Electric furnace production management system based on big data

Abstract

The present invention relates to the technical field of data management, specifically an electric furnace production management system based on big data. The system includes: a real-time temperature analysis module extracts the current and historical temperature data of the electric furnace from the electric furnace sensors, calculates the change trend data and time series attributes of the electric furnace temperature, constructs a temperature control model, and generates the analysis result of the electric furnace temperature trend. In the present invention, the Gaussian process regression is used to predict the temperature of the electric furnace, making the temperature control more precise, which helps to maintain the optimal temperature conditions during the material processing process, thereby improving the product quality and energy efficiency. In addition, the kernel density estimation method is adopted to analyze the statistical distribution of production parameters such as temperature, pressure and current, effectively identifying outliers from historical and real-time data, reducing the risk of production interruption and equipment failure caused by parameter deviation. Significantly enhances the stability and reliability of the production line. By combining historical adjustment records and real-time data analysis, the operation parameter adjustment strategy of the electric furnace is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and particularly to an electric furnace production management system based on big data. Background Art

[0002] The technical field of data management involves collecting, storing, protecting, validating, and processing data to ensure the accessibility, reliability, and timeliness of data. This field utilizes a variety of technologies and systems, including database management systems, data warehouses, data lakes, and various data processing frameworks. It conducts the transformation process of data from its original form to useful information and is widely applied in business intelligence, data analysis, information recovery, and data-driven decision support systems. It helps organizations optimize their business processes, improve decision-making quality, and at the same time ensure data security and compliance.

[0003] Among them, the electric furnace production management system refers to a set of systems used for short-process electric furnace steelmaking, which is used to monitor and control various stages in the steelmaking process, thereby improving production efficiency and product quality. The system usually includes real-time data collection, processing, and analysis functions, which help operators adjust production parameters and react to changes in production in real time, thus ensuring the stability of the steelmaking process and optimizing output. Through the electric furnace production management system, steel mills can manage the production process more effectively, reduce energy consumption, lower production costs, and at the same time improve the consistency and quality of the final product.

[0004] Existing electric furnace production management systems have limitations in dealing with highly variable and complex production environments. Especially in terms of response to temperature fluctuations and anomaly detection, it is difficult to accurately predict and adjust complex production variables, often resulting in a lag in response and the inability to adjust production parameters in a timely manner, which affects production efficiency and product quality. In addition, the deficiencies of traditional systems in anomaly detection, such as the failure to effectively identify hidden patterns and trends in data, often lead to misjudgment or missed judgment of abnormal conditions, not only increasing the risk of production interruption, but also possibly causing serious equipment damage and safety accidents, and improvement is needed. Summary of the Invention

[0005] The purpose of the present invention is to solve the drawbacks existing in the prior art and propose an electric furnace production management system based on big data.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions: The electric furnace production management system based on big data includes:

[0007] The real-time temperature analysis module extracts the current and historical temperature data of the electric furnace from the electric furnace sensors, calculates the change trend data and time series attributes of the electric furnace temperature, constructs a temperature control model, and generates the analysis result of the electric furnace temperature trend;

[0008] The Gaussian temperature prediction module utilizes the analysis result of the temperature trend of the electric furnace, predicts the future temperature fluctuation of the electric furnace through Gaussian process regression, adjusts the heating power and heat preservation time according to the prediction result, and obtains the adjusted temperature data of the electric furnace;

[0009] The production parameter monitoring module collects the pressure and current data of the electric furnace, combines the adjusted temperature data of the electric furnace, analyzes the statistical distribution of the temperature parameters of the electric furnace using the kernel density estimation method, identifies potential deviations and evaluates the possibility of the existence of outliers, and generates an anomaly detection result;

[0010] The dynamic adjustment and optimization module combines the anomaly detection result and the historical adjustment record, evaluates the current production line state, and performs real-time optimization on the operating parameters of the electric furnace to obtain the production optimization result of the electric furnace.

