Algorithm for optimizing factory production plan by predicting and monitoring photovoltaic, energy storage, waste heat power generation and electrical load in cement plant

Abstract

The invention relates to an integrated energy management system of a cement plant, in particular to an algorithm for optimizing a factory production plan by predicting and monitoring photovoltaic, energy storage, waste heat power generation and electrical load in the cement plant. The algorithm comprises a data collection unit, and the data collection unit is used for collecting all data needed by the algorithm; the data processing unit is used for processing abnormal values and missing values in the data collection unit; the prediction and monitoring unit comprises a photovoltaic power generation prediction module, a waste heat recovery monitoring module and an electrical load prediction module, the photovoltaic power generation prediction module is used for predicting photovoltaic power generation capacity, the waste heat recovery monitoring module is used for determining recovery potential and efficiency of waste heat, and the electrical load prediction module is used for predicting future electrical load; the production optimization unit adopts a target optimization model to achieve the purposes of energy cost minimization, environmental influence minimization and production efficiency maximization; according to the algorithm, fine management of energy consumption of the cement plant can be realized, the energy utilization efficiency is improved, and the production cost is reduced.

Description

Technical Field

[0001] This invention relates to a comprehensive energy management system for cement plants, specifically, to an algorithm for optimizing factory production plans by predicting and monitoring photovoltaic power generation, energy storage, waste heat power generation, and electricity load in cement plants. Background Technology

[0002] With the global energy crisis and environmental problems becoming increasingly severe, the cement industry, as one of the high-energy-consuming and high-emission industries, faces enormous pressure to conserve energy and reduce emissions. Traditional production planning often relies on experience-based judgment, lacking accurate forecasting and control of energy consumption, leading to energy waste and low production efficiency.

[0003] In cement plants, production planning algorithms are optimized by predicting and monitoring photovoltaic power generation, energy storage, waste heat power generation, and electricity load. By integrating photovoltaic power generation forecasting, waste heat recovery monitoring, and electricity load forecasting, energy efficiency can be improved and waste reduced. Furthermore, based on the forecasting and monitoring data, production plans can be adjusted to achieve the goals of minimizing energy costs, minimizing environmental impact, and maximizing production efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide an algorithm for optimizing factory production planning in cement plants by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, so as to solve the problems of lack of accurate prediction and control of energy consumption, which leads to energy waste and low production efficiency as mentioned in the background art.

[0005] To achieve the above objectives, the present invention aims to provide an algorithm for optimizing factory production planning in cement plants by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, including:

[0006] The data collection unit includes an energy storage device data collection module, a power generation data collection module, and a recycling data collection module. The power generation data collection module is used to collect data on sunshine hours, solar radiation intensity, and temperature. The recycling data collection module is used to collect data on temperature, flow rate, pressure, power consumption, flue gas composition, and ambient humidity during the production process.

[0007] The data processing unit includes an outlier processing module and a missing value processing module. The outlier processing module is used to identify and remove outliers in the data collected by the data collection unit, and the missing value processing module is used to fill in the missing values ​​in the data collected by the data collection unit.

[0008] The prediction and monitoring unit includes a photovoltaic power generation prediction module, a waste heat recovery monitoring module, and an electricity load prediction module. The photovoltaic power generation prediction module predicts photovoltaic power generation based on data collected by the power generation data collection module. The waste heat recovery monitoring module determines the waste heat recovery potential and efficiency based on data collected by the recovery data collection module. The electricity load prediction module predicts future electricity load based on historical electricity consumption data and production plans using a multivariate nonlinear regression model.

