CN-121997118-A - Method for processing real-time data fault time value based on interpolation algorithm

CN121997118ACN 121997118 ACN121997118 ACN 121997118ACN-121997118-A

Abstract

The invention discloses a processing method of a real-time data fault time value based on an interpolation algorithm, which comprises the following steps of step 1, pre-configuring an attribute set for each data acquisition point in a real-time database, step 2, acquiring the real-time data of the data acquisition points in step 1, judging whether the current data is the fault time value based on a preset rule, step 3, selecting the interpolation algorithm for the fault time value based on a multi-level decision mechanism when judging the fault time value, step 4, calculating a correction value according to the selected interpolation algorithm by utilizing effective historical data, step 5, verifying the validity of the correction value calculated in step 4, updating a quality code after verification and outputting the correction value, and step 6, storing the original fault time value and the correction value in an associated mode.

Inventors

LI YUGUI
LIU TENG
Sheng Xiangxin
YAO YU
WANG XUEYAN
GUO ZIYING
FAN GUOQIANG
GUO QIJUN

Assignees

中国大唐集团科学技术研究总院有限公司
北京庚顿数据科技有限公司
北京中恒瑞翔能源科技有限公司
大唐可再生能源试验研究院有限公司

Dates

Publication Date: 20260508
Application Date: 20251226

Claims (10)

