Data quality analysis and efficiency improvement evaluation method

Abstract

The invention discloses a data quality analysis and efficiency improvement evaluation method. The method comprises the following steps: A, collecting data including unit operation data and wind power prediction tower data in a wind power plant; b, data cleaning: removing messy code data and over-limit data in a non-numerical form; c, data quality analysis; and D, efficiency improvement analysis: after data quality analysis, carrying out efficiency improvement analysis, and correcting an efficiency improvement analysis result according to a historical data analysis result. Data quality judgment mainly aims at several key factors, including wind direction, control parameters, wind speed, power and acquisition equipment, influencing the generating capacity, the judgment is carried out according to the control theory of the wind turbine generator, the judgment is simple and easy to understand and easy to operate, and meanwhile, the provided efficiency-improving evaluation method avoids the data quality problem as much as possible, so that the evaluation result is more persuasive; and a more objective and popularized method is provided for evaluating the power-increasing and efficiency-improving technical improvement effect of the wind turbine generator.

Description

Technical Field

[0001] The present invention relates to the technical field of data quality analysis and efficiency improvement evaluation, and particularly relates to a data quality analysis and efficiency improvement evaluation method. Background Art

[0002] With the continuous increase in installed capacity, the number of units with an operating life exceeding 10 years has gradually increased. With the development of technology, there are more and more projects for power augmentation transformation of in-service units, including but not limited to: blade type: installation of vortex generators, increase in impeller diameter; main control type: yaw correction, parameter self-seeking, etc. And the effect evaluation after power augmentation has become the focus of common concern for owners and technical renovation parties. The existing effect evaluation methods are as follows: 1) Install a wind measurement tower to measure the wind speed-power curve before and after the technical renovation for evaluation; 2) Compare the power curves before and after the technical renovation by using the method of calibrating the wind speed in the nacelle with an airborne nacelle radar; 3) Evaluate by comparing the power curve and power generation of the technical renovation unit by applying SCADA data. However, the above effect evaluation methods have the following problems: 1) Adding a wind measurement tower has high costs, a long cycle, and limited number of evaluation units; 2) Adding an airborne lidar has high costs; 3) The changes in the unit's hardware, software, acquisition equipment, etc. during the time period between before and after the technical renovation are not considered, and these changes will all affect the efficiency improvement results.

[0003] The Chinese patent "Method, System, Equipment and Medium for Evaluating the Power Generation Increase Rate before and after Unit Transformation" with the publication number: CN 114399402 A provides a method based on the analysis of SCADA system data. This method comprehensively considers various factors during the evaluation period before and after the technical renovation, but does not consider the situation between the evaluation time period before the renovation and the renovation time point.

[0004] Currently, the commonly used evaluation methods for technical renovation and efficiency improvement are single-dimensional comparison methods: comparing the data of the unit before and after the technical renovation, or two-dimensional comparison methods: comparing the data before and after the technical renovation of the reference unit and the technical renovation unit. These two methods only focus on the data quality of the data used for analysis, while ignoring the changes in the unit's data during the period when the data is used for analysis. For example, due to the adjustment and replacement of wind speed measurement equipment, the zero adjustment of wind direction measurement equipment, the change of control parameters, the change of power data, and some abnormal operations of the acquisition system, etc. These changes will all affect the power generation of the unit, and the key point of the power augmentation effect evaluation is exactly the evaluation of power generation. Therefore, it is necessary to identify the interference in this aspect during the efficiency improvement analysis and compensate it in the evaluation result to make the evaluation result truly reflect the effect of the technical renovation factor. Summary of the Invention

[0005] In view of the technical problems mentioned in the background art, the present invention provides a data quality analysis and efficiency improvement evaluation method.

[0006] The above technical objectives of the present invention are achieved through the following technical solutions: A data quality analysis and efficiency evaluation method, comprising the following steps:

[0007] A. Collect data:

[0008] Unit operation data: The data collection period is at least half a year after the technical transformation and at least one year before the technical transformation; the data tags include: wind speed, power, generator speed, pitch angle, total power generation, power loss, unit operation flag, yaw error, absolute wind direction, atmospheric temperature and other unit operation-related data; the units involved are all units in the field;

[0009] Wind power prediction tower data in wind farms: wind speed and direction data at different levels;

[0010] B. Data cleaning: Eliminate garbled data and over-limit data in non-numeric form; over-limit data include: data other than wind speed 0-50m / s, data of power 0-1.3 times the rated power, data other than pitch angle -5 to 95 degrees, and data other than speed 0 to 1.2 times the rated speed;

[0011] C. Data quality analysis;

[0012] D. Efficiency analysis: After data quality analysis, efficiency analysis is conducted. The results of the efficiency analysis need to be revised based on the results of historical data analysis.

