A hierarchical power supply user division method based on multi-dimensional information characterization

Abstract

The present invention relates to a hierarchical power supply user division method based on multi-dimensional information characterization, comprising the following steps: step 1, constructing a user voltage tolerance curve based on a label system and a portrait technology; step 2, based on the user voltage tolerance curve in step 1, quantifying the impact of the sag considering the severity of the user's sag, the sag loss and the load importance level; step 3, based on the quantified results of the impact of the sag considering the severity of the user's sag, the sag loss and the load importance level in step 2, using a multi-dimensional information comprehensive evaluation method to divide the hierarchical power supply users. The present invention can intuitively and accurately characterize the severity of the impact of the voltage sag on the user, thereby providing a reference for the hierarchical power supply and prevention and control work of sensitive users.

Description

Technical Field

[0001] The present invention belongs to the technical field of power quality analysis, and relates to a method for dividing power supply users into hierarchical groups, and in particular to a method for dividing power supply users into hierarchical groups based on multi-dimensional information characterization. Background Art

[0002] Voltage sag refers to the voltage disturbance phenomenon in which the RMS voltage at a certain point in the power supply system drops suddenly and recovers in a short time when a system short circuit fault occurs, a large-capacity induction motor is started, or a transformer is put into operation. As modern power electronics technology, automatic control technology, etc. gradually penetrate and integrate into the power system, equipment failures and shutdowns, production line interruptions caused by voltage sags have caused huge economic losses to users and enterprises. Voltage sag has become an unavoidable power quality problem in the power grid, but some measures can be taken to prevent or control it to reduce or suppress its severity. To this end, the present invention proposes a hierarchical power supply user division method based on multi-dimensional information characterization, which is of great significance for the subsequent prevention and control of voltage sags.

[0003] The severity assessment of voltage sag events is usually divided into two perspectives for research and analysis: the grid side and the user side. However, the ultimate goal of the assessment is to quantify the changes in the user's own operating status after the sag event is perceived by the user's equipment. In order to characterize the sag sensitivity of their user equipment, many experts and scholars at home and abroad have conducted a large number of tests on typical sensitive equipment and obtained their respective voltage tolerance characteristic curves. However, with the rapid development of modern high-end manufacturing users, it is time-consuming and laborious to test each device to obtain its voltage tolerance curve, and the engineering practicality is not strong.

[0004] Users usually contain many sensitive devices, and their voltage tolerance curves are different. The voltage tolerance curve of the user can be represented by a voltage tolerance curve with an uncertain area. At present, there are methods based on probability evaluation and fuzzy evaluation to quantify the severity of sag events, but they have the disadvantages of large amount of calculation and poor practicality. In addition, the economic losses caused by voltage sag events have also become an important factor affecting their governance. In addition to the direct economic losses caused to user equipment, there are also indirect economic losses such as corporate image, social impact, and load importance level.

[0005] Therefore, it is urgent to conduct a scientific and reasonable risk assessment of voltage sag events, and guide the tiered power supply to sensitive users based on the assessment results to achieve high-quality and reliable electricity supply for sensitive users.

[0006] After searching, no public documents of the prior art that are identical or similar to the present invention were found. Summary of the invention

[0007] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a hierarchical power supply user classification method based on multi-dimensional information characterization, which can intuitively and accurately characterize the severity of the impact of voltage sag on users, and thus provide a reference for the hierarchical power supply and preventive management work of sensitive users.

[0008] The present invention solves the practical problem by adopting the following technical solutions:

[0009] A method for dividing hierarchical power supply users based on multi-dimensional information characterization comprises the following steps:

[0010] Step 1: Construct a user voltage tolerance curve based on the label system and portrait technology;

[0011] Step 2: Based on the user voltage tolerance curve in step 1, the sag impact degree is quantified considering the user sag severity, sag loss and load importance level;

[0012] Step 3: Based on the quantitative results of the impact of temporary sag considering the severity of temporary sag, temporary sag loss and load importance level of users in step 2, a multi-dimensional information comprehensive evaluation method is used to divide the hierarchical power supply users.

[0013] Moreover, the specific steps of step 1 include:

[0014] (1) Construct a label feature set of sensitive users based on the device hierarchical method;

[0015] (2) Obtaining user voltage tolerance curve based on feature optimization and weight calculation

[0016] According to the label feature set obtained in step (1), the relevant features of the voltage tolerance curve are firstly selected, and then the feature selection is performed through the maximum correlation and minimum redundancy criterion to obtain the optimal feature set of the user voltage tolerance curve. On this basis, the feature normalization and feature weight coefficient calculation are performed, and the feature severity is sorted. The corresponding user voltage tolerance curve is described and characterized according to the set threshold;

[0017] The feature optimization strategy is proposed: firstly, the feature quantities related to the user voltage sag tolerance curve are preliminarily selected, and the selection result is recorded as the original feature set T = {t1, t2, ..., t N}; Then the maximum correlation and minimum redundancy criteria are used to perform feature optimization;

[0018] After discretizing the features, we get the i-th feature t i The entropy of is shown in formula (1):

[0019]

[0020] Where N iCharacteristic t i The number of intervals; M u Characteristic t i The number of samples falling in the uth interval; M is the total number of samples.

[0021] The information entropy of the user voltage tolerance curve d is shown in formula (2):

[0022]

[0023] Where N d M is the number of types of user voltage tolerance curves; v is the number of samples of the vth category.

