A method and device for cutting a thermal power unit operating condition

Abstract

A method and device for cutting a thermal power unit operation condition, the method comprising: obtaining a thermal power unit operation data, and fitting a main line according to the operation data; calculating one or more characteristic curves according to the determined main line curve and operation data; selecting one or more characteristic curves according to different operation conditions to combine, to obtain a trend characteristic combination curve; and performing a cutting processing operation on the thermal power unit operation condition according to the trend characteristic combination curve.

Description

Technical Field

[0001] This application relates to the field of thermal power unit operation optimization technology, and more particularly to a method and apparatus for dividing the operating conditions of thermal power units. Background Technology

[0002] As a typical process industry, thermal power plants possess a large amount of direct production process data. The thermal power unit operation data condition segmentation algorithm segments operating conditions into selectable time periods and algorithms. Through efficient analysis of parameter settings and algorithm selection for operating conditions, condition segmentation can be performed on single or combined operating conditions. Logical operations can be performed between operating conditions, with AND and OR relationships involved in the calculation. This allows for the acquisition of operating condition information and its time periods under certain conditions, and the results can be visualized. Operators and maintenance personnel can instantly and conveniently connect to the real-time historical data storage server based on the power plant's real-time operating conditions. They can retrieve process data, filter out information with reference value or very similar operating conditions, and comprehensively analyze and fully utilize this data through condition segmentation and visualization analysis tools. Based on the calculation results, operating parameters can be evaluated and corrected to optimize control strategies. When optimal operating conditions are identified through operating condition analysis tools, and further analysis, judgment, and decision-making are conducted by operators and experts, not only can the unit operate more safely and efficiently, but it will also achieve the goal of reducing manpower and increasing efficiency. When abnormal operating conditions are identified through operating condition analysis tools, quantitative analysis and fault warning of the main parameters will be conducted to minimize safety issues and bring certain economic benefits.

[0003] Currently, there are roughly two types of algorithms for feature extraction and operating condition segmentation of time-series data: one is to obtain the original data curve trend by performing polynomial fitting on the original data, then perform statistical analysis and feature extraction on the fitted curve before segmenting the operating conditions; the other is to use a moving window approach to segment and calculate the historical data trend and features. With the development of big data storage technology, the amount and growth rate of data have reached a considerable level. Directly performing polynomial fitting on the original data curve will encounter computational bottlenecks and cannot guarantee the fitting effect and accuracy. Since the moving window algorithm, which segments historical data time periods manually or custom-definedly, is only effective for regression modeling of time-series data of a given length, it cannot guarantee the rationality of the data time period segmentation and has a large trend extraction error. Therefore, it cannot obtain good curve trend features and segmentation effects, and the segmentation of operating conditions for the running data cannot be guaranteed to be accurate. Summary of the Invention

[0004] This application provides a method and apparatus for segmenting the operating conditions of thermal power units. The method obtains curves that meet the combined characteristics by calculating the curve characteristic index, and performs segmentation based on specific characteristic parameters and filters out the operating conditions that meet the conditions.

[0005] This application provides a method for dividing the operating conditions of a thermal power unit, the method comprising:

[0006] Obtain operating data of thermal power units and determine the main curve based on the operating data;

[0007] Calculate one or more characteristic curves based on the determined main curve and operating data;

[0008] One or more characteristic curves can be selected and combined according to different working conditions to obtain trend characteristic combination curves;

[0009] The operating conditions of thermal power units are segmented based on the combined trend feature curves.

[0010] In one exemplary embodiment, determining the main curve based on the operating data includes:

[0011] Data where the difference between two adjacent fitting points is less than a first predetermined threshold, and the difference between the value of each fitting point and the original running data is less than a second predetermined threshold, are fitted according to the fitting formula.

[0012] Based on all main values ​​corresponding to the minimum value of the fitted function j Determine the principal curve of the fit;

[0013] The fitting formula is as follows:

[0014]

[0015] In the above formula, F represents the fitted principal curve main. j The function, main j It is the j-th element of the main curve Main, x j It is the j-th element of the running data sequence X, adjust k It is the first parameter that adjusts the smoothness of the fit. j -x j) 2 It is the distance between the fitted principal curve and the running data. j -main j-1 ) 2 +(main j+1 -main j ) 2 It is the distance between two adjacent fitting points on the fitting principal curve.

[0016] In one exemplary embodiment, the step of determining the minimum value of the fitted function based on all main... j Determine the principal curve for fitting, including:

[0017] Calculate the fitting function F for main j The partial derivatives are equal to 0, thus yielding the coefficient matrix C;

[0018] The fitted principal curve Main is determined based on the coefficient matrix C and the running data X;

[0019] Where Main = linefit(C,X), and linefit is the least squares fitting function.