[0013] Preferably, in step C, the data quality analysis adopts a method of sliding forward comparison of monthly data, and qualitatively analyzes the availability of data of the technical transformation unit based on the analysis results, and quantifies its impact so as to compensate for it in the evaluation results: the method of sliding forward comparison of monthly data is to compare the data of the month before the technical transformation node and the data of the two months before the technical transformation, and compare the data of multiple months in a month-by-month manner according to the time reduction;

[0014] The comparison dimensions include: wind direction data consistency, control parameter consistency, wind speed data consistency, wind direction equipment calibration consistency, power data consistency; mainly through the unit characteristic relationship, the characteristics include: wind speed-power relationship, power-pitch angle relationship, wind speed-generator speed relationship, generator speed-power relationship, power curves in different yaw angle ranges, wind power prediction tower wind speed and unit wind speed correlation, wind direction distribution, etc.

[0015] Wind direction data consistency: includes two aspects: one is that after monthly analysis of historical data, if the wind direction distribution of a single unit has not changed, it means that the unit has not been re-calibrated to the north during this period. In addition, whether the units of the entire wind farm have been re-calibrated to the north during installation: through horizontal comparison, and at the same time, the wind power prediction tower wind direction is corrected;

[0016] Consistency of control parameters: Changes in control parameters include changes in rotational speed parameters, changes in control PI parameters, changes in pitch angle control strategies, etc.; qualitative analysis is carried out through the scatter plot of rotational speed - power and power - pitch angle. Since the relationship between rotational speed - power and power - pitch angle is not affected by the environment, etc., and their relationship is entirely the effect of control parameters. The comparison of monthly data of rotational speed - power sliding forward is mainly to find out whether there are changes in the generator rotational speed and PI parameter control strategies. The comparison of monthly data of power - pitch angle sliding forward is mainly to find out whether there are changes in the optimal pitch angle control strategy. Its changes will affect both the power generation and the power curve, and will also affect the evaluation effect. By comparing historical data month by month, it can be observed from the scatter plot whether the data characteristics have changed between the evaluation data segment before the technical transformation and the data evaluation segment after the technical transformation. If there are changes, they need to be considered and corrected during the evaluation;

[0017] Consistency of wind speed data: The correlation between the data of the anemometer tower and the data of the unit is mainly used, supplemented by the statistics of the correlation between the technical transformation unit and the reference unit;

[0018] Consistency of wind direction equipment calibration: That is, whether the wind direction measuring equipment is re - zeroed or changed, etc.: The wind speed - power curve is compared in intervals according to the yaw error;

[0019] Consistency of power data: Changes in power characteristics caused by internal abnormalities or abnormal operations in the data acquisition system are judged by the power generation to determine whether the power data characteristics have changed.

[0020] Preferably, in the step C, the data quality analysis adopts the method of comparing the power curve sliding forward month by month and the method of comparing the power generation sliding forward month by month;

[0021] The method of comparing the power curve sliding forward month by month means that after the data is secondarily cleaned using the quartile method, the fitted wind speed - power curve is obtained, and the ratio C of the power generation of the two months for comparison is calculated respectively using the wind frequency distribution of each of the previous and next months. 1综合 =E 综合月2 / E 综合月1 , because the wind resources are seasonal, there will be differences in the calculation results of single - unit monthly forward sliding. The same - month results of all units in the whole field are regular. Therefore, the average value of the calculated ratios of all units in the whole field per month is statistically calculated. If this value exceeds the threshold, and the threshold is [-1%, 1%], then the wind speed - power characteristic curve of this unit has changed. If this unit is a technical transformation unit, this situation needs to be considered in the results. If this unit is a non - technical transformation unit, this unit cannot be used as a reference unit;

[0022] The method of sliding forward comparison of power generation month by month: According to the monthly power generation and power loss data of each unit in the collected statistical data, the monthly power generation of each unit is obtained. The same formula C is used for this value as in the previous step. 1综合 =E 综合月2 / E 综合月1 Compare with the method.

[0023] Preferably, in step D, the efficiency improvement method of the technical transformation unit is determined by referring to the unit, and the formula is as follows:

[0024] C 11 =((E 1后 -E 2后 ) / (E 1前 -E 2前 )-1)*100%

[0025] Among them C 11 To increase the proportion of power generation for the technical transformation effect;

[0026] E 1后 : Power generation of the transformed unit after technical transformation;

[0027] E 2后 : Power generation of the reference unit after technical transformation;

[0028] E 1前 : Power generation of the upgraded unit before upgrading;

[0029] E 2前 : Power generation of the reference unit before technical transformation;

[0030] There are two ways to calculate power generation: power curve wind frequency distribution calculation method and statistical actual power generation calculation method;

[0031] The settlement results of the method of calculating power generation using power curves and actual wind frequency distribution need to be revised based on the evaluation results of historical data.