[0024] The i-th feature t i The joint information entropy with the user voltage tolerance curve d is shown in formula (3):

[0025]

[0026] Where M uv t i The number of samples that fall in the uth interval and whose user voltage tolerance curve is v.

[0027] The i-th feature t i The mutual information with the user voltage tolerance curve d is shown in formula (4):

[0028] I(t i ,d)=H(t i )+H(d)-H(t i ,d) (4)

[0029] The maximum correlation index D(S,d) between the optimal feature set and the user voltage tolerance curve d is shown in formula (5):

[0030]

[0031] In the formula, S and N s is the optimal feature set and the number of features it contains;

[0032] The correlation coefficient is used to calculate the minimum redundancy index R(S), as shown in equations (6) and (7):

[0033]

[0034]

[0035] Where, t i and t j are the i-th and j-th features in the optimal feature set S respectively; cov(t i ,t j) is the covariance of the two features; ρ ti and ρ tj They are respectively the characteristics t i and t j The standard deviation of titj is the correlation coefficient between two features, and its value range is [-1,1].

[0036] Combining the above two indicators, we can get the maximum correlation and minimum redundancy criterion, as shown in formula (8):

[0037]

[0038] In the formula, I mRMR is the maximum related minimum redundancy index.

[0039] Formula (9) is used to unify the dimensions of different features and measure the importance of features:

[0040]

[0041] In the formula, x ij f ij The normalized value of f ij is the jth label value of user i; set f i min is the minimum value among the jth labels; f i max is the maximum value among the jth labels.

[0042] On this basis, the entropy weight method is used to calculate the entropy value of each label and obtain the corresponding weight of each dimensional label, as shown in formula (10):

[0043]

[0044] For the adjacent j-1th and jth labels, let their importance ratio be r j , as shown in formula (11):

[0045]

[0046] Assuming that the label sequence of a certain dimension in the label system is j0 to j1, the weight of the j1th label in this dimension is as shown in formula (12):

[0047]

[0048] When s = j1-1, j1-2, ..., j0, the weight of the sth label is as shown in formula (13):

[0049] w s =r s+1 w s+1 (13)

[0050] The feature weight coefficients are sorted according to their numerical values, and an artificial threshold can be set according to the accuracy of the voltage tolerance curve described by the user; when the accumulated feature weight coefficient is greater than the set threshold, the closest curve is selected; if it is less than this threshold, it is necessary to add feature quantities with smaller weight coefficients until the accumulated feature weight coefficient is greater than the threshold;

[0051] Then the expression of the voltage tolerance curve characterized by the characteristic label of the sensitive device can be set as shown in formula (14):

[0052] L = f(s i w si ),i=1,2,.... (14)

[0053] Where L is the expression of the voltage tolerance curve of the sensitive equipment based on the characteristics and characteristic weights, s i is the i-th feature label; w si is the weight of the i-th feature label.

[0054] When the sensitive user is composed of b sensitive devices connected in series, the expression of its voltage tolerance curve is shown in formula (15). When the sensitive user is composed of b sensitive devices connected in parallel, the expression of its voltage tolerance curve is shown in formula (16):

[0055]

[0056]

[0057] In the formula, H c is the expression of the voltage tolerance curve of the cth sensitive user, t i is the ith time period into which the voltage tolerance curves of b sensitive devices are divided according to the change of the inflection points, and d is the total number of time periods into which the voltage tolerance curves of b sensitive devices are divided according to the change of the inflection points.

[0058] When the sensitive user is composed of q sensitive devices connected in series and p sensitive devices connected in parallel (q+p=b), the expression of its voltage tolerance curve is shown in formula (17):

[0059]

[0060] Moreover, the specific steps of step 2 include:

[0061] (1) Considering the distribution of voltage sag events, the severity of voltage sag is evaluated and a user sag severity index is constructed to quantify the severity of the impact of voltage sag on users.

[0062] The proportion of these temporary drop events to all temporary drop events is defined as the user temporary drop severity index, as shown in formula (19), which is used to quantify the severity of the impact of temporary drop events on users:

[0063]

[0064]

[0065] Where IMSDI is the probability value of the severity of the sag event in the uncertainty area; T is the duration; U is the sag amplitude; U max , U min , T max , T min is the boundary value of the user voltage tolerance curve with uncertain areas; UVSSI is the user sag severity index; N3 is the number of sags that characterize user failures caused by sag events in the uncertain area; N4 is the number of sag events in the abnormal area; N is the total number of sag events; N k It is the total number of times users are affected by voltage sag.

[0066] (2) Construct user temporary sag loss index and load importance level index to evaluate the impact of temporary sag and quantify the severity of the impact of temporary sag on users;

[0067] Within a certain monitoring period, the user's direct economic loss is shown in formula (20):

[0068]

[0069] In the formula, E sag. (k,j) represents the economic loss (10,000 yuan) caused by the k-th sensitive equipment under the user when the j-th voltage sag event occurs; M is the total number of sensitive equipment types under the user.

[0070] The indirect economic loss of the user is shown in formula (21), and the loss consequence depends on the degree to which the voltage deviates from the allowable range:

[0071]

[0072] Where: E sag_potential Indicates the indirect economic loss value of the user / ten thousand yuan; U y is the residual voltage amplitude of the yth sag; U allow K is the tolerable voltage limit under normal working conditions; 1.x is the maximum value of the potential economic loss function of voltage sag at node x; 2 is the sensitivity parameter of potential economic loss.

