Comprehensive energy operation and maintenance management system based on smart park

Abstract

The invention discloses a comprehensive energy operation and maintenance management system based on a smart park, which belongs to the technical field of energy operation and maintenance management and comprises a data acquisition and exchange module, a data analysis and prediction module, an auxiliary decision module, an intelligent scheduling module, a micro-grid management module, an operation and maintenance simulation module, an operation and maintenance management module and a real-time feedback module. Characteristics related to the energy demand are selected through the Pearson's correlation coefficient matrix of the data analysis and prediction module, the model is constructed through the technologies of difference, autoregression, moving average and the like, the trend and periodic changes of the energy demand can be effectively captured, the weighted model result can accurately predict the future energy demand, and the prediction accuracy is improved. The park energy management is facilitated, various decision-making schemes are generated through the auxiliary decision-making module according to the result of the data analysis and prediction module and through predefined decision-making rules and logics, the schemes are evaluated according to preset evaluation standards, decisions are provided for park managers, and the decision-making efficiency and quality are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy operation and maintenance management, and specifically refers to a comprehensive energy operation and maintenance management system based on a smart park. Background Art

[0002] With the rapid development of smart cities, smart parks, as an important part of urban development, have increasingly prominent demands for intelligent and efficient energy operation and maintenance management systems. Traditional energy management systems have problems such as incomplete data collection, inaccurate energy consumption analysis, and low operation and maintenance management efficiency, which make it difficult to meet the requirements of modern smart parks for energy utilization efficiency.

[0003] However, the existing energy operation and maintenance management still has certain defects. The correlation between the selected features of the existing energy operation and maintenance management system and energy demand affects the prediction accuracy. It does not fully consider the nonlinear characteristics of energy demand and the non-stationarity of the data, resulting in the model being unable to adapt well to different energy supply and demand fields, inaccurate prediction results, lacking an effective auxiliary decision-making mechanism, and unable to automatically generate multiple decision-making plans based on the prediction results. Therefore, a comprehensive energy operation and maintenance management system based on smart parks is proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a comprehensive energy operation and maintenance management system based on a smart park to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a comprehensive energy operation and maintenance management system based on a smart park, comprising a data acquisition and exchange module, a data analysis and prediction module, an auxiliary decision module, an intelligent scheduling module, a microgrid management module, an operation and maintenance simulation module, an operation and maintenance management module and a real-time feedback module;

[0006] The data acquisition and exchange module, the data analysis and prediction module, the auxiliary decision module, the intelligent scheduling module, the microgrid management module, the operation and maintenance simulation module and the real-time feedback module are wirelessly connected;

[0007] The data analysis and prediction module is used to analyze and predict energy demand;

[0008] The specific operation method of the data analysis and prediction module includes the following steps:

[0009] A1. Preprocess the energy data obtained by the data acquisition and exchange module and select energy features through the Pearson correlation coefficient matrix;

[0010] A2. Construct an autoregressive model based on difference, autoregression and moving average, and evaluate the model parameters by maximum likelihood evaluation method;

[0011] A3. Forecast energy demand based on selected energy characteristics;

[0012] A4. Combining the autoregressive model results with the energy demand results for weighting;

[0013] A5. Forecast future energy demand after weighting;

[0014] A6. Analyze the predicted results with the actual data.

[0015] Wherein, A1 selects energy features according to the Pearson correlation coefficient matrix, and assumes that the energy data is X, and the energy data X includes i features and n samples, where X = [x1, x2, ..., x i ], x i represents the energy characteristic variable, assuming that the energy demand variable is Y, the implementation formula is:

[0016]

[0017] In the formula, Represents the energy characteristic variable x i and energy demand variables, x i and i represent the value of the ith sample of the source characteristic variable and the value of the ith sample of the energy demand variable, respectively. and Represents the mean of energy characteristic variable x and energy demand variable y.