1. The method for processing the real-time data fault time value based on the interpolation algorithm is characterized by comprising the following steps of: Step 1, modeling and attribute configuration of a real-time database, namely pre-configuring an attribute set for each data acquisition point in the real-time database; step 2, collecting real-time data of the data collection points in the step 1 and judging whether the current data is a fault time value or not based on a preset rule; step 3, an interpolation algorithm decision is carried out, namely when the fault time value is determined, an interpolation algorithm is selected for the fault time value based on a multi-level decision mechanism; step 4, interpolation calculation, namely calculating a correction value by using the effective historical data according to the interpolation algorithm selected in the step 3; Step 5, verifying and outputting the result, namely verifying the reasonability of the correction value calculated in the step 4, and updating and outputting the quality code after verification; and 6, storing the original fault time value and the corrected value passing verification in the step 5 in a correlated manner.
2. The method for processing real-time data fault duration based on interpolation algorithm according to claim 1, wherein the attribute set in step 1 includes a data type, a normal value range, a fault duration processing policy, an interpolation quality code identifier, a data step attribute, an effective data set size, a default interpolation algorithm, a fault quality code definition, and an acquisition period.
3. The method for processing the real-time data fault duration based on the interpolation algorithm according to claim 1, wherein the specific process of the step 2 is as follows: Step 2.1, real-time data acquisition, namely acquiring real-time data comprising a numerical value field and a quality code field from industrial field equipment through an industrial bus interface according to a preset acquisition period; and 2.2, performing parallel data verification and fault judgment, namely after the real-time data in the step 2.1 are obtained, performing verification and analysis operation simultaneously, and comprehensively judging whether the current real-time data is a fault time value or not based on the result, wherein the specific steps are as follows: step 2.2.1, checking the validity of the data, namely comparing the numerical value field of the real-time data with a normal value range predefined for the data point and checking the validity of the data format of the data; 2.2.2, analyzing a quality code field which conforms to an industry standard protocol in the real-time data to acquire a represented data acquisition state; 2.2.3, judging whether the current real-time data is a fault time value or not based on a preset rule; and 2.3, triggering the flow, namely executing the operation of the subsequent step 3 when the fault time value is determined.
4. A method of processing real-time data fault duration based on interpolation algorithm according to claim 3, wherein the preset rules of step 2.2.3 comprise at least one of the following conditions: the value of the real-time data exceeds a threshold range calculated based on the normal value range; The quality code of the real-time data is matched with the fault quality code identification; The real-time data is in an invalid floating point number format.
5. The method for processing real-time data fault duration based on interpolation algorithm according to claim 1, wherein the interpolation algorithm in step 3 comprises a linear interpolation algorithm, a forward neighbor interpolation algorithm, a multi-state weight voting interpolation algorithm, a weighted moving average interpolation algorithm, an adaptive exponential smoothing interpolation algorithm, and a mutation suppression interpolation algorithm.
6. The method for processing real-time data fault duration based on interpolation algorithm according to claim 5, wherein the specific process of selecting the interpolation algorithm for the fault duration based on the multi-level decision mechanism in step 3 is as follows: Step 3.1, screening quality codes, namely defining an effective data set according to the size setting of the effective data set, wherein the effective data set is specifically as follows: (1) Wherein, the The set of valid data is represented as such, The data representing the instant i is displayed, The selection condition is that the quality code of the moment data is GOOD, i represents a certain moment, t represents the effective sampling moment before the current fault moment, n represents a plurality of past moments continuously, and the value is the effective data set size in the acquisition point attribute; If the number of the screened effective historical data is smaller than a preset threshold value, adopting a default interpolation algorithm configured for the acquisition point, wherein the default interpolation algorithm is any one of a linear difference algorithm and a forward adjacent interpolation algorithm, and otherwise, entering a next-stage step attribute judging flow; step 3.2, judging the step attribute, namely selecting a forward adjacent interpolation algorithm if the step attribute of the acquisition point is ON, and entering a dynamic data classification flow if the step attribute of the acquisition point is OFF; step 3.3, dynamic data classification, namely selecting an algorithm according to the data type of the acquisition point; If the data type is the digital quantity type, the data type is further divided into BOOL and INT, wherein if the data type is BOOL, a forward adjacent interpolation algorithm is selected, and if the data type is INT, a multi-state weight voting interpolation algorithm is selected; And if the model is of an analog quantity type, calculating the relative change rate and the variation coefficient of the historical effective data, and dividing the data into a stable type, a fluctuation type and a mutation type based on the calculation results of the relative change rate and the variation coefficient, wherein a weighted moving average interpolation algorithm is selected if the model is of the stable type, an adaptive exponential smoothing interpolation algorithm is selected if the model is of the fluctuation type, and a mutation suppression interpolation algorithm is selected if the model is of the mutation type.
7. The method for processing real-time data fault duration based on interpolation algorithm according to claim 6, wherein the calculating the relative change rate and the variation coefficient of the historical effective data in step 3.3 divides the data into stationary type, fluctuation type and mutation type based on the calculation result of the relative change rate and the variation coefficient, and the specific procedures are as follows: Calculating a relative change rate and a variation coefficient, and fixing at least 5 nearest effective data points by adopting a sliding window; the relative rate of change was calculated as follows: (2) Wherein, the The average value is represented by a value of, The i-th value is represented, i represents a position index, and n represents the total number of numerical values; (3) Wherein, the Represents the average value, r% represents the relative rate of change, 、 The last time value and the last time value in the valid data point are respectively represented, Representing the time difference between two data points, t representing the last instant in the valid data points; the coefficient of variation is calculated as follows: (4) Wherein, the Represents standard deviation, n represents total number of numerical values, i represents position index, The value of the i-th is indicated, Represents an average value; (5) Wherein CV represents the coefficient of variation, The standard deviation is indicated as such, Represents an average value; And (3) classification judgment: Stationary coefficient of variation CV < C Percent, and relative rate of change r% < The%/sec; The wave type is one satisfying any one of the following conditions, and specifically comprises the following steps: Coefficient of variation C %≤CV<C % ; Relative rate of change: The ratio of%/second is less than or equal to r% < The%/sec; the mutant type is one satisfying any one of the following conditions, and specifically comprises the following: Coefficient of variation, CV > C % ; The relative change rate is r percent or more The%/sec.
8. The method for processing a real-time data fault duration based on an interpolation algorithm according to claim 6, wherein the linear difference algorithm has a specific calculation formula as follows: (6) Wherein r is the slope of the gradient, , An interpolated value representing the time of failure, A value representing the last time instant in the valid data point, t representing the last time instant in the valid data point; The forward adjacent interpolation algorithm has the following specific calculation formula: (7) Wherein, the An interpolated value representing the time of failure, A value representing the last time in the valid data point, t representing the value of the last time in the valid data point; The multi-state weight voting interpolation algorithm comprises the following specific calculation formula: (8) Wherein, the An interpolated value representing the time of failure, t representing the last time in the valid data points, Representing the corresponding input parameters that enable the entire summation expression to reach a maximum; n represents the number of historical data used for analysis; the weight of the ith historical state is represented, and the weight adopts a time attenuation mode; Is a Croneck function when When s is used, the number of the components is, When =1 When it is not equal to s, =0; I represents a position index; The specific calculation formula of the weighted moving average interpolation algorithm is as follows: (9) Wherein, the An interpolated value representing the time of failure, The value representing the last moment in the valid data point, Representing the value of the 2 nd time in the valid data point, A value representing the 3 rd time instant in the valid data point, The value at the 4 th-last time in the valid data points, A value representing the 5 th time from the last valid data point; t represents the last moment in the valid data points; The adaptive exponential smoothing interpolation algorithm has the following specific calculation formula: (10) Wherein, the The predicted value at time t +1 is indicated, The predicted value at the time t is indicated, The actual observed value at the time t is represented, and t represents the last time in the effective data points; representing the smooth coefficient, the value range is 0.3-less Less than or equal to 0.5, and is specifically as follows: (11) wherein CV represents the coefficient of variation, C Represents the minimum coefficient of variation, C Representing the maximum coefficient of variation; the mutation suppression interpolation algorithm comprises the following specific calculation formula: (12) Wherein, the An interpolated value representing the time of failure, The value representing the last moment in the valid data point, A value representing the 2 nd time in the valid data point, A value representing the 3 rd time of the last time in the valid data point, t representing the last time in the valid data point; median (..) represents the median function, taking the median; Is an attenuation factor, a factor between 0 and 1, specifically as follows: (13) Wherein max (.+ -.) represents a maximum value, min (.+ -.) represents a minimum value, and r% represents a relative change rate; The spatial consistency check function is expressed, specifically a discriminant function, and is output as one of True and False, specifically as follows: (14) Wherein, the The value representing the last moment in the valid data point, A value representing the 2 nd time in the valid data point; Representing a variation threshold; r% represents the relative rate of change, Representing the relative rate of change of the close proximity of the acquisition point of strong correlation; =true if the change rate difference between the local and neighboring devices is less than the threshold Judging that the spaces are consistent; =false, if the variation rate difference is equal to or greater than the threshold value And judging that the spaces are not consistent.
9. The method for processing a real-time data fault duration based on an interpolation algorithm according to claim 1, wherein the rationality verification in step 5 includes: checking whether the correction value is in a normal value range or not; And verifying trend consistency, namely calculating the deviation between the corrected value and the historical effective data trend, and if the deviation is smaller than a preset threshold value, verifying.
10. The method for processing real-time data fault duration based on interpolation algorithm according to claim 1, wherein the data double-track storage in step 6 is characterized in that an original fault duration record with a fault quality code and a correction value record with an interpolation quality code are stored in association with the same time stamp, and different query interfaces are provided, a default query returns the correction value record, and a specified query can return the original fault duration record.

