Data analysis electricity consumption prediction service system

Abstract

The invention discloses a data analysis electricity consumption prediction service system, and relates to the technical field of electricity consumption prediction. The data analysis electricity consumption prediction service system comprises the following specific system processes: S1, the system starts to enter database extraction, data cleaning is carried out after data extraction, data prediction is carried out after data cleaning, and a prediction result is obtained after data prediction is finished; analyzing the prediction result after the prediction result is obtained, forming customer classification after the prediction result is analyzed, and finally ending the system; s2, judging whether 24 o'clock exists or not during data cleaning, if the 24 o'clock electric quantity exists, performing user number electric quantity cleaning, and then obtaining a customer electric quantity summary; s3, if not, metering point electric quantity reading cleaning is carried out by using 96 o'clock reading, then metering point electric quantity summarization is carried out, then household number electric quantity summarization is carried out, and finally customer electric quantity summarization is obtained; and S4, data prediction: firstly entering customer data query, and performing type prediction on customer data.

Description

Technical Field

[0001] The present invention relates to the technical field of electricity consumption prediction, and in particular to a data analysis electricity consumption prediction service system. Background Art

[0002] The allowable deviation range of the electricity consumption side of the wholesale market is -10% to 10%, and the electricity within the allowable deviation range is settled according to the contract electricity price. The electricity that exceeds the allowable deviation range is settled according to the contract electricity price but with a surcharge assessment. The deviation electricity assessment will increase the cost of the power sales company and affect its profitability. Based on the historical data of the customer side, the classification regression analysis of the electricity load is carried out to complete the electricity forecast work and steadily promote the company's market-oriented electricity sales business. Therefore, there is an urgent need for an electricity forecast service system that can improve the electricity forecast. Summary of the invention

[0003] The purpose of the present invention is to solve at least one of the technical problems existing in the prior art and to provide a data analysis power consumption prediction service system that can solve the above-mentioned problem.

[0004] To achieve the above object, the present invention provides the following technical solutions: a data analysis power consumption prediction service system, including the following specific system process S1: the system starts to enter the database extraction, performs data cleaning after data extraction, enters the data prediction after data cleaning, obtains the prediction result after the data prediction is completed, analyzes the prediction result after the prediction result is obtained, forms customer classification after the prediction result is analyzed, and finally the system ends;

[0005] S2: When cleaning the data, determine whether it is 24 o'clock. If it is 24 o'clock, clean the account number electricity, and then obtain the customer electricity summary;

[0006] S3: No 24 points, use 96 points to clean the meter point electricity indication, then summarize the meter point electricity, then summarize the household electricity, and finally get the customer electricity summary;

[0007] S4: Data prediction, first enter the customer data query and predict the type of customer data.

[0008] Preferably, the prediction types are divided into two categories. The first type is an informer model, which uses a php script called by java, then writes data in data or file mode, and then writes the parameters backward.

[0009] Preferably, the prediction types are divided into two categories. The second type of machine prediction, such as double gray, enters data cleaning and extreme value judgment after judgment, and also enters the prediction black box after judgment.

[0010] Preferably, after the two types of S4 prediction types enter the prediction black box, they both go through the following system process: enter the prediction black box, analyze the abnormalities of the prediction data, save the prediction results when the analysis result is no, enter the prediction results, analyze the prediction results, form customer classification, and end the system.

[0011] Preferably, enter the prediction black box, analyze the predicted data anomaly, and when the analysis result is yes, perform secondary cleaning, delete the added data, and predict the data anomaly again through the prediction black box until the result is no.

[0012] Preferably, it also includes a weather and classification factor parameter database, and the weather data, customer data and other databases are pushed into the weather and classification factor parameter database by the intelligent network for storage.

[0013] Preferably, the prediction structure is saved, and the parameters of the relevant influencing factors are retained and also stored in the weather and classification factor parameter database.

[0014] Preferably, the formed customer type is also entered into the weather and classification factor parameter database for storage.

[0015] Compared with the prior art, the present invention has the following beneficial effects:

[0016] (1) Improve the data analysis power consumption forecasting service system. Through this system, customers can be classified and power consumption can be predicted more accurately. In this way, data collection, processing, analysis, sharing and other services for power trading can be provided to the power sales business based on the predicted power consumption, so as to improve the operating efficiency and professional strength of the company's business and achieve the optimal configuration of power resources on the customer side and the power plant side. By evaluating these factors, the overall situation of market risks can be obtained, and corresponding risk management strategies can be formulated, stable market rules can be formulated, market supervision can be strengthened, and market transparency can be improved to reduce the probability and impact of market risks. This can improve the security and stability of the power trading system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention is further described below in conjunction with the accompanying drawings and embodiments:

[0018] Figure 1 It is a schematic diagram of the system flow structure of the present invention. DETAILED DESCRIPTION

[0019] This section will describe in detail the specific embodiments of the present invention. The preferred embodiments of the present invention are shown in the accompanying drawings. The purpose of the accompanying drawings is to supplement the description of the text part of the specification with graphics, so that people can intuitively and vividly understand each technical feature and the overall technical solution of the present invention, but it cannot be understood as a limitation on the scope of protection of the present invention.