[0018] Among them, A2 constructs an autoregressive model through difference, autoregression and moving average, and assumes that the difference parameter, autoregression parameter and moving average parameter are d, p and q respectively, and performs ADF stationarity test on the pre-processed data. If the sequence is not stable enough, the non-stationary sequence is converted into a stationary sequence through difference;

[0019] The moving average parameter and autoregressive parameter are determined by autocorrelation and partial autocorrelation respectively. The number of differences is determined by the autoregressive parameter p value of the ADF test. The autoregressive model is constructed according to the difference parameter d, the autoregressive parameter p and the moving average parameter q. The formula is:

[0020] Δ d y t =c+φ p Δ d y t-p +θ q ∈ t-q +∈ t ,

[0021] In the formula, Δ d y t represents the difference value of the current time after d times of difference, c represents the constant term, φ prepresents the autoregressive coefficient, θ q represents the moving average coefficient, ∈ t-q Represents the error at the previous time point, ∈ t Indicates the error at the current time point;

[0022] The parameters of the autoregressive model are evaluated by the maximum likelihood evaluation method, and the implementation formula is:

[0023] ,

[0024] In the formula, L(θ|y) represents the probability of the observed data y occurring under the given parameter θ, T represents the number of time points of the observed data, and μ t represents the expected value of the autoregressive model,

[0025] It represents the residual sum of squares at all time points t, measuring the difference between the actual observed value and the value predicted by the autoregressive model.

[0026] A3 performs energy demand forecasting based on the selected energy characteristics and defines the energy demand forecasting model as h t =tanh(W xh x t +W hh h t-1 +b h ), h t represents the hidden state at time step t, W xh represents the weight matrix input to the hidden layer, x t represents the input vector at time step t, W hh represents the weight matrix from hidden layer to hidden layer, b h represents the bias term of the hidden layer;

[0027] According to the hidden state h t Predict the output at the next time point:

[0028]

[0029] In the formula, W hy represents the weight matrix from the hidden layer to the output layer, b y Represents the bias term of the output layer;

[0030] The difference between the predicted result and the true value of the energy demand forecasting model is measured by the mean square error, and the implementation formula is:

[0031]

[0032] In the formula, y t represents the actual output value, represents the predicted output value,

[0033] It represents the square of the difference between the true value and the predicted value.

[0034] Among them, A4 combines the autoregressive model results with the energy demand results for weighting, and the implementation formula is:

[0035]

[0036] In the formula, represents the predicted value at time t, Δ d y t represents the predicted value of the autoregressive model at time point t, represents the predicted value of the energy demand forecasting model, w1 and w2 represent the weights of the autoregressive model and the energy demand forecasting model respectively.

[0037] By combining the autoregressive model and the energy demand forecasting model and combining the two in a weighted manner, the accuracy and stability of the forecast are further improved. Through time series analysis methods such as differencing, autoregression and moving average, it is able to process non-stationary data and convert non-stationary sequences into stationary sequences through differencing, which enhances the adaptability and flexibility of the model. The parameters of the autoregressive model are determined by the maximum likelihood evaluation method, which can more accurately reflect the inherent laws of the data and improve the model's ability to fit complex energy data.

[0038] Among them, the auxiliary decision-making module receives the data and results of the data analysis and prediction module, and evaluates the results obtained by the data analysis and prediction module through the set threshold rules through predefined decision rules and logic, generates preliminary suggestions based on the evaluation results and needs, generates multiple decision plans based on the preliminary suggestions, and evaluates the decision plans according to predetermined evaluation criteria.

[0039] The data acquisition and exchange module collects energy data information monitored by the park equipment in real time, performs preliminary data processing on the collected data information, and transmits the data to the data analysis and prediction module via wireless connection;

[0040] The operation and maintenance simulation module receives the prediction results and decision solutions of the data analysis and prediction module and the auxiliary decision-making module, and performs operation and maintenance simulation according to the prediction results and decision solutions.

[0041] The intelligent scheduling module generates a scheduling plan according to the decision-making scheme of the auxiliary decision-making module, evaluates the feasibility of the scheduling plan, and performs preliminary verification after generating the scheduling plan;

[0042] The operation and maintenance simulation module and the intelligent scheduling module perform operation and maintenance simulation and generate scheduling plans based on the prediction results and decision-making plans respectively. The feasibility and effectiveness of the scheduling plan are ensured through preliminary verification and evaluation. The microgrid management module allocates resources and generates control instructions according to the scheduling plan to ensure that the equipment is executed according to the plan, thereby improving the efficiency and accuracy of operation and maintenance management.

