Parameter Processing Method and Device for Compressed Air Energy Storage Device

Abstract

A parameter processing method and device for a compressed air energy storage device, belonging to the technical field of electrical digital data processing. For the attributes of pressure values and flow values, it carefully analyzes the relationship between pressure values and flow values. Through compliant parsing operations, it accurately determines the interference amount of each pressure value within the selected value cluster. Based on this interference amount, it constructs a special anti-interference method for each pressure value, which includes dynamically determining its cleaning frequency and the level of unit capacity at different cleaning frequencies to achieve dynamic and active cleaning. Through such a method, the situation of high-pressure values being wrongly removed can be efficiently avoided, thereby obtaining a more accurate cleaning value, improving the accuracy and correctness of the future point pressure value estimation, and achieving accurate detection of the parameter status at the inlet of the turbine.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric digital data processing, and particularly relates to a parameter processing method and device for a compressed air energy storage device. Background Art

[0002] Compressed-Air Energy Storage (CAES) refers to a energy storage method in which electrical energy is used to compress air during the low-load period of the power grid, and the high-pressure air is sealed in abandoned mines, sunken undersea gas storage tanks, caves, expired oil and gas wells or newly built gas storage wells, and the compressed air is released during the high-load period of the power grid to drive a turbine for power generation.

[0003] When a compressed air energy storage device actually operates, as mentioned in the prior art solution with the patent publication number "CN107769233B" and the patent name "A small-capacity compressed air energy storage device and its grid connection method", a pressure sensor is often set at the inlet of its turbine to detect the pressure value at the inlet of the turbine.

[0004] On the other hand, the pressure value sampled by the pressure sensor is often subject to external electromagnetic interference and is prone to generating interference values. Therefore, it is necessary to clean the pressure value to remove the interference values. Currently, the method for cleaning the pressure value is often the Gaussian filtering method. However, since the cleaning method of the Gaussian filtering method defines a fixed unit capacity for all the pressure values to be cleaned, the unit capacity is the number of adjacent sampled pressure values after the sampling time point of the defined pressure value, and the interference removal method of the Gaussian filtering method uses a fixed unit capacity for all the pressure values to be cleaned. Such a fixed processing method often causes misidentification when facing high pressure values at the inlet of the turbine. Because high pressure values are often very similar to interference values, they are often misidentified as interference values and removed. This not only weakens the accuracy and completeness of the pressure value, but also causes a large error between the estimated value of the pressure value at a future time point derived from the estimation model constructed based on such pressure values and the actual situation, and cannot accurately detect the parameter status at the inlet of the turbine. Summary of the Invention

[0005] To solve the defects in the prior art, the present invention proposes a parameter processing method and device for a compressed air energy storage device, effectively avoiding the defects of the Gaussian filtering method in the prior art, which uses a fixed unit capacity for all the pressure values to be cleaned at the inlet of the turbine of the compressed air energy storage device, often causing misidentification, weakening the accuracy and completeness of the pressure value, causing a large error between the estimated value of the pressure value at a future time point derived from the estimation model constructed based on such pressure values and the actual situation, and unable to accurately detect the parameter status at the inlet of the turbine.

[0006] The present invention employs the following technical solutions.

[0007] A parameter processing method for a compressed air energy storage device, comprising:

[0008] A pressure sensor and a flow sensor respectively sample the pressure value and the flow value of the gas at the inlet of the turbine of the compressed air energy storage device, and transmit the pressure value and the flow value to an industrial control computer for processing;

[0009] The method of transmitting the pressure value and the flow value to an industrial control computer for processing comprises:

[0010] Step 1, receiving the pressure value and the flow value at the inlet of the turbine during a defined period sampled by the sampling;

[0011] Step 2, performing several cleanings on all the pressure values, and actively adjusting the unit capacity of the pressure value before each cleaning;

[0012] Step 3, performing cleaning according to the adjusted unit capacity of all the pressure values, and detecting the operating condition at the inlet of the turbine according to all the pressure values after cleaning.

[0013] Preferably, in Step 1, the pressure value and the flow value are respectively obtained by sampling through the pressure sensor and the flow sensor. The starting sampling time points and the sampling speed of the pressure value and the flow value are the same. The sampling speed is 2 s / time. The sampling time of the pressure value and the flow value is defined as 30 min, and the sampling time is used as the defined period.

[0014] Preferably, in Step 2, the method of actively adjusting the unit capacity of the pressure value comprises:

[0015] Step 2-1, obtaining the final interference amount of each pressure value;

[0016] Step 2-2, determining whether the pressure value needs to be included in this cleaning according to the level of the final interference amount of each pressure value. If there is a need, actively adjust the unit capacity of the pressure value by using the final interference amount of each pressure value. If there is no need, this pressure value does not perform this cleaning.

[0017] Preferably, Step 2-1 specifically comprises:

[0018] Step 2-1-1, estimating the initial interference amount of the pressure value through the error between each pressure value and the average of its adjacent numerical cluster, and combining the irrelevance between the pressure value and the flow value corresponding to its sampling time point;

[0019] Step 2-1-2, calculating the final interference amount of each pressure value according to the difference between the initial interference amount of each pressure value and the initial interference amount of the pressure values within its adjacent numerical cluster, and combining the confidence level of the difference.

[0020] Preferably, in step 2-1-1, the method for estimating the initial interference amount includes: selecting the numerical cluster in the vicinity of this pressure value, that is, taking this pressure value as the midpoint, and selecting pressure values respectively before and after its sampling time point, and jointly forming the numerical cluster in the vicinity of this pressure value with the pressure value taken as the midpoint;

[0021] The operation equation for the initial interference amount of each pressure value is:

[0022] ;

[0023] In the equation, is the initial interference amount of the th pressure value, is the value of the th pressure value, is the mean of the numerical cluster in the vicinity of the th pressure value, is the irrelevance between the th pressure value and the flow value corresponding to the sampling time point of the th pressure value, is to perform standardization on using the Z-score standardization method.