[0032] Preferably, in step D, the actual power generation increase ratio before and after the technical transformation is C 21 , the improvement ratio C calculated by the power generation calculated by the power curve before and after the technical transformation 22 Then the actual power generation increase ratio and the power curve calculate the difference C of power generation 23 =C 21 -C 22 , C 23 This means the increase in power generation due to non-efficiency-enhancing transformation. If C 23 >=0, and C 21 and C 22 The value of is greater than 0 after correction based on the quantitative results of historical data analysis, and this efficiency-enhancing technical reform is effective.

[0033] C 21 Calculation method of

[0034] C 22 Calculation method of

[0035] E 21前 represents the actual power generation + loss power counted by the monitoring system before the technical transformation of the unit;

[0036] E 21后 represents the actual power generation + loss power counted by the monitoring system after the technical transformation of the unit;

[0037] E 200 represents the power generation calculated from the designed power curve of the unit to be technically transformed and the wind frequency distribution during the period before the technical transformation of the wind power prediction tower;

[0038] E 201 represents the power generation calculated from the designed power curve of the unit to be technically transformed and the wind frequency distribution during the period after the technical transformation of the wind power prediction tower;

[0039] E 22前 The calculation method is the power generation of the fitted power curve after the technical transformation of the unit and the actual wind frequency distribution of the unit before the technical transformation;

[0040] E 22后 The calculation method is the power generation of the fitted power curve after the technical transformation of the unit and the actual wind frequency distribution of the unit after the technical transformation;

[0041] E 22前 and E 22后 For the wind speed data involved in the calculation, the wind speed of the unit or the wind speed of the anemometer tower is selected according to the data quality analysis results.

[0042] In summary, the present invention mainly has the following beneficial effects: The present invention provides a data quality analysis and efficiency improvement evaluation method. The judgment of data quality is mainly aimed at several key factors affecting power generation: wind direction, control parameters, wind speed, power, and acquisition equipment, and is carried out based on the control theory of wind turbines, which is simple, easy to understand, and easy to operate. At the same time, the provided efficiency improvement evaluation method tries to avoid data quality problems as much as possible, making the evaluation results more persuasive; it provides a more objective and popularizable method for evaluating the effect of power increase and efficiency improvement of wind turbines. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is the overall flow chart of the present invention;

[0044] Figure 2 is the logic diagram of data quality analysis; DETAILED DESCRIPTION OF THE INVENTION

[0045] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] See also Figure 1 and Figure 2 The present invention provides a data quality analysis and efficiency evaluation method, comprising the following steps:

[0047] A. Collect data:

[0048] Unit operation data: The data collection period is at least half a year after the technical transformation and at least one year before the technical transformation (preferably three years); data tags include: wind speed, power, generator speed, pitch angle, total power generation, power loss, unit operation flag, yaw error, absolute wind direction, atmospheric temperature and other unit operation-related data; the units involved are all units in the field;

[0049] Wind power prediction tower data in wind farms: wind speed and direction data at different levels;

[0050] B. Data cleaning: Eliminate garbled data and over-limit data in non-numeric form; over-limit data include: data other than wind speed 0-50m / s, data of power 0-1.3 times the rated power, data other than pitch angle -5 to 95 degrees, and data other than speed 0 to 1.2 times the rated speed; eliminate blank data; eliminate duplicate time data: only one duplicate time data caused by data storage and export is required;

[0051] C. Data quality analysis: Analyze the unit characteristic data by using the monthly data sliding forward comparison method from the technical transformation time node to the present, and conduct data quality analysis; the availability of wind direction data; whether the control parameters have changed; whether the wind speed parameters have changed; whether the wind direction measurement equipment has been recalibrated or changed; changes in power characteristics caused by abnormalities or abnormal operations within the data acquisition system, that is, conduct wind direction data consistency, control parameter consistency, wind speed data consistency, wind direction equipment calibration consistency, and power data consistency analysis;

[0052] The monthly data sliding forward method is to compare the data of the month before the technological transformation node and the data of the two months before the technological transformation, and compare multiple months by decreasing month by month in time. For example, if the technological transformation node is 01 / 01 / 2022, then the data of December 2021 (01 / 12 / 2021 - 12 / 31 / 2021) and November 2021 (01 / 11 / 2021 - 11 / 30 / 2021) are compared before the technological transformation, the data of December 2021 minus one month (01 / 11 / 2021 - 11 / 30 / 2021) and November 2021 minus one month (01 / 10 / 2021 - 10 / 31 / 2021) are compared, … the data of December 2021 minus N months and November 2021 minus N months are compared, that is, on the basis of 01 / 12 / 2021 - 12 / 31 / 2021 and 01 / 11 / 2021 - 11 / 30 / 2021, they are simultaneously slid and decreased, and the data of N + 1 pairs of two months before and after are compared.