[0043] The microgrid management module receives the scheduling plan generated by the intelligent scheduling module, allocates and configures corresponding resources according to the scheduling plan, assigns tasks to corresponding devices and generates control instructions, and the devices execute according to the control instructions.

[0044] The operation and maintenance management module receives information from the operation and maintenance simulation module and the microgrid management module, comprehensively manages the simulation results and execution status and records them, and sends the records to the real-time feedback module;

[0045] The real-time feedback module receives information from the operation and maintenance management module, and feeds back the received information to the intelligent scheduling module in real time. The intelligent scheduling module dynamically adjusts the scheduling plan according to the real-time feedback information.

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

[0047] 1. In order to solve the problems of low energy prediction accuracy and operation and maintenance efficiency and improve energy prediction efficiency and operation and maintenance efficiency, the present invention selects features related to energy demand through the Pearson correlation coefficient matrix of the data analysis prediction module, and constructs a model through technologies such as difference, autoregression and moving average, which can effectively capture the trend and cyclical changes of energy demand. The weighted model results can accurately predict future energy demand, which is helpful for park energy management;

[0048] 2. The present invention combines the autoregressive model and the energy demand forecasting model, and combines the two in a weighted manner, thereby further improving the accuracy and stability of the forecast. Through time series analysis methods such as difference, autoregression and moving average, it can process non-stationary data, and convert non-stationary sequences into stationary sequences through difference, thereby enhancing the adaptability and flexibility of the model. The parameters of the autoregressive model are determined by the maximum likelihood evaluation method, which can more accurately reflect the inherent laws of the data and improve the model's ability to fit complex energy data.

[0049] 3. The present invention generates multiple decision-making schemes through the auxiliary decision-making module according to the results of the data analysis and prediction module, through predefined decision-making rules and logic, and evaluates the schemes according to predetermined evaluation criteria, so as to provide decision-making support for park managers and help improve decision-making efficiency and quality;

[0050] 4. The present invention uses the operation and maintenance simulation module and the intelligent scheduling module to perform operation and maintenance simulation and generate a scheduling plan according to the prediction results and decision-making plan, respectively. The feasibility and effectiveness of the scheduling plan are ensured through preliminary verification and evaluation. The microgrid management module performs resource allocation and control instruction generation according to the scheduling plan to ensure that the equipment is executed according to the plan, thereby improving the efficiency and accuracy of operation and maintenance management;

[0051] 5. Through accurate energy demand forecasting and intelligent scheduling management, the present invention can reasonably allocate energy resources within the park to avoid energy waste and excessive consumption. At the same time, combined with operation and maintenance simulation and real-time feedback mechanism, it can continuously optimize energy use plans and improve energy utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a structural schematic diagram of the comprehensive energy operation and maintenance management system based on the smart park of the present invention;

[0053] Figure 2 The operation process of the data analysis and prediction module of the comprehensive energy operation and maintenance management system based on the smart park of the present invention Figure 1 ;

[0054] Figure 3 The operation process of the data analysis and prediction module of the comprehensive energy operation and maintenance management system based on the smart park of the present invention Figure 2 . DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Example

[0057] See also Figure 1-Figure 3 As shown, the present invention provides a technical solution: including a data acquisition and exchange module, a data analysis and prediction module, an auxiliary decision module, an intelligent scheduling module, a microgrid management module, an operation and maintenance simulation module, an operation and maintenance management module and a real-time feedback module;

[0058] The data acquisition and exchange module, the data analysis and prediction module, the auxiliary decision module, the intelligent scheduling module, the microgrid management module, the operation and maintenance simulation module and the real-time feedback module are wirelessly connected;

[0059] The data analysis and prediction module is used to analyze and predict energy demand;

[0060] The specific operation method of the data analysis and prediction module includes the following steps:

[0061] A1. Preprocess the energy data obtained by the data acquisition and exchange module and select energy features through the Pearson correlation coefficient matrix;

[0062] A2. Construct an autoregressive model based on difference, autoregression and moving average, and evaluate the model parameters by maximum likelihood evaluation method;

[0063] A3. Forecast energy demand based on selected energy characteristics;

[0064] A4. Combining the autoregressive model results with the energy demand results for weighting;

[0065] A5. Forecast future energy demand after weighting;

[0066] A6. Analyze the predicted results with the actual data.

