Electric quantity monitoring system for intelligent park

Abstract

The invention discloses an electric quantity monitoring system for an intelligent park, and the system comprises a data collection module which is used for extracting evaluation indexes from a historical database; the power distribution module is used for obtaining a power prediction value of the sub-park in a prediction period and predicted power consumption in a monitoring period through a power prediction model constructed based on a neural network model; and determining the power demand priority of the sub-park according to the evaluation index, and constructing a multi-target optimization model based on the electric quantity predicted value of the sub-park to obtain a power distribution scheme, and the electric quantity monitoring and regulation and control module is used for obtaining the actual electricity consumption of the sub-park in the monitoring period by using monitoring equipment, comparing the actual electricity consumption of the sub-park in the previous period with the predicted electricity consumption of the sub-park in the corresponding period, and adjusting an electric power distribution scheme according to the electricity demand priority of the sub-park and a comparison result. The invention relates to the field of park management, and solves the technical problem that a power distribution scheme in an existing electric quantity monitoring system is not intelligent enough.

Description

Technical Field

[0001] The invention belongs to the field of intelligent park management, relates to power monitoring and power distribution technology, and specifically is an power monitoring system for an intelligent park. Background Art

[0002] Smart parks are modern industrial parks that use modern information technology and intelligent systems to integrate various resources and services within the park to achieve efficient, green and sustainable development. Smart parks are characterized by highly integrated management and service systems. Through the Internet of Things technology, big data analysis and artificial intelligence, they realize functions such as energy management automation, intelligent security monitoring, precise environmental monitoring, and transportation and logistics optimization. In smart parks, power monitoring and power distribution are the key to ensuring energy efficiency, controlling costs, reducing environmental pollution and improving user satisfaction. Accurate power monitoring not only helps to identify and reduce unnecessary energy waste and achieve the goal of energy conservation and emission reduction, but also effectively controls power costs and optimizes operational efficiency. In addition, reasonable power distribution can minimize the impact on the environment and meet the requirements of sustainable development.

[0003] However, existing power monitoring systems usually rely on simple historical data analysis to predict future power demand, which can easily lead to inaccurate prediction results. When an imbalance in power supply and demand is detected, the existing system often takes a long time to adjust the power distribution plan, resulting in a slow response to energy management in the park and an inability to quickly respond to emergencies or demand fluctuations. In addition, traditional systems lack flexibility and dynamic optimization capabilities, making it difficult to achieve real-time, personalized power resource allocation, limiting the potential of smart parks in terms of energy efficiency improvement and cost control. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an electric power monitoring system for a smart park, which is used to solve the technical problem that the power distribution scheme in the existing electric power monitoring system is not accurate and intelligent enough.

[0005] To achieve the above object, the present invention provides an electric quantity monitoring system for a smart park, comprising:

[0006] Data collection module: used to collect calendar data, meteorological data and sub-park characteristic data of each day, and extract evaluation indicators from the historical database;

[0007] Power distribution module: used to input the calendar data, meteorological data and sub-park characteristic data of each day within the forecast period into the power forecast model, obtain the daily power consumption forecast values ​​of several sub-parks, and sum the daily power consumption forecast values ​​within the forecast period and the monitoring period respectively, to obtain the power forecast values ​​of several sub-parks and the predicted power consumption of several sub-parks; wherein the power forecast model is constructed based on the historical database and the neural network model, and the forecast period duration is longer than the monitoring period duration; and,

[0008] Determine the power demand priorities of several sub-parks according to the evaluation indicators, and build a multi-objective optimization model based on the power forecast values ​​of the several sub-parks to obtain a power distribution plan;

[0009] Power monitoring and control module: Use monitoring equipment to obtain the actual power consumption of several sub-parks during the monitoring period, and compare the actual power consumption of several sub-parks in the previous period with the predicted power consumption of the sub-parks in the corresponding period to obtain the comparison result, and adjust the power distribution plan according to the power demand priority of the sub-park and the comparison result.