[0020] In the description of the present invention, it should be understood that descriptions involving orientations, such as up, down, front, back, left, right, etc., and orientations or positional relationships indicated are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0021] In the description of the present invention, "greater than", "less than", "exceed" etc. are understood as not including the number itself, and "above", "below", "within" etc. are understood as including the number itself. If there is a description of "first" or "second", it is only used for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0022] In the description of the present invention, unless otherwise clearly defined, terms such as setting, installing, connecting, etc. should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meanings of the above terms in the present invention based on the specific content of the technical solution.

[0023] See also Figure 1 , the present invention provides a technical solution: a data analysis power consumption prediction service system, including the following specific system process S1: the system starts to enter the database extraction, after the data extraction, the data is cleaned, after the data is cleaned, the data prediction is entered, after the data prediction is completed, the prediction result is obtained, the prediction result is analyzed, after the prediction result is analyzed, the customer classification is formed, and finally the system ends;

[0024] S2: When cleaning the data, determine whether it is 24 o'clock. If it is 24 o'clock, clean the account number electricity, and then obtain the customer electricity summary;

[0025] S3: No 24 points, use 96 points to clean the meter point electricity indication, then summarize the meter point electricity, then summarize the household electricity, and finally get the customer electricity summary;

[0026] S4: Data prediction, first enter customer data query and predict the type of customer data;

[0027] There are two types of predictions. The first type is the informer model, which uses a php script called by Java, then writes data in data or file mode, and then writes parameters backwards.

[0028] There are two types of predictions. The second type is machine prediction, such as double gray, which then enters data cleaning and extreme value judgment, and also enters the prediction black box after judgment;

[0029] After entering the prediction black box, the two types of S4 prediction types will go through the following system process: enter the prediction black box, analyze the abnormality of the prediction data, save the prediction result if the analysis result is negative, enter the prediction result, analyze the prediction result, form customer classification, and end the system;

[0030] Enter the prediction black box and analyze the predicted data anomalies. When the analysis result is yes, perform secondary cleaning, delete the added data, and predict the data anomalies through the prediction black box again until the result is no.

[0031] Furthermore, a data analysis power consumption prediction service system also includes a weather and classification factor parameter database, and the weather data, customer data and other databases are pushed into the weather and classification factor parameter database for storage by the intelligent network;

[0032] Among them, the prediction structure is saved, and the parameters of related influencing factors are also saved in the weather and classification factor parameter database;

[0033] Among them, the customer type is also entered into the weather and classification factor parameter database for storage.

[0034] Embodiment 1: Data fitting and classification rules,

[0035] 1. Fitting rules for 96-point indication data

[0036] 1. Fitting zero point and 96 points

[0037] ① If data exists at both zero and 96 points, then pass it.

[0038] ② When it is determined that at least one point is empty, query the data of the two days before and after and splice the data of the day. If any of the two days does not exist, use all-0 data for temporary supplementation, and interpolate the missing zero or 96-point data.

[0039] Assume that the zero point is C, the data point before the zero point is A, and the data point after the zero point is B, then the calculation formula is:

[0040] Point A reading + (point B reading - point A reading) / (distance between AB) × (distance between AC) = point C reading;

[0041] 2. Single point missing at non-zero point and 96 points

[0042] ① Calculate the mean insertion. Assuming the defect is C, the data point before the defect is A, and the data point after the defect is B, the calculation formula is:

[0043] Point A indication + (point B indication - point A indication) / (distance between AB) = point C indication.

[0044] 3. Continuous drop in points within one day

[0045] ① Fitting by using the power ratio value;

[0046] Assume that there are seven consecutive points A, B, C, D, E, F, and G in one day, among which point A and G have data, and other points are missing, then first calculate the proportion of the five points B, C, D, E, and F in the historical data of the past 5 days, for example, the proportion between AB = the difference between points AB in the past 5 days / 5, so assuming that the difference ratios of B, C, D, E, and F are 0.04, 0.05, 0.09, 0.07, and 0.05 respectively;

[0047] Then the indication at point B = the indication at point A + (the indication at point G - the indication at point A) x the ratio of B value / (the sum of the ratios of B+C+D+E+F);

[0048] The data is considered in three dimensions: working days, non-working days, and holidays. When the initial data has no proportion, it cannot be corrected. When there is no data in the past five days, the data of the last five days of the same period are used for correction.

[0049] 4. When there is no data fitting between zero point and 96 point (initial data and end data)

[0050] For some eligible customers, the most recent data with the same value is used to fill in the gaps;

[0051] 5. If the fitted power is missing for the whole day and is also missing yesterday or the next day, it cannot be corrected and needs to be manually collected and fitted.