[0024] Preferably, in step 2-1-1, the irrelevance between the th pressure value and the flow value corresponding to the sampling time point of the th pressure value

[0025] ;

[0026] In the equation, is the sequence code ratio of the th pressure value within its numerical cluster in the vicinity, is the sequence code ratio of the flow value corresponding to the sampling time point of the th pressure value within its numerical cluster in the vicinity, is the defined additional value.

[0027] Preferably, step 2-1-1 further includes: estimating the initial interference amount of the flow value through the error between each flow value and the mean of its numerical cluster in the vicinity, plus the irrelevance between the flow value and the pressure value corresponding to its sampling time point. The operation equation for the initial interference amount of the flow value is:

[0028] ;

[0029] In the equation, is the initial interference amount of the th flow value, is the The value of a flow value is the mean of the near-domain numerical cluster of the th flow value is the irrelevance between the pressure value corresponding to the sampling time point of the th flow value and the th flow value is to perform standardization on using the Z-score standardization method

[0030] Preferably, in step 2-1-2, the operation equation for the final interference amount of the pressure value is:

[0031] ;

[0032] In the equation, is the final interference amount of the th pressure value is the number of pressure values outside the pressure value itself in the near-domain numerical cluster of the th pressure value is the sequence code of the pressure values outside the pressure value itself in the near-domain numerical cluster of the th pressure value is the starting interference amount of the th pressure value is the starting interference amount of the th pressure value outside the pressure value itself in the near-domain numerical cluster of the th pressure value represents the difference between the starting interference amount of the th pressure value and the starting interference amount of the th pressure value in its near-domain numerical cluster is the credible amount of, and its operation equation is:

[0033] ;

[0034] In the equation, is the outlier amount of the flow value corresponding to the sampling time point of the th pressure value is the time interval between the th pressure value and the th pressure value outside the pressure value itself in its near-domain numerical cluster

[0035] Preferably, in step 2-2, a critical value of the final interference amount is defined, and before each cleaning, each pressure value is compared according to the final interference amount of the pressure value and the critical value of the final interference amount. If the final interference amount of the pressure value is higher than the critical value of the final interference amount, there is a need to add this cleaning. Conversely, if the final interference amount of the pressure value is not higher than the critical value of the final interference amount, there is no need to add this cleaning for this pressure value.

[0036] Preferably, in step 2-2, the operation equation for actively adjusting the unit capacity is:

[0037] ;

[0038] In the equation, is the Euler number, is the current cleaning frequency of the th pressure value, is the th pressure value when adding the nd cleaning, the actively adjusted unit capacity, is the unit capacity predefined for all pressure values, is the th pressure value before adding the th cleaning, the final interference amount, is the th pressure value before adding the th cleaning, the final interference amount.

[0039] Preferably, in step 3, the method for detecting the operating condition of the inlet of the turbine according to all the pressure values after cleaning includes:

[0040] Regarding all the pressure values after cleaning as past pressure values, then regarding the past pressure values as the dependent variable and the corresponding sampling time points of the past pressure values as the independent variable, using the least squares method to obtain the regression equation of the past pressure values, and obtaining the estimated values of the pressure values at each sampling time point after the defined period according to the regression equation;

[0041] According to the difference between the estimated values and the measured values of the pressure values at each time point after the defined period, determine whether the condition of the inlet of the turbine is abnormal, and thus achieve the detection of the condition of the inlet of the turbine. The method is:

[0042] Calculate the difference between the measured value and the estimated value of the pressure value at each time point after the defined period; if the difference between the measured value and the estimated value of the pressure value at a certain time point is higher than the predefined difference critical value, the pressure value at this time point is abnormal; if the difference is not higher than the difference critical value, the pressure value at this time point is not abnormal;

[0043] Sum up whether the pressure values at each time point after the defined period are abnormal. After the defined period, starting from the time point when the first abnormal pressure value appears, sum up the number of abnormal pressure values within the defined time period; if the ratio of the number of abnormal pressure values to the total number of all pressure values within the defined time period exceeds the defined ratio threshold, the condition at the inlet of the turbine is in an abnormal condition, and if the ratio does not exceed the defined ratio threshold, the condition at the inlet of the turbine is in a normal condition.

[0044] A parameter processing device for a compressed air energy storage device, comprising:

[0045] A pressure sensor and a flow sensor are provided at the inlet of the turbine of the compressed air energy storage device, and both the pressure sensor and the flow sensor are connected to an industrial control computer;

[0046] The modules running on the industrial control computer include:

[0047] A receiving module, which is used to receive the pressure value and flow value at the inlet of the turbine during the defined period sampled and transmitted;

[0048] An adjustment module, which is used to perform several cleanings on all pressure values, and actively adjust the unit capacity of the pressure values before each cleaning;

[0049] A detection module, which is used to perform cleaning according to the unit capacity after adjusting all pressure values, and perform condition detection on the inlet of the turbine according to all the cleaned pressure values.

[0050] The beneficial effects of the present invention are that, compared with the prior art, the technical effects of the present invention include:

[0051] Regarding the attributes of the pressure value and the flow value, carefully analyze the relationship between the pressure value and the flow value. Through compliant analysis operations, accurately determine the interference amount of each pressure value within the selected value cluster. According to this interference amount, construct a special anti-interference method for each pressure value, which includes automatically determining its cleaning frequency and the level of the unit capacity at different cleaning frequencies, achieving automatic and active cleaning. Through such methods, it is possible to effectively avoid the situation where high-pressure values are wrongly removed, thereby obtaining more accurate cleaning values, improving the accuracy and correctness of the prediction of future time point pressure values, and enabling accurate detection of the parameter conditions at the inlet of the turbine. Description of the Drawings

[0052] Figure 1 is a flowchart of the parameter processing method for the compressed air energy storage device described in the present invention;

[0053] Figure 2 is a module structure diagram of the industrial control computer described in the present invention. Detailed Embodiments

[0054] To make the objectives, technical solutions and advantages of the present invention clearer, the following will, in conjunction with the accompanying drawings in the embodiments of the present invention, clearly and completely describe the technical solutions of the present invention. The embodiments described in this application are only some embodiments of the present invention, rather than all embodiments. Based on the spirit of the present invention, other embodiments obtained by those skilled in the art without creative efforts all fall within the protection scope of the present invention.