[0010] By leveraging big data and machine learning technologies, we can accurately predict future electricity demand and achieve more efficient power resource management. We can formulate power distribution plans based on the forecast results and dynamically adjust the distribution plans through real-time monitoring data to ensure the optimal balance between power supply and demand. By precisely controlling and optimizing power distribution, we can help improve energy efficiency and reduce unnecessary energy waste, thereby reducing operating costs and environmental impact.

[0011] Furthermore, the energy efficiency ratio is expressed by the formula Calculated; among them, EER i Represents the energy efficiency ratio of the ith sub-park, Pout i Represents the output power of the ith sub-park, Pin i represents the input power of the ith sub-park;

[0012] It should be noted that the input power and output power of each sub-park are obtained through the intelligent control system equipped in the park.

[0013] Furthermore, the average power factor of the production equipment is calculated by the formula Calculated, where PF i represents the average power factor of the production equipment in the ith sub-park, PF iq represents the power factor of the qth production equipment in the ith sub-park, and Q represents the number of production equipment in the ith sub-park;

[0014] It should be noted that the power factor of the production equipment in each sub-park is obtained through the intelligent control system equipped in the park.

[0015] Furthermore, the power consumption prediction model is constructed based on the historical database and the neural network model, including the following steps:

[0016] A1, collect sub-park characteristic data, sub-park historical power consumption data, historical calendar data and historical meteorological data to obtain model construction data; wherein the sub-park characteristic data includes the sub-park identifier, building area, number of production equipment, average power factor of production equipment, number of personnel and personnel density; the sub-park historical power consumption data is daily power consumption; the historical calendar data includes working days, holidays, and activity days; the historical meteorological data includes temperature and sunshine time;

[0017] A2, converting the model building data into a unified format and performing data cleaning to obtain pre-processed data;

[0018] A3, construct a random forest model, and divide the preprocessed data into a training set and a validation set, use the training set and the validation set to train and validate the random forest model, iteratively adjust the model according to the validation results, and finally obtain an electricity prediction model whose input is calendar data, meteorological data and characteristic data of sub-parks, and whose output is daily electricity consumption data of several sub-parks.

[0019] By integrating multi-source data to establish a power forecasting model, the accuracy of the model's prediction can be improved. The addition of sub-park characteristic data helps the model understand the specific energy consumption patterns of different sub-parks, thereby making more personalized and accurate power forecasts, helping the park to allocate resources more efficiently and optimize operational efficiency.

[0020] Furthermore, determining the power demand priorities of the plurality of sub-parks according to the evaluation index comprises the following steps:

[0021] B1, using the range standardization formula The evaluation index is subjected to data standardization; wherein x ij is the original value of the jth evaluation index of the ith sub-park, min(x j ) and max(x j ) are the minimum and maximum values ​​of the jth evaluation index in all sub-parks, z ij is the jth evaluation index value of the ith sub-park after range standardization;

[0022] B2, through the formula p ij =w ij × ij The data after range standardization is weighted and standardized to obtain standardized data; where w ij is the weight of the jth evaluation index value of the ith sub-park, pij is the weighted and standardized value of the jth evaluation index of the ith sub-park; the weight of each evaluation index is determined by the expert scoring method;

[0023] B3, use the TOPSIS evaluation method to process the standardized data and obtain the priority ranking of electricity demand of each sub-park.

[0024] In the process of preprocessing the evaluation indicators, weights are assigned to each evaluation indicator through the expert scoring method, which reflects the importance of different indicators on electricity demand, making the evaluation system more scientific and objective; the comprehensive range standardization and weighted standardization processing helps to improve the accuracy of the judgment of sub-park electricity demand and provide solid data support for the subsequent allocation and optimization of power resources.