[0009] As a further improvement to this technical solution, the power generation data collection module includes a sunshine recorder, a solar radiation meter, and a temperature sensor, which are used to acquire sunshine hours, solar radiation intensity, and key temperature characteristic data, respectively. The recovery data collection module includes a temperature sensor, a flow meter, a pressure sensor, an electricity meter, a gas analyzer, and a hygrometer. The temperature sensor is used to collect kiln temperature, cooler outlet temperature, clinker temperature, flue gas temperature, waste heat boiler inlet and outlet water temperature, heat exchanger inlet and outlet temperature, and ambient temperature. The flow meter is used to collect clinker mass flow rate, flue gas volume flow rate, waste heat boiler steam flow rate, and water flow rate. The pressure sensor is used to collect kiln pressure, flue gas pressure, and waste heat boiler steam pressure. The electricity meter, the gas analyzer, and the hygrometer are used to collect power consumption data, flue gas composition data, and ambient humidity data, respectively.

[0010] As a further improvement to this technical solution, the outlier processing module is used to identify and remove outliers in the data from the power generation data collection module and the data recovery collection module. Specifically, the outlier processing module is as follows:

[0011]

[0012] Where σ represents the standard deviation of the dataset, X represents a single measurement and data point, Z represents the Z-score, which is the standard score of data point x, and μ represents the mean of the dataset.

[0013] As a further improvement to this technical solution, the missing value processing module, based on the data processed by the outlier processing module, is used to fill in the missing values ​​in the power generation data collection module and the recycling data collection module. Specifically, the missing value processing module is as follows:

[0014]

[0015] in, Represents missing values; x j represents non-missing values; N represents the number of non-missing values.

[0016] As a further improvement to this technical solution, the photovoltaic power generation prediction module includes a power generation prediction model and a training model;

[0017] The power generation prediction model is used to predict photovoltaic power generation, and the power generation prediction model is as follows:

[0018] y = β0 + β1x1 + β2x2 + ... + β n x n +∈

[0019] Where y represents the predicted photovoltaic power generation; x1, x2, ..., x n Represents the selected characteristics of sunshine duration, solar radiation intensity, and temperature; β0 is the intercept term; β1, β2, ..., β n Weights representing sunshine duration, solar radiation intensity, and temperature characteristics; ∈ represents error;

[0020] The training model is based on historical power generation data to train the power generation prediction model, in order to find the optimal β1, β2, ..., β n The value, the training model specifically is:

[0021]

[0022] Where m represents the number of samples; y i This represents the actual power generation of the i-th sample. The model represents the predicted power generation for the i-th sample.

[0023] As a further improvement to this technical solution, the waste heat recovery monitoring module specifically comprises:

[0024]

[0025] Where η_total represents the total waste heat recovery efficiency; Q_recovered represents the actual recovered waste heat; m -i c represents the mass flow rate of clinker within i time periods; -i This represents the specific heat capacity of clinker within the i-th time period; ρ represents the temperature drop of the clinker during the i-th time period; -j V represents the density of the flue gas during the j-th time period; -j Cp represents the volumetric flow rate of flue gas during the j-th time period; -j This represents the isobaric specific heat capacity of the flue gas during the j-th time period; K represents the temperature drop of the flue gas during the j-th time period; K_loss represents the equipment heat loss correction factor; K_season represents the seasonal climate correction factor.

[0026] As a further improvement to this technical solution, the multivariate nonlinear regression model is specifically as follows:

[0027]

[0028] in, The predicted electricity load represents the load at time point t; α0 represents the intercept term; α i The autoregressive coefficient representing historical electricity load; L t-i The representative is the historical electricity load at point ti; β j The coefficient representing the production planning variable; P t-j Production plan variable representing time point tj; γ k Representing other explanatory variables X t-k The coefficients include weekday / weekend indicators and seasonal factors; δ ij The coefficients of the interaction term representing historical electricity load and production plan variables; ∈ t represents the error term; n represents the lag order of historical electricity load; m represents the lag order of the production plan variable; o represents the lag order of other explanatory variables;

[0029] Where, α i β j γ k and δ ij The model parameters are estimated based on historical data, and the model is fitted and validated. The specific steps are as follows:

[0030] S7.1 Determine the lag order n, m, and o of the model, and whether interaction terms need to be introduced, through correlation function and partial autocorrelation function analysis;

[0031] S7.2 Use the least squares method to estimate the model parameters;

[0032] S7.3. Use R-squared to evaluate the goodness of fit of the model;

[0033] S7.4 Use the fitted model to predict future electricity load.