[0073] The total economic loss of users caused by voltage sag consists of direct economic loss and indirect economic loss, which is defined as the user voltage sag loss index, as shown in formula (22):

[0074] E sag =E sag_direct +E sag_potential (twenty two)

[0075] In the formula, E sag It is the economic loss caused by the temporary drop in power consumption within a certain period of time (such as one year) (ten thousand yuan / year). The larger the value is, the greater the impact of the temporary drop on the user.

[0076] The present invention defines the economic loss value of the user per 10,000 yuan of gross domestic product, as shown in formula (23), which represents the relative economic loss of the user when affected by a temporary drop event.

[0077] ψ=E sag / E ∑ (twenty three)

[0078] In the formula, E ∑ is the user's gross production value within a certain period of time (such as one year) (ten thousand yuan / year); Ψ is the user's relative economic loss per ten thousand yuan of gross production value.

[0079] The present invention defines a load importance level index according to different user load application scenarios: extremely important load is 1, important load is 0.5, and ordinary load is 0.

[0080] Moreover, the specific steps of step 3 include:

[0081] (1) Radar chart of user comprehensive temporary drop level based on multi-dimensional characterization:

[0082] Based on the user voltage tolerance curve obtained in step 1, the probability of the user being affected by the sag event based on the tolerance curve is obtained. Then, the user sag severity index, user sag loss index and load importance level index in step 2 are comprehensively considered to form a comprehensive sag level radar chart composed of these four dimensions, so as to realize the visualization of the scoring results of different users under different influencing factors;

[0083] (2) Method for classifying users of power supply based on comprehensive evaluation of user temporary power supply drop

[0084] According to the user comprehensive temporary sag level diagram drawn in step 3 (1), the governance classification is carried out based on the comprehensive evaluation of user temporary sag, and then the sensitive users with different temporary sag levels are powered in different levels.

[0085] Advantages and beneficial effects of the present invention:

[0086] 1. The present invention constructs a user voltage tolerance curve label system based on the idea of ​​stratification and grading, and realizes a comprehensive and objective selection of influencing factors; the user voltage tolerance curve under the corresponding threshold is obtained based on the feature optimization strategy, weight coefficient calculation and electrical logic relationship, which is more in line with objective reality; considering the different impacts of users on temporary sag events under different tolerance curves, the probability of users being affected by temporary sag events is defined, which can more reasonably evaluate the severity of voltage sag events.

[0087] 2. The present invention defines a user sag severity index by taking into account the differences in severity of sag events in the uncertainty region of the tolerance curve, making its evaluation result more accurate and effective; it defines a user sag loss index by taking into account the economic losses of users caused by sag events, and defines a load importance level index due to different user application scenarios, which is conducive to the prevention and control of voltage sags.

[0088] 3. The present invention comprehensively considers four dimensions, namely, the probability of users within the tolerance curve being affected by temporary sag events, the severity of user temporary sags, user temporary sag losses and load importance level, to construct a comprehensive scoring diagram for user temporary sags, which intuitively and accurately depicts the severity of the impact of user voltage sags, can guide the implementation of the user-level power supply user classification method, and can achieve the goal of maximizing governance benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 It is a flow chart of a method for dividing hierarchical power supply users based on multi-dimensional information characterization in a specific implementation mode of the present invention;

[0090] Figure 2 is a schematic diagram of a tag feature set of a sensitive user's voltage sag tolerance capability in a specific embodiment of the present invention;

[0091] Figure 3 is a voltage sag tolerance curve diagram of a sensitive user in a specific embodiment of the present invention;

[0092] Figure 4 is a critical failure curve diagram depicted in an uncertain region in a specific embodiment of the present invention;

[0093] Figure 5 It is a radar chart of the comprehensive temporary drop level of sensitive users in the specific implementation mode of the present invention. DETAILED DESCRIPTION

[0094] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0095] A hierarchical power supply user division method based on multi-dimensional information characterization, such as Figure 1 As shown, the following steps are included:

[0096] Step 1: Construct a user voltage tolerance curve based on the label system and portrait technology;

[0097] The specific steps of step 1 include:

[0098] (1) Constructing a label feature set of sensitive users based on device stratification and classification method

[0099] Taking into account that sensitive users are composed of devices according to a certain functional logic, the "layered and graded" method can be used to obtain the label feature set of the corresponding sensitive users, so as to achieve comprehensive and rapid selection of the factors affecting voltage sag on sensitive users.

[0100] In this embodiment, a sensitive user usually contains many sensitive devices, which are spatially combined into integrated devices with different functions according to a certain electrical logic relationship. These integrated devices are combined to obtain a series of links, which form a production line in the order of the links, and then form a complete industrial user. Based on this, the present invention describes the four levels of equipment level → link level → production line level → user level in accordance with the idea of ​​"layering and grading", obtains the label feature set of the voltage sag tolerance capability of the corresponding sensitive user, and realizes comprehensive and rapid selection of the factors affecting the voltage sag on the sensitive user.