[0025] Furthermore, the use of the TOPSIS evaluation method to process the standardized data includes the following steps:

[0026] C1, arrange the standardized data in rows to construct a standardized matrix; each row of the matrix represents the values ​​of several evaluation indicators corresponding to a park, and each column represents the values ​​of an evaluation indicator corresponding to several sub-parks;

[0027] C2, the maximum value of energy efficiency ratio, the minimum value of electricity cost and carbon emission are taken as the positive ideal solution, and the minimum value of energy efficiency ratio, the maximum value of electricity consumption and carbon emission are taken as the negative ideal solution;

[0028] C3, using the formula Calculate the distance from each sub-park to the positive ideal solution and the distance to the negative ideal solution in, represents the positive ideal solution of the jth evaluation index, represents the negative ideal solution of the jth evaluation index, n represents the number of evaluation indexes and is a positive integer;

[0029] C4, using the formula Calculate the closeness C of several sub-parks i , and sort the proximity to obtain the electricity demand priority of each sub-park.

[0030] Furthermore, a multi-objective optimization model is constructed based on the power forecast values ​​of the plurality of sub-parks, including the following steps:

[0031] D1, determine the electricity distribution objectives as minimizing carbon emissions and minimizing electricity costs, and define the carbon emissions function f1(x) and the electricity cost function f2(x);

[0032] D2, count the current power supply S of the park, and according to the constraints The NSGA-II algorithm is used to solve the minimum f1(x) and the minimum f2(x) to obtain the power distribution plan for each park; where x i represents the power consumption of the ith sub-park, y i It represents the predicted power consumption of the ith sub-park during the prediction period, and N represents the number of sub-parks.

[0033] Furthermore, the carbon emission function f1(x) is expressed as: Among them, α im is the carbon emission factor of the mth energy source in the ith sub-park, x im represents the amount of electricity used by the mth energy source in the i-th sub-park, η im represents the conversion efficiency of the mth energy source in the ith sub-park, k i It represents the electricity operation efficiency coefficient of the ith sub-park, and s(t) represents the seasonal adjustment factor.

[0034] Furthermore, the expression of the seasonal adjustment factor s(t) is: Among them, σ represents the amplitude adjustment coefficient, which determines the relative size of seasonal changes, t represents the month, and φ represents the peak month of electricity consumption in the i-th sub-park in the previous year.

[0035] Furthermore, the expression of the power cost function f2(x) is: Among them, c i represents the electricity cost coefficient of the ith sub-park, p i represents the fixed electricity cost of the ith sub-park, and β is the cost growth coefficient, which represents the rate at which the unit cost increases when electricity usage increases.

[0036] Taking minimizing carbon emissions and electricity costs as the dual goals of electricity distribution, it takes into account both economic benefits and environmental protection, and meets the requirements of sustainable development. It also fine-tunes electricity use through seasonal adjustment factors, taking into account the electricity consumption characteristics and peak-valley changes in different seasons, thereby enhancing the adaptability of the multi-objective optimization model and the reliability of the electricity distribution plan.

[0037] Further, adjusting the power distribution plan according to the power demand priorities of the sub-parks and the comparison results includes:

[0038] E1, calculate the difference between the actual power consumption of several sub-parks in the previous cycle and the predicted power consumption in the corresponding cycle, obtain the power consumption difference, and determine whether the power consumption difference is a negative number; if so, keep the current power allocation plan; if not, jump to E2;

[0039] E2: Determine whether there is abnormal power consumption in each sub-park based on the monitoring node; if yes, start the alarm; if not, jump to E3;

[0040] E3, according to formula E i =1.2e i Calculate the power supplement value of several sub-parks and distribute the power of several sub-parks according to the priority of power demand; where E i represents the power supplement value of the ith sub-park, e i Represents the power consumption difference of the i-th sub-park.