[0034] As a further improvement to this technical solution, the relevant function of S7.1 is specifically as follows:

[0035]

[0036] Where ACF represents the correlation function; Y t Represents the observed values ​​of the time series at time point t; Cov represents the covariance; Var represents the variance; k represents the lag order;

[0037] Specifically, the partial autocorrelation function in S7.1 is as follows:

[0038]

[0039] Where PACF represents the partial autocorrelation function Y t Represents the observed values ​​of the time series at time point t; Cov represents the covariance; Var represents the variance; k represents the lag order;

[0040] Specifically, the least squares method in S7.2 is as follows:

[0041]

[0042] in, Represents the model parameter vector, including α i β j γ k and δ ij L t This represents the actual observed electrical load. This represents the predicted electricity load based on parameter θ; T represents the total number of observed data.

[0043] Specifically, the R-squared in S7.3 is:

[0044]

[0045] Among them, R 2 L represents the coefficient of determination; T represents the length of the time series; L represents the coefficient of determination. t This represents the actual observed electricity load value at time point t; This represents the predicted electricity load value at time point t; This represents the average value of the actual electrical load. The expression is This represents the summation of all terms from t=1 to t=T; This represents the squared difference between the predicted value and the actual observed value. This represents the squared difference between the actual observed value and its mean.

[0046] As a further improvement to this technical solution, the target optimization model is specifically as follows:

[0047]

[0048] Where Z represents the total cost, which is minimized; T represents the total number of time periods; c buy This represents the unit cost of purchasing additional electricity from the grid; y t The predicted photovoltaic power generation at time t; The predicted electrical load at time t; η_total t The predicted waste heat recovery energy represents the energy available at time t. This represents the amount of electricity purchased from the power grid; + indicates a positive value. env E represents the environmental impact cost per unit of carbon dioxide emissions; t c represents the carbon dioxide emissions at time t; prod G represents the production cost per unit of non-renewable energy consumed; t Production costs.

[0049] As a further improvement to this technical solution, the target optimization model has constraints, specifically:

[0050]

[0051] Among them, G grid，t k represents the amount of electricity purchased from the grid at time t; CO2 k represents the coefficient that converts energy consumption into carbon dioxide emissions. cost This represents the coefficient that converts energy consumption into production costs; This represents all t.

[0052] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0053] In cement plants, an algorithm for optimizing production planning is developed by predicting and monitoring photovoltaic (PV) power generation, waste heat recovery, and electricity load. By integrating PV power generation forecasting, waste heat recovery monitoring, and electricity load forecasting, the algorithm enables monitoring and management of energy use in cement plants, contributing to efficient energy utilization and reduced production costs. Based on the predicted data, a target optimization model comprehensively considers energy costs, environmental impact, and production efficiency. Through intelligent algorithms, it dynamically adjusts production plans to minimize energy costs, environmental impact, and maximize production efficiency. This helps cement plants reduce operating costs and environmental footprint while ensuring production quality. This patent, by integrating clean energy and control technologies, promotes the green transformation of the industry and has high application value and market prospects. Attached Figure Description

[0054] Figure 1 This is a flowchart illustrating the overall process of the present invention.

[0055] The meanings of the labels in the diagram are as follows:

[0056] 1. Data collection unit; 2. Data processing unit; 3. Prediction and monitoring unit; 4. Production optimization unit. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Example

[0059] Please see Figure 1 As shown, an algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load is provided. The algorithm includes a data collection unit 1, which includes an energy storage data collection module, a power generation data collection module, and a recycling data collection module. The power generation data collection module is used to collect data on sunshine hours, solar radiation intensity, and temperature. The recycling data collection module is used to collect data on temperature, flow rate, pressure, power consumption, flue gas composition, and ambient humidity during the production process.