[0053] Wind direction data consistency: Since the final efficiency improvement result is analyzed based on the data of the main wind direction, it is necessary to judge the quality of the main wind direction data of the unit. First, perform secondary data cleaning: exclude the abnormal wind direction data caused by the abnormal alignment of the unit due to human operation: eliminate it through the quartile method based on the relationship between power and yaw error, and then statistically analyze the wind direction distribution of each unit in the wind farm month by month, and make a horizontal comparison of the monthly statistical results to check whether the difference in the main wind direction of each unit in the same month exceeds the threshold. The threshold is [-22.5, 22.5]. If it exceeds this threshold, it is judged that the main wind direction data of the unit is unavailable. If this situation only occurs in a certain month and does not occur in other months, and the month in which it occurs does not belong to the time of pre- and post-technological transformation evaluation, it can be ignored. If this situation still exists until the technological transformation time node, then the wind direction data in the post-technological transformation evaluation data needs to be corrected. The correction method is to offset the wind direction data of the abnormal month with the wind direction data before the abnormal month. If the main wind direction of a certain unit is inconsistent with that of other units, then offset the wind direction data of this unit with the main wind direction data of other units; in addition, offset and correct the wind direction data of all units in the wind farm with the wind direction of the wind power prediction tower.

[0054] Control parameter consistency: Changes in control parameters include changes in rotational speed parameters, control PI parameters, pitch angle control strategies, etc.; qualitative analysis is carried out through speed-power scatter plots and power-pitch angle. Since the speed-power relationship and power-pitch angle relationship are not affected by the environment, their relationships are entirely the result of control parameters. By using the method of sliding forward monthly data before the technical transformation node to draw scatter plots simultaneously, it is convenient and intuitive to find out whether their characteristics have changed. If there is an obvious deviation in the data of two months from the comparison scatter plot, it indicates that the control parameters have changed. Or through statistical data analysis, the generator speed is segmented every 100 rpm, and then the average value of the power within each speed segment is obtained after cleaning the data using the quartile method for comparison. If the percentage of the mean difference in the mean exceeds the threshold, and the threshold is [-5%, 5%], it is considered that the parameters have changed, and this change will affect the power generation of the unit. Therefore, it should be considered in the efficiency improvement evaluation. If this situation exists in all units of the whole plant, then in the efficiency improvement evaluation of the technical transformation unit, evaluation method one is selected, and the reference unit method is used to ensure that the influence of this aspect is excluded from the technical transformation effect. If only the technical transformation unit has this situation, then compare the percentage change in the power generation of other units that have not undergone this change before and after the node where this change occurs, screen the data according to the quartile method, and calculate the average value C of the screened data. 月 , and then statistically calculate the percentage change C’ in the power generation of the technical transformation unit. 月改 , use their difference C 修 = C’ 月改 - C 月 to correct the final efficiency improvement analysis;

[0055] The monthly comparison of power-pitch angle data by sliding forward is mainly to find out whether the optimal pitch angle control strategy has changed. Its change will affect both the power generation and the power curve, and will also affect the evaluation effect. By comparing historical data month by month, it can be observed from the scatter plot whether the data characteristics have changed between the pre-technical transformation evaluation data segment and the post-technical transformation data evaluation segment. If there is a change, it needs to be considered and corrected during the evaluation. The specific method:

[0056] Extract the wind speed-power curves of the changed month and the previous month, then calculate the power generation, calculate the power generation increase ratio, and at the same time calculate the difference from the average value of the increase ratios of other units that have not undergone characteristic changes, and use the difference to correct the final power generation efficiency improvement ratio.

[0057] Wind speed data consistency: The correlation between the anemometer tower data and the unit data is mainly supplemented by the statistics of the correlation between the technical transformation units and the non-technical transformation units;

[0058] Use wind tower data for analysis: First, filter out the data of the main wind direction, and analyze the wind speed correlation between the technical transformation unit and the wind tower data by sliding the data forward month by month: that is, perform a straight line fit between the wind speed of the technical transformation unit and the wind speed data of the wind tower to obtain the fitting parameters a, b and the correlation coefficient r 2 If the difference between the two-month comparisons exceeds a certain limit, it is considered that the wind speed characteristics have changed. The threshold is: a: [-0.1, 0.1], r 2 :[-0.3,0.3], if the wind speed trend is decreasing, the wind speed applied in method 1 and method 2 will be changed to the wind speed of the wind tower, avoiding the change of wind speed characteristics of the technical transformation unit.