[0020] In one exemplary embodiment, the characteristic curves include: a rise / fall characteristic curve, an amplitude characteristic curve, and a frequency characteristic curve.

[0021] In one exemplary embodiment, calculating one or more characteristic curves based on the determined main curve and operating data includes:

[0022] When the characteristic curve is a rising or falling feature, the rising or falling index curve is determined based on the difference sequence of the main curve within the preset time window, and the rising or falling speed index curve is determined based on the fitted slope sequence of the main curve within the preset time window.

[0023] The lifting characteristic curves include the lifting index curve and the lifting speed index curve.

[0024] In one exemplary embodiment, calculating one or more characteristic curves based on the determined main curve and operating data includes:

[0025] When the characteristic curve is an amplitude characteristic, the process of calculating the amplitude exponential sequence based on the determined principal curve and operating data is as follows:

[0026] Based on the running data and the fitting formula, the fluctuation curve is calculated using a second parameter for adjusting the fitting smoothness, wherein the parameter for adjusting the fitting smoothness is smaller than the parameter for adjusting the fitting smoothness.

[0027] The amplitude curve is determined based on the amplitude difference between the principal curve and the fluctuation curve;

[0028] The maximum absolute value of the amplitude curve within the preset time window is used as the amplitude index sequence.

[0029] In one exemplary embodiment, calculating one or more characteristic curves based on the determined main curve and operating data includes:

[0030] When the characteristic curve is a frequency characteristic, the process of calculating the frequency index sequence based on the determined main curve and operating data is as follows:

[0031] Determine the number of times the wave curve Wave intersects with the main curve Main within a preset time window, and use the determined number as the frequency index sequence.

[0032] In one exemplary embodiment, the step of selecting one or more characteristic curves according to different working conditions and combining them to obtain a trend characteristic combination curve includes:

[0033] Select one or more characteristic curves according to different working conditions;

[0034] Map the value of each characteristic curve to a preset value range;

[0035] Each characteristic curve is then numbered to obtain a trend characteristic combination curve.

[0036] In one exemplary embodiment, the operation of segmenting the operating conditions of thermal power units based on the trend feature combination curve includes:

[0037] Within the time window of the working conditions to be segmented, select the characteristic curve segments that meet the conditions from the trend characteristic combination curves;

[0038] The selected characteristic curve segments are used as the curves to display the current operating conditions.

[0039] This application also provides a device for dividing the operating conditions of a thermal power unit. The device includes a memory and a processor. The memory is used to store a program for dividing the operating conditions of a thermal power unit, and the processor is used to read and execute the program for dividing the operating conditions of a thermal power unit, and to execute the method described in any of the above embodiments.

[0040] Compared with related technologies, this application provides a method and apparatus for segmenting the operating conditions of thermal power units. The method includes: acquiring operating data of the thermal power unit and determining a main curve based on the operating data; calculating one or more characteristic curves based on the determined main curve and the operating data; selecting one or more characteristic curves to combine according to different operating conditions to obtain a trend characteristic combination curve; and performing segmentation processing on the operating conditions of the thermal power unit based on the trend characteristic combination curve. This application obtains curves that satisfy the combined characteristics by calculating curve characteristic indices, and performs segmentation based on specific characteristic parameters and filters out operating conditions that meet the conditions.

[0041] Other features and advantages of this application will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the application. Other advantages of this application can be realized and obtained by means of the solutions described in the description and the accompanying drawings. Attached Figure Description

[0042] The accompanying drawings are used to provide an understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0043] Figure 1 This is a flowchart of the method for dividing the operating conditions of thermal power units according to an embodiment of this application;

[0044] Figure 2 This is a schematic diagram of the operating condition curve trend in an example embodiment;

[0045] Figure 3 This is a schematic diagram of the main curve extracted from the operating condition curve in an example embodiment;

[0046] Figure 4 This is a schematic diagram of the characteristic index of the rise and fall of the operating condition curve in an example embodiment;

[0047] Figure 5 This is a schematic diagram of the acceleration / deceleration rate index of the operating condition curve in an example embodiment;

[0048] Figure 6 This is a schematic diagram of the comprehensive characteristic index of the operating condition curve in an example embodiment;

[0049] Figure 7 This is a schematic diagram of the operating condition curve segmentation in an example embodiment;

[0050] Figure 8 This is a schematic diagram of a thermal power unit operating condition switching device according to an embodiment of this application. Detailed Implementation

[0051] This application describes several embodiments, but these descriptions are exemplary and not restrictive, and it will be apparent to those skilled in the art that many more embodiments and implementations are possible within the scope of the embodiments described herein. Although many possible combinations of features are shown in the drawings and discussed in the detailed description, many other combinations of the disclosed features are also possible. Unless specifically limited, any feature or element of any embodiment may be used in combination with, or may replace, any feature or element of any other embodiment.