[0060] Data processing unit 2 includes an outlier processing module and a missing value processing module. The outlier processing module is used to identify and remove outliers in the data collected by data collection unit 1, and the missing value processing module is used to fill in missing values ​​in the data collected by data collection unit 1.

[0061] The prediction and monitoring unit 3 includes a photovoltaic power generation prediction module, a waste heat recovery monitoring module, and an electricity load prediction module. The photovoltaic power generation prediction module is used to predict photovoltaic power generation based on data collected by the power generation data collection module. The waste heat recovery monitoring module is used to determine the waste heat recovery potential and efficiency based on data collected by the recovery data collection module. The electricity load prediction module is used to predict future electricity load based on historical electricity consumption data and production plans, using a multivariate nonlinear regression model.

[0062] Production optimization unit 4, based on the data predicted by prediction and monitoring unit 3, adopts a target optimization model to achieve the goals of minimizing energy costs, minimizing environmental impact, and maximizing production efficiency.

[0063] The power generation data collection module includes a sunshine recorder, a solar radiation meter, and a temperature sensor, which are used to acquire key characteristic data such as sunshine duration, solar radiation intensity, and temperature, respectively. The recovery data collection module includes a temperature sensor, a flow meter, a pressure sensor, an electricity meter, a gas analyzer, and a hygrometer. The temperature sensor is used to collect data on kiln temperature, cooler outlet temperature, clinker temperature, flue gas temperature, waste heat boiler inlet and outlet water temperature, heat exchanger inlet and outlet temperatures, and ambient temperature. The flow meter is used to collect data on clinker mass flow rate, flue gas volume flow rate, waste heat boiler steam flow rate, and water flow rate. The pressure sensor is used to collect data on kiln pressure, flue gas pressure, and waste heat boiler steam pressure. The electricity meter, gas analyzer, and hygrometer are used to collect data on power consumption, flue gas composition, and ambient humidity, respectively.

[0064] The outlier handling module is used to identify and remove outliers from the power generation data collection module and the data recovery data collection module. Specifically, the outlier handling module is as follows:

[0065]

[0066] Where Z represents the standardized value, also known as the Z-score; σ represents the standard deviation of the dataset; X represents a single measurement and data point; μ represents the mean of the dataset, which is the arithmetic mean of all data points; by subtracting the mean μ from the original data point X and then dividing by the standard deviation σ, we can obtain the position information of the data point relative to the entire dataset. If the absolute value of the data point Z is greater than 3, it is considered an outlier.

[0067] The missing value handling module, based on the data processed by the outlier handling module, is used to fill in missing values ​​in the power generation data collection module and the recycling data collection module. Specifically, the missing value handling module is as follows:

[0068]

[0069] in, Represents missing values; x j Represents non-missing values; N represents the number of non-missing values, i.e., the number of data points used to calculate the mean during imputation. This represents a summation operation over all non-missing data points, which adds up all available observations; This represents the arithmetic mean of the non-missing data points, obtained by dividing the sum by the total number of non-missing data points; the formula estimates missing values ​​by calculating the mean of the non-missing data.