[0041] Further, judging whether each sub-park has abnormal power consumption according to the monitoring node includes the following steps:

[0042] Collect the maximum voltage value V, maximum current value I, and standard frequency f of the sub-park for normal operation;

[0043] Record and calculate the average voltage value V', average current value I', average frequency value f' and current average power factor PF' of the monitoring node within the preset monitoring period;

[0044] Calculate the anomaly score according to the formula where λ d (d=1,2,3,4) represents the weight coefficient of each parameter;

[0045] It is determined whether the abnormality score is greater than the abnormality threshold; if yes, the sub-park has abnormal power consumption; if no, the sub-park has no abnormal power consumption.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] The use of big data and machine learning technology to accurately predict electricity demand improves energy efficiency; through a multi-objective optimization model that minimizes carbon emissions and electricity costs, it not only pursues maximum economic benefits, but also pays attention to environmental protection and reduces the burden on the environment; the power monitoring and control module can monitor and record the actual power consumption of each sub-park in real time, compare it with the predicted value, adjust the power distribution plan, and detect abnormal power consumption, quickly discover and respond to abnormal power consumption, and improve management flexibility and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 A schematic diagram of the working process of a power monitoring system for a smart park provided by the present invention;

[0050] Figure 2 A schematic diagram of a process for determining the power demand priorities of several sub-parks provided by the present invention;

[0051] Figure 3 A schematic diagram of the workflow of the power monitoring and control module provided by the present invention. DETAILED DESCRIPTION

[0052] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] See also Figure 1-Figure 3 The embodiment of the present invention provides a power monitoring system for a smart park, including:

[0054] Data collection module: used to collect calendar data, meteorological data and sub-park characteristic data of each day, and extract evaluation indicators from the historical database;

[0055] Power distribution module: used to input calendar data, meteorological data and sub-park characteristic data of each day within the forecast period into the power forecast model, obtain the daily power consumption forecast values ​​of several sub-parks, sum the daily power consumption forecast values ​​within the forecast period and the monitoring period, respectively, to obtain the power forecast values ​​of several sub-parks and the predicted power consumption of several sub-parks; wherein the power forecast model is constructed based on the historical database and the neural network model, and the forecast period duration is longer than the monitoring period duration; and,

[0056] Determine the power demand priorities of several sub-parks based on the evaluation indicators, and build a multi-objective optimization model based on the power forecast values ​​of several sub-parks to obtain a power distribution plan;

[0057] Power monitoring and control module: Use monitoring equipment to obtain the actual power consumption of several sub-parks during the monitoring period, and compare the actual power consumption of several sub-parks in the previous period with the predicted power consumption of several sub-parks in the corresponding period to obtain the comparison results, and adjust the power distribution plan according to the power demand priority of the sub-park and the comparison results.

[0058] In this embodiment, the smart park is equipped with an intelligent control system, which can obtain the input power, output power and power factor of each sub-park and then calculate the power factor of the production equipment through the formula Calculate the energy efficiency ratio EER of the i-th sub-park i And the average power factor PF of the production equipment i ; Among them, Pouti Represents the output power of the ith sub-park, Pin i represents the input power of the ith sub-park, PF iq It represents the power factor of the qth production equipment in the ith sub-park, and Q represents the number of production equipment in the ith sub-park.

[0059] Predicting future power consumption is a key part of efficient energy management in smart parks, which can make full use of the park's intelligent technology to achieve reasonable resource allocation. Therefore, in order to accurately predict power consumption in the future, big data and machine learning technologies are used to process historical multi-source data to prepare for subsequent efficient power distribution plans.

[0060] Specifically, we first use big data technology to collect a large amount of data needed to build a power forecasting model, including sub-park characteristic data, sub-park historical power consumption data, historical calendar data, and historical meteorological data.