[0055] As Figure 1 shown, a method for processing parameters of a compressed air energy storage device according to the present invention includes:

[0056] A pressure sensor and a flow sensor respectively sample the pressure value and the flow value of the gas at the inlet of the turbine of the compressed air energy storage device, and transmit the pressure value and the flow value to an industrial control computer for processing;

[0057] The method of transmitting the pressure value and the flow value to the industrial control computer for processing runs on the industrial control computer and includes:

[0058] Step 1, receiving the pressure value and the flow value at the inlet of the turbine in a defined time period sampled;

[0059] The industrial control computer receives the pressure value and the flow value at the inlet of the turbine of the compressed air energy storage device sampled. These two types of parameters present the main parameter conditions at the inlet of the turbine.

[0060] Therefore, by carefully analyzing the relationship between the pressure value and the flow value, for example, when the flow rate increases, the pressure usually also increases. Such a relationship can help estimate the conditions at the inlet of the turbine under different working conditions, so as to accurately detect the parameter conditions. Long-term detection of such pressure values and flow values can estimate the parameter change trend at the inlet of the turbine and formulate a maintenance plan for the turbine in advance.

[0061] In a preferred but non-limiting embodiment of the present invention, in Step 1, the pressure value and the flow value are respectively obtained by sampling through the pressure sensor and the flow sensor. The starting sampling time points and the sampling speeds of the pressure value and the flow value are the same. The sampling speed is 2 s / time. Define the sampling time of the pressure value and the flow value as 30 min, and regard the sampling time of the first time as the defined time period. Therefore, a pressure value and a flow value can be obtained at each sampling time point, ensuring the temporal coordination of all the sampled pressure values and flow values, which is conducive to later analysis of the pressure value and the flow value.

[0062] Step 2, perform several cleanings on all the pressure values. Before each cleaning, actively adjust the unit capacity of the pressure value; the unit capacity is the number of pressure values of the adjacent sampling after the sampling time point of the defined pressure value , performing cleaning on the pressure value means sending the pressure value and the pressure values adjacent to it after the sampling time point into the Gaussian filtering method to perform cleaning on the pressure value. If the number of pressure values adjacent to it after the sampling time point is less than , then obtain the adjacent pressure values sampled before the sampling time point of the pressure value to make up pressure values. Then send the pressure value and the obtained pressure values into the Gaussian filtering method to perform cleaning on the pressure value. The cleaning method is the Gaussian filtering method. The number of times can be determined according to specific requirements.

[0063] In a preferred but non-limiting embodiment of the present invention, in step 2, the method for actively adjusting the unit capacity of the pressure value includes:

[0064] Step 2-1, obtain the final interference amount of each pressure value;

[0065] In a preferred but non-limiting embodiment of the present invention, step 2-1 specifically includes:

[0066] Step 2-1-1, estimate the initial interference amount of the pressure value through the error between each pressure value and the mean of its adjacent numerical cluster, and combining the irrelevance between the pressure value and the flow value corresponding to its sampling time point;

[0067] In a preferred but non-limiting embodiment of the present invention, in step 2-1-1, the method for estimating the initial interference amount includes: select the adjacent numerical cluster of the pressure value, that is, take the pressure value as the midpoint, and select pressure values before and after its sampling time point respectively, and form the adjacent numerical cluster of the pressure value together with the pressure value taken as the midpoint, is a defined value, and its value can be , that is, take the pressure value, the pressure values sampled at the sampling time points adjacent to the sampling time point of the pressure value before sampling time points and the pressure values adjacent to the sampling time point of the pressure value after pressure values, a total of two hundred and one pressure values as the adjacent numerical cluster of the pressure value; and if there are no pressure values before a pressure value, then the insufficient number is selected after the pressure value. Just like if there are fifty pressure values before a pressure value and there are still fifty insufficient, then take another fifty pressure values after the pressure value, and a total of one hundred and fifty pressure values are obtained after the pressure value. If the number of pressure values after the pressure value is less than one hundred, similarly, supplement the insufficient number of pressure values before the pressure value.

[0068] The operation equation for the initial interference amount of each pressure value is:

[0069] ;

[0070] In the equation, is the starting interference amount of the th pressure value, is the value of the th pressure value, is the mean of the near-domain numerical cluster of the th pressure value, is the irrelevance between the th pressure value and the flow value corresponding to the sampling time point of the th pressure value, is to perform standardization on using the Z-score standardization method. is the error between the pressure value and the mean of its near-domain numerical cluster.

[0071] In this equation, represents the difference between the pressure value corresponding to the th pressure value and the mean of the pressure values in its near-domain numerical cluster. The higher is, the more significant the th pressure value is. It is more likely to be a true abnormal value of high pressure caused by changes in the parameters of the turbine's inlet itself, or it may also be an interference value. To more precisely identify the specific situation, it is necessary to analyze the irrelevance between the th pressure value and the flow value corresponding to its sampling time point. According to the rule that an increase in the gas flow at the turbine's inlet will cause an increase in the gas pressure, so when the flow increases, it often causes an increase in the gas pressure at the turbine's inlet. If the coherence between a pressure value and the flow value corresponding to its sampling time point is lower (the irrelevance is higher), it indicates that this pressure value is more likely to be an interference value rather than a true abnormal value caused by changes in the parameters of the turbine's inlet itself.