[0011] As a further solution of the present invention, the steps for obtaining the change trend data and time series attributes of the temperature of the electric furnace are specifically as follows:

[0012] Extract the current and historical temperature data from the electric furnace sensor, and set the electric furnace temperature data set as , and adopt the formula:

[0013]

[0014] Calculate the temperature change rate at the th time point to obtain the temperature change data;

[0015] Among them, to , represents the temperature at the th time point, represents the th time point, represents the temperature change rate at the th time point, is the attenuation factor, represents the natural exponent;

[0016] Based on the temperature change data, adopt the formula:

[0017]

[0018] Calculate the change trend data at the th time point through the moving average method combined with the weighting factor to obtain the change trend data of the temperature of the electric furnace;

[0019] Among them, represents the change trend data at the th time point, represents a time point, and j represents the index of each time point within the time interval from to , where the time interval is from represents the length of the average interval, is the weighting factor;

[0020] Based on the change trend data of the electric furnace temperature and the electric furnace temperature data set , the formula is used:

[0021]

[0022] to calculate the time series attribute of the th time point by combining the original data and the trend data ;

[0023] Among them, represents the time series attribute of the th time point by combining the original data and the trend data, and are the weight coefficients, used to adjust the sensitivity to the difference between the two, represents the absolute difference between the original data and the trend data.

[0024] As a further solution of the present invention, the steps for obtaining the analysis result of the electric furnace temperature trend are specifically as follows:

[0025] Use the time series attribute of the th time point , construct a temperature control model, and use the formula:

[0026]

[0027] to analyze the relationship between the time series attribute and the temperature, and output the temperature trend value ;

[0028] Among them, represents the temperature trend value, that is, the output of the model at the time point , is the model coefficient, used to convert the time series attribute into a temperature control trend, is the intercept, representing the baseline temperature level;

[0029] Based on the temperature trend value , the formula is used:

[0030]

[0031] Output the dynamic change of the temperature trend by integrating the model, and output the analysis result of the electric furnace temperature trend ;

[0032] Among them, is the scaling factor, used to adjust the model output.

[0033] As a further solution of the present invention, the acquisition steps of the predicted future temperature fluctuation of the electric furnace are specifically as follows:

[0034] Based on the analysis result of the electric furnace temperature trend , use the formula:

[0035]

[0036] Perform electric furnace temperature prediction to obtain the future temperature fluctuation data of the electric furnace ;

[0037] Among them, is the future temperature fluctuation data of the electric furnace, represents the predicted time point, represents the current time point, is the length parameter of time, is the scale parameter of the Gaussian process, adjusts the periodic fluctuation of temperature prediction, represents the duration of the period, used to calculate the parameters of the periodic function, The function is used to simulate the periodic change of temperature over time;

[0038] Based on the future temperature fluctuation data of the electric furnace , combined with the current heating demand, use the formula:

[0039]

[0040] Adjust the prediction result to obtain the temperature adjustment value ;

[0041] Among them, is the temperature adjustment value, and respectively represent the target and current temperatures, is the adjustment factor, used to adjust according to the square of the temperature difference.

[0042] As a further solution of the present invention, the acquisition steps of the adjusted electric furnace temperature data are specifically as follows:

[0043] Based on the future temperature fluctuation data of the electric furnace and the obtained temperature adjustment value , use the formula:

[0044]

[0045] Evaluate the demand for the current electric furnace heating power to obtain the adjusted heating power ;

[0046] Among them, represents the adjusted heating power, represents the current heating power, is the sensitivity coefficient for adjusting the power according to the temperature difference, controls the curvature of the influence of the temperature difference on the power adjustment, is the current temperature;

[0047] Based on the adjusted heating power and the temperature adjustment value , use the formula:

[0048]

[0049] Adjust the heat preservation time and integrate to obtain the adjusted electric furnace temperature data and ;

[0050] Among them, represents the adjusted heat preservation time, is the current heat preservation time, is the coefficient according to the temperature difference and the power adjustment time, is the target temperature.