[0052] This application includes and contemplates combinations of features and elements known to those skilled in the art. The embodiments, features, and elements disclosed in this application may also be combined with any conventional features or elements to form a unique inventive scheme as defined by the claims. Any feature or element of any embodiment may also be combined with features or elements from other inventive schemes to form another unique inventive scheme as defined by the claims. Therefore, it should be understood that any feature shown and / or discussed in this application may be implemented individually or in any suitable combination. Therefore, the embodiments are not limited except by the limitations imposed by the appended claims and their equivalents. Furthermore, various modifications and changes may be made within the scope of the appended claims.

[0053] Furthermore, in describing representative embodiments, the specification may have presented methods and / or processes as a specific sequence of steps. However, the method or process should not be limited to the specific order of steps described herein, to the extent that it does not depend on such a specific order. As will be understood by those skilled in the art, other sequences of steps are also possible. Therefore, the specific order of steps set forth in the specification should not be construed as a limitation of the claims. Moreover, the claims concerning the method and / or process should not be limited to the steps performed in the written order, and those skilled in the art will readily understand that these orders can be varied and still remain within the spirit and scope of the embodiments of this application.

[0054] This disclosure provides a method for segmenting the operating conditions of thermal power units, such as... Figure 1 As shown, the method includes steps S100-S130;

[0055] S100. Obtain the operating data of the thermal power unit and fit the main line based on the operating data;

[0056] S110. Calculate one or more characteristic curves based on the determined main curve and operating data;

[0057] S120. Select one or more characteristic curves according to different working conditions to combine them to obtain trend characteristic combination curves;

[0058] S130. Perform segmentation processing on the operating conditions of thermal power units based on the trend feature combination curve.

[0059] In one exemplary embodiment, operating data of a thermal power unit is acquired, and a trend line representing the original operating condition data curve is determined. The process of determining the trend line curve based on the operating data includes:

[0060] Data where the difference between two adjacent fitting points is less than a first predetermined threshold, and the difference between the value of each fitting point and the original running data is less than a second predetermined threshold, are fitted according to the fitting formula.

[0061] Based on the minimum value of the fitted function (F), all main j Determine the principal curve of the fit;

[0062] The fitting formula is as follows:

[0063]

[0064] In the above formula, F represents the fitted principal curve main. j The function, main j It is the j-th element of the main curve Main, x j It is the j-th element of the running data sequence X, adjust k It is the first parameter that adjusts the smoothness of the fit. j -x j) 2 It is the distance between the fitted principal curve and the running data. j -main j-1 ) 2 +(main j+1 -main j ) 2 It is the distance between two adjacent fitting points on the fitting principal curve.

[0065] In one exemplary embodiment, based on all main values ​​corresponding to the minimum value of the fitting function (F) j Determine the principal curve for fitting, including:

[0066] Calculate the fitting function F for main j The partial derivatives are equal to 0, thus yielding the coefficient matrix C;

[0067] The fitted principal curve Main is determined based on the coefficient matrix C and the running data X;

[0068] Where Main = linefit(C,X), and linefit is the least squares fitting function.

[0069] In one exemplary embodiment, the characteristic curve may include: a rise and fall characteristic curve, an amplitude characteristic curve, and a frequency characteristic curve. The characteristic curve is not specifically limited and may be other characteristic curves.

[0070] In one exemplary embodiment, one or more characteristic curves are calculated based on the determined main curve and operating data, including: when the characteristic curve is a rising or falling feature, determining the rising or falling index curve based on the difference sequence of the main curve within a preset time window, and determining the rising or falling speed index curve based on the fitted slope sequence of the main curve within the preset time window; wherein, the rising or falling feature curve includes the rising or falling index curve and the rising or falling speed index curve.

[0071] In one exemplary embodiment, one or more characteristic curves are calculated based on the determined main curve and operating data, including:

[0072] When the characteristic curve is an amplitude characteristic, the process of calculating the amplitude exponential sequence (Range) based on the determined principal curve and operating data is as follows:

[0073] Based on the running data and the fitting formula, the wave curve is calculated using the second parameter for adjusting the fitting smoothness, wherein the parameter for adjusting the fitting smoothness is smaller than the parameter for adjusting the fitting smoothness.

[0074] The amplitude curve (RangeC) is determined based on the amplitude difference between the principal curve and the fluctuation curve.

[0075] The maximum absolute value of the amplitude curve (RangeC) within the preset time window is taken as the amplitude exponential sequence (Range).