[0072] In a preferred but non-limiting embodiment of the present invention, in step 2-1-1, the equation for calculating the irrelevance between the th pressure value and the flow value corresponding to the sampling time point of the th pressure value is:

[0073] ;

[0074] In the equation, is the sequence code ratio of the th pressure value within its near-domain numerical cluster, is the sequence code ratio of the flow value corresponding to the sampling time point of the th pressure value within its near-domain numerical cluster, is a defined additional value, can be , which is to avoid ​A situation of zero occurs. It is the difference between the pressure value and the sequence code ratio within the corresponding near-domain numerical cluster of the flow rate value, which can estimate the irrelevance between these two types of parameters. The higher the difference between the sequence code ratios, the lower the coherence and the higher the irrelevance between the pressure value and its corresponding flow rate value. This pressure value is more likely to be an interference value. Introducing an additional value can avoid the calculation obstacles caused when the sequence code ratios are the same and improve the accuracy of later estimation. The sequence code ratio is to arrange the values within the corresponding near-domain numerical cluster in ascending order, and then assign corresponding sequence codes to each value in the near-domain numerical cluster one by one starting from 1 in ascending order. The sequence code ratio of the corresponding value is the sequence code divided by the number of values in the near-domain numerical cluster.

[0075] Next, the near-domain numerical cluster of the flow rate value is segmented in the same way as the pressure value, and then the initial interference amount of each flow rate value is estimated in the same way.

[0076] In a preferred but non-limiting embodiment of the present invention, step 2-1-1 further includes: estimating the initial interference amount of the flow rate value through the error between each flow rate value and the mean of its near-domain numerical cluster, plus the irrelevance between the flow rate value and the pressure value corresponding to its sampling time point. The initial interference amount is proportional to both the error and the irrelevance. The calculation equation for the initial interference amount of the flow rate value is:

[0077] ;

[0078] In the equation, is the initial interference amount of the th flow rate value, is the value of the th flow rate value, is the mean of the near-domain numerical cluster of the th flow rate value, is the irrelevance between the th flow rate value and the pressure value corresponding to the sampling time point of the th flow rate value, is to perform standardization on using the Z-score standardization method. is the error between the flow rate value and the mean of its near-domain numerical cluster.

[0079] In this equation, represents the error of the th flow rate value corresponding to the mean of the flow rate values in its near-domain numerical cluster. The higher , the The more significant a flow value is, the more likely it is to be a true abnormal value of high flow caused by changes in the parameters at the inlet of the turbine itself, or it may also be an abnormal value caused by interference values. To more precisely identify the specific situation, analyze the irrelevance between the second flow value and the pressure value corresponding to its sampling time point , and , if the coherence between a flow value and the pressure value corresponding to its sampling time point is lower (the irrelevance is higher), it indicates that this flow value is more likely to be an interference value rather than a true abnormal value caused by changes in the parameters at the inlet of the turbine itself.

[0080] The above initial interference amounts of both pressure values and flow values mainly involve the significance of the parameters and the coherence between the two types of parameters. The higher the initial interference amount of a parameter, the more likely it is to be an interference value.

[0081] Step 2-1-2: Calculate the final interference amount of each pressure value based on the difference between the initial interference amount of each pressure value and the initial interference amounts of the pressure values within its near-domain numerical cluster, combined with the confidence level of the difference.

[0082] In the above method, the initial interference amounts of each pressure value have been calculated. However, the initial interference amount only analyzes based on the near-domain numerical cluster in a short time period and cannot fully represent the global properties of the pressure value. If only relying on the pressure values (near-domain numerical cluster) in a short time period around a pressure value, the global trend in a long time period is often ignored, resulting in inaccurate analysis values. For example, when comparing a pressure value with its near-domain numerical cluster, the error is not small and it is often an outlier interference value. However, when compared with all pressure values, the global error with all pressure values is not high and it cannot be determined as an interference value. Therefore, the above initial interference amount needs to be further improved to obtain a more accurate final interference amount for each pressure value.

[0083] To improve the accuracy of estimating the outlier amount of each pressure value, the calculation method of the outlier amount needs to be further improved. Calculate the initial interference amounts of all pressure values within the near-domain numerical cluster of this pressure value according to the method in Step 2-1-1. Then, calculate the difference between the initial interference amount of this pressure value and the initial interference amounts of other pressure values in its near-domain numerical cluster. The higher the difference, the more significant the comparison of the outlier amount of this pressure value with the outlier amounts of other values. And analyze the confidence level of this difference. If the difference between the initial interference amount of a pressure value and the initial interference amounts of other pressure values in its near-domain numerical cluster is higher and the confidence level of this difference is greater, then this pressure value is more likely to be a true interference value.

[0084] Therefore, by combining the difference between each pressure value and the starting interference amount of other pressure values in its adjacent numerical cluster, and combining the confidence amount of this difference, the final interference amount of this pressure value is calculated. In a preferred but non-limiting embodiment of the present invention, in step 2-1-2, the calculation equation for the final interference amount of the pressure value is:

[0085] ;

[0086] In the equation, is the final interference amount of the th pressure value, is the number of pressure values outside this pressure value itself in the adjacent numerical cluster of the th pressure value, is the serial number of the pressure value outside this pressure value itself in the adjacent numerical cluster of the th pressure value (the serial numbers of the pressure values outside this pressure value itself in the adjacent numerical cluster of the th pressure value are: arrange the pressure values outside this pressure value itself in the adjacent numerical cluster of the th pressure value in the order of their sampling time points, and then sequentially assign corresponding serial numbers starting from one and incrementing by one to each pressure value according to this order), is the starting interference amount of the th pressure value, is the starting interference amount of the th pressure value outside this pressure value itself in the adjacent numerical cluster of the th pressure value, represents the difference between the starting interference amount of the th pressure value and the starting interference amount of the th pressure value in its adjacent numerical cluster. The higher this difference, the higher the difference between the outlier amount of the th pressure value and the starting interference amounts of other pressure values in its adjacent region, indicating that the th pressure value is more likely to be an interference value or an outlier, and its final interference amount is also higher, is the confidence amount of , and its calculation equation is:

[0087] ;

[0088] In the equation, is the outlier amount of the flow value corresponding to the sampling time point of the th pressure value (the outlier amount of this flow value can be the starting interference amount of this flow value), is the time interval between the th pressure value and the th pressure value outside this pressure value itself in its adjacent numerical cluster (the The time interval between a pressure value and the inner cluster of its neighboring values outside this pressure value and the pressure value is: the absolute value of the difference between the sampling time points of the pressure value and the sampling time point of this pressure value). The higher the value of the confidence measure, the higher the referenceability of the difference between the starting interference amount of the pressure value and the starting interference amount of the pressure value within its neighboring value cluster.