[0051] As a further solution of the present invention, the obtaining steps of analyzing the statistical distribution of the electric furnace temperature parameters are specifically as follows:

[0052] Obtain the adjusted electric furnace temperature data and , and collect the electric furnace pressure and current data to construct an initial data set , use the formula:

[0053]

[0054] Integrate the data for each electric furnace operation cycle to generate an electric furnace operation data set ;

[0055] Among them, and respectively represent the th observation values of the pressure and current data, and are the th observation values of the adjusted electric furnace temperature and heat preservation time, is the index, indicating the observation period;

[0056] Based on the electric furnace operation dataset , calculate the dataset of the temperature of the electric furnace to obtain the kernel density analysis result ;

[0057] Based on the kernel density analysis result , use the formula:

[0058]

[0059] to calculate the dispersion degree of the density function , and reveal the statistical distribution of the electric furnace temperature parameters;

[0060] wherein, represents the dispersion degree of the density function, is the estimated density function of the mean value, representing the average probability density of all observed values, represents the integration operation for calculating the entire function over all to obtain the sum, is the differential element, representing the infinitesimal change amount of during the integration process.

[0061] As a further solution of the present invention, the steps for obtaining the anomaly detection result are specifically as follows:

[0062] Take the dispersion degree of the density function as the threshold criterion, and compare it with the density function of each data point according to the formula:

[0063]

[0064] to calculate the deviation degree of the th data point;

[0065] wherein, is the density value of the th data point in the kernel density estimation, is the average value of the density values of all data points, adjusts the sensitivity of the deviation degree calculation, represents the deviation degree of the th data point;

[0066] Based on the deviation degree of the th data point, use the formula:

[0067]

[0068] Calculate the ratio of the deviation degree of each data point to the variance of the density function to obtain the anomaly degree of the th data point ;

[0069] wherein, is the pi, used to normalize the result of to the interval [0, 1], represents the anomaly degree of the th data point;

[0070] Based on the anomaly degree of the th data point , use the formula:

[0071]

[0072] By setting the anomaly degree threshold , screen the data points with anomaly degree higher than , mark them as potential deviations or outliers, and generate the anomaly detection result ;

[0073] wherein, ' is the preset anomaly degree threshold, is the coefficient for adjusting the threshold, is the original data point, is the anomaly detection result, that is, the set of outliers.

[0074] As a further solution of the present invention, the steps for obtaining the optimized result of the electric furnace production are specifically as follows:

[0075] Collect the historical adjustment record H, integrate it with the anomaly detection result , and according to the formula:

[0076]

[0077] Evaluate the operating status of the current production line and calculate the evaluated production line status index ;

[0078] wherein, represents the number of data points participating in the calculation, represents the value of the abnormal data point, represents the value of the historical data point, represents the evaluated production line status index;

[0079] According to the evaluated production line status index , according to the formula:

[0080]

[0081] Adjust the operating parameters of the electric furnace in real time to obtain an adjustment function ;

[0082] Among them, used to adjust the influence of real-time feedback, is the operating parameter of the previous cycle, is the adjustment function, that is, the operating parameter after adjustment in the current cycle;

[0083] Based on the adjustment function , use the formula:

[0084]

[0085] Perform optimization operations to obtain an optimized result function , and generate an optimized result for the electric furnace production;

[0086] Among them, is the average value of, used to standardize the comparison between the current operating parameter and the historical performance, represents the optimized result function, used to reflect the influence of the increase in the operating parameter on the production result, is an adjustment term, used to compare the current operating parameter with its average .

[0087] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0088] In the present invention, Gaussian process regression is used to predict the temperature of the electric furnace, making the temperature control more accurate, which helps to maintain the optimal temperature conditions during the material processing process, thereby improving the product quality and energy efficiency. In addition, the kernel density estimation method is used to analyze the statistical distribution of production parameters such as temperature, pressure, and current, effectively identifying outliers from historical and real-time data, reducing the risk of production interruption and equipment failure caused by parameter deviation. Significantly enhances the stability and reliability of the production line. By combining historical adjustment records and real-time data analysis, the adjustment strategy of the operating parameters of the electric furnace is optimized, ensuring the continuity and efficiency of the steelmaking process, while reducing energy consumption and production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 is the system flow chart of the present invention;

[0090] Figure 2 is the flow chart of the acquisition steps of the change trend data and time series attributes of the temperature of the electric furnace of the present invention;

[0091] Figure 3 This is a flowchart of the steps for obtaining the analysis results of the electric furnace temperature trend of the present invention;

[0092] Figure 4 This is a flowchart of the steps for obtaining the prediction of the future temperature fluctuation of the electric furnace of the present invention;

[0093] Figure 5 This is a flowchart of the steps for obtaining the adjusted electric furnace temperature data of the present invention;

[0094] Figure 6 This is a flowchart of the steps for obtaining the statistical distribution of the analysis of the electric furnace temperature parameters of the present invention;

[0095] Figure 7 This is a flowchart of the steps for obtaining the anomaly detection results of the present invention;

[0096] Figure 8 This is a flowchart of the steps for obtaining the production optimization results of the electric furnace of the present invention. Detailed implementation manners

[0097] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0098] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the present invention. In addition, in the description of the present invention, the meaning of "a plurality" is two or more unless otherwise specifically defined.