[0101] The voltage sag tolerance of sensitive equipment should be considered from two perspectives: voltage sag characteristics and load characteristics, so as to construct a voltage sag tolerance label system for sensitive equipment. Among them, voltage sag characteristics should include sag amplitude, duration, phase jump, waveform point, sag type and sag shape; load characteristics should include load rate, equipment consequence status and tolerance curve inflection point and other information. Links are formed by a series of sensitive equipment or integrated equipment arranged and combined rationally in space, and should include link information and link response information under sag events. Among them, link information includes link name, link function, link importance level, link sag sensitive mechanism type, manufacturer, link model, link parameters (rated voltage, rated current, rated capacity, working efficiency), wiring method, number of equipment in the link and series-parallel mode; link response information under sag events includes link commissioning status, link tolerance curve and link industrial parameter recovery time and other information. After a single or multiple links are connected, a complete production line can be formed. A production line-level voltage sag tolerance characteristic label system is established based on the production line information and the response information of the production line to the sag event, including information such as the production line name, production line function, connection method of links within the production line, production line consequence status, production line perceived sag time, and production line tolerance curve. Based on the above analysis, the voltage sag tolerance level of a user as a whole, part, or component can be obtained. In addition, in order to qualitatively analyze the sag performance between users, a voltage sag tolerance characteristic label system can be constructed from the user side, including information such as user name, user function, connection method between production lines within the user, user tolerance curve, user consequence status, and user perceived sag time. In summary, a label feature system for the voltage sag tolerance capability of sensitive users can be constructed, such as Figure 2 shown.

[0102] (2) Obtaining user voltage tolerance curve based on feature optimization and weight calculation

[0103] According to the label feature set obtained in step (1), the relevant features of the voltage tolerance curve are first preliminarily selected, and then the features are optimized through the maximum correlation and minimum redundancy criterion to obtain the optimal feature set of the user voltage tolerance curve. On this basis, feature normalization and feature weight coefficient calculation are performed, and feature severity is sorted. The corresponding user voltage tolerance curve is described and characterized according to the set threshold.

[0104] In this embodiment, the voltage sag tolerance feature label set in A1 is obtained through detailed analysis of sensitive users, but there are many types of them. If all of them are adopted, the operation is complicated and not conducive to practical engineering applications. At the same time, there is a redundant relationship between these features, which may result in over-evaluation.

[0105] Therefore, the present invention proposes a feature optimization strategy: firstly, the feature quantities related to the user voltage sag tolerance curve are preliminarily selected, and the selection result is recorded as the original feature set T = {t1, t2, ..., t N}; Then the maximum relevance and minimum redundancy criteria are used to perform feature optimization.

[0106] After discretizing the features, we get the i-th feature t i The entropy of is shown in formula (1):

[0107]

[0108] Where N i Characteristic t i The number of intervals; M u Characteristic t i The number of samples falling in the uth interval; M is the total number of samples.

[0109] The information entropy of the user voltage tolerance curve d is shown in formula (2):

[0110]

[0111] Where N d M is the number of types of user voltage tolerance curves; v is the number of samples of the vth category.

[0112] The i-th feature t i The joint information entropy with the user voltage tolerance curve d is shown in formula (3):

[0113]

[0114] Where M uv t i The number of samples that fall in the uth interval and whose user voltage tolerance curve is v.

[0115] The i-th feature t i The mutual information with the user voltage tolerance curve d is shown in formula (4):

[0116] I(t i ,d)=H(t i )+H(d)-H(t i ,d) (4)

[0117] The maximum correlation index D(S,d) between the optimal feature set and the user voltage tolerance curve d is shown in formula (5):

[0118]

[0119] In the formula, S and N sis the optimal feature set and the number of features it contains.

[0120] The present invention uses the correlation coefficient to calculate the minimum redundancy index R(S), as shown in formulas (6) and (7):

[0121]

[0122]

[0123] Where, t i and t j are the i-th and j-th features in the optimal feature set S respectively; cov(t i ,t j ) is the covariance of the two features; ρ ti and ρ tj They are respectively the characteristics t i and t j The standard deviation of titj is the correlation coefficient between two features, and its value range is [-1,1].

[0124] Combining the above two indicators, we can get the maximum correlation and minimum redundancy criterion, as shown in formula (8):

[0125]

[0126] In the formula, I mRMR is the maximum related minimum redundancy index.

[0127] Considering that the number of initial features of the user's voltage sag tolerance capability is not large, the present invention uses the traversal method to obtain the global optimal feature solution S. Since the feature labels in the optimal feature set have different dimensions and importance, in order to accurately characterize the user's voltage sag tolerance capability, formula (9) is used to unify the dimensions of different features and measure the feature importance.

[0128]

[0129] In the formula, x ij f ij The normalized value of f ij is the jth label value of user i; set f i min is the minimum value among the jth labels; f i max is the maximum value among the jth labels.

[0130] On this basis, the entropy weight method is used to calculate the entropy value of each label and obtain the corresponding weight of each dimensional label, as shown in formula (10):

[0131]

[0132] For the adjacent j-1th and jth labels, let their importance ratio be r j , as shown in formula (11):

[0133]

[0134] Assuming that the label sequence of a certain dimension in the label system is j0 to j1, the weight of the j1th label in this dimension is as shown in formula (12):

[0135]

[0136] When s = j1-1, j1-2, ..., j0, the weight of the sth label is as shown in formula (13):

[0137] w s =r s+1 w s+1 (13)

[0138] The feature weight coefficients are sorted according to their numerical values, and an artificial threshold can be set according to the accuracy of the user's description of the voltage tolerance curve. When the accumulated feature weight coefficient is greater than the set threshold, the closest curve is selected; if it is less than this threshold, it is necessary to add feature quantities with smaller weight coefficients until the accumulated feature weight coefficient is greater than the threshold. The expression of the voltage tolerance curve characterized by the feature label of the sensitive device can be set as shown in formula (14):

[0139] L = f(s i w si ),i=1,2,.... (14)

[0140] Where L is the expression of the voltage tolerance curve of the sensitive equipment based on the characteristics and characteristic weights, s i is the i-th feature label; w si is the weight of the i-th feature label.