[0070] The photovoltaic power generation prediction module includes a power generation prediction model and a training model;

[0071] The power generation prediction model is used to predict photovoltaic power generation. The specific power generation prediction model is as follows:

[0072] y = β0 + β1x1 + β2x2 + ... + β n x n +∈

[0073] Where y represents the predicted photovoltaic power generation; x1, x2, ..., x n Represents the selected characteristics of sunshine duration, solar radiation intensity, and temperature; β0 is the intercept term; β1, β2, ..., β n The weights represent sunshine duration, solar radiation intensity, and temperature characteristics; ∈ represents error; this formula analyzes β1, β2, ..., β n The weights are used to predict photovoltaic power generation y;

[0074] The training model is based on historical power generation data to train the power generation prediction model, which is used to find the optimal β1, β2, ..., β n The value, and the specific training model are as follows:

[0075]

[0076] Where MSE represents mean squared error, measuring the accuracy of the model's predictions; m represents the sample size, i.e., the total number of observations used to evaluate the model; y i This represents the actual power generation of the i-th sample. The model represents the predicted power generation for the i-th sample; the MSE is obtained by averaging the squared prediction errors of all samples, thus reflecting the overall level of deviation between the model's predicted value and the actual value.

[0077] The waste heat recovery monitoring module is as follows:

[0078]

[0079] Where η_total represents the total waste heat recovery efficiency; Q_recovered represents the actual recovered waste heat; m -i c represents the mass flow rate of clinker within i time periods; -i This represents the specific heat capacity of clinker within the i-th time period; ρ represents the temperature drop of the clinker during the i-th time period; ρ-j represents the density of the flue gas during the j-th time period; V -j Cp represents the volumetric flow rate of flue gas during the j-th time period; -j This represents the isobaric specific heat capacity of the flue gas during the j-th time period; K represents the temperature drop of flue gas during the j-th time period; K_loss represents the equipment heat loss correction factor; K_season represents the seasonal climate correction factor; by calculating the total heat energy released by clinker and flue gas, taking into account equipment heat loss and seasonal climate effects, and comparing it with the actual recovered waste heat, the efficiency of the waste heat recovery system is evaluated.

[0080] The specific multivariate nonlinear regression model is as follows:

[0081]

[0082] in, The predicted electricity load represents the load at time point t; α0 represents the intercept term, which represents the expected electricity load value when all independent variables are zero; α i The autoregressive coefficient representing historical electricity load; L t-i The representative is the historical electricity load at point ti; β j The coefficient representing the production planning variable; P t-j Production plan variable representing time point tj; γ k Representing other explanatory variables X t-k The coefficients include weekday / weekend indicators and seasonal factors; δ ij The coefficients of the interaction term representing historical electricity load and production plan variables; ∈ t The term represents the error term; n represents the lag order of historical electricity load; m represents the lag order of the production plan variable; o represents the lag order of other explanatory variables; by adjusting the parameter α... i β j γ k and δ ij This allows for the prediction of electricity load. As close as possible to the actual observed electrical load L t ;

[0083] Where, α i β j γ k and δ ij The model parameters are estimated based on historical data, and the model is fitted and validated. The specific steps are as follows:

[0084] S7.1 Determine the lag order n, m, and o of the model, and whether interaction terms need to be introduced, through correlation function and partial autocorrelation function analysis;

[0085] S7.2 Use the least squares method to estimate the model parameters;

[0086] S7.3. Use R-squared to evaluate the goodness of fit of the model;

[0087] S7.4 Use the fitted model to predict future electricity load.

[0088] The relevant functions in S7.1 are as follows:

[0089]

[0090] Where ACF represents the correlation function; Y t Y represents the observed value of the time series at time point t; Cov represents the covariance; Var represents the variance; k represents the lag order; the autocorrelation function ACF(k) calculates the value of the data point Y at the current time in the time series. t And the data points Y k steps earlier t-k The linear correlation between them has been standardized by the variance of the data points themselves. If the value of ACF(k) is close to 1, it indicates that Y... t and Y t-k There is a strong positive correlation between them; if it is close to -1, there is a strong negative correlation; if it is close to 0, it means that there is almost no linear correlation between them.