[0099] Please refer to Figure 1 , the electric furnace production management system based on big data includes:

[0100] The real-time temperature analysis module extracts the current and historical temperature data of the electric furnace from the electric furnace sensor, calculates the change trend data and time series attributes of the electric furnace temperature, constructs a temperature control model, and generates the analysis results of the electric furnace temperature trend;

[0101] The Gaussian temperature prediction module uses the analysis results of the electric furnace temperature trend, predicts the future temperature fluctuation of the electric furnace through Gaussian process regression, and adjusts the heating power and heat preservation time according to the prediction results to obtain the adjusted electric furnace temperature data;

[0102] The production parameter monitoring module collects the pressure and current data of the electric furnace, combines the adjusted temperature data of the electric furnace, analyzes the statistical distribution of the temperature parameters of the electric furnace using the kernel density estimation method, identifies potential deviations and evaluates the possibility of the existence of outliers, and generates anomaly detection results;

[0103] The dynamic adjustment and optimization module combines the anomaly detection results and historical adjustment records, evaluates the current production line status, and performs real-time optimization on the operating parameters of the electric furnace to obtain the electric furnace production optimization results.

[0104] The electric furnace temperature trend analysis results include the comparison of the high and low points of the electric furnace historical temperature, the temperature values at key time points, and the periodic temperature change characteristics. The adjusted electric furnace temperature data includes the highest temperature, the lowest temperature, and the temperature fluctuation range at future time points of the electric furnace. The anomaly detection results include the statistical number of temperature anomalies, the anomaly duration, and their potential impact on the production process. The electric furnace production optimization results include the optimal heating rate, the heat preservation time, and the expected energy-saving effect adjusted based on real-time data.

[0105] Please refer to Figure 2 , the steps for obtaining the change trend data and time series attributes of the electric furnace temperature are specifically as follows:

[0106] Extract the current and historical temperature data from the electric furnace sensor, and set the electric furnace temperature data set as , and use the formula:

[0107]

[0108] Calculate the temperature change rate at the th time point to obtain the temperature change data;

[0109] Among them, to , represents the temperature at the th time point, represents the th time point, represents the temperature change rate at the th time point, is the attenuation factor used to adjust the sensitivity of the temperature change, represents the natural exponent used to modulate the non-linear effect of the difference;

[0110] Based on the temperature change data, use the formula:

[0111]

[0112] Calculate the change trend data at the , the changing trend data of the electric furnace temperature is obtained;

[0113] Among them, represents the changing trend data at the -th time point, represents the time point, and j represents the index of each time point within the time interval from to . represents the length of the average interval, is the weighting factor, reflecting the relative importance of each time point;

[0114] Based on the changing trend data of the electric furnace temperature and the electric furnace temperature data set , the formula:

[0115]

[0116] is used to calculate the time series attribute of the -th time point combining the original data and the trend data ;

[0117] Among them, represents the time series attribute of the -th time point combining the original data and the trend data, and are the weight coefficients used to balance the influence of the original temperature data and the trend data , strengthen the sensitivity to the difference between the two, represents the absolute difference between the original data and the trend data, which is used to highlight anomalies or drastic changes.

[0118] The calculation process is as follows:

[0119] Assume that the temperature data obtained from the electric furnace sensor, and the time point is .

[0120] For the formula:

[0121]

[0122] Among them, is the attenuation factor, set to 0.05. This factor helps to adjust the influence brought by rapid temperature changes.

[0123] For :

[0124]

[0125] Similarly calculate .

[0126] As a result, an array of temperature change rates is obtained .

[0127] The calculation of the short-term trend sets the average interval , and the time weighting factor .

[0128] For the formula:

[0129]

[0130] where , takes values of 2 and 3, and the weighting factor : .