[0061] in,

[0062] Sub-park characteristic data is used to help the model identify specific energy consumption patterns of different sub-parks, including sub-park identifiers, building areas, number of people, and population density; sub-park historical electricity consumption data is daily electricity consumption, which is used to help the model learn the electricity demand of different sub-parks; historical calendar data is used to help the model identify the peak electricity consumption period of the sub-park based on time information, including working days, holidays, and event days, and "0, 1" is used to represent different states. For example, Wednesday of a certain week is a rest day for Labor Day, and there are no activities in the park, then the calendar data of that day is represented as "010", that is, it is not a working day, it is a holiday, and it is not an event day; historical meteorological data includes temperature and sunshine time, which is used to help the model understand the impact of weather changes on electricity demand;

[0063] Then, the above data is converted into a unified format and cleaned to obtain preprocessed data. A random forest model is constructed using a machine learning algorithm, and the preprocessed data is divided into a training set and a validation set. The training set and the validation set are used to train and validate the random forest model. The model is iteratively adjusted according to the validation results to finally obtain a power prediction model.

[0064] Then, the calendar data, meteorological data and sub-park characteristic data of each day within the forecast period are input; the daily power consumption forecast values ​​of several sub-parks are obtained; the daily power consumption forecast values ​​are summed according to the forecast period and the monitoring period, and the power consumption forecast values ​​of several sub-parks and several predicted power consumptions are obtained.

[0065] In a smart park, different sub-parks have different functions and different contributions to the park, so it is necessary to determine the priority of power distribution. In this embodiment, the TOPSIS evaluation method is used to prioritize each sub-park; specifically,

[0066] First, we need to preprocess the evaluation indicators in the historical database: using the range standardization formula Standardize the data of each evaluation index; ij is the original value of the jth evaluation index of the ith sub-park, min(x j ) and max(x j ) are the minimum and maximum values ​​of the jth evaluation index in all sub-parks, z ij is the jth evaluation index value of the ith sub-park after the range standardization process; then the expert scoring method is used to determine the weight of each evaluation index, and the formula p ij =w ij × ij The data after range standardization is weighted and standardized to obtain standardized data; where w ij is the weight of the jth evaluation index value of the ith sub-park, p ij It is the weighted and standardized value of the jth evaluation index of the ith sub-park;

[0067] Then, the standardized data was processed using the TOPSIS evaluation method to obtain the priority ranking of several sub-parks: the standardized data was constructed into a standardized matrix; the maximum value of the energy efficiency ratio, the minimum value of electricity cost and carbon emissions were taken as the positive ideal solution, and the minimum value of the energy efficiency ratio, the maximum value of electricity consumption and carbon emissions were taken as the negative ideal solution; and the formula was used to calculate the optimal solution. Calculate the distance from each sub-park to the positive ideal solution and the distance to the negative ideal solution in, represents the positive ideal solution of the jth evaluation index, represents the negative ideal solution of the jth evaluation index, n represents the number of evaluation indexes and is a positive integer; then use the formula Calculate the closeness C of several sub-parks i , and sort them according to the proximity to obtain the priority ranking of electricity demand of each sub-park.

[0068] For example, suppose a smart park has three sub-parks A, B, and C, where the carbon emissions of Park A in the previous year were 10 tons, the electricity cost was 100,000 yuan, and the energy efficiency ratio was 0.8; the carbon emissions of Park B in the previous year were 12 tons, the electricity cost was 90,000 yuan, and the energy efficiency ratio was 0.7; the carbon emissions of Park C in the previous year were 11 tons, the electricity cost was 120,000 yuan, and the energy efficiency ratio was 0.9; the weights of each scoring indicator obtained after expert scoring are: carbon emissions = 0.2, electricity cost = 0.3, energy efficiency ratio = 0.5;

[0069] Then, the standardized matrix obtained after weighted standardization of the indicator data of each sub-park is: The maximum value of the energy efficiency ratio, the minimum value of electricity cost and carbon emissions are taken as the positive ideal solution, and the minimum value of the energy efficiency ratio, the maximum value of electricity consumption and carbon emissions are taken as the negative ideal solution, and the positive ideal solution A is obtained. + =(0.5,0,0) and negative ideal solution A - =(0,0.3,0.2); then the distance from each sub-park to the positive ideal solution is obtained according to the distance calculation formula and the distance to the negative ideal solution

[0070] Then, the closeness C of several sub-parks is obtained according to the closeness calculation formula i : The priority order based on proximity is C>A>B.