[0076] In one exemplary embodiment, calculating one or more characteristic curves based on the determined main curve and operating data includes:

[0077] When the characteristic curve is a frequency characteristic, the process of calculating the frequency index sequence based on the determined main curve and operating data is as follows:

[0078] Determine the number of times the wave curve Wave intersects with the main curve Main within a preset time window, and use the determined number as the frequency index sequence.

[0079] In one exemplary embodiment, the step of selecting one or more characteristic curves according to different working conditions and combining them to obtain a trend characteristic combination curve includes:

[0080] Select one or more characteristic curves according to different working conditions;

[0081] Map the value of each characteristic curve to a preset value range;

[0082] Each characteristic curve is then numbered to obtain a trend characteristic combination curve.

[0083] In one exemplary embodiment, the operation of segmenting the operating conditions of thermal power units based on the trend feature combination curve includes:

[0084] Within the time window of the working conditions to be segmented, select the characteristic curve segments that meet the conditions from the trend characteristic combination curves;

[0085] The selected characteristic curve segments are used as the curves to display the current working conditions.

[0086] This embodiment addresses the prominent issues of insufficient data utilization and waste of data resources currently faced in the operation of thermal power plants, and provides an algorithm for segmenting operating conditions of thermal power unit operating data:

[0087] The algorithm segments operating conditions by using curve feature indices for classification, primarily relying on these indices for judgment and selection. This algorithm identifies the main trend line representing the original operating condition data curves and establishes an error function between the trend line and the original curves. It extracts the main trend line and calculates feature indices from the original data curves, optimizing computational performance. The algorithm first calculates curve feature indices by setting parameters to obtain curves that satisfy combined characteristics. Then, through a visual parameter interface, it segments and filters operating conditions based on specific feature parameters.

[0088] Example 1

[0089] Step 1. Obtain operating data of thermal power units, such as Figure 2 As shown, the main curve is fitted based on the running data, as shown below. Figure 3 As shown; where the fitted main curve is a curve with time as the horizontal axis and the values ​​of the running data as the vertical axis.

[0090] In this step, the steps for fitting the principal curve include:

[0091] Step 101. Fit the data where the difference between the values ​​of two adjacent fitting points is less than a first predetermined threshold, and the difference between the value of each fitting point and the original running data is less than a second predetermined threshold, according to the fitting formula.

[0092] Step 102. Based on all main values ​​corresponding to the minimum value of the fitted function (F) j Determine the principal curve of the fit;

[0093] The fitting formula is as follows:

[0094]

[0095] In the above formula, F represents the fitted principal curve main. j The function, main jIt is the j-th element of the main curve Main, x j It is the j-th element of the running data sequence X, adjust k It is the first parameter that adjusts the smoothness of the fit. j -x j) 2 It is the distance between the fitted principal curve and the running data. j -main j-1 ) 2 +(main j+1 -main j ) 2 It is the distance between two adjacent fitting points on the fitting principal curve.

[0096] Step 103. Based on all main values ​​corresponding to the minimum value of the fitted function (F) j Determine the principal curve for fitting, including:

[0097] (1) Calculate the fitting function F for main j The partial derivatives are equal to 0, yielding the coefficient matrix C. The specific calculation is as follows:

[0098]

[0099]

[0100]

[0101] …

[0102]

[0103] The coefficient matrix C is obtained as follows:

[0104]

[0105] In the coefficient matrix above, all blank spaces are 0.

[0106] (2) Determine the fitted principal curve Main based on the coefficient matrix C and the running data X;

[0107] Main = linefit(C,X), where linefit is the least squares fitting function.

[0108] adjust k It is the first parameter to adjust the smoothness of the fit. Adjusting adjust_k can obtain a main line with a smoothness similar to the mean main line, but the properties of the fitted main line curve are better.

[0109] The parameter used to adjust the mean principal curve is the moving window, denoted as ob_level. Using ob_level as a parameter to adjust the principal curve will yield better results. In this case, ob_level can be called the observation level. For example, when ob_level = 60, the interval between two points in sequence X is 1 second, then the fitted principal curve represents the trend of a 1-minute level curve; when ob_level = 3600, the fitted principal curve represents the trend of an hourly level curve. Adjusting ob_level yields the principal curves at different observation levels.

[0110] Step 2. Calculate one or more characteristic curves based on the determined main curve and operating data.

[0111] In this step, the characteristic curves may include: a rise / fall characteristic curve, an amplitude characteristic curve, and a frequency characteristic curve. In this embodiment, multiple characteristic curves can be generated by combining several calculated characteristics. For example, three characteristics may be calculated: rise / fall characteristic, amplitude characteristic, and frequency characteristic.