[0141] When the sensitive user is composed of b sensitive devices connected in series, the expression of its voltage tolerance curve is shown in formula (15). When the sensitive user is composed of b sensitive devices connected in parallel, the expression of its voltage tolerance curve is shown in formula (16):

[0142]

[0143]

[0144] In the formula, H c is the expression of the voltage tolerance curve of the cth sensitive user, t iis the ith time period into which the voltage tolerance curves of b sensitive devices are divided according to the change of the inflection points, and d is the total number of time periods into which the voltage tolerance curves of b sensitive devices are divided according to the change of the inflection points.

[0145] When the sensitive user is composed of q sensitive devices connected in series and p sensitive devices connected in parallel (q+p=b), the expression of its voltage tolerance curve is shown in formula (17):

[0146]

[0147] Step 2: Based on the user voltage tolerance curve in step 1, the sag impact degree is quantified considering the user sag severity, sag loss and load importance level;

[0148] The specific steps of step 2 include:

[0149] (1) Considering the distribution of voltage sag events, the severity of voltage sag is evaluated and a user sag severity index is constructed to quantify the severity of the impact of voltage sag on users.

[0150] Taking into account the influence of external factors such as user model and weather, the user voltage tolerance curve is usually expressed as a sag characteristic curve with an uncertain area. When a sag event occurs, all events in the fault area and some events in the uncertain area will cause sensitive users to shut down and even cause a chain reaction of other users, and the severity of the sag is relatively large. Based on this, the present invention defines the proportion of these sag events to all sag events as a user sag severity index to characterize the overall distribution of sag events.

[0151] In this embodiment, the user voltage sag tolerance curve can be obtained according to the method mentioned in step 1, such as Figure 3 As shown. However, due to the influence of external factors such as user model and weather, the user voltage tolerance curve should be expressed as a sag characteristic curve with an uncertain area: when the sag event is in the normal area, the probability value of affecting the user can be considered to be 0; when the sag event is in the abnormal area, the probability value of causing user failure can be considered to be 1; when the sag event is in the uncertain area, the degree to which the user is affected by it is within the fluctuating interval value. Therefore, the present invention proposes to classify the severity level of the sag events in the uncertain area, so as to qualitatively analyze the degree to which the user is affected by the sag in the uncertain area.

[0152] First, a reasonable threshold should be selected as the basis for qualitative analysis of severity. max or U min When the temporary drop event occurs, it is very likely to cause equipment failure. At this time, it can be considered that the coordinate point (T max ,U min ​) is the critical failure point. At this time, the failure probability value of the sag event can be calculated according to formula (18), and then the critical failure curve can be characterized, as shown in Figure 4 As shown, the distribution area of ​​severe temporary drop events in the uncertain area can be obtained. These temporary drop events and all temporary drop events in the fault area will cause sensitive users to shut down and even cause a chain reaction of other users, and the user consequence status will change, and the user will be greatly affected. Based on this, the present invention defines the proportion of these temporary drop events to all temporary drop events as the user temporary drop severity index, as shown in formula (19), which is used to quantify the severity of the impact of temporary drop events on users.

[0153]

[0154]

[0155] Where IMSDI is the probability value of the severity of the sag event in the uncertainty area; T is the duration; U is the sag amplitude; U max , U min , T max , T min is the boundary value of the user voltage tolerance curve with uncertain areas; UVSSI is the user sag severity index; N3 is the number of sags that characterize user failures caused by sag events in the uncertain area; N4 is the number of sag events in the abnormal area; N is the total number of sag events; N k It is the total number of times users are affected by voltage sag.

[0156] (2) Construct user temporary sag loss index and load importance level index to evaluate the impact of temporary sag and quantify the severity of the impact of temporary sag on users;

[0157] The user's intuitive perception of the degree of impact of voltage sag is the economic loss caused by the voltage sag event, including direct economic loss and indirect economic loss. In addition, when the spatial position (application scenario) of the user (or the sensitive electrical equipment among the users) is different, the social impact caused is also completely different, that is, it can be equivalent to the importance level of the user's load. When in an important position, the higher the importance level of the same user (or the sensitive electrical equipment among the users), the more serious the impact caused by the voltage sag event. Based on this, the present invention takes into account the difference between the economic loss of the user's temporary sag and the application scenario, and constructs a user temporary sag loss index and a load importance level index.

[0158] In this embodiment, the user's intuitive perception of the degree of impact of voltage sag is the economic loss caused by the voltage sag event, including direct economic loss and indirect economic loss. Direct economic loss refers to the economic loss caused by production interruption, product damage, defectiveness, and waste of the user's sensitive equipment due to voltage sag. The size of the loss depends on factors such as the severity of a single voltage sag and the sensitivity of the equipment. Within a certain monitoring period, the user's direct economic loss is shown in formula (20):

[0159]

[0160] In the formula, E sag. (k,j) represents the economic loss (10,000 yuan) caused by the k-th sensitive equipment under the user when the j-th voltage sag event occurs; M is the total number of sensitive equipment types under the user.