[0071] In order to reduce electricity costs and environmental pollution, this embodiment takes minimizing carbon emissions and minimizing electricity costs as the electricity allocation goals, and defines the carbon emissions function and electricity cost function

[0072] Among them, α im is the carbon emission factor of the mth energy source in the ith sub-park, x im represents the amount of electricity used by the mth energy source in the i-th sub-park, η im represents the conversion efficiency of the mth energy source in the ith sub-park, k i represents the electricity operation efficiency coefficient of the ith sub-park, s(t) represents the seasonal adjustment factor, and its expression is: Among them, σ represents the amplitude adjustment coefficient, which determines the relative size of seasonal changes, t represents the month, and φ represents the peak month of electricity consumption in the i-th sub-park in the previous year;

[0073] c i represents the electricity cost coefficient of the ith sub-park, p irepresents the fixed electricity cost of the ith sub-park, β is the cost growth coefficient, which indicates the rate at which the unit cost increases when the electricity usage increases;

[0074] Then count the current power supply S of the park and calculate the power supply S according to the constraints. The NSGA-II algorithm is used to solve the minimum f1(x) and the minimum f2(x) to obtain the power distribution plan for each park; where x i represents the power consumption of the ith sub-park, y i It represents the predicted power consumption of the ith sub-park during the prediction period, and N represents the number of sub-parks.

[0075] The power distribution plan is obtained based on historical data and prediction models. In the actual production process, there are often bad power usage phenomena such as power theft and leakage. Therefore, it is necessary to use real-time power monitoring to compare the difference between the actual power and the predicted power, determine which sub-park may have insufficient power, and then further investigate the bad power usage phenomenon. After the problem is solved, power will be redistributed according to the power usage in the next monitoring cycle; if there is no bad power usage phenomenon, power will be supplemented and allocated according to the difference and sub-park priority ranking.

[0076] Specifically, the actual power consumption of each sub-park in the previous monitoring cycle is first counted, and then the difference is made with the predicted power consumption of the corresponding cycle to compare the two. If the actual power consumption is less than the predicted power consumption, it means that the current power allocation plan can ensure the normal operation of the sub-park; if the actual power consumption is greater than the predicted power consumption, it means that the sub-park may have abnormal power consumption during the monitoring period;

[0077] Next, determine whether the sub-park with actual power consumption greater than the predicted power consumption has abnormal power consumption behavior: collect the maximum voltage value V, maximum current value I, and standard frequency f of the sub-park during normal operation; then record and calculate the average voltage value V', average current value I', average frequency value f' and the current average power factor PF' of the monitoring node within the preset monitoring period; then calculate the abnormal score according to the formula where λ d (d=1, 2, 3, 4) represents the weight coefficient of each parameter; judge whether the abnormal score is greater than the abnormal threshold; if yes, the sub-park has abnormal power consumption; if no, the sub-park has no abnormal power consumption;

[0078] For sub-parks without abnormal power consumption, according to formula E i =1.2e i Calculate the power supplement value of several sub-parks and distribute the power of several sub-parks according to the priority of power demand; where E i represents the power supplement value of the ith sub-park, e iIt represents the power consumption difference of the i-th sub-park. For sub-parks with abnormal power consumption, an alarm will be activated. After the staff checks the possible problems of the alarmed monitoring nodes one by one and solves them accordingly, they will record and compare the difference between the actual power consumption and the predicted power consumption in the current monitoring cycle at the beginning of the next monitoring cycle. After ensuring that there is no abnormal power consumption in the sub-park, a new round of power adjustment will be carried out.