[0059] The wind speed correlation analysis is performed on the technically upgraded units and other units in the wind farm that have not been technically upgraded. The technically upgraded units and five other units are randomly selected, and the wind speed correlation coefficients a, b and correlation coefficient r of the selected units and the remaining units in the wind farm are calculated month by month. 2 , using the statistical results of the five units for two months before and after 2 The average value of the difference is used as the standard. If the difference between the two months before and after the technical transformation unit exceeds this average value, it means that the wind speed characteristics of the unit have changed. It is necessary to use the a and b values ​​of the unit with the highest correlation coefficient before the change and the technical transformation unit to correct the wind speed data after the change.

[0060] Wind direction equipment calibration consistency: that is, whether the wind direction measuring equipment has been recalibrated or changed, etc.: The wind speed-power curve is compared by yaw error in different intervals; the power generation and total power generation ratio within different yaw error ranges of the units with technical transformation and the units without technical transformation are analyzed, and the analysis results of each month before the technical transformation time node are statistically analyzed. The average value of the longitudinal statistical results of each unit is calculated. If it exceeds the average value of +-1%, it is considered that the wind direction equipment of this unit has changed in this month, and this unit cannot be used for technical transformation effect analysis. If this unit happens to be a technical transformation unit, then when evaluating the effect of the technical transformation, method 1 is used to calculate the difference between the calculated efficiency improvement ratio and this deviation value as the final efficiency improvement ratio, and method 2 is C 23 The value of is not affected, because both the power generation and power curves can reflect its increased value, and the difference between the two cancels it out.

[0061] Power data consistency analysis: Changes in power data characteristics caused by abnormalities or abnormal operations within the data acquisition system: First, perform a time series analysis on the total power generation data, and then use the power generation to determine whether its power data characteristics have changed;

[0062] Judgment using total power generation data: First, determine whether there are steps or power outages in the total power generation data through time series data. In the absence of such situations, the total power generation data can be used to review the power data. The method is as follows: Perform a sliding difference on the total power generation time series data. First, calculate the time difference between two time points. If the time is 10 minutes, subtract the power generation at the previous data time point from the power generation at the next data time point, and record the result as EE 10 , which represents the integral of the second-level power data over time within 10 minutes. Then multiply the power value at this moment by 10 minutes and divide by 60, and the result E 10 is the integral of the 10-minute average power over time corresponding to this 10 minutes. (EE 10 - E 10 ) / EE 10 The deviation does not exceed [-5%, 5%]. If it exceeds, it means this power value is abnormal. If the number of consecutive abnormal data points is less than 5 and the non-consecutive state appears, it can be excluded. If the power data of 30 consecutive points is abnormal, the quotient of the total power generation value divided by time needs to be used instead. This method is used when screening data in the subsequent evaluation methods 1 and 2.

[0063] Another comprehensive evaluation method is the way of sliding and comparing the power curve month by month forward, and the way of sliding and comparing the power generation month by month forward;

[0064] The way of sliding and comparing the power curve month by month forward means that after applying the quartile method to the data for secondary cleaning, the fitted wind speed-power curve is obtained. Calculate the ratio C of the power generation of the two months for comparison using the wind frequency distributions of the previous and next months respectively 1综合 = E 综合月2 / E 综合月1 . Since the wind resources are seasonal, the calculation results of single-unit sliding forward month by month will be different. The whole-field units have regularity in the same month. Therefore, calculate the average value of the calculated ratios of the whole-field units each month. If this value exceeds the threshold, the threshold is [-1%, 1%], then the wind speed-power characteristic curve of this unit has changed. If this unit is a technical renovation unit, this situation needs to be considered in the results. If this unit is not a technical renovation unit, this unit cannot be used as a reference unit;

[0065] The way of sliding and comparing the power generation month by month forward: Calculate the expected power generation of each unit each month based on the power generation and loss power data of each unit each month in the collected statistical data. Compare this value according to the same formula C 1综合 = E 综合月2 / E 综合月1 and the method.

[0066] D. Efficiency analysis: After data quality analysis, efficiency analysis is conducted. The results of the efficiency analysis need to be revised based on the results of historical data analysis.

[0067] There are two methods for efficiency analysis:

[0068] Method 1: Determine the efficiency improvement method of the technical transformation unit by referring to the unit. The formula is as follows:

[0069] C 11 =((E 1后 -E 2后 ) / (E 1前 -E 2前 )-1)*100%

[0070] Among them C 11 To increase the proportion of power generation for the technical transformation effect;

[0071] E 1后 : Power generation of the transformed unit after technical transformation;

[0072] E 2后 : Power generation of the reference unit after technical transformation;

[0073] E 1前 : Power generation of the upgraded unit before upgrading;

[0074] E 2前 : Power generation of the reference unit before technical transformation;

[0075] There are two ways to calculate power generation: power curve wind frequency distribution calculation method and statistical actual power generation calculation method;