[0161] The indirect economic loss of the user is shown in formula (21), which reflects the potential increase in costs for the user when the voltage deviates from the normal value. The loss consequences depend on the degree to which the voltage deviates from the allowable range.

[0162]

[0163] Where: E sag_potential Indicates the indirect economic loss value of the user / ten thousand yuan; U y is the residual voltage amplitude of the yth sag; U allow K is the tolerable voltage limit under normal working conditions; 1.x is the maximum value of the potential economic loss function of voltage sag at node x; 2 is the sensitivity parameter of potential economic loss.

[0164] The total economic loss of users caused by voltage sag consists of direct economic loss and indirect economic loss, which is defined as the user voltage sag loss index, as shown in formula (22):

[0165] E sag =E sag_direct +E sag_potential (twenty two)

[0166] In the formula, E sag It is the economic loss (10,000 yuan / year) caused by a temporary drop event that occurs to the user within a certain period of time (such as one year). The larger its value is, the greater the impact of the temporary drop on the user. This indicator can, to a certain extent, represent the overall economic loss of the user when it is affected by a temporary drop event, but it is not related to the overall production and operation status of the user. For this reason, the present invention defines the economic loss value of the user per 10,000 yuan of gross production value, as shown in formula (23), which represents the relative economic loss of the user when it is affected by a temporary drop event.

[0167] ψ=E sag / E∑ (twenty three)

[0168] In the formula, E ∑ is the user's gross production value within a certain period of time (such as one year) (ten thousand yuan / year); Ψ is the user's relative economic loss per ten thousand yuan of gross production value.

[0169] In addition, when users are located in different spatial positions, that is, in different application scenarios, the social impact caused by the same temporary drop event on users may be completely different, which can be equivalent to the difference in user load importance levels. The more important the user's position is, that is, the higher the importance level, the more serious the impact of the temporary drop event. Based on this, the present invention defines a load importance level index according to different user load application scenarios: extremely important load is 1, important load is 0.5, and ordinary load is 0.

[0170] Step 3: Based on the quantitative results of the impact of temporary sag considering the severity of temporary sag, temporary sag loss and load importance level of users in step 2, a multi-dimensional information comprehensive evaluation method is used to divide the hierarchical power supply users.

[0171] The specific steps of step 3 include:

[0172] (1) Radar chart of user comprehensive temporary drop level based on multi-dimensional characterization:

[0173] Based on the user voltage tolerance curve obtained in step 1, the probability of the user being affected by the temporary sag event based on the tolerance curve is obtained. Then, considering the user temporary sag severity index, user temporary sag loss index and load importance level index in step 2, a comprehensive temporary sag level radar chart composed of these four dimensions is formed to realize the visualization of the scoring results of different users under different influencing factors.

[0174] (2) Method for classifying users of power supply based on comprehensive evaluation of user temporary power supply drop

[0175] According to the user comprehensive temporary sag level diagram drawn in step 3 (1), the governance classification is carried out based on the comprehensive evaluation of user temporary sag, and then the sensitive users with different temporary sag levels are powered in different levels.

[0176] In this embodiment, in step 3 of a method for classifying power supply users based on multi-dimensional information of the present invention, the method for classifying power supply users based on comprehensive evaluation of multi-dimensional information includes:

[0177] (1) Radar chart of user comprehensive temporary drop level based on multi-dimensional characterization

[0178] The voltage tolerance curves of different user devices are different, and the difference between the voltage tolerance curves can be used to quantitatively characterize the impact of voltage sag events on user devices. For sag events, the probability of a sag amplitude greater than 0.9 pu or a duration greater than 1 min is very small, and the impact on the device is also very small, so sag events in this range are usually no longer considered. Therefore, the present invention proposes that the user voltage tolerance curve index is characterized by the ratio of the area of ​​the abnormal area to the area of ​​the two-dimensional space composed of the sag amplitude of [0,0.9] / pu and the duration of [0,1000] / ms, as shown in formula (24):

[0179]

[0180] In the formula, S i is the area of ​​the abnormal region of the user voltage tolerance curve; S is the area of ​​the entire two-dimensional space plane.

[0181] On this basis, the user sag severity index obtained by formula (19), the user economic loss index per 10,000 yuan of turnover obtained by formula (23), the user voltage tolerance curve index obtained by formula (24) and the load importance level index are comprehensively considered to draw a comprehensive sag level radar chart composed of four dimensional information, so as to realize the visualization of the evaluation results of different users under multiple influencing factors, such as Figure 5 shown.

[0182] (2) Method for classifying users of power supply based on comprehensive evaluation of user temporary power supply drop

[0183] From step 3 (1), although the scores of each dimension information can be known, the result of the joint action of multiple dimensions has not yet been clearly quantified or intuitively perceived. Therefore, the present invention presents the effect of the joint action of multiple dimensions from two perspectives: First, comprehensively consider the score results of each dimension in step 3 (1), and add them up to obtain the user's comprehensive score, as shown in formula (25).

[0184]

[0185] In the formula, S is the comprehensive score of the user's multi-dimensional information. The larger the value, the more severe the temporary drop event the user has to endure, and the more likely the user is to have a failure.