[0079] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0080] Working principle of the present invention:

[0081] Firstly, a power forecasting model is built based on the historical database and neural network, and the evaluation index is extracted from the historical database through the data collection module; then, in the power distribution module, the calendar data, meteorological data and characteristic data of the sub-park are input into the power forecasting model to obtain the power forecast value of each sub-park within the forecast period and the predicted power consumption within the monitoring period, and the power demand priority of each sub-park is determined according to the evaluation index, and a multi-objective optimization model is established to formulate a power distribution plan; finally, in the power monitoring and control module, the actual power consumption is recorded by the monitoring equipment, and then the actual power consumption is compared with the predicted power consumption to determine whether there is a power shortage, and to check the abnormal power consumption nodes, and finally the power distribution plan is adjusted according to the priority and comparison results.

[0082] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A power monitoring system for a smart park, characterized in that: include: Data collection module: used to collect calendar data, meteorological data and sub-park characteristic data of each day, and extract evaluation indicators from the historical database; Power distribution module: used to input calendar data, meteorological data and sub-park characteristic data of each day within the forecast period into the power forecast model, obtain daily power consumption forecast values ​​of several sub-parks, sum the daily power consumption forecast values ​​within the forecast period and the monitoring period, respectively, to obtain power forecast values ​​of several sub-parks and predicted power consumption of several sub-parks; wherein the power forecast model is constructed based on the historical database and the neural network model, and the forecast period duration is longer than the monitoring period duration; and, Determine the power demand priorities of several sub-parks according to the evaluation indicators, and build a multi-objective optimization model based on the power forecast values ​​of the several sub-parks to obtain a power distribution plan; Power monitoring and control module: Use monitoring equipment to obtain the actual power consumption of several sub-parks during the monitoring period, and compare the actual power consumption of several sub-parks in the previous period with the predicted power consumption of the sub-parks in the corresponding period to obtain the comparison result, and adjust the power distribution plan according to the power demand priority of the sub-park and the comparison result.

2. The power monitoring system for a smart park according to claim 1 is characterized in that: The power prediction model is constructed based on the historical database and the neural network model, and includes the following steps: A1, collect sub-park characteristic data, sub-park historical power consumption data, historical calendar data and historical meteorological data to obtain model construction data; wherein the sub-park characteristic data includes the sub-park identifier, building area, number of production equipment, average power factor of production equipment, number of personnel and personnel density; the sub-park historical power consumption data is daily power consumption; the historical calendar data includes working days, holidays, and activity days; the historical meteorological data includes temperature and sunshine time; A2, converting the model building data into a unified format and performing data cleaning to obtain pre-processed data; A3, construct a random forest model, and divide the preprocessed data into a training set and a validation set, use the training set and the validation set to train and validate the random forest model, iteratively adjust the model according to the validation results, and finally obtain an electricity prediction model whose input is calendar data, meteorological data and characteristic data of sub-parks, and whose output is daily electricity consumption data of several sub-parks.

3. The power monitoring system for a smart park according to claim 1, characterized in that: Determining the power demand priorities of the plurality of sub-parks according to the evaluation index comprises the following steps: B1, using the range standardization formula The evaluation index is subjected to data standardization; wherein x ij is the original value of the jth evaluation index of the ith sub-park, min(x j ) and max(x j ) are the minimum and maximum values ​​of the jth evaluation index in all sub-parks, z ij is the jth evaluation index value of the ith sub-park after range standardization; B2, through the formula p ij =w ij × ij The data after range standardization is weighted and standardized to obtain standardized data; where w ij is the weight of the jth evaluation index value of the ith sub-park, p ij It is the weighted and standardized value of the jth evaluation index of the ith sub-park, and the weight of the evaluation index is determined by the expert scoring method; B3, use the TOPSIS evaluation method to process the standardized data and obtain the priority ranking of electricity demand of each sub-park.