Description

Method for processing real-time data fault time value based on interpolation algorithm Technical Field The invention belongs to the technical field of real-time data processing of power stations, and particularly relates to a processing method of a real-time data fault time value based on an interpolation algorithm. Background In the field of industrial monitoring of power systems, new energy power stations and the like, the systems rely on massive real-time data, such as telemetry data and remote signaling data, to make operation decisions. However, the devices such as the sensor, the instrument, the signal transmission channel and the like may malfunction, so that a "malfunction time value" of the collected data point occurs, and the malfunction time value refers to the collected measurement point value under the condition of malfunction of the measurement point signal measurement processing transmission device such as sensor malfunction, instrument malfunction, signal measurement and processing device malfunction, signal transmission channel malfunction and the like. If the abnormal data are directly applied to monitoring, analysis and diagnosis models, the accuracy of system judgment and the stability of operation are seriously affected. The existing data interpolation or filling technology has the defects that firstly, special attributes of electric power data, such as remote signaling/telemetry difference, quality code and step characteristic, are not fully considered, secondly, the algorithm is single in selection and is not adapted according to the dynamic characteristics of the data, so that the interpolation precision is insufficient, thirdly, a physical rule verification mechanism of an interpolation result is lacked. Thus, there is a need for a method that can dynamically, accurately, and in real-time process fault-duration values in conjunction with industrial data features. Disclosure of Invention The invention aims to provide a processing method of a real-time data fault time value based on an interpolation algorithm, which solves the problems of insufficient interpolation precision caused by single algorithm selection and unadapted data dynamic characteristics in the prior art. The technical scheme adopted by the invention is that the method for processing the real-time data fault time value based on the interpolation algorithm comprises the following steps: Step 1, modeling and attribute configuration of a real-time database, namely pre-configuring an attribute set for each data acquisition point in the real-time database; Step 2, collecting and detecting faults in real time, namely collecting real-time data of the data collection points in the step 1, and judging whether the current data is a fault time value or not based on preset rules; step 3, an interpolation algorithm decision is carried out, namely when the fault time value is determined, an interpolation algorithm is selected for the fault time value based on a multi-level decision mechanism; step 4, interpolation calculation, namely calculating a correction value by using the effective historical data according to the interpolation algorithm selected in the step 3; Step 5, verifying and outputting the result, namely verifying the reasonability of the correction value calculated in the step 4, and updating and outputting the quality code after verification; and 6, storing the original fault time value and the corrected value passing verification in the step 5 in a correlated manner. The present invention is also characterized in that, In the step 1, the attribute set comprises a data type, a normal value range, a fault time value processing strategy, an interpolation quality code identifier, a data step attribute, an effective data set size, a default interpolation algorithm, a fault quality code definition and a collection period. The specific process of the step 2 is as follows: Step 2.1, real-time data acquisition, namely acquiring real-time data comprising a numerical value field and a quality code field from industrial field equipment through an industrial bus interface according to a preset acquisition period; And 2.2, performing parallel data verification and fault judgment, namely after the real-time data in the step 2.1 are acquired, performing verification and analysis operation simultaneously, and comprehensively judging whether the current real-time data is a fault time value or not based on the result, wherein the specific steps are as follows: Step 2.2.1, checking the validity of the data, namely comparing a numerical value field of real-time data with a normal value range predefined for the data point and checking the validity of a data format of the data; 2.2.2, analyzing quality code fields conforming to an industry standard protocol in real-time data to acquire a represented data acquisition state; 2.2.3, judging whether the current real-time data is a fault time value or not based on a preset rule; and 2.3, trigge