[0186] The second is to draw a circle with a radius of 100, and take the score value of each dimensional information as the starting point, and depict it at 90° intervals in sequence, so that the distribution of the user's comprehensive sag level map in the virtual circle can be obtained. In order to intuitively depict the score of each dimensional information, three virtual circles with radii of 75, 50 and 25 can be drawn according to the 75% probability value, 50% probability value and 25% probability value. At this time, the score results of the four dimensional information fall in different circles, with a total of 16 distribution situations. The score of each dimensional information is set to be between 0 and 100. When the four dimensional information is located in the virtual circle with a radius of 25, the user is less affected by the sag event, and the normal power supply requirements can be adopted for the user; when the four dimensional information is located in the virtual circle with a radius of 50, the voltage sag has a greater impact on the user, and the power supply requirements are higher; when the four dimensional information is located in the virtual circle with a radius of 75, the voltage sag has a greater impact on the user, and the power supply requirements are high; when the four dimensional information exceeds the virtual circle with a radius of 75, the voltage sag has a great impact on the user, and the power supply requirements are high. Based on this, users can be connected to four different types of power supply according to the size of the comprehensive score of the four-dimensional information, as shown in Table 1. This can achieve the purpose of hierarchical power supply and maximize the governance benefits, and is also conducive to the effective implementation of temporary sag governance and prevention work.

[0187] Table 1 is a hierarchical power supply table for sensitive users in a specific implementation manner of the present invention:

[0188] Table 1

[0189]

[0190] The present invention proposes a hierarchical power supply user division method based on multi-dimensional information characterization, which obtains the label feature set of the voltage sag tolerance capability of sensitive users according to the hierarchical and graded method, thereby realizing the comprehensive and objective selection of influencing factors; obtains the user voltage tolerance curve under the corresponding threshold based on the feature optimization strategy, weight coefficient calculation and electrical logic relationship, which is more in line with objective reality; establishes the critical fault limit considering the distribution of sag events, and then quantifies the severity of user sag, making its evaluation result more accurate and effective; comprehensively considers the economic loss of users and the social impact of sag, establishes the user sag loss index and the load importance level index, which is conducive to the prevention and control of voltage sag; characterizes the comprehensive sag level of users based on multi-dimensional information, guides the implementation of the hierarchical power supply user division method, and can achieve the purpose of maximizing the control benefit.

[0191] Through the above description, the basic functions of the hierarchical power supply user division method based on multi-dimensional information characterization of the present invention have been explained. The present invention obtains the user voltage tolerance curve based on label feature design, feature optimization, quantification, etc.; defines the user temporary sag severity index and the user temporary sag loss index, and reasonably quantifies the impact of temporary sag events on users; constructs a user temporary sag comprehensive scoring chart based on multi-dimensional information characterization, which can intuitively and vividly describe the user temporary sag level, which is conducive to the prevention and control of voltage sag events, and has theoretical value and practical significance.

[0192] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0193] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0194] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0195] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1The steps for the functions specified in one or more boxes.