[0089] Here, the confidence measure is elaborated as follows:

[0090] (1) Regarding the time interval between the pressure value and the pressure value within its neighboring value cluster, the lower the time interval, the closer the pressure value and the pressure value within its neighboring value cluster in terms of time point, and the pressure value and the pressure value within its neighboring value cluster in terms of the difference in starting interference amount is more credible, and its referenceability is also higher. The lower the time interval between two pressure values, the closer the two pressure values are in terms of time point, and the more coherent and similar the changes between the two pressure values are. Therefore, if the time interval between two pressure values is not large, the difference in starting interference amount between the two pressure values is more credible because such a difference represents a real numerical difference rather than a sudden fluctuation. Therefore, the referenceability of such a difference is better.

[0091] (2) Regarding the starting interference amount of the flow value corresponding to the sampling time point of the pressure value, the lower this starting interference amount, the more accurate the calculated value is, and its referenceability is also stronger. If the starting interference amount of the flow value corresponding to the sampling time point of the pressure value is not high, it indicates that the change of the flow value around this sampling time point is very stable. Therefore, the calculated is more accurate. In such a situation, it can well represent the real outlier situation rather than the non-real outlier situation formed by the fluctuation of the flow value itself. Therefore, the referenceability of such a difference is better.

[0092] Since the neighboring value cluster of each pressure value contains the pressure value itself and each of its previous and subsequent Adjacent pressure values. Therefore, when calculating the initial interference amount of all pressure values in the local numerical cluster of a pressure value, it is essentially based on a wider pressure value range. When analyzing the initial interference amount of all pressure values in the local numerical cluster of a pressure value, a larger number of corresponding pressure values are involved, which can make the finally calculated interference amount more comprehensively reflect the global attributes of the pressure value. Moreover, the calculation of the final interference amount involves the outlier difference and its credibility, which is conducive to more accurate identification of interference values and improves the credibility of later estimation.

[0093] Step 2-2: Determine whether the pressure value needs to be included in this cleaning according to the level of the final interference amount of each pressure value. If there is a need, actively adjust the unit capacity of the pressure value using the final interference amount of each pressure value. If there is no need, the pressure value does not perform this cleaning.

[0094] Before adjusting the unit capacity of each pressure value, since the interference amounts of the approximate interference values represented by different pressure values are different, and the cleaning frequencies required for them are also different. In step 2-2, the cleaning frequency of each pressure value is accurately determined at the beginning.

[0095] In a preferred but non-limiting embodiment of the present invention, in step 2-2, a critical value of the final interference amount needs to be defined. Here, the critical value can be defined as , and this value can be determined according to specific requirements. Before each cleaning of each pressure value, compare the final interference amount of the pressure value with the critical value of the final interference amount. If the final interference amount of the pressure value is higher than the critical value of the final interference amount, it indicates that the interference value of the pressure value has not been completely removed, and the cleaning function is not good, so there is still a need to include this pressure value in this cleaning to improve the accuracy of the pressure value. On the contrary, if the final interference amount of the pressure value is not higher than the critical value of the final interference amount, it indicates that the cleaning function of the pressure value has achieved the purpose and can represent the true situation of the pressure value. At this time, this pressure value does not need to be included in this cleaning to avoid over-treatment.

[0096] Before each cleaning, for each pressure value, according to this compliance confirmation rule, obtain its most appropriate cleaning frequency to ensure that when efficiently removing interference values, the original attributes of the pressure value are maintained as much as possible.

[0097] Then, before each cleaning, for all pressure values that need to be included in this cleaning, actively adjust their unit capacities according to the level of their final interference amounts to achieve a flexible and active cleaning process.

[0098] When performing the first cleaning at a certain pressure value, if the final interference amount is not low, it indicates that the probability of being interfered by the interference value is not small. In this situation, to more effectively remove the interference value, a certain unit capacity is required because a certain unit capacity can introduce a larger amount of pressure value for cleaning, thereby increasing the function of eliminating the interference value and improving the accuracy and reliability of the pressure value. When a pressure value is in several cleaning cycles, according to the rule that the higher the final interference amount, the higher the required unit capacity. In addition, the cleaning function of the pressure value after the previous cleaning also needs to be considered. If the cleaning function of the pressure value after the previous cleaning is not good, for example, after the previous cleaning, the final interference amount of the pressure value is still not low and there is no significant change compared with before, it indicates that the previous cleaning did not completely remove the interference value. Then, a higher unit capacity is required to perform a more detailed cleaning during this cleaning. Through such methods, the true efficacy of each cleaning can be accurately estimated, providing a basis for adjusting the level of the unit capacity for the subsequent cleaning.

[0099] According to the above rules, actively adjust the level of the unit capacity for each pressure value at different cleaning frequencies to ensure that during the cleaning process, not only can the appropriate unit capacity be flexibly selected according to the interference value attribute of the pressure value to effectively remove the interference value, but also the true change attribute of the pressure value can be well maintained, providing pressure value conditions with excellent accuracy for the subsequent pressure value analysis.

[0100] In a preferred but non-limiting embodiment of the present invention, in step 2-2, the operation equation for actively adjusting the unit capacity is:

[0101] ;

[0102] In the equation, is the Euler number, is the current cleaning frequency of the th pressure value. is the main parameter for determining which operation method to use. Different cleaning frequencies correspond to different operation rules, showing the flexible matching of the pressure value processing process. is the th pressure value when adding the th cleaning. The actively adjusted unit capacity after adjustment is the main purpose of this operation. Whether it is accurately determined will determine the cleaning function. is the unit capacity predefined for all pressure values. Its value can be ten, which provides an initial capacity for the operation, ensuring that all values have a common initial unit capacity without the influence of other factors. is the th pressure value. The final interference amount before adding the It shows the interference value situation of the current pressure value when cleaning is about to be performed, which is the main basis for determining the adjustment of the unit capacity level. is the th pressure value and the last interference amount before the th cleaning. By comparing it with the current interference amount, the cleaning function of the previous cleaning can be estimated. and have the same calculation method, only the cleaning frequency is different. has its corresponding numerical meaning in the equation and is used to adjust the weight of the unit capacity level according to the change of the interference amount.