[0067] Wherein, A1 selects energy features according to the Pearson correlation coefficient matrix, and assumes that the energy data is X, and the energy data X includes i features and n samples, where X = [x1, x2, ..., x i ], x i represents the energy characteristic variable, assuming that the energy demand variable is Y, the implementation formula is:

[0068]

[0069] In the formula, Represents the energy characteristic variable x i and energy demand variables, x i and i represent the value of the ith sample of the source characteristic variable and the value of the ith sample of the energy demand variable, respectively. and Represents the mean of energy characteristic variable x and energy demand variable y.

[0070] Among them, A2 constructs an autoregressive model through difference, autoregression and moving average, and assumes that the difference parameter, autoregression parameter and moving average parameter are d, p and q respectively, and performs ADF stationarity test on the pre-processed data. If the sequence is not stable enough, the non-stationary sequence is converted into a stationary sequence through difference;

[0071] The moving average parameter and autoregressive parameter are determined by autocorrelation and partial autocorrelation respectively. The number of differences is determined by the autoregressive parameter p value of the ADF test. The autoregressive model is constructed according to the difference parameter d, the autoregressive parameter p and the moving average parameter q. The formula is:

[0072] Δd y t =c+φ p Δ d y t-p +θ q ∈ t-q +∈ t ,

[0073] In the formula, Δ d y t represents the difference value of the current time after d times of difference, c represents the constant term, φ p represents the autoregressive coefficient, θ q represents the moving average coefficient, ∈ t-q Represents the error at the previous time point, ∈ t Indicates the error at the current time point;

[0074] The parameters of the autoregressive model are evaluated by the maximum likelihood evaluation method, and the implementation formula is:

[0075]

[0076] In the formula, L(θ|y) represents the probability of the observed data y occurring under the given parameter θ, T represents the number of time points of the observed data, and μ t represents the expected value of the autoregressive model, It represents the residual sum of squares at all time points t, measuring the difference between the actual observed value and the value predicted by the autoregressive model.

[0077] A3 performs energy demand forecasting based on the selected energy characteristics and defines the energy demand forecasting model as h t =tanh(W xh x t +W hh h t-1 +b h ), h t represents the hidden state at time step t, W xh represents the weight matrix input to the hidden layer, x t represents the input vector at time step t, W hh represents the weight matrix from hidden layer to hidden layer, b h represents the bias term of the hidden layer;

[0078] According to the hidden state h t Predict the output at the next time point:

[0079]

[0080] In the formula, W hy represents the weight matrix from the hidden layer to the output layer, b yRepresents the bias term of the output layer;

[0081] The difference between the predicted result and the true value of the energy demand forecasting model is measured by the mean square error, and the implementation formula is:

[0082]

[0083] In the formula, y t represents the actual output value, represents the predicted output value,

[0084] It represents the square of the difference between the true value and the predicted value.

[0085] Among them, A4 combines the autoregressive model results with the energy demand results for weighting, and the implementation formula is:

[0086]

[0087] In the formula, represents the predicted value at time t, Δ d y t represents the predicted value of the autoregressive model at time point t, represents the predicted value of the energy demand forecasting model, w1 and w2 represent the weights of the autoregressive model and the energy demand forecasting model respectively.

[0088] By combining the autoregressive model and the energy demand forecasting model and combining the two in a weighted manner, the accuracy and stability of the forecast are further improved. Through time series analysis methods such as differencing, autoregression and moving average, it is able to process non-stationary data and convert non-stationary sequences into stationary sequences through differencing, which enhances the adaptability and flexibility of the model. The parameters of the autoregressive model are determined by the maximum likelihood evaluation method, which can more accurately reflect the inherent laws of the data and improve the model's ability to fit complex energy data.

[0089] Among them, the auxiliary decision-making module receives the data and results of the data analysis and prediction module, and evaluates the results obtained by the data analysis and prediction module through the set threshold rules through predefined decision rules and logic, generates preliminary suggestions based on the evaluation results and needs, generates multiple decision plans based on the preliminary suggestions, and evaluates the decision plans according to predetermined evaluation criteria.