[0131] Similarly calculate .

[0132] As a result, a short-term trend array is obtained .

[0133] Using the previous data and the assumed model parameters, the time series attributes can be calculated

[0134] For the formula:

[0135]

[0136] Assume , , .

[0137] where :

[0138]

[0139]

[0140] As a result, a time series attribute array is obtained .

[0141] Please refer to Figure 3 , and the specific steps for obtaining the analysis results of the electric furnace temperature trend are as follows:

[0142] The time series attributes at the th time point are used to construct a temperature control model, and the formula is adopted:

[0143]

[0144] Analyze the relationship between the time series attributes and the temperature, and output the temperature trend value ;

[0145] Among them, represents the temperature trend value, that is, the output of the model at the time point ; is the model coefficient, used to convert the time series attributes into the temperature control trend, is the intercept, representing the baseline temperature level;

[0146] Based on the temperature trend value , the formula is adopted:

[0147]

[0148] The dynamic change of the temperature trend is output by integrating the model, and the analysis result of the electric furnace temperature trend is output ;

[0149] Among them, is the scaling factor, used to adjust the model output to match the current temperature control requirements.

[0150] The calculation process is as follows:

[0151] For the formula:

[0152]

[0153] The time series attributes are input into this model. is the model coefficient, assumed to be obtained through historical data analysis and set to (this value is assumed to be the standard conversion factor obtained in the past electric furnace production). is the intercept, representing the baseline temperature level, set to (this value is assumed to be based on the basic temperature state of the electric furnace without materials or operations).

[0154] Calculate the model output :

[0155]

[0156] The result represents the expected output temperature trend of the electric furnace control model under the current input conditions.

[0157] For the formula:

[0158]

[0159] —the scaling factor, adjusted according to the actual temperature control requirements of the electric furnace, set to (This value is assumed to be the normalized output, directly reflecting the output of the model as the final trend analysis result).

[0160] Generate the electric furnace temperature trend analysis result :

[0161]

[0162] Result Represents the final temperature trend analysis result of the electric furnace temperature control model under the given time series attributes, and can be directly used to judge and adjust the temperature settings of the electric furnace.

[0163] Please refer to Figure 4 , and the specific steps for obtaining the prediction of the future temperature fluctuation of the electric furnace are as follows:

[0164] Based on the electric furnace temperature trend analysis result , use the formula:

[0165]

[0166] Conduct the electric furnace temperature prediction to obtain the future temperature fluctuation data of the electric furnace ;

[0167] Among them, represents the prediction time point, represents the current time point, which is used to calculate the square of the time interval and is reflected in the prediction in exponential form; is the length parameter of time, which controls the influence of the time interval on the prediction; is the scale parameter of the Gaussian process, which determines the change range of the entire prediction; adjusts the periodic fluctuation of the temperature prediction, enabling the model to adapt to the periodic temperature change; represents the duration of this periodicity, which is used to calculate the parameters of the periodic function; The function is used to simulate the periodic change of temperature over time;

[0168] Based on the future temperature fluctuation data of the electric furnace , combined with the current heating demand, use the formula:

[0169]

[0170] Adjust the prediction result to obtain the temperature adjustment value ;

[0171] Among them, and respectively represent the target and current temperatures, is the adjustment factor, which is used for fine-tuning according to the square of the temperature difference, so as to more precisely control the sensitivity of the temperature adjustment.

[0172] The calculation process is as follows:

[0173] For the formula:

[0174]

[0175] where, is the predicted target time point, assumed to be hours. is the current time point, assumed to be hours. is the length parameter of time, controlling the influence of the time interval on the prediction, assumed hours. is the scale parameter of the Gaussian process, determining the variation range of the whole prediction, assumed . is the period adjustment coefficient, enabling the model to adapt to the periodic temperature change, assumed . is the period length, representing the duration of the periodic change, assumed hours. is the current temperature trend analysis result, given from the previous analysis .

[0176] Calculate the square of the time interval:

[0177]

[0178] Calculate the exponential part:

[0179]

[0180] Calculate the sine part:

[0181]

[0182] Calculate :

[0183]

[0184] For the formula:

[0185]

[0186] where, is the result from step 1, . is the target temperature, assumed to be degrees. is the current actual temperature, assumed to be degrees. is an adjustment factor, which is fine-tuned according to the square of the temperature difference. Assume that .