[0091] Specifically, the partial autocorrelation function in S7.1 is as follows:

[0092]

[0093] Where PACF represents the partial autocorrelation function Y t Y represents the observed value of the time series at time point t; Cov represents the covariance; Var represents the variance; k represents the lag order; the partial autocorrelation function PACF(k) measures the current value Y of the time series after excluding all intermediate lag effects. t The value Y before k steps t-k The strength of the linear relationship between them;

[0094] The least squares method in S7.2 is as follows:

[0095]

[0096] in, Represents the model parameter vector, including α i β j γ k and δ ij L t This represents the actual observed electrical load. This represents the predicted electricity load based on parameter θ; T represents the total number of observed data. Represents the residual sum of squares (RSS), which measures the model's predicted values. Compared with the actual power load L tThe sum of squares of the differences between them. Here, t is each time point in the time series, from 1 to T, and T is the total number of observations or data points; we minimize the sum of squares of the prediction errors at all time points by adjusting the parameter θ, so that the model's predictions are closer to the actual observed electricity load data, thereby improving the model's prediction accuracy and reliability;

[0097] Specifically, R-squared in S7.3 is:

[0098]

[0099] Among them, R 2 represents the coefficient of certainty, ranging from 0 to 1. The closer the value is to 1, the better the model's predictive performance, meaning the model explains more data variation; T represents the length of the time series, i.e., the total number of data points; L t This represents the actual observed electricity load value at time point t; This represents the predicted electricity load value at time point t; This represents the average value of the actual electrical load. The expression is This represents the summation of all terms from t=1 to t=T; This represents the squared difference between the predicted value and the actual observed value. R represents the squared difference between the actual observed value and its mean; 2 It is an important indicator for evaluating the model's fit and predictive ability. For multivariate nonlinear regression models, it helps us understand the extent to which the model can capture and explain the patterns of electricity load changes.

[0100] The objective optimization model is as follows:

[0101]

[0102] Where Z represents the total cost, which is the objective of the entire optimization model. This value is minimized by adjusting the decision variables; T represents the total number of time periods, meaning that the optimization model considers the total cost from the first time point to the last time point; c buy This represents the unit cost of purchasing additional electricity from the grid. This cost is incurred when the cement plant's own power generation (including photovoltaic power generation and waste heat recovery) is insufficient to meet its electricity load. t The predicted photovoltaic power generation at time t represents the amount of electricity generated by the cement plant itself through photovoltaic panels. η_total represents the predicted electricity load at time t, i.e., the cement plant's projected electricity demand at that time. t The predicted waste heat recovery energy represents time t, which can be converted from the waste heat recovery system in the cement production process. This represents the amount of electricity the cement plant needs to purchase from the grid at time t. The plus sign indicates a positive value, ensuring that a positive number (i.e., the amount of electricity to be purchased) is only calculated when the electricity load exceeds the supply of renewable energy. env The environmental impact cost per unit of carbon dioxide emissions is related to the cement plant's environmental responsibility and potential carbon tax or carbon trading costs; prod This represents the production cost per unit of non-renewable energy consumption, and this cost is directly linked to the cement plant's production efficiency and economic benefits; C t Production cost E t and C t The carbon dioxide emissions and production costs at time t are directly related to the additional electricity the cement plant purchases from the grid.

[0103] The objective optimization model has constraints, specifically:

[0104]

[0105] Among them, G grid，t k represents the amount of electricity purchased from the grid at time t; CO2 k represents the coefficient that converts energy consumption into carbon dioxide emissions. cost This represents the coefficient that converts energy consumption into production costs; The first constraint represents the predicted power load of the cement plant for all t; it ensures that at any time t, the predicted power load of the cement plant is... Not exceeding the predicted photovoltaic power generation y at that moment t Predicting the total energy recovered from waste heat, η_total t Electricity G purchased from the grid grid，t The sum; the second constraint defines the carbon dioxide emissions E at each time t. t The third constraint quantifies the production cost G at each time t. t .