[0103] In this equation, shows the cleaning function of the th pressure value added to the th cleaning. If the cleaning function of the th cleaning is better, should be higher. When performing the th cleaning, the unit capacity should be reduced. If the cleaning function of the th cleaning is not good, then should be lower. When performing the th cleaning, the unit capacity should be increased. Therefore, through constructs the inverse relationship between the cleaning function of the th cleaning and the level of the unit capacity during the th cleaning.

[0104] In this equation, mainly involves the cleaning function of the th pressure value during the th cleaning and the last interference amount of the th pressure value at present, that is:

[0105] When is not small, it indicates that the last interference amount of the th pressure value before the th cleaning is not low, that is, the interference amplitude of the current pressure value by the interference value is large, and a relatively large unit capacity is required to effectively eliminate the interference value and reduce the effect of the interference value.

[0106] And if is not high, then is not small, indicating that the cleaning function of the previous cleaning is not good, and the last interference amount of the th pressure value has not decreased significantly after the previous cleaning. At this time, is not small, which indicates that the unit capacity should be increased during the th cleaning to more effectively remove the interference value; and if is not small, then is not high, indicating that the cleaning function of the previous cleaning is good. After the previous cleaning, the last interference amount of the th pressure value has been significantly reduced. At this time, even if is not small, but the value will not be too high, which means that in the case of good cleaning function in the previous cleaning, there is no need to increase the unit capacity too much to avoid removing too many values that represent the true change attributes of the pressure value.

[0107] At is not high, it indicates that the last interference amount of the th pressure value before the th cleaning is not high. In terms of maintaining the correctness of the pressure value, it is expected that the unit capacity does not need to be too high to avoid too many pressure values being cleaned. And if is not high, then is not small, indicating that the cleaning function of the previous cleaning is not good. The last interference amount of the th pressure value after the previous cleaning has not been significantly reduced. But because the value is not high, at this time the value will not be too high, which means that in the case of good cleaning function in the previous cleaning, there is no need to increase the unit capacity too much to avoid removing too many values that represent the true change attributes of the pressure value; and if is not small, then is not high, indicating that the cleaning function of the previous cleaning is good. The last interference amount of the th pressure value after the previous cleaning has been significantly reduced, and the value is not high. At this time the value will be even lower, which more clearly indicates that in the current situation of small interference amount and good cleaning function in the previous cleaning, the unit capacity that is not high should be maintained to greatly maintain the true change attributes of the pressure value.

[0108] Through the method of actively adjusting the unit capacity in step 2-2, before each cleaning, according to the change attributes of the pressure value, such as the interference amount and its change trend, the unit capacity of each pressure value can be adjusted flexibly. Such flexible adjustment enables the subsequent cleaning process to be closer to the true situation of the pressure value, greatly improving the accuracy of cleaning. For example, in the detection of the parameter situation at the inlet of the turbine, the pressure values sampled at different times are often interfered by interference values of different amplitudes and types. Through such active unit capacity adjustment, appropriate anti-interference methods can be applied according to different pressure value situations, making the pressure value after cleaning more accurately represent the true parameter situation at the inlet of the turbine. Based on such pressure values, more accurate detection of the parameter situation at the inlet of the turbine can be carried out, which is conducive to the efficient implementation of the monitoring of the inlet of the turbine.

[0109] Step 3: Perform cleaning based on the adjusted unit capacity of the overall pressure value, and perform condition detection on the inlet of the turbine based on the overall pressure value after cleaning.

[0110] Use the Gaussian filtering method to perform cleaning on the overall pressure values added in each cleaning. When performing cleaning each time, perform cleaning to remove interference based on the adjusted unit capacity of each pressure value. This can not only efficiently remove interference values, but also maintain the trend and main attributes of the pressure values. Finally, the pressure values after removing interference are obtained. The pressure values after removing interference contain the true values of the changes in the parameters of the inlet of the turbine itself. Based on this true value, accurate detection of the parameter conditions at the inlet of the turbine can be performed in the later stage.

[0111] In a preferred but non-limiting embodiment of the present invention, in Step 3, the method for performing condition detection on the inlet of the turbine based on the overall pressure value after cleaning includes:

[0112] Take the overall pressure value after cleaning as the past period pressure value. Then, take the past period pressure value as the dependent variable and the corresponding sampling time point of the past period pressure value as the independent variable, and use the least squares method to obtain the regression equation of the past period pressure value. Based on the regression equation, obtain the estimated values of the pressure values at each sampling time point after the defined period; define the next sampling time point after the defined period as , The next sampling time point of .

[0113] Based on the difference between the estimated values and the measured values (the measured value is the pressure value sampled by the pressure sensor) of the pressure values at each time point after the defined period, confirm whether the condition at the inlet of the turbine is abnormal, and thus achieve the condition detection of the inlet of the turbine. The method is as follows:

[0114] Calculate the difference between the measured value and the estimated value of the pressure values at each time point after the defined period (this difference is the absolute value of the quantity obtained by subtracting the estimated value from the measured value); if the difference between the measured value and the estimated value of the pressure value at a certain time point is higher than the pre-defined difference critical value, the pressure value at this time point is abnormal; if the difference is not higher than the difference critical value, the pressure value at this time point is not abnormal; in this application, the defined difference critical value between the measured value and the estimated value of the pressure value can be , if The difference between the estimated value and the measured value of the pressure value at the time point is higher than , it is determined that the pressure value at the time point is abnormal. On the contrary, the pressure value at the Use the overall pressure values at the current time point and its previous samplings as the new past pressure values, and execute Steps 1 to 3 again. Use the obtained interference-removed past pressure values as the dependent variable and the corresponding sampling time points of the past pressure values as the independent variable. Apply the least squares method to obtain the regression equation of the past pressure values, and obtain the estimated value of the pressure value at the time point according to the regression equation, and then confirm whether the pressure value at the time point is abnormal; for the next sampling time point after the time point, use the overall pressure values at the current time point and its previous samplings as the new past pressure values, and execute Steps 1 to 3 again. Use the obtained interference-removed past pressure values as the dependent variable and the corresponding sampling time points of the past pressure values as the independent variable. Apply the least squares method to obtain the regression equation of the past pressure values, and obtain the estimated value of the pressure value at the time point according to the regression equation, and then confirm whether the pressure value at the time point is abnormal, and continue to loop in this way to achieve efficient detection of whether the pressure value is abnormal.

[0115] Sum up the situations of whether the pressure values at each time point after the defined time period are abnormal. After the defined time period, starting from the time point when the first abnormal pressure value appears, sum up the number of abnormal pressure values within the defined time. If the ratio of the number of abnormal pressure values to the total number of all pressure values within the defined time exceeds the defined ratio critical value, the situation at the inlet of the turbine is in an abnormal situation. If the ratio does not exceed the defined ratio critical value, the situation at the inlet of the turbine is in a normal situation. The defined time in this application can be 60s, and the defined ratio critical value can be 20%. The defined time and the defined ratio critical value can be set according to specific requirements. For example, starting from the time point, continuously sample all the pressure values within 60s. Within these 60s, if the number of abnormal pressure values (the absolute value of the quantity obtained by subtracting the estimated value from the measured value is higher than the pressure value) exceeds 20% of the total number of all pressure values within these 60s, it is determined that the situation at the inlet of the turbine is in an abnormal situation, and there is a risk of failure at the inlet of the turbine. It is advisable to use corresponding methods. For example, check the inlet of the turbine and deal with the corresponding risks. If the magnetic mesh filter at the inlet of the turbine is blocked, perform dredging and other treatment methods. If the number of abnormal pressure values within these 60s does not exceed 20% of the total number of all pressure values within these 60s, it is determined that the situation at the inlet of the turbine is in a normal situation. The normal situation means that the inlet of the turbine is operating correctly. When the situation at the inlet of the turbine is in an abnormal situation, the abnormal situation message can also be displayed on the display screen of the industrial control computer to inform the staff to deal with the corresponding risks.

[0116] For any random point in time after the defined period, starting from that point in time, the pressure value can be continuously sampled for 60 s. According to the same method, it is estimated whether the condition at the inlet of the turbine is in an abnormal condition, so as to achieve efficient detection of the parameter condition at the inlet of the turbine.

[0117] Therefore, by performing steps 1 to 3 in this application, the final interference amount of each pressure value can be estimated more accurately, improving the accuracy of abnormal detection. The calculation of the final interference amount involves outlier discrimination and its confidence amount, which is conducive to more accurately identifying interference values and improving the credibility of estimation. By performing efficient estimation on the pressure value after removing interference and performing abnormal condition estimation based on the difference between the estimated value and the measured value, the parameter changes at the inlet of the turbine can be efficiently detected, and thus the risks that may occur at the inlet of the turbine can be efficiently detected and handled.

[0118] As Figure 2 shown, a parameter processing device for a compressed air energy storage device according to the present invention includes:

[0119] A pressure sensor and a flow sensor are provided at the inlet of the turbine of the compressed air energy storage device, and both the pressure sensor and the flow sensor are connected to the industrial control computer; the pressure sensor and the flow sensor are used to respectively sample the pressure value and the flow value of the gas at the inlet of the turbine of the compressed air energy storage device, and transmit the pressure value and the flow value to the industrial control computer for processing;

[0120] The modules running on the industrial control computer include:

[0121] A receiving module, which is used to receive the pressure value and the flow value at the inlet of the turbine during the defined period sampled and transmitted;

[0122] An adjustment module, which is used to perform several cleanings on all pressure values, and actively adjust the unit capacity of the pressure value before each cleaning;

[0123] A detection module, which is used to perform cleaning according to the unit capacity after adjusting all pressure values, and perform condition detection on the inlet of the turbine according to all pressure values after cleaning.

[0124] The beneficial effects of the present invention are that, compared with the prior art, the technical effects of the present invention include:

[0125] Regarding the attributes of the pressure value and the flow value, carefully analyze the relationship between the pressure value and the flow value. Through compliant parsing operations, accurately determine the interference amount of each pressure value within the selected numerical cluster. Based on this interference amount, construct a special anti-interference method for each pressure value, which includes automatically determining its cleaning frequency and the level of unit capacity at different cleaning frequencies, achieving automatic active cleaning. Through such methods, the situation of high pressure values being wrongly removed can be efficiently avoided, thereby obtaining a more accurate cleaning value, improving the accuracy and correctness of the future-time pressure value estimation, and enabling accurate detection of the parameter conditions at the inlet of the turbine.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not 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 implementation manners 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 parameter processing method for a compressed air energy storage device, characterized in that, Including: The pressure sensor and the flow sensor respectively sample the pressure value and the flow value of the gas at the inlet of the turbine of the compressed air energy storage device, and transmit the pressure value and the flow value to the industrial control computer for processing; The method of transmitting the pressure value and the flow value to the industrial control computer for processing includes: Step 1, receiving the pressure value and the flow value at the inlet of the turbine in the defined period sampled; Step 2, performing several cleanings on all the pressure values. Before each cleaning, actively adjust the unit capacity of the pressure value; Step 3, performing cleaning according to the adjusted unit capacity of all the pressure values, and performing condition detection on the inlet of the turbine according to all the pressure values after cleaning; In Step 2, the method of actively adjusting the unit capacity of the pressure value includes: Step 2-1, obtaining the final interference amount of each pressure value; Step 2-2, determining whether the pressure value needs to be included in this cleaning according to the level of the final interference amount of each pressure value. If there is a need, actively adjust the unit capacity of the pressure value by using the final interference amount of each pressure value. If there is no need, this pressure value does not perform this cleaning; Step 2-1 specifically includes: Step 2-1-1, estimating the initial interference amount of the pressure value through the error between each pressure value and the average of its adjacent numerical cluster, combined with the irrelevance between the pressure value and the flow value corresponding to its sampling time point; Step 2-1-2, calculating the final interference amount of each pressure value according to the difference between the initial interference amount of each pressure value and the initial interference amount of the pressure values within its adjacent numerical cluster, combined with the confidence level of the difference; In Step 2-2, the operation equation for actively adjusting the unit capacity is: ； In the equation, is the Euler number, is the current cleaning frequency of the th pressure value, is the th pressure value, and the unit capacity after active adjustment when adding the th cleaning, is the unit capacity predefined for all pressure values, is the th pressure value, and the last interference amount before adding the th cleaning, is the th pressure value, and the last interference amount before adding the th cleaning.