[0187] Calculate the square of the temperature difference:

[0188]

[0189] Calculate :

[0190]

[0191] represents the temperature fluctuation during prediction, which is calculated to be approximately . is the final temperature adjustment value, which is approximately , indicating that the heating power needs to be adjusted according to the actual demand to reach the target temperature after prediction degrees.

[0192] Please refer to Figure 5 , and the specific steps for obtaining the adjusted electric furnace temperature data are as follows:

[0193] Based on the future temperature fluctuation data of the electric furnace The obtained temperature adjustment value , use the formula:

[0194]

[0195] Evaluate the demand for the current heating power of the electric furnace to obtain the adjusted heating power ;

[0196] Among them, represents the adjusted heating power, represents the current heating power, is the sensitivity coefficient for adjusting the power according to the temperature difference, controls the curvature of the influence of the temperature difference on the power adjustment to ensure faster adjustment at large temperature differences, is the current actual temperature.

[0197] Based on the adjusted heating power and the temperature adjustment value , use the formula:

[0198]

[0199] Adjust the holding time, and integrate to obtain the adjusted electric furnace temperature data and ;

[0200] Among them, represents the adjusted holding time, is the current holding time, is the coefficient for adjusting time according to the temperature difference and power, is the target temperature, ensuring that the adjustment time reflects appropriate sensitivity and responds quickly to the ideal state.

[0201] The calculation process is as follows:

[0202] For the formula:

[0203]

[0204] , assume kW is the current heating power, based on the common settings of a typical medium-sized electric furnace. °C is the current actual temperature, in line with the common operating temperature for steel melting. is the power adjustment coefficient, ensuring that the adjustment is not too drastic. is the parameter for adjusting the temperature difference sensitivity, providing a smooth adjustment dynamics.

[0205] Calculation of the temperature difference:

[0206] Calculation of the logarithmic logic function:

[0207]

[0208] The value of is very large, and it should be considered that this value may cause numerical overflow in practical applications. Here, a large number is used for approximation, such as .

[0209] Calculation of the new heating power:

[0210]

[0211] Since the logarithmic logic part is very large, the influence is minimized, resulting in almost no change in the actual heating power.

[0212] For the formula:

[0213]

[0214] Assume minutes, is the current holding time, based on the standard operating cycle. is the time adjustment factor, ensuring a moderate time adjustment. °C is the target temperature. kW is the result from step 1.

[0215] Calculation of the absolute value of the target temperature difference:

[0216]

[0217] Adjustment calculation of heat preservation time:

[0218]

[0219] Therefore, the adjusted heat preservation time is approximately 42 minutes.

[0220] Conclusion: Maintaining approximately 100 kilowatts shows that the logarithmic logic function is effective in preventing over-adjustment when the temperature difference is large. Approximately 42 minutes indicates that the heat preservation time needs to be increased to reach the target temperature according to the new temperature prediction and the current power.

[0221] Please refer to Figure 6 , and the specific steps for obtaining the statistical distribution of the electric furnace temperature parameters are as follows:

[0222] Obtain the adjusted electric furnace temperature data and , and collect the electric furnace pressure and current data to construct an initial data set , and use the formula:

[0223]

[0224] Integrate the data for each electric furnace operation cycle to generate an electric furnace operation data set ;

[0225] Among them, and respectively represent the th observation values of the pressure and current data, and are the th observation values of the adjusted electric furnace temperature and heat preservation time, is the index indicating the specific observation period;

[0226] Based on the electric furnace operation data set , calculate the kernel density of the electric furnace temperature in the data set to obtain the kernel density analysis result ;

[0227] Based on the kernel density analysis result , use the formula:

[0228]

[0229] Calculate the discreteness of the density function , revealing the statistical distribution of the electric furnace temperature parameters;

[0230] Among them, represents the degree of dispersion (variance) of the density function, which is used to measure the central tendency and dispersion tendency in the statistical distribution. is the estimated density function The mean value represents the average probability density of all observed values. This analysis helps to identify abnormal fluctuations and deviations in the temperature distribution. represents the integral operation, which is used to calculate the entire function over all to obtain the total variance. is the differential element, which represents the infinitesimal change amount of during the integration process;

[0231] The calculation process is as follows:

[0232] Suppose there are data for three different operating cycles of the electric furnace:

[0233] Cycle 1: Pressure psi, current A, adjusted temperature °C, holding time minutes.