[0112] Among them, the rise and fall characteristics describe the rise and fall of the curve and the speed of rise and fall; the amplitude characteristics describe the changes in the amplitude of the curve's vibration; and the frequency characteristics describe the changes in the frequency of the curve's vibration.

[0113] Step 3. Calculate the rise and fall characteristics.

[0114] The rise and fall characteristic curves include the rise and fall exponential curve and the rise and fall speed exponential curve. When the characteristic curve is a rise and fall characteristic, the rise and fall exponential curve is determined based on the difference sequence of the main curve within a preset time window, and the rise and fall speed exponential curve is determined based on the fitted slope sequence of the main curve within the preset time window.

[0115] The process of calculating the rise and fall characteristics is as follows:

[0116] 1. The rise and fall characteristics of the main line are described by the difference within a preset time window. This difference sequence is the rise and fall index sequence, denoted as Lift, where the i-th element is called the i-th rise and fall index, denoted as lift. i This preset time window is called the lifting interval, denoted as lift_interval. i A value greater than 0 indicates an upward movement, lift i A value less than 0 indicates a decrease, while a larger absolute value indicates a more pronounced increase or decrease.

[0117] The formula for calculating the ascending / descending exponent sequence is:

[0118] lift_interval = ob_level\20

[0119] lift i =main i-main i-lift_interval

[0120] Among them, main i It is the i-th element of the main line, main i-lift_interval It is the i-lift_interval-th element of the main line, and the lifting characteristic index of the working condition curve calculated above is as follows: Figure 4 As shown.

[0121] 2. The fitted slope within the preset time window of the main curve is used as the acceleration / deceleration index. The slope sequence is called the acceleration / deceleration index sequence, denoted as Slope. The i-th element is called the i-th acceleration / deceleration index, denoted as Slopei. This interval is called the acceleration / deceleration interval, denoted as slope_interval. A slopei greater than 0 indicates an increase, and a slopei less than 0 indicates a decrease. The larger the absolute value, the faster the increase or decrease. The acceleration / deceleration index of the working condition curve is as follows: Figure 5 As shown.

[0122] The formula for calculating the exponential sequence of acceleration and deceleration is:

[0123] slope_interval = ob_level\20

[0124]

[0125] Among them, main i It is the i-th element of the main line, main i-slope_interval+j It is the i-slope_interval+j-th element of the main line.

[0126] Step 4. When the characteristic curve is an amplitude characteristic, calculate the amplitude index sequence (Range) and amplitude rise / fall index (Range_lift) based on the determined main curve and running data.

[0127] Step 41. The specific calculation process for the amplitude exponential sequence (Range) is as follows:

[0128] Step 411. Based on the running data and the fitting formula, calculate the wave curve using the second parameter to adjust the fitting smoothness.

[0129] In this step, the parameter for adjusting the fitting smoothness is smaller than the parameter for adjusting the fitting smoothness.

[0130] The parameter for adjusting the fit smoothness mentioned above is adjust_k = 2*(4log2(ob_level / 15)-1);

[0131] The second parameter for adjusting the smoothness of the fit is ob_level / 5. This is used as the fitting parameter for the small smoothness curve and is called the amplitude observation level, denoted as range_level.

[0132] range_level = ob_level / 5

[0133] Preferably, the original value (operating data of thermal power units) can be used as this small smoothness curve, that is, range_leve=0. This small smoothness curve is called the wave curve and denoted as Wave.

[0134] Step 412. Determine the amplitude curve (RangeC) based on the amplitude difference between the main curve and the fluctuation curve; take the maximum absolute value of the amplitude curve (RangeC) within the preset time window as the amplitude exponential sequence (Range).

[0135] In this step, the amplitude of the wave oscillation around the main wave is called the amplitude. The difference between the main wave and the wave is called the amplitude curve, denoted as RangeC. The maximum absolute value of RangeC within a time window is the amplitude exponent sequence, denoted as Range, where the i-th element is the amplitude exponent, denoted as range. i This time window is called the amplitude interval, denoted as range_interval.

[0136] Wave=main_fit(X,range_level)

[0137] The main_fit() function is the main line fitting function.

[0138] RangeC = Main — Wave

[0139] range_interval = ob_level\20

[0140] range i =max([RangeC i RangeC i+1 RangeC i+2 ,…,RangeC i+range_interval-1 ])

[0141] Step 42. The specific calculation process for the amplitude rise / fall index (Range_lift) is as follows:

[0142] The characteristic that describes the amplitude rise and fall is the amplitude rise and fall index, denoted as Range_lift.

[0143] Treating the amplitude sequence as the original value sequence, the amplitude rise and fall index is obtained by calculating the rise and fall index in the same way.