Claims

1. A method for classifying power supply users based on multi-dimensional information characterization, characterized by: The following steps are involved: Step 1: Construct a user voltage tolerance curve based on the label system and portrait technology; Step 2: Based on the user voltage tolerance curve in step 1, the sag impact degree is quantified considering the user sag severity, sag loss and load importance level; Step 3: Based on the quantitative results of the impact of power sag considering the severity of power sag, power sag loss and load importance level of users in step 2, the hierarchical power supply users are divided by using a multi-dimensional information comprehensive evaluation method; The specific steps of step 1 include: (1) Construct a label feature set of sensitive users based on the device hierarchical method; (2) Obtaining user voltage tolerance curve based on feature optimization and weight calculation According to the label feature set obtained in step (1), the relevant features of the voltage tolerance curve are firstly selected, and then the feature selection is performed through the maximum correlation and minimum redundancy criterion to obtain the optimal feature set of the user voltage tolerance curve. On this basis, the feature normalization and feature weight coefficient calculation are performed, and the feature severity is sorted. The corresponding user voltage tolerance curve is described and characterized according to the set threshold; The feature optimization strategy is proposed: firstly, the feature quantities related to the user voltage sag tolerance curve are preliminarily selected, and the selection result is recorded as the original feature set T = {t1, t2, ..., t N }; Then the maximum correlation and minimum redundancy criteria are used to perform feature optimization; After discretizing the features, we get the i-th feature t i The entropy of is shown in formula (1): Where N i Characteristic t i The number of intervals; M u Characteristic t i The number of samples falling in the uth interval; M is the total number of samples; The information entropy of the user voltage tolerance curve d is shown in formula (2): Where N d M is the number of types of user voltage tolerance curves; v is the number of samples of the vth category; The i-th feature t i The joint information entropy with the user voltage tolerance curve d is shown in formula (3): Where M uv t i The number of samples falling in the uth interval and with a user voltage tolerance curve of v; The i-th feature t i The mutual information with the user voltage tolerance curve d is shown in formula (4): I(t i ,d)=H(t i )+H(d)-H(t i ,d) (4) The maximum correlation index D(S,d) between the optimal feature set and the user voltage tolerance curve d is shown in formula (5): In the formula, S and N s is the optimal feature set and the number of features it contains; The correlation coefficient is used to calculate the minimum redundancy index R(S), as shown in equations (6) and (7): Where, t i and t j are the i-th and j-th features in the optimal feature set S respectively; cov(t i ,t j ) is the covariance of the two features; ρ ti and ρ tj They are the characteristics t i and t j The standard deviation of titj is the correlation coefficient of two features, and its value range is [-1,1]; Combining the above two indicators, we can get the maximum correlation and minimum redundancy criterion, as shown in formula (8): In the formula, I mRMR is the maximum related minimum redundancy index; Formula (9) is used to unify the dimensions of different features and measure the importance of features; In the formula, x ij f ij The normalized value of f ij is the jth label value of user i; set is the minimum value among the jth labels; is the maximum value among the jth labels; On this basis, the entropy weight method is used to calculate the entropy value of each label and obtain the corresponding weight of each dimensional label, as shown in formula (10): For the adjacent j-1th and jth labels, let their importance ratio be r j , as shown in formula (11): Assuming that the label sequence of a certain dimension in the label system is j0 to j1, the weight of the j1th label in this dimension is as shown in formula (12): When s = j1-1, j1-2, ..., j0, the weight of the sth label is as shown in formula (13): w s =r s+1 w s+1 (13) The feature weight coefficients are sorted according to their numerical values, and an artificial threshold can be set according to the accuracy of the voltage tolerance curve described by the user; when the accumulated feature weight coefficient is greater than the set threshold, the closest curve is selected; if it is less than this threshold, it is necessary to add feature quantities with smaller weight coefficients until the accumulated feature weight coefficient is greater than the threshold; Then the expression of the voltage tolerance curve characterized by the characteristic label of the sensitive device can be set as shown in formula (14): Where L is the expression of the voltage tolerance curve of the sensitive equipment based on the characteristics and characteristic weights, s i is the i-th feature label; w si is the weight of the i-th feature label; When the sensitive user is composed of b sensitive devices connected in series, the expression of its voltage tolerance curve is shown in formula (15). When the sensitive user is composed of b sensitive devices connected in parallel, the expression of its voltage tolerance curve is shown in formula (16): In the formula, H c is the expression of the voltage tolerance curve of the cth sensitive user, t i is the i-th time period divided by the voltage tolerance curve of b sensitive devices according to the change of the inflection point, and d is the total number of time periods divided by the voltage tolerance curve of b sensitive devices according to the change of the inflection point; When the sensitive user is composed of q sensitive devices connected in series and p sensitive devices connected in parallel (q+p=b), the expression of its voltage tolerance curve is shown in formula (17): The specific steps of step 2 include: (1) Considering the distribution of voltage sag events, the severity of voltage sag is evaluated and a user sag severity index is constructed to quantify the severity of the impact of voltage sag on users. The proportion of these temporary drop events to all temporary drop events is defined as the user temporary drop severity index, as shown in formula (19), which is used to quantify the severity of the impact of temporary drop events on users; Where IMSDI is the probability value of the severity of the sag event in the uncertainty area; T is the duration; U is the sag amplitude; U max , U min , T max , T min is the boundary value of the user voltage tolerance curve with uncertain areas; UVSSI is the user sag severity index; N3 is the number of sags that characterize user failures caused by sag events in the uncertain area; N4 is the number of sag events in the abnormal area; N is the total number of sag events; N k The total number of times users are affected by voltage sag; (2) Construct user temporary sag loss index and load importance level index to evaluate the impact of temporary sag and quantify the severity of the impact of temporary sag on users; Within a certain monitoring period, the user's direct economic loss is shown in formula (20): In the formula, E sag. (k,j) represents the economic loss caused by the k-th sensitive equipment of the user when the j-th voltage sag event occurs / ten thousand yuan; M is the total number of sensitive equipment types of the user; The indirect economic loss of the user is shown in formula (21), and the loss consequence depends on the degree to which the voltage deviates from the allowable range: Where: E sag_potential Indicates the indirect economic loss value of the user / ten thousand yuan; U y is the residual voltage amplitude of the yth sag; U allow K is the tolerable voltage limit under normal working conditions; 1.x is the maximum value of the potential economic loss function of voltage sag at node x; 2 is the sensitivity parameter of potential economic losses; The total economic loss of users caused by voltage sag consists of direct economic loss and indirect economic loss, which is defined as the user voltage sag loss index, as shown in formula (22): AND sag =And sag_direct +E sag_potential (22) In the formula, E sag It is the economic loss caused by the temporary drop in power consumption within a certain period of time (such as one year) (ten thousand yuan / year). The larger the value is, the greater the impact of the temporary drop on the user. The present invention defines the economic loss value of the user per 10,000 yuan of gross domestic product, as shown in formula (23), which represents the relative economic loss of the user when affected by a temporary drop event; ψ=E sag / AND ∑ (23) In the formula, E ∑ is the user's gross production value in a certain period of time (such as one year) (ten thousand yuan / year); Ψ is the user's relative economic loss per ten thousand yuan of gross production value; The present invention defines the load importance level index according to different user load application scenarios: extremely important load is 1, important load is 0.5, and ordinary load is 0; The specific steps of step 3 include: (1) Radar chart of user comprehensive temporary drop level based on multi-dimensional characterization: Based on the user voltage tolerance curve obtained in step 1, the probability of the user being affected by the sag event based on the tolerance curve is obtained. Then, the user sag severity index, user sag loss index and load importance level index in step 2 are comprehensively considered to form a comprehensive sag level radar chart composed of these four dimensions, so as to realize the visualization of the scoring results of different users under different influencing factors; (2) Method for classifying users of power supply based on comprehensive evaluation of user temporary power supply drop According to the user comprehensive temporary sag level diagram drawn in step 3 (1), the governance classification is carried out based on the comprehensive evaluation of user temporary sag, and then the sensitive users with different temporary sag levels are powered in different levels.

Images