[0234] Cycle 2: Pressure psi, current A, adjusted temperature °C, holding time minutes.

[0235] Cycle 3: Pressure psi, current A, adjusted temperature °C, holding time minutes.

[0236] Kernel density estimation calculation:

[0237] The result of represents the kernel density estimate value of the electric furnace temperature at

[0238] Suppose there is at the kernel density estimate value of °C, (which is a hypothetical value), calculate the variance of the density function:

[0239]

[0240] The variance calculation can be processed through a discrete approximation:

[0241] Suppose ( 's mean value) is the simple average of all values.

[0242] Further assume that at °C, °C, °C there are also estimated values (also assumed values): , .

[0243] First, calculate the mean value :

[0244]

[0245] Then, calculate the variance :

[0246]

[0247]

[0248] The result of

[0249] Please refer to Figure 7 , and the specific steps to obtain the anomaly detection result are as follows:

[0250] Take the dispersion degree of the density function as the threshold criterion, and compare it with the density function of each data point according to the formula:

[0251]

[0252] Calculate the deviation degree of the -th data point;

[0253] Among them, is the density value of the -th data point in the kernel density estimation, is the average value of the density values of all data points, adjusts the sensitivity of the deviation degree calculation, ' represents the deviation degree of the -th data point;

[0254] Based on the deviation degree of the -th data point, use the formula:

[0255]

[0256] Calculate the ratio of the deviation degree of each data point to the variance of the density function to obtain the outlier degree of the th data point ;

[0257] where is the pi, used to normalize the result to the interval [0, 1], represents the outlier degree of the th data point. The closer the value is to 1, the more likely the point is an outlier;

[0258] Based on the outlier degree of the th data point , use the formula:

[0259]

[0260] By setting the outlier degree threshold , filter the data points with an outlier degree higher than , mark them as potential deviations or outliers, and generate the outlier detection result ;

[0261] where ' is the preset outlier degree threshold, is the coefficient to adjust the threshold, is the original data point, is the outlier detection result, that is, the set of outliers.

[0262] The calculation process is as follows:

[0263] For the formula:

[0264]

[0265] Suppose there are the following data points and their kernel density estimates:

[0266] Data points: , , ;

[0267] Density values: , , ;

[0268] Suppose the average density ;

[0269] The variance of the density function is supposed to be ;

[0270] Balance coefficient ;

[0271] For the degree of deviation ' is calculated as follows:

[0272]

[0273] The calculated degree of deviation indicates that there is a large deviation between the density value of

[0274] For the formula:

[0275]

[0276] Use the degree of deviation ;

[0277]

[0278] The calculation result shows that: Indicates is relatively high in the outlier score, and the closer it is to 1, the more likely it is to be an outlier.

[0279] For the formula:

[0280]

[0281] Assume , , then;

[0282]

[0283] Since is not greater than , therefore is not marked as an anomaly.

[0284] Indicates that under the given threshold setting, no data points exceed the outlier standard.

[0285] Please refer to Figure 8 , the specific steps for obtaining the optimization results of electric furnace production are as follows:

[0286] Collect the historical adjustment records H and integrate them with the anomaly detection results , and according to the formula:

[0287]

[0288] Evaluate the operating status of the current production line and calculate the evaluated production line status index ;

[0289] Among them, represents the number of data points involved in the calculation, represents the value of the abnormal data point, represents the value of the historical data point, represents the production line status index after comprehensive evaluation;

[0290] According to the production line status index after evaluation , according to the formula:

[0291]

[0292] Adjust the operating parameters of the electric furnace in real time to obtain the adjustment function ;

[0293] Among them, is used to adjust the influence of real-time feedback, is the operating parameter of the previous cycle, is the adjustment function, that is, the operating parameter after adjustment in the current cycle.