[0144] The principal line fitted using the amplitude sequence is called the amplitude principal line, denoted as Range_main, and the fitting parameters use the amplitude observation level ob_level. The interval for calculating the amplitude rise and fall exponent is called the amplitude rise and fall interval, denoted as range_lift_interval.

[0145] Amplitude main line:

[0146] Range_main=main_fit(Range,ob_level)

[0147] Amplitude fluctuation range:

[0148] range_lift_interval=range_interval

[0149] Amplitude fluctuation index:

[0150] range_lift i =range_main i -range_main i-range_lift_interval

[0151] Step 5. When the characteristic curve is a frequency characteristic, calculate the frequency index sequence and the frequency rise / fall index.

[0152] Step 51. The process of calculating the frequency index sequence based on the determined main curve and operating data is as follows:

[0153] The number of times the wave curve Wave intersects the main curve Main within a preset time window is determined, and this determined number is used as the frequency exponential sequence. That is, the number of times the wave curve Wave crosses the main curve Main within a certain interval is called the frequency exponential sequence, which is also the number of times the amplitude curve RangeC alternates between positive and negative values, denoted as Frequency. This interval is called the frequency interval, denoted as frequency_interval, and the i-th element of the frequency exponential sequence is called the frequency exponent, denoted as frq. i .

[0154] frequency_interval=ob_level\20

[0155] frq i =[rangec i-frequency_interval+1 ,

[0156] rangec i-frequency_interval+2 ,…,rangec i .count(~*~[-1]<0)

[0157] Among them, rangec i It is the i-th element of RangeC, count() is the counting function, ~ is the element of the loop sequence, ~[-1] is the element before the current element in the loop, ~*~[-1]<0 means that the two adjacent elements have different signs, that is, the number of times Wave crosses Main.

[0158] Step 52. Calculate the frequency rise / fall index based on the determined main curve and operating data.

[0159] In this step, the frequency lift index, denoted as Frequency_lift, is a feature describing the frequency lift characteristics. Treating the frequency sequence as the original value (thermal power unit operating data) sequence, the lift index is calculated using the same method. The specific calculation method is as follows:

[0160] The principal line fitted using the frequency sequence is called the frequency principal line, denoted as Frequency_main, and the fitting parameters use the frequency observation level ob_level. The interval for calculating the amplitude rise and fall index is called the amplitude rise and fall interval, denoted as frequency_lift_interval, which is equal to the amplitude rise and fall interval range_lift_interval.

[0161] Vibration frequency main line:

[0162] Frequency_main=main_fit(Frequency,ob_level)

[0163] Frequency rise and fall range:

[0164] frequency_lift_interval=frequency_interval

[0165] Frequency rise / fall index:

[0166] frequency_lift i =frequency_main i -frequency_main i-range_lift_interval

[0167] Step 6. Select one or more characteristic curves according to different working conditions and combine them to obtain a trend characteristic combination curve. This combination process includes:

[0168] Step 61. Select one or more characteristic curves according to different working conditions;

[0169] For example, in the case of oscillating divergence, two characteristic curves, the amplitude rise and fall index and the frequency rise and fall index, are selected to divide the working condition.

[0170] Step 62. Map the value of each characteristic curve to a preset value range;

[0171] Step 621. Projecting Features: When calculating features, the dimensions of each feature are not the same, making it difficult to set their value ranges when filtering these features. Therefore, it is necessary to project the feature values ​​into a unified range. For example, the unified range could be [-1, 1], where -1 represents the minimum and 1 represents the maximum.

[0172] Step 622. The specific mapping process is as follows: Features with positive and negative values: Assume the maximum value of the feature is fmax and the minimum value is fmin. Project [fmin, 0] onto [-1, 0], and [0, fmax] onto [0, 1]. The functional relationship between feature f and the selection parameter a is:

[0173]

[0174] The characteristic that it can only take positive values:

[0175]

[0176] Once the features are projected to [-1,1], the parameters become easier to set.

[0177] Step 623. Number each characteristic curve to obtain the trend characteristic combination curve; as shown in the appendix. Figure 6 The diagram shows the comprehensive characteristic index of the working condition curves in Figures a, b, c, and d.

[0178] Step 7. The operation of segmenting the operating conditions of thermal power units based on the trend characteristic combination curve includes:

[0179] Step 71. Within the time window of the working conditions to be segmented, select the characteristic curve segments that meet the conditions from the trend characteristic combination curves;

[0180] This step involves feature filtering, which includes:

[0181] 1) Select the feature or combination of features to be used. For example, select the lifting feature and extract the feature to be filtered, lift, and the original value to form a two-dimensional table T.