4. The power monitoring system for a smart park according to claim 3 is characterized in that: The method of processing the standardized data using the TOPSIS evaluation method comprises the following steps: C1, arrange the standardized data in rows to construct a standardized matrix; each row of the matrix represents the values ​​of several evaluation indicators corresponding to a park, and each column represents the values ​​of an evaluation indicator corresponding to several sub-parks; C2, the maximum value of energy efficiency ratio, the minimum value of electricity cost and carbon emission are taken as the positive ideal solution, and the minimum value of energy efficiency ratio, the maximum value of electricity consumption and carbon emission are taken as the negative ideal solution; C3, using the formula Calculate the distance from each sub-park to the positive ideal solution and the distance to the negative ideal solution in, represents the positive ideal solution of the jth evaluation index, represents the negative ideal solution of the jth evaluation index, n represents the number of evaluation indexes and is a positive integer; C4, using the formula Calculate the closeness C of several sub-parks i , and sort the proximity to obtain the electricity demand priority of each sub-park.

5. The power monitoring system for a smart park according to claim 1, characterized in that: Building a multi-objective optimization model based on the power forecast values ​​of the plurality of sub-parks includes the following steps: D1, determine the electricity distribution objectives as minimizing carbon emissions and minimizing electricity costs, and define the carbon emissions function f1(x) and the electricity cost function f2(x); D2, count the current power supply S of the park, and according to the constraints The NSGA-II algorithm is used to solve the minimum f1(x) and the minimum f2(x) to obtain the power distribution plan for each park; where x i represents the power consumption of the ith sub-park, y i It represents the predicted power consumption of the ith sub-park during the prediction period, and N represents the number of sub-parks.

6. The power monitoring system for a smart park according to claim 5, characterized in that: The expression of the carbon emission function f1(x) is: Among them, α im is the carbon emission factor of the mth energy source in the ith sub-park, x im represents the amount of electricity used by the mth energy source in the i-th sub-park, η im represents the conversion efficiency of the mth energy source in the ith sub-park, k i It represents the electricity operation efficiency coefficient of the ith sub-park, and s(t) represents the seasonal adjustment factor.

7. The power monitoring system for a smart park according to claim 6, characterized in that: The expression of the seasonal adjustment factor s(t) is: Among them, σ represents the amplitude adjustment coefficient, which determines the relative size of seasonal changes, t represents the month, and φ represents the peak month of electricity consumption in the i-th sub-park in the previous year.

8. The power monitoring system for a smart park according to claim 5, characterized in that: The expression of the power cost function f2(x) is: Among them, c i represents the electricity cost coefficient of the ith sub-park, p i represents the fixed electricity cost of the ith sub-park, and β is the cost growth coefficient, which represents the rate at which the unit cost increases when electricity usage increases.

9. The power monitoring system for a smart park according to claim 1, characterized in that: The power distribution plan is adjusted according to the power demand priorities of the sub-parks and the comparison results, including: E1, calculate the difference between the actual power consumption of several sub-parks in the previous cycle and the predicted power consumption in the corresponding cycle, obtain the power consumption difference, and determine whether the power consumption difference is a negative number; if so, keep the current power allocation plan; if not, jump to E2; E2: Determine whether there is abnormal power consumption in each sub-park based on the monitoring node; if yes, start the alarm; if not, jump to E3; E3, according to formula E i =1.2e i Calculate the power supplement value of several sub-parks and distribute the power of several sub-parks according to the priority of power demand; where E i represents the power supplement value of the ith sub-park, e i Represents the power consumption difference of the i-th sub-park.

10. The power monitoring system for a smart park according to claim 9, characterized in that: The method of judging whether each sub-park has abnormal power consumption according to the monitoring node includes the following steps: Collect the maximum voltage value V, maximum current value I, and standard frequency f of the sub-park for normal operation; Record and calculate the average voltage value V', average current value I', average frequency value f' and current average power factor PF' of the monitoring node within the preset monitoring period; Calculate the anomaly score according to the formula where λ d (d=1,2,3,4) represents the weight coefficient of each parameter; It is determined whether the abnormality score is greater than the abnormality threshold; if yes, the sub-park has abnormal power consumption; if no, the sub-park has no abnormal power consumption.

Images