[0076] The settlement results of the method of calculating power generation using power curve and actual wind frequency distribution need to be corrected according to the evaluation results of historical data:

[0077] First, the data is cleaned twice: data outside the main wind direction is screened out, data with abnormal power data is screened out, and if the wind speed data is abnormal, the corrected wind speed data is used. Then, the power curve is fitted for all units of the same model in the field, that is, the wind speed is divided into bins according to 0.5m / s, and the quartile method is used to eliminate abnormal data. Then, the power curves of the technically modified units and the reference units before and after the technical modification are fitted, and the wind frequency distribution of the units is statistically calculated. Then, the corresponding power generation is calculated using the power generation calculation formula:

[0078]

[0079] AEP – Annual Electricity Production;

[0080] Nh – number of hours of the analysis period;

[0081] N —— The number of intervals;

[0082] V i —— The standardized average wind speed of the i-th interval;

[0083] P i —— The standardized average output power of the i-th interval;

[0084] F(V) —— The wind speed frequency distribution function.

[0085] The power curve fitting in this step includes the power curve fitting of all units during the comparison period before the technical transformation, the power curve fitting of units without abnormal key factors found in the year before the technical transformation, and the power curve fitting of units after the technical transformation;

[0086] Selection of reference units: 1) Conduct a correlation analysis on the wind speeds of the units with technical transformation and those without technical transformation during the period before the technical transformation when no abnormal key factors are found, and obtain the correlation coefficient r 2 ; 2) According to the power curve fitting of units without abnormal key factors found in the year before the technical transformation in the previous step, determine the deviation between the power values of the power curve of the units with technical transformation and those of other units at each wind speed section, and calculate the ratio of this deviation to the power value; 3) Prioritize selecting units with a larger wind speed correlation coefficient and a smaller power deviation ratio as reference units, and then select units with both the correlation coefficient and the power deviation ratio in the middle position. If no reference units meeting the above requirements are found, prioritize using the top three units with a smaller ratio of power curve deviation values as the reference unit combination, and the power generation of the reference units of the units before and after the technical transformation corresponding to this is the average value of the three units;

[0087] The power generation improvement ratio calculated in this way needs to be corrected according to the results of the forward analysis of monthly data sliding, and the corrected value is used as the actual improvement result of the unit;

[0088] Applying the actual power generation calculation method can avoid the impact caused by changes in wind speed equipment. At the same time, if there is non-technical transformation improvement in other aspects according to the analysis of historical data, the results need to be corrected. The selection method of reference units is the same as above, specifically as follows:

[0089] Data screening: Screen out data outside the main wind direction, screen out data with abnormal power data, screen out data with power curtailment, and screen out data with yaw error not within the range of +-10 degrees;

[0090] Calculation of actual comparison power generation: Extract data points where the power of the units with technical transformation and the reference units at the same moment before or after the technical transformation is greater than 0 and the power generation value calculated by taking the difference in the total power generation between the upper and lower moments is within the range of [0, P N / 6], P NIt represents the rated power of the unit, and then groups are made every 100 kW according to their respective power values. The data points of the retrofitted units and non-retrofitted units in the same group are extracted to ensure that the power characteristics of the retrofitted units and the retrofitted units before are not too different and do not affect the results.

[0091] Sum the power and power generation values of the extracted data points to obtain the following values:

[0092] P 1后 : The sum of the powers of the retrofitted units that meet the requirements after retrofitting, E 1后 =P 1后 *10 / 60;

[0093] P 2后 : The sum of the powers of the reference units that meet the requirements after retrofitting, E 2后 =P 2后 *10 / 60;

[0094] P 1前 : The sum of the powers of the retrofitted units that meet the requirements before retrofitting, E 1前 =P 1前 *10 / 60;

[0095] P 2前 : The sum of the powers of the reference units that meet the requirements before retrofitting, E 2前 =P 2前 *10 / 60;

[0096] E 11后 : The sum of the power generations of the retrofitted units that meet the requirements after retrofitting, E 1后 =E 11后 ;

[0097] E 21后 : The sum of the power generations of the reference units that meet the requirements after retrofitting, E 2后 =E 21后 ;

[0098] E 11前 : The sum of the power generations of the retrofitted units that meet the requirements before retrofitting, E 1前 =E 11前 ;

[0099] E 21前 : The sum of the power generations of the reference units that meet the requirements before retrofitting, E 1前 =E 21前 ;

[0100] The power generations calculated above can all be used to substitute into the formula for calculation, and the results are the same.

[0101] This method avoids the influence of abnormal wind speed and power, and the influence on the control algorithm and the correction of the wind direction device needs to be corrected according to the quantization value that slides forward month by month with the data.