[0182] 2) Select the range of feature values, such as [-1, 0] to represent a descending curve, and project it into the value of lift according to the feature projection method. That is, liftmin <= lift <= 0. Where liftmin = F(-1), 0 = F(0).

[0183] 3) Select curve segments that meet the criteria:

[0184] om_index_all=T.pselect(liftmin<=lift<=0),

[0185] Where om_index_all is the index of all curves that meet the conditions --- not the key points.

[0186] 4) Select the duration range of the curve segment, such as [300, 600] which indicates a duration of five to ten minutes.

[0187] om_index_group=om_index.group(~-#)

[0188] Wherein, om_index_group is the index group of the curve, ~ is the current member of om_index during the loop, # is the index of the current member of om_index during the loop, and group(~-#) means that consecutive indexes are grouped together, indicating that rows in T that consecutively meet the filtering conditions are grouped together.

[0189] Step 72. Divide the working conditions according to the selected characteristic curve segments.

[0190] In this step, the working conditions are segmented based on the selected characteristic curve segments, resulting in a schematic diagram of the working condition curve segmentation, as shown below. Figure 7 As shown in figures a and b.

[0191] To select two different characteristic curves (e.g., a curve that first decreases and then stabilizes), follow these steps:

[0192] 1) Filter out curve indices for various features using the feature filtering method.

[0193] om_indexd is the index of the curve segment of the descending feature selected by the feature selection method, and om_indexs is the index of the curve segment of the stationary feature selected.

[0194] 2) Add numbering to the curve indexes for each feature, for example: descending segment 1, descending segment 2... stationary segment 1, lower stationary segment 2...

[0195] 3) Sort by curve index, i.e., sort by the index of each curve segment.

[0196] om_index_seq_sort=om_index_seq.sort(seq)

[0197] Among them, om_index_seq_sort is the result of om_idex_seq sorting.

[0198] 4) Group two adjacent characteristic curve segments together: om_index_seq_sort_group = om_index_seq_sort.group(idx-#)

[0199] Among them, om_index_seq_sort_group is the grouping result based on whether the labels are adjacent.

[0200] 5) Filter out groups where the interval between two adjacent curve segments is sufficiently small.

[0201] The allowed interval between the two curves is p (e.g., 60, representing 1 minute). Select curve segments whose length is equal to the number of curve segments to be filtered and whose interval between the two curves is less than p.

[0202] 6) Merge characteristic curve indices within the same group

[0203] The final index of the curve segment is the sum of the first index of the first curve segment within a group and the last index of the last curve segment within the group.

[0204] 7) Retrieve data by index, where T_om = om_seg_index. (T(~)) is a piecewise curve that meets the conditions.

[0205] The method for dividing the operating conditions of thermal power units implemented in this embodiment has the following technical effects:

[0206] 1. High accuracy: Compared with the previous simple working condition segmentation and filtering algorithm, it has been greatly improved in terms of rationality and accuracy, and can completely extract the set of different curve features from the original curve.

[0207] 2. High timeliness: Because the feature curve algorithm based on the least squares method takes into account both logical operation functions and computational performance, it is fast in terms of computation timeliness, query and display.

[0208] 3. High scalability: Thermal power unit operating condition data is a multi-point, large-capacity, high-density dataset, which presents certain difficulties in terms of data mining value and utilization. Therefore, it can be easily extended to other industries and other types of datasets.

[0209] Example 2

[0210] The specific implementation methods for segmenting the oscillating divergence condition include:

[0211] 1. For the oscillating divergence condition, two characteristic curves are selected: amplitude rise and fall index and frequency rise and fall index.

[0212] 2. Map the values ​​of the two characteristic curves, amplitude rise / fall index and frequency rise / fall index, to a preset value range;

[0213] 3. Assign feature numbers to each feature curve to obtain the trend feature combination curve (the specific operation here is to assign the amplitude rise and fall index feature number as range_lift and the frequency rise and fall index feature number as frequency_lift).

[0214] 4. Within the time window of the working conditions to be segmented, select the characteristic curve segments that meet the conditions from the trend characteristic combination curves;

[0215] The specific implementation process of the screening is as follows: screening is performed on the amplitude rise and fall index and the frequency rise and fall index respectively; curve segments that meet the conditions for amplitude rise and fall index are identified and indexed; curve segments that meet the conditions for frequency rise and fall index are identified and indexed.

[0216] 5. Among multiple curve segments that meet the conditions, determine that both are within the same time window. Both meet the conditions, that is, within the same time window, the logical calculation relationship between the two is "AND" (i.e., "AND" or "&"). When dividing the working conditions, both indices represent the "rising interval". This time period is the oscillation and divergence working condition.