[0090] The data acquisition and exchange module collects energy data information monitored by the park equipment in real time, performs preliminary data processing on the collected data information, and transmits the data to the data analysis and prediction module via wireless connection;

[0091] The operation and maintenance simulation module receives the prediction results and decision solutions of the data analysis and prediction module and the auxiliary decision-making module, and performs operation and maintenance simulation according to the prediction results and decision solutions.

[0092] The intelligent scheduling module generates a scheduling plan according to the decision-making scheme of the auxiliary decision-making module, evaluates the feasibility of the scheduling plan, and performs preliminary verification after generating the scheduling plan;

[0093] The operation and maintenance simulation module and the intelligent scheduling module perform operation and maintenance simulation and generate scheduling plans based on the prediction results and decision-making plans respectively. The feasibility and effectiveness of the scheduling plan are ensured through preliminary verification and evaluation. The microgrid management module allocates resources and generates control instructions according to the scheduling plan to ensure that the equipment is executed according to the plan, thereby improving the efficiency and accuracy of operation and maintenance management.

[0094] The microgrid management module receives the scheduling plan generated by the intelligent scheduling module, allocates and configures corresponding resources according to the scheduling plan, assigns tasks to corresponding devices and generates control instructions, and the devices execute according to the control instructions.

[0095] The operation and maintenance management module receives information from the operation and maintenance simulation module and the microgrid management module, comprehensively manages the simulation results and execution status and records them, and sends the records to the real-time feedback module;

[0096] The real-time feedback module receives information from the operation and maintenance management module, and feeds back the received information to the intelligent scheduling module in real time. The intelligent scheduling module dynamically adjusts the scheduling plan according to the real-time feedback information.

[0097] In this example, specifically: A2 constructs an autoregressive model through difference, autoregression and moving average, and assumes that the difference parameter, autoregression parameter and moving average parameter are d, p and q respectively. ADF stationarity test is performed on the pre-processed data. If the sequence is not stable enough, the non-stationary sequence is converted into a stationary sequence through difference;

[0098] The moving average parameter and autoregressive parameter are determined by autocorrelation and partial autocorrelation respectively. The number of differences is determined by the autoregressive parameter p value of the ADF test. The autoregressive model is constructed according to the difference parameter d, the autoregressive parameter p and the moving average parameter q. The formula is:

[0099] Δ d y t =c+φ p Δ d y t-p +θ q ∈ t-q +∈ t ,

[0100] In the formula, Δ d yt represents the difference value of the current time after d times of difference, c represents the constant term, φ p represents the autoregressive coefficient, θ q represents the moving average coefficient, ∈ t-q Represents the error at the previous time point, ∈ t Indicates the error at the current time point;

[0101] The parameters of the autoregressive model are evaluated by the maximum likelihood evaluation method, and the implementation formula is:

[0102]

[0103] In the formula, L(θ|y) represents the probability of the observed data y occurring under the given parameter θ, T represents the number of time points of the observed data, and μ t represents the expected value of the autoregressive model,

[0104] It represents the residual sum of squares at all time points t, measuring the difference between the actual observed value and the value predicted by the autoregressive model.

[0105] In this example, specifically: A3 performs energy demand forecasting based on the selected energy characteristics, and defines the energy demand forecasting model as t=tanh(W xh x t +W hh h t-1 +b h ), h t represents the hidden state at time step t, W xh represents the weight matrix input to the hidden layer, x t represents the input vector at time step t, W hh represents the weight matrix from hidden layer to hidden layer, b h represents the bias term of the hidden layer;

[0106] According to the hidden state h t Predict the output at the next time point:

[0107]

[0108] In the formula, W hy represents the weight matrix from the hidden layer to the output layer, b y Represents the bias term of the output layer;

[0109] The difference between the predicted result and the true value of the energy demand forecasting model is measured by the mean square error, and the implementation formula is:

[0110]

[0111] In the formula, y trepresents the actual output value, represents the predicted output value,

[0112] It represents the square of the difference between the true value and the predicted value.

[0113] In this example, specifically: A4 combines the autoregressive model result with the energy demand result for weighting, and the implementation formula is:

[0114]

[0115] In the formula, represents the predicted value at time t, Δ d y t represents the predicted value of the autoregressive model at time point t, represents the predicted value of the energy demand forecasting model, w1 and w2 represent the weights of the autoregressive model and the energy demand forecasting model respectively.