[0102] Method 2: The actual power generation increase ratio C before and after the technical transformation 21 , the increase ratio C calculated from the power generation calculated by the power curve before and after the technical transformation 22 , then the difference C between the actual power generation increase ratio and the power generation calculated by the power curve 23 = C 21 - C 22 , C 23 That is, it represents the power generation increase ratio caused by non-efficiency improvement transformation. If C 23 >= 0, and the values of C 21 and C 22 are both greater than 0 after being corrected by applying the historical data analysis quantization result, take C 22 the corrected result as the power generation increase ratio. If C 23 < 0, there is no improvement effect.

[0103] Calculation method of C 21 :

[0104] Calculation method of C 22 :

[0105] E 21前 represents the actual power generation + lost power counted by the monitoring system before the technical transformation of the unit;

[0106] E 21后 represents the actual power generation + lost power counted by the monitoring system after the technical transformation of the unit;

[0107] E 200 represents the power generation calculated from the designed power curve of the unit after the technical transformation and the wind frequency distribution in the time period before the technical transformation of the wind power prediction tower. The method is the same as before;

[0108] E 201 represents the power generation calculated from the designed power curve of the unit after the technical transformation and the wind frequency distribution in the time period after the technical transformation of the wind power prediction tower. The method is the same as before;

[0109] E 22前 The calculation method is the power generation of the fitted power curve after the technical transformation of the unit and the actual wind frequency distribution before the technical transformation of the unit. The method is the same as before;

[0110] E 22后 The calculation method is the power generation of the fitted power curve after the technical transformation of the unit and the actual wind frequency distribution after the technical transformation of the unit. The method is the same as before;

[0111] E 22前 and E 22后When calculating, the wind speed data involved is selected as the wind speed of the unit or the wind speed of the anemometer tower according to the data quality analysis results;

[0112] This calculation method avoids the influence on the qualitative analysis and evaluation result C when data characteristics such as wind speed, power, control parameters, and wind direction equipment calibration change. 23 If the actual power generation data is abnormal, the result needs to be corrected by using the quantitative analysis of historical data.

[0113] The above has described the embodiments of the present invention in detail with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, without departing from the principle and spirit of the present invention, various changes, modifications, substitutions, and variations to these embodiments still fall within the protection scope of the present invention.

Claims

1. A method for data quality analysis and efficiency improvement evaluation, characterized in that: The following steps are involved: A. Collect data: Unit operation data: The data collection period is at least half a year after the technical transformation and at least one year before the technical transformation; the data tags include: wind speed, power, generator speed, pitch angle, total power generation, power loss, unit operation flag, yaw error, absolute wind direction, atmospheric temperature and other unit operation-related data; the units involved are all units in the field; Wind power prediction tower data in wind farms: wind speed and direction data at different levels; B. Data cleaning: Eliminate garbled data and over-limit data in non-numeric form; over-limit data include: data other than wind speed 0-50m / s, data of power 0-1.3 times the rated power, data other than pitch angle -5 to 95 degrees, and data other than speed 0 to 1.2 times the rated speed; C. Data quality analysis; D. Efficiency analysis: After data quality analysis, efficiency analysis is conducted. The results of the efficiency analysis need to be revised based on the results of historical data analysis.

2. The method for data quality analysis and efficiency improvement evaluation according to claim 1, characterized in that: In step C, the data quality analysis adopts the method of sliding forward comparison of monthly data, and qualitatively analyzes the availability of data of the technical transformation unit based on the analysis results, and quantifies its impact so as to compensate for it in the evaluation results: the method of sliding forward comparison of monthly data is to compare the data of the month before the technical transformation node and the data of the two months before the technical transformation, and compare the data of multiple months in time on a monthly basis; The comparison dimensions include: wind direction data consistency, control parameter consistency, wind speed data consistency, wind direction equipment calibration consistency, power data consistency; mainly through the unit characteristic relationship, the characteristics include: wind speed-power relationship, power-pitch angle relationship, wind speed-generator speed relationship, generator speed-power relationship, power curves in different yaw angle ranges, wind power prediction tower wind speed and unit wind speed correlation, wind direction distribution, etc. Wind direction data consistency: includes two aspects: one is that after monthly analysis of historical data, if the wind direction distribution of a single unit has not changed, it means that the unit has not been re-calibrated to the north during this period. In addition, whether the units of the entire wind farm have been re-calibrated to the north during installation: through horizontal comparison, and at the same time, the wind power prediction tower wind direction is corrected; Control parameter consistency: Control parameter changes include speed parameter changes, PI parameter changes, pitch angle control strategy changes, etc. Qualitative analysis is performed through speed-power scatter plots and power-pitch angles, because the speed-power relationship and power-pitch angle relationship are not affected by the environment, etc., and their relationship is entirely the effect of control parameters. The comparison of the speed-power data sliding forward on a monthly basis is mainly to find out whether the generator speed and PI parameter control strategy have changed. The comparison of the power-pitch angle data sliding forward on a monthly basis is mainly to find out whether the optimal pitch angle control strategy has changed. Its change will affect both the power generation and the power curve, and will affect the evaluation effect. By comparing the historical data on a monthly basis, it can be observed from the scatter plot whether its data characteristics have changed between the evaluation data segment before the technical transformation and the data evaluation segment after the technical transformation. If changes have occurred, they need to be considered and corrected during the evaluation; Wind speed data consistency: mainly based on the correlation between wind tower data and unit data, supplemented by statistics on the correlation between technical transformation units and reference units; Wind direction equipment calibration consistency: whether the wind direction measuring equipment has been recalibrated or changed, etc.: compare the wind speed-power curve according to the yaw error interval; Power data consistency: The power characteristics change due to abnormalities or abnormal operations within the data acquisition system. The power generation is used to determine whether the power data characteristics have changed.