2. The parameter processing method for a compressed air energy storage device according to claim 1, wherein In Step 1, the pressure value and the flow value are respectively obtained by sampling through the pressure sensor and the flow sensor. The initial sampling time points and the sampling speed of the pressure value and the flow value are the same. The sampling speed is 2 s / time. Define the sampling time of the pressure value and the flow value as 30 min, and regard the sampling time as the defined period.

3. The parameter processing method for a compressed air energy storage device according to claim 2, wherein In step 2-1-1, the method for estimating the initial interference amount includes: selecting the numerical cluster in the vicinity of this pressure value, that is, taking this pressure value as the midpoint, and selecting pressure values before and after its sampling point, and jointly forming the numerical cluster in the vicinity of this pressure value with the pressure value taken as the midpoint; The operation equation for the initial interference amount of each pressure value is: ； Within the equation, is the starting interference amount of the th pressure value, is the value of the th pressure value, is the mean of the numerical cluster in the vicinity of the th pressure value, is the irrelevance between the th pressure value and the flow value corresponding to the sampling time point of the th pressure value, is to perform standardization on using the Z-score standardization method; In step 2-1-1, the irrelevance between the flow rate values corresponding to the sampling points of the pressure values of the th and th pressure values is represented by the following operation equation: ； In the equation, is the sequence code ratio of the -th pressure value within its adjacent numerical cluster, is the sequence code ratio of the flow value corresponding to the sampling time point of the -th pressure value within its adjacent numerical cluster, is the defined additional value.

4. The parameter processing method for a compressed air energy storage device according to claim 3, wherein Step 2-1-1 also includes: estimating the initial interference amount of the flow value through the error between each flow value and the average of its adjacent numerical cluster, plus the irrelevance between the flow value and the pressure value corresponding to its sampling time point. The operation equation for the initial interference amount of the flow value is: ； Inside the equation, is the starting interference amount of the th flow value, is the value of the th flow value, is the mean of the numerical cluster in the near domain of the th flow value, is the irrelevance between the th flow value and the pressure value corresponding to the sampling time point of the th flow value, is to perform standardization on using the Z-score standardization method; In Step 2-1-2, the operation equation for the final interference amount of the pressure value is: ； Within the equation, is the last interference amount of the th pressure value, is the number of pressure values outside the pressure value itself within the local numerical cluster of the th pressure value, is the sequence code of the pressure values outside the pressure value itself within the local numerical cluster of the th pressure value, is the starting interference amount of the th pressure value, is the starting interference amount of the th pressure value among the pressure values outside the pressure value itself within the local numerical cluster of the th pressure value, represents the difference between the starting interference amount of the th pressure value and the starting interference amount of the th pressure value within its local numerical cluster, is the credible amount of , and its operation equation is: ； Within the equation, is the outlier of the flow value corresponding to the sampling time point of the th pressure value, is the time interval between the th pressure value and the th pressure value outside the pressure value within its neighboring numerical cluster.

5. The parameter processing method for a compressed air energy storage device according to claim 4, wherein In Step 2-2, a critical value of the final interference amount needs to be defined. Before each cleaning of each pressure value, compare the final interference amount of the pressure value with the critical value of the final interference amount. If the final interference amount of the pressure value is higher than the critical value of the final interference amount, there is a need to be included in this cleaning. On the contrary, if the final interference amount of the pressure value is not higher than the critical value of the final interference amount, this pressure value has no need to be included in this cleaning.

6. The parameter processing method for a compressed air energy storage device according to claim 5, wherein, In Step 3, the method of performing condition detection on the inlet of the turbine according to all the pressure values after cleaning includes: Take all the pressure values after cleaning as the previous period pressure values. Then, take the previous period pressure values as the dependent variable and the corresponding sampling time points of the previous period pressure values as the independent variable, and use the least squares method to obtain the regression equation of the previous period pressure values. According to the regression equation, obtain the estimated values of the pressure values at each sampling time point after the defined period; Based on the difference between the estimated values and the measured values of the pressure values at each time point after the defined period, determine whether the condition at the inlet of the turbine is abnormal, and thus complete the detection of the condition at the inlet of the turbine. The method is as follows: Calculate the difference between the measured value and the estimated value of the pressure value at each time point after the defined period; if the difference between the measured value and the estimated value of the pressure value at a certain time point is higher than the predefined difference critical value, the pressure value at this time point is abnormal; if the difference does not exceed the difference critical value, the pressure value at this time point is not abnormal; Sum up the abnormal conditions of the pressure values at each time point after the defined period. After the defined period, starting from the time point when the first abnormal pressure value appears, sum up the number of abnormal pressure values within the defined time; if the ratio of the number of abnormal pressure values to the total number of all pressure values within the defined time exceeds the defined ratio critical value, the condition at the inlet of the turbine is in an abnormal state; if the ratio does not exceed the defined ratio critical value, the condition at the inlet of the turbine is in a normal state.

Images