[0116] Working principle: By collecting energy data information monitored by various devices in the park in real time, the collected data is preliminarily processed, and the processed data is transmitted to the data analysis and prediction module through wireless connection. The Pearson correlation coefficient matrix selects features that are highly correlated with energy demand from the pre-processed data;

[0117] An autoregressive model is constructed according to the difference, autoregressive and moving average methods, and the model parameters are evaluated by the maximum likelihood evaluation method. Energy demand is predicted according to the selected energy characteristics and the defined energy demand prediction model. The prediction results of the autoregressive model are weighted with the results of the energy demand prediction model, and future energy demand is predicted according to the weighted results. The prediction results are compared and analyzed with the actual data to evaluate the performance of the prediction model. The auxiliary decision-making module receives the output results of the data analysis and prediction module, and generates preliminary suggestions based on the predefined decision rules and logic, combined with the results of the data analysis and prediction module, and evaluated through the set threshold rules. A variety of decision plans are generated based on the preliminary suggestions, and these plans are evaluated according to the predetermined evaluation criteria to select the optimal plan. The operation and maintenance simulation module receives the data analysis and prediction module and the auxiliary decision-making module The prediction results and decision-making plans are used to simulate the operation and maintenance according to the prediction results and decision-making plans, and the effects and possible risks of the decision-making plans in actual operations are evaluated. The intelligent scheduling module generates a specific scheduling plan according to the decision-making plan of the auxiliary decision-making module, evaluates the feasibility of the scheduling plan, and performs preliminary verification. The microgrid management module receives the scheduling plan generated by the intelligent scheduling module, allocates and configures the corresponding resources according to the plan, generates control instructions to the corresponding equipment according to the assigned tasks, and the equipment performs operations according to the instructions. The operation and maintenance management module receives information from the operation and maintenance simulation module and the microgrid management module, comprehensively manages the simulation results and execution status, and records them, and sends the recorded information to the real-time feedback module. The real-time feedback module feeds back the recorded information of the operation and maintenance management module to the intelligent scheduling module in real time, and the intelligent feedback module dynamically adjusts the scheduling plan according to the feedback information.

[0118] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0119] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. The integrated energy operation and maintenance management system based on the smart park is characterized by: It includes data acquisition and exchange module, data analysis and prediction module, decision support module, intelligent dispatch module, microgrid management module, operation and maintenance simulation module, operation and maintenance management module and real-time feedback module; The data acquisition and exchange module, the data analysis and prediction module, the auxiliary decision module, the intelligent scheduling module, the microgrid management module, the operation and maintenance simulation module and the real-time feedback module are wirelessly connected; The data analysis and prediction module is used to analyze and predict energy demand; The specific operation method of the data analysis and prediction module includes the following steps: A1. Preprocess the energy data obtained by the data acquisition and exchange module and select energy features through the Pearson correlation coefficient matrix; A2. Construct an autoregressive model based on difference, autoregression and moving average, and evaluate the model parameters by maximum likelihood evaluation method; A3. Forecast energy demand based on selected energy characteristics; A4. Combining the autoregressive model results with the energy demand results for weighting; A5. Forecast future energy demand after weighting; A6. Analyze the predicted results with the actual data.

2. The integrated energy operation and maintenance management system based on the smart park according to claim 1 is characterized by: A1 selects energy features according to the Pearson correlation coefficient matrix. Let the energy data be X. The energy data X includes i features and n samples, where X = [x1, x2, ..., x i ], x i represents the energy characteristic variable, assuming that the energy demand variable is Y, the implementation formula is: In the formula, Represents the energy characteristic variable x i and energy demand variables, x i and i represent the value of the ith sample of the source characteristic variable and the value of the ith sample of the energy demand variable, respectively. and Represents the mean of energy characteristic variable x and energy demand variable y.