3. The method for data quality analysis and efficiency improvement evaluation according to claim 1, characterized in that: In step C, the data quality analysis adopts a method of sliding forward comparison of power curves month by month and a method of sliding forward comparison of power generation month by month; The power curve is compared by sliding forward month by month, that is, the data is cleaned twice by applying the quartile method to obtain the fitted wind speed-power curve, and the ratio of the power generation of the two months before and after is calculated by using the wind frequency distribution of the two months before and after. 1综合 =E 综合月2 / E 综合月1 ,Since wind resources are seasonal, the calculation results of a single unit sliding forward month by month will be different. The same month for all units in the field has regularity, so the average value of the calculation ratio of all units in the field is calculated each month. If this value exceeds the threshold, the threshold is [-1%, 1%], then the wind speed-power characteristic curve of this unit changes. If this unit is a technical transformation unit, this situation needs to be considered in the result. If this unit is a non-technical transformation unit, this unit cannot be used as a reference unit; The method of sliding forward comparison of power generation month by month: According to the monthly power generation and power loss data of each unit in the collected statistical data, the monthly power generation of each unit is obtained. The same formula C is used for this value as in the previous step. 1综合 =E 综合月2 / E 综合月1 Compare with the method.

4. The method for data quality analysis and efficiency improvement evaluation according to claim 1, characterized in that: In step D, the efficiency improvement method of the technical transformation unit is determined by referring to the unit, and the formula is as follows: C 11 = ((E 1后 - E 2后 ) / (E 1前 - E 2前 ) - 1) * 100% Among which C 11 is the increased proportion of power generation due to the technical renovation effect; E 1后 : The power generation of the technical transformation unit after technical transformation; E 2后 : The power generation of the reference unit after the technical transformation; E 1前 : The power generation of the technical transformation unit before the technical transformation; E 2前 : The power generation of the reference unit before the technical transformation; There are two ways to calculate power generation: power curve wind frequency distribution calculation method and statistical actual power generation calculation method; The settlement results of the method of calculating power generation using power curves and actual wind frequency distribution need to be revised based on the evaluation results of historical data.

5. The method for data quality analysis and efficiency improvement evaluation according to claim 1, characterized in that: In the said step D, the improvement ratio C of the actual power generation before and after the technical transformation 21 , the improvement ratio C calculated from the power generation calculated by the power curve before and after the technical transformation 22 , then the difference C between the improvement ratio of the actual power generation and the power generation calculated by the power curve 23 = C 21 - C 22 , C 23 That is, it represents the improvement ratio of power generation caused by non-efficiency improvement transformation. If C 23 >= 0, and the values of C 21 and C 22 are both greater than 0 after being corrected by the quantitative analysis results of the application historical data, this efficiency improvement technical transformation is effective. C 21 Calculation method: C 22 Calculation method: E 21前 Indicates the actual power generation + loss of power counted by the monitoring system before the unit's technical transformation; E 21后 Represents the actual power generation + loss power counted by the monitoring system after the unit's technical transformation; E 200 It represents the generated electricity calculated from the designed power curve of the technical renovation unit and the wind frequency distribution during the period before the technical renovation of the wind power prediction tower. E 201 Indicates the generated electricity calculated from the designed power curve of the technical renovation unit and the wind frequency distribution during the time period after the technical renovation of the wind power prediction tower; E 22前 The calculation method is the generated electricity of the fitted power curve after the unit's technological transformation and the actual wind frequency distribution of the unit before the technological transformation; E 22后 The calculation method is the generated electricity of the fitted power curve after the unit's technological transformation and the actual wind frequency distribution of the unit after the technological transformation; E 22前 and E 22后 When calculating, the wind speed data involved is selected as the wind speed of the unit or the anemometer tower wind speed according to the data quality analysis results.

Images