[0052] 2. 24 and 96-point power fitting rules

[0053] 1. The basic rules of 24-point power are consistent with the 96-point indication rules 1 to 6;

[0054] 2. Eliminate erroneous negative power values ​​and set them to 0 (caused by historical data errors);

[0055] 3. Eliminate the power exceeding 100 and set it to 0 (the error caused by continuous data interruption, and the value of 0 is used to supplement the power calculation for the interrupted days);

[0056] 4. The power outage or consecutive power outages will be filled with the same value of the most recent power (the new version of the system adds a filling mark).

[0057] 3. Classification rules

[0058] 1. Customer stability rules

[0059] ① Compare the percentage deviation between the customer's 24-point history and forecast for one day;

[0060] Example: Deviation = (forecast - history) ÷ history × 100%;

[0061] ② When there are at least 21 points below 24 points with a deviation within 10%, the customer is considered stable on that day;

[0062] ③ Within 30 days >= 25 days stable ==> Customer stable mark;

[0063] Within 30 days>=16&&<25 days==> The customer is relatively stable;

[0064] <16 within 30 days ==> Customer unstable mark.

[0065] 2. Customer classification rules

[0066] ① Calculate the sum of 11:00 to 22:00 in 10 days as A value, the sum of 23:00 to 11:00 as B value, and the sum of all days in 10 days as C value;

[0067] ②A value>0.65×C value==>Work from 9 am to 9 pm;

[0068] ③A value < 0.35 × C value ==> night work;

[0069] ④ Other working hours that are approximately the same shall be deemed as full-day work.

[0070] 3. Customer cycle rules

[0071] ① Extract 30 days of customer data and make separate judgments on each point from 1 to 24;

[0072] For example: Take the second point of a certain day for judgment, and if the value is <0.8×(average within 30 days) or the value is >1.3×(average within 30 days), it is judged that the change of day and point is abnormal;

[0073] ② The cumulative number of abnormal changes for 4 to 12 days determines that the customer's point change is in a periodic or unstable state;

[0074] ③ Determine the time point of the customer cycle and the day of the week the cycle falls on;

[0075] ④ If the customer is in a cycle state for more than 10 points out of a total of 24 points, the customer will be marked as a cycle.

[0076] Furthermore, the system can classify customers and predict electricity more accurately, so that the data collection, processing, analysis, sharing and other services of electricity trading can be provided for the electricity sales business according to the predicted electricity, so as to improve the operating efficiency and professional strength of the company's business and achieve the optimal configuration of power resources on the customer side and the power plant side. By evaluating these factors, the overall situation of market risks can be obtained, and corresponding risk management strategies can be formulated, stable market rules can be formulated, market supervision can be strengthened, market transparency can be improved, etc., so as to reduce the probability and impact of market risks. This can improve the security and stability of the power trading system.

[0077] The embodiments of the present invention are described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Various changes can be made within the knowledge scope of ordinary technicians in the technical field without departing from the purpose of the present invention.

Claims

1. A data analysis power consumption prediction service system, characterized by: The system includes the following specific process S1: the system starts to enter the database extraction, performs data cleaning after data extraction, enters the data prediction after data cleaning, obtains the prediction result after the data prediction is completed, analyzes the prediction result after the prediction result is obtained, forms the customer classification after the prediction result is analyzed, and finally the system ends; S2: When cleaning the data, determine whether it is 24 o'clock. If it is 24 o'clock, clean the account number electricity, and then obtain the customer electricity summary; S3: No 24 points, use 96 points to clean the meter point electricity indication, then summarize the meter point electricity, then summarize the household electricity, and finally get the customer electricity summary; S4: Data prediction, first enter the customer data query and predict the type of customer data.

2. A data analysis power consumption prediction service system according to claim 1, characterized in that: The prediction types are divided into two categories. The first type is the informer model, which uses a php script called by java, then writes data in data or file mode, and then writes the parameters backward.

3. The data analysis power consumption prediction service system according to claim 1, characterized in that: There are two types of predictions. The second type is machine prediction, such as double gray, which then enters data cleaning and extreme value judgment, and also enters the prediction black box after judgment.

4. The data analysis power consumption prediction service system according to claim 1, characterized in that: After entering the prediction black box, both types of S4 prediction types will go through the following system process: enter the prediction black box, analyze the abnormalities of the prediction data, save the prediction results when the analysis result is no, enter the prediction results, analyze the prediction results, form customer classification, and end the system.

5. The data analysis power consumption prediction service system according to claim 1, characterized in that: Enter the prediction black box and analyze the predicted data anomalies. When the analysis result is yes, perform secondary cleaning, delete the added data, and predict the data anomalies through the prediction black box again until the result is no.

6. The data analysis power consumption prediction service system according to claim 1, characterized in that: It also includes weather and classification factor parameter databases. Weather data, customer data and other databases are pushed into the weather and classification factor parameter database for storage by the intelligent network.

7. A data analysis power consumption prediction service system according to claim 6, characterized in that: Save the forecast structure, retain the relevant influencing factor parameters and save them in the weather and classification factor parameter database.

8. The data analysis power consumption prediction service system according to claim 1, characterized in that: The customer type is also entered into the weather and classification factor parameter database for storage.

Images