3. According to the comprehensive energy operation and maintenance management system based on the smart park according to claim 1, Features: A2, constructing an autoregressive model through difference, autoregression and moving average, assuming that the difference parameter, autoregression parameter and moving average parameter are d, p and q respectively, and performing ADF stationarity test on the pre-processed data. If the sequence stationarity is insufficient, the non-stationary sequence is converted into a stationary sequence through difference; The moving average parameter and autoregressive parameter are determined by autocorrelation and partial autocorrelation respectively. The number of differences is determined by the autoregressive parameter p value of the ADF test. The autoregressive model is constructed according to the difference parameter d, the autoregressive parameter p and the moving average parameter q. The formula is: D d y t =c+φ p D d y t-p +θ q ∈ t-q +∈ t , In the formula, Δ d y t represents the difference value of the current time after d differences, c represents the constant term, φ p represents the autoregressive coefficient, θ q represents the moving average coefficient, ∈ t-q Represents the error at the previous time point, ∈ t Indicates the error at the current time point; The parameters of the autoregressive model are evaluated by the maximum likelihood evaluation method, and the implementation formula is: In the formula, L(θ|y) represents the probability of the observation data y occurring under the given parameter θ, T represents the number of time points of the observation data, and μ|t represents the expected value of the autoregressive model. It represents the residual sum of squares at all time points t, measuring the difference between the actual observed value and the value predicted by the autoregressive model.

4. The integrated energy operation and maintenance management system based on the smart park according to claim 1 is characterized in that: A3, based on the selected energy characteristics, performs energy demand forecasting and defines the energy demand forecasting model as h t =tanh(W xh x t +W hh h t-1 +b h ), h t represents the hidden state at time step t, W xh represents the weight matrix input to the hidden layer, x t represents the input vector at time step t, W hh represents the weight matrix from hidden layer to hidden layer, b h represents the bias term of the hidden layer; According to the hidden state h t Predict the output at the next time point: In the formula, W hy represents the weight matrix from the hidden layer to the output layer, b y Represents the bias term of the output layer; The difference between the predicted result and the true value of the energy demand forecasting model is measured by the mean square error, and the implementation formula is: In the formula, y t represents the actual output value, represents the predicted output value, Represents the square of the difference between the true value and the predicted value.

5. The integrated energy operation and maintenance management system based on the smart park according to claim 1 is characterized in that: A4 combines the autoregressive model results with the energy demand results for weighting, and the implementation formula is: In the formula, represents the predicted value at time t, Δ d y t represents the predicted value of the autoregressive model at time point t, represents the predicted value of the energy demand forecasting model, w1 and w2 represent the weights of the autoregressive model and the energy demand forecasting model respectively.

6. The integrated energy operation and maintenance management system based on the smart park according to claim 1 is characterized by: The auxiliary decision-making module receives the data and results of the data analysis and prediction module, and through predefined decision rules and logic, evaluates the results obtained by the data analysis and prediction module through the set threshold rules, generates preliminary suggestions based on the evaluation results and needs, generates multiple decision plans based on the preliminary suggestions, and evaluates the decision plans according to predetermined evaluation criteria.

7. The integrated energy operation and maintenance management system based on the smart park according to claim 1 is characterized in that: The data acquisition and exchange module collects energy data information monitored by the park equipment in real time, performs preliminary data processing on the collected data information, and transmits the data to the data analysis and prediction module via wireless connection; The operation and maintenance simulation module receives the prediction results and decision solutions of the data analysis and prediction module and the auxiliary decision-making module, and performs operation and maintenance simulation according to the prediction results and decision solutions.

8. The integrated energy operation and maintenance management system based on the smart park according to claim 1 is characterized by: The intelligent scheduling module generates a scheduling plan according to the decision-making scheme of the auxiliary decision-making module, evaluates the feasibility of the scheduling plan, and performs preliminary verification after generating the scheduling plan; The microgrid management module receives the scheduling plan generated by the intelligent scheduling module, allocates and configures corresponding resources according to the scheduling plan, assigns tasks to corresponding devices and generates control instructions, and the devices execute according to the control instructions.

9. The integrated energy operation and maintenance management system based on the smart park according to claim 1 is characterized by: The operation and maintenance management module receives information from the operation and maintenance simulation module and the microgrid management module, comprehensively manages the simulation results and execution status and records them, and sends the records to the real-time feedback module; The real-time feedback module receives information from the operation and maintenance management module, and feeds back the received information to the intelligent scheduling module in real time. The intelligent scheduling module dynamically adjusts the scheduling plan according to the real-time feedback information.

Images