[0217] This application also provides a device for switching the operating conditions of a thermal power unit, such as... Figure 8 As shown, the device includes a memory 810 and a processor 820; the memory is used to store a program for dividing the operating conditions of thermal power units, and the processor is used to read and execute the program for dividing the operating conditions of thermal power units, and execute the method described in any of the above embodiments.

[0218] It will be understood by those skilled in the art that all or some of the steps, systems, or apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all components may be implemented as software executed by a processor, such as a digital signal processor or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

Claims

1. A method for dividing the operating conditions of a thermal power unit, characterized in that, The method includes: Obtain operating data of thermal power units and determine the main curve based on the operating data; Calculate one or more characteristic curves based on the determined main curve and operating data; One or more characteristic curves can be selected and combined according to different working conditions to obtain trend characteristic combination curves; The operating conditions of thermal power units are segmented based on the combined trend feature curves. The step of determining the main curve based on the operating data includes: Data where the difference between two adjacent fitting points is less than a first predetermined threshold, and the difference between the value of each fitting point and the original running data is less than a second predetermined threshold, are fitted according to the fitting formula. Based on all main values ​​corresponding to the minimum value of the fitted function j Determine the principal curve of the fit; The fitting formula is as follows: In the above formula, F represents the fitted principal curve main. j The function, main j It is the j-th element of the main curve Main, x j It is the j-th element of the running data sequence X, adjust k It is the first parameter that adjusts the smoothness of the fit. j -x j) 2 It is the distance between the fitted principal curve and the running data. j -main j-1 ) 2 +(main j+1 -main j ) 2 It is the distance between two adjacent fitting points on the fitting principal line; The step of selecting one or more characteristic curves according to different working conditions and combining them to obtain a trend characteristic combination curve includes: Select one or more characteristic curves according to different working conditions; Map the value of each characteristic curve to a preset value range; Each characteristic curve is then numbered to obtain a trend characteristic combination curve.

2. The method for dividing the operating conditions of thermal power units according to claim 1, characterized in that, The main values ​​corresponding to the minimum value of the fitting function j Determine the principal curve for fitting, including: Calculate the fitting function F for main j The partial derivatives are equal to 0, thus yielding the coefficient matrix C; The fitted principal curve Main is determined based on the coefficient matrix C and the running data X; In this function, Main=linefit(C,X), where linefit is the least squares fitting function.

3. The method for dividing the operating conditions of thermal power units according to claim 1, characterized in that, The characteristic curves include: rise and fall characteristic curve, amplitude characteristic curve, and frequency characteristic curve.

4. The method for dividing the operating conditions of thermal power units according to claim 1, characterized in that, The calculation of one or more characteristic curves based on the determined main curve and operating data includes: When the characteristic curve is a rising or falling feature, the rising or falling index curve is determined based on the difference sequence of the main curve within the preset time window, and the rising or falling speed index curve is determined based on the fitted slope sequence of the main curve within the preset time window. The lifting characteristic curves include the lifting index curve and the lifting speed index curve.

5. The method for dividing the operating conditions of thermal power units according to claim 1, characterized in that, The calculation of one or more characteristic curves based on the determined main curve and operating data includes: When the characteristic curve is an amplitude characteristic, the process of calculating the amplitude exponential sequence based on the determined principal curve and operating data is as follows: Based on the running data and the fitting formula, the fluctuation curve is calculated using a second parameter for adjusting the fitting smoothness, wherein the parameter for adjusting the fitting smoothness is smaller than the parameter for adjusting the fitting smoothness. The amplitude curve is determined based on the amplitude difference between the principal curve and the fluctuation curve; The maximum absolute value of the amplitude curve within the preset time window is used as the amplitude index sequence.

6. The method for dividing the operating conditions of thermal power units according to claim 5, characterized in that, The calculation of one or more characteristic curves based on the determined main curve and operating data includes: When the characteristic curve is a frequency characteristic, the process of calculating the frequency index sequence based on the determined main curve and operating data is as follows: Determine the fluctuation curve within the preset time window Wave With the main curve Main The number of crossovers is used as the frequency index sequence.

7. The method for dividing the operating conditions of thermal power units according to claim 1, characterized in that, The operation of segmenting the operating conditions of thermal power units based on the trend feature combination curve includes: Within the time window of the working conditions to be segmented, select the characteristic curve segments that meet the conditions from the trend characteristic combination curves; The selected characteristic curve segments are used as the curves to display the current operating conditions.

8. A device for dividing the operating conditions of a thermal power unit, characterized in that, The device comprises: a memory and a processor; characterized in that the memory is used to store a program for dividing the operating conditions of thermal power units, and the processor is used to read and execute the program for dividing the operating conditions of thermal power units, and execute the method according to any one of claims 1-7.

Images