[0106] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An algorithm for optimizing factory production planning in cement plants through prediction and monitoring of photovoltaic power generation, energy storage, waste heat power generation, and electricity load, characterized by: include: The data collection unit (1) includes an energy storage data collection module, a power generation data collection module and a recycling data collection module. The power generation data collection module is used to collect data on sunshine hours, solar radiation intensity and temperature. The recycling data collection module is used to collect data on temperature, flow rate, pressure, power consumption, flue gas composition and ambient humidity during the production process. The data processing unit (2) includes an outlier processing module and a missing value processing module. The outlier processing module is used to identify and remove outliers in the data collected by the data collection unit (1), and the missing value processing module is used to fill in the missing values ​​in the data collected by the data collection unit (1). The prediction and monitoring unit (3) includes a photovoltaic power generation prediction module, a waste heat recovery monitoring module and an electricity load prediction module. The photovoltaic power generation prediction module is used to predict photovoltaic power generation based on the data collected by the power generation data collection module. The waste heat recovery monitoring module is used to determine the waste heat recovery potential and efficiency based on the data collected by the recovery data collection module. The electricity load prediction module is used to predict future electricity load based on historical electricity consumption data and production plans using a multivariate nonlinear regression model. The production optimization unit (4) uses a target optimization model based on the data predicted by the prediction and monitoring unit (3) to achieve the goals of minimizing energy costs, minimizing environmental impact, and maximizing production efficiency.

2. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 1, is characterized in that: The power generation data collection module includes a sunshine recorder, a solar radiation meter, and a temperature sensor, which are used to acquire sunshine duration, solar radiation intensity, and key temperature characteristic data, respectively. The recovery data collection module includes a temperature sensor, a flow meter, a pressure sensor, an electricity meter, a gas analyzer, and a hygrometer. The temperature sensor is used to collect kiln temperature, cooler outlet temperature, clinker temperature, flue gas temperature, waste heat boiler inlet and outlet water temperature, heat exchanger inlet and outlet temperatures, and ambient temperature. The flow meter is used to collect clinker mass flow rate, flue gas volume flow rate, waste heat boiler steam flow rate, and water flow rate. The pressure sensor is used to collect kiln pressure, flue gas pressure, and waste heat boiler steam pressure. The electricity meter, the gas analyzer, and the hygrometer are used to collect power consumption data, flue gas composition data, and ambient humidity data, respectively.

3. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 1, is characterized in that: The outlier processing module is used to identify and remove outliers from the power generation data collection module and the data recovery collection module. Specifically, the outlier processing module is: Where σ represents the standard deviation of the dataset, X represents a single measurement and data point, Z represents the Z-score, which is the standard score of data point X, and μ represents the mean of the dataset.

4. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 1, is characterized in that: The missing value processing module, based on the data processed by the outlier processing module, is used to fill in missing values ​​in the data of the power generation data collection module and the data recovery collection module. Specifically, the missing value processing module is as follows: in, Represents missing values; x j represents non-missing values; N represents the number of non-missing values.

5. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 1, is characterized in that: The photovoltaic power generation prediction module includes a power generation prediction model and a training model; The power generation prediction model is used to predict photovoltaic power generation, and the power generation prediction model is as follows: y=β0+β1x1+β2x2+...+β n x n +∈ Where y represents the predicted photovoltaic power generation; x1, x2, ..., x n Represents the selected characteristics of sunshine duration, solar radiation intensity, and temperature; β0 is the intercept term; β1, β2, ..., β n Weights representing sunshine duration, solar radiation intensity, and temperature characteristics; ∈ represents error; The training model is based on historical power generation data to train the power generation prediction model, in order to find the optimal β1, β2, ..., β n The value, the training model specifically is: Where m represents the number of samples; y i This represents the actual power generation of the i-th sample. The model represents the predicted power generation for the i-th sample.

6. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 1, is characterized in that: The waste heat recovery monitoring module is specifically: Where η-total represents the total waste heat recovery efficiency; Q-recovered represents the actual recovered waste heat; m -i c represents the mass flow rate of clinker within i time periods; -i ΔT represents the specific heat capacity of clinker within the i-th time period; clinker_i ρ represents the temperature drop of the clinker during the i-th time period; -j V represents the density of the flue gas during the j-th time period; -j Cp represents the volumetric flow rate of flue gas during the j-th time period; -j This represents the isobaric specific heat capacity of the flue gas during the j-th time period; K represents the temperature drop of the flue gas during the j-th time period; K_loss represents the equipment heat loss correction factor; K_season represents the seasonal climate correction factor.

7. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 1, is characterized in that: The specific multivariate nonlinear regression model is as follows: in, The predicted electricity load represents the load at time point t; α0 represents the intercept term; α i The autoregressive coefficient representing historical electricity load; L t-i The representative is the historical electricity load at point ti; β j The coefficient representing the production planning variable; P t-j Production plan variable representing time point tj; γ k Representing other explanatory variables X t-k The coefficients include weekday / weekend indicators and seasonal factors; δ ij The coefficients of the interaction term representing historical electricity load and production plan variables; ∈ t represents the error term; n represents the lag order of historical electricity load; m represents the lag order of the production plan variable; o represents the lag order of other explanatory variables; Where, α i β j γ k and δ ij The model parameters are estimated based on historical data, and the model is fitted and validated. The specific steps are as follows: S7.1 Determine the lag order n, m, and o of the model, and whether interaction terms need to be introduced, through correlation function and partial autocorrelation function analysis; S7.2 Use the least squares method to estimate the model parameters; S7.

3. Use R-squared to evaluate the goodness of fit of the model; S7.4 Use the fitted model to predict future electricity load.

8. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 7, is characterized in that: The relevant functions in S7.1 are as follows: Where ACF represents the correlation function; Y t Represents the observed values ​​of the time series at time point t; Cov represents the covariance; Var represents the variance; k represents the lag order; Specifically, the partial autocorrelation function in S7.1 is as follows: Where PACF represents the partial autocorrelation function Y t Represents the observed values ​​of the time series at time point t; Cov represents the covariance; Var represents the variance; k represents the lag order; Specifically, the least squares method in S7.2 is as follows: in, Represents the model parameter vector, including α i β j γ k and δ ij L t This represents the actual observed electrical load. This represents the predicted electricity load based on parameter θ; T represents the total number of observed data. Specifically, the R-squared in S7.3 is: Among them, R 2 L represents the coefficient of determination; T represents the length of the time series; L represents the coefficient of determination. t This represents the actual observed electricity load value at time point t; This represents the predicted electricity load value at time point t; This represents the average value of the actual electrical load. The expression is This represents the summation of all terms from t=1 to t=T; This represents the squared difference between the predicted value and the actual observed value. This represents the squared difference between the actual observed value and its mean.

9. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 7, is characterized by: The target optimization model is specifically as follows: Where Z represents the total cost, which is minimized; T represents the total number of time periods; c buy This represents the unit cost of purchasing additional electricity from the grid; y t The predicted photovoltaic power generation at time t; The predicted electrical load at time t; η_total t The predicted waste heat recovery energy represents the energy available at time t. This represents the amount of electricity purchased from the power grid; + indicates a positive value. env E represents the environmental impact cost per unit of carbon dioxide emissions; t c represents the carbon dioxide emissions at time t; prod C represents the production cost per unit of non-renewable energy consumed; t Production costs.

10. The algorithm for optimizing factory production planning in a cement plant by predicting and monitoring photovoltaic, energy storage, waste heat power generation, and electricity load, as described in claim 9, is characterized in that: The objective optimization model has constraints, specifically: Among them, G grid，t k represents the amount of electricity purchased from the grid at time t; CO2 k represents the coefficient that converts energy consumption into carbon dioxide emissions. cost This represents the coefficient that converts energy consumption into production costs; This represents all t.

Images