reactive power compensation device of the substation is responsible for reactive power compensation; when the load of the calcium carbide furnace is higher than 50%, the reactive power compensation of the substation is withdrawn; The dynamic reactive power compensation device includes a disconnector QS. One side of the disconnector QS is connected to a voltage regulator T. One side of the voltage regulator is connected to a reactor L and a lightning arrester FV. The lightning arrester FV is successively connected in series with a capacitor C, a bridge differential current transformer ΔI, and an earthing switch QE2.

2. The 110KV voltage regulation type large-capacity dynamic reactive power compensation method for a substation according to claim 1, wherein, An earthing switch QE1 is also provided between the disconnector QS and the voltage regulator T.

3. The 110KV voltage regulating type large-capacity dynamic reactive power compensation method for a substation according to claim 1, characterized in that The load of the calcium carbide furnace is obtained by real-time acquisition to determine the real-time load.

4. The 110KV voltage regulating type large-capacity dynamic reactive power compensation method for a substation according to claim 1, characterized in that, The load of the calcium carbide furnace is predicted by an intelligent model. If it is predicted that the load of the calcium carbide furnace will drop below 50%, the dynamic reactive power compensation device is started to adjust the reactive power output capacity of the calcium carbide furnace. At the same time, the real-time load of the calcium carbide furnace is monitored in real time. If no mutation behavior occurs in the calcium carbide furnace within the set time period T2, the dynamic reactive power compensation device is stopped from adjusting the reactive power output capacity of the calcium carbide furnace.

5. The 110KV voltage regulating type large-capacity dynamic reactive power compensation method for a substation according to claim 1, wherein The specific method for predicting the load of the calcium carbide furnace by the intelligent model is as follows: The historical load data is the relevant parameters of several calcium carbide furnaces and their maximum change values during the first T1 time period before the mutation behavior occurs. Here, T1 is a preset value; the relevant parameters include furnace temperature, furnace pressure, voltage, current, power factor, and start / stop furnace frequency; The collected historical load data is divided into a training set and a validation set. The training set is used as input to train the intelligent model by means of a long short-term memory network (LSTM). After training, the validation set is used to verify its accuracy rate. When the accuracy rate exceeds the set ratio R1, it means that the intelligent model is available; Use the intelligent model to predict the load data of the calcium carbide furnace. When it is predicted that the load of the calcium carbide furnace will drop below 50%, the dynamic reactive power compensation device is started to adjust the reactive power output capacity of the calcium carbide furnace. At the same time, the real-time load of the calcium carbide furnace is monitored in real time. If no mutation behavior occurs in the calcium carbide furnace within the set time period T2, the dynamic reactive power compensation device is stopped from adjusting the reactive power output capacity of the calcium carbide furnace.

6. The 110KV voltage regulating type large-capacity dynamic reactive power compensation method for a substation according to claim 1, characterized in that, The specific method for predicting the load of the calcium carbide furnace by the intelligent model is as follows: ST1: Mark all relevant parameters of the calcium carbide furnace as target parameters; ST2: Mark each behavior of the load of the calcium carbide furnace dropping below 50% as a mutation behavior, and obtain all the target parameters that have changed during the first T1 time period before the mutation behavior occurs. Here, T1 is a preset value; here, the numerical change of the target parameter is at least greater than X1, and X1 is a preset value; ST3: Then obtain all the target parameters and their occurrence ratios in each mutation behavior, ST4: Obtain the occurrence ratios of all the target parameters, screen according to the occurrence ratios, determine the screening line according to the standard deviation of the occurrence ratios and the distribution of the values in the occurrence ratios, and mark the target parameters with occurrence ratios exceeding the screening line as object parameters; ST5: Conduct a numerical constancy analysis on the object parameters. Determine the mutation value based on the maximum change value of the numerical variation of the object parameters during the time period T1 before the occurrence of the mutation behavior. Determine the variation line according to the distribution of the mutation values; obtain the variation line of each object parameter. ST6: Conduct real-time monitoring on the object parameters of the calcium carbide furnace. When any object parameter changes and the change value exceeds the variation line, automatically monitor the time period T1 at this time. If more than a set proportion R2 of the object parameters change and the change values exceed the corresponding variation lines within this time period, a pre-start signal is automatically generated. ST7: When the pre-start signal is generated, start the dynamic reactive power compensation device to adjust the reactive power output capacity of the calcium carbide furnace. At the same time, monitor the real-time load of the calcium carbide furnace in real time. If no mutation behavior occurs in the calcium carbide furnace within the set time period T2, stop the dynamic reactive power compensation device from adjusting the reactive power output capacity of the calcium carbide furnace.

7. The 110KV voltage regulating type large-capacity dynamic reactive power compensation method for a substation according to claim 6, characterized in that The specific method for screening object parameters in ST4 is as follows: First, obtain the occurrence proportion of the target parameters, automatically obtain the mean value of all occurrence proportions, and then calculate the standard deviation of all occurrence proportions using a formula. When the standard deviation is lower than the preset value X2, mark the mean value of the occurrence proportions as the screening line at this time. If the standard deviation is not lower than X2, automatically obtain the number of target parameters with occurrence proportions greater than the mean value, mark it as the upper number, and mark the number of target parameters with occurrence proportions less than the mean value as the lower number. When the upper number exceeds the lower number, automatically mark the value obtained by multiplying the mean value by 1.15 as the screening line at this time, otherwise mark the mean value as the screening line. Mark the target parameters with occurrence proportions exceeding the screening line as object parameters.

8. The 110KV voltage regulating type large-capacity dynamic reactive power compensation method for a substation according to claim 6, characterized in that, The specific method for numerical constancy analysis in ST5 is as follows: Select any object parameter, obtain the maximum change value of its corresponding numerical variation during each time period T1 before the occurrence of the mutation behavior, and mark it as the mutation value. Obtain several mutation values Bj, where j = 1,..., n. Then automatically obtain the average value P of Bj, and calculate the stability value W of Bj using a formula. The specific calculation formula is: When the value of W does not exceed the preset value X3, mark the minimum value in Bj at this time as the variation line of the corresponding object parameter. When the value of W exceeds the preset value X3, sort Bj in descending order according to |Bj - P|, then separately screen out the corresponding Bj where Bj - P is less than zero, sort them according to the original data, select one Bj in turn, delete it each time a Bj is selected, and then recalculate the value of W until the value of W does not exceed the preset value X3. At this time, mark the minimum value in the remaining Bj as the variation line of the object parameter. If the value of W still exceeds X3 after deleting all the screened data, automatically mark the mean value as the variation line.

Images