[0294] Based on the adjustment function , use the formula:

[0295]

[0296] Perform optimization calculations to obtain the optimization result function , and generate the optimization result of the electric furnace production;

[0297] Among them, is the average value of, used to standardize the comparison between the current operating parameter and the historical performance, represents the optimized electric furnace production result, reflects the influence of the increase in the operating parameter on the production result, is an adjustment term that compares the current operating parameter with its long-term average and adjusts the response sensitivity in this way. When is greater than , the denominator becomes larger, making the whole function value tend to 1, indicating that the production state is close to optimization; when is less than or equal to , the function value decreases, indicating that further optimization is needed.

[0298] The calculation process is as follows:

[0299] For the formula:

[0300]

[0301] Suppose there are 3 data points and the anomaly detection results and historical adjustment records are as follows respectively:

[0302]

[0303] The calculation process is as follows:

[0304]

[0305]

[0306]

[0307]

[0308] The calculated status indicator indicates that the current production line status is very close to the historical records, indicating that the adjustment strategy maintains good continuity and stability.

[0309] For the formula:

[0310]

[0311] Suppose . (The adjustment sensitivity coefficient is medium). (The parameter value of the previous cycle).

[0312]

[0313] The obtained indicates that the operating parameters of the electric furnace in the current cycle are adjusted based on the previous settings and the current status evaluation to ensure the continuity and stability of production.

[0314] For the formula:

[0315]

[0316] . (Long-term average value).

[0317]

[0318] Finally, it indicates that the optimized production output is very high, reflecting a great improvement in production efficiency.

[0319] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An electric furnace production management system based on big data, characterized in that, The system includes: The real-time temperature analysis module extracts the current and historical temperature data of the electric furnace from the electric furnace sensors, calculates the change trend data and time series attributes of the electric furnace temperature, constructs a temperature control model, and generates the electric furnace temperature trend analysis result; The Gaussian temperature prediction module uses the electric furnace temperature trend analysis result to predict the future temperature fluctuation of the electric furnace through Gaussian process regression, and adjusts the heating power and heat preservation time according to the prediction result to obtain the adjusted electric furnace temperature data; The specific obtaining step of predicting the future temperature fluctuation of the electric furnace is as follows: Based on the analysis results of the electric furnace temperature trend , use the formula: Perform electric furnace temperature prediction to obtain future temperature fluctuation data of the electric furnace ; Among them, is the future temperature fluctuation data of the electric furnace, represents the predicted time point, represents the current time point, is the length parameter of time, is the scale parameter of the Gaussian process, adjusts the periodic fluctuation of temperature prediction, represents the duration of the period, used to calculate the parameters of the periodic function, The function is used to simulate the periodic change of temperature over time; Based on the future temperature fluctuation data of the electric furnace , combined with the current heating demand, use the formula: Adjust the prediction result to obtain the temperature adjustment value ; Among them, is the temperature adjustment value, and represent the target and current temperatures respectively, is the adjustment factor, which is used to make adjustments according to the square of the temperature difference; The production parameter monitoring module collects the pressure and current data of the electric furnace, combines the adjusted electric furnace temperature data, analyzes the statistical distribution of the electric furnace temperature parameters using the kernel density estimation method, identifies potential deviations and evaluates the possibility of the existence of outliers, and generates an anomaly detection result; The dynamic adjustment and optimization module combines the anomaly detection result and the historical adjustment record, evaluates the current production line status, and performs real-time optimization on the operating parameters of the electric furnace to obtain the electric furnace production optimization result.

2. The electric furnace production management system based on big data according to claim 1, wherein, The specific obtaining step of the change trend data and time series attributes of the electric furnace temperature is as follows: Extract the current and historical temperature data from the electric furnace sensor. Let the electric furnace temperature data set be , and use the formula: Calculate the temperature change rate at the th time point to obtain temperature change data; Among them, to , represents the temperature at the th time point, represents the th time point, represents the temperature change rate at the th time point, is the attenuation factor, represents the natural exponent; Based on the temperature change data, the formula is adopted: Calculate the change trend data at the th time point by combining the moving average method with a weighting factor , and obtain the change trend data of the electric furnace temperature; Among them, represents the change trend data at the th time point, represents the time point, j represents the index of each time point within the time interval from to , represents the length of the average interval, is the weighting factor; Based on the change trend data of the electric furnace temperature and the electric furnace temperature data set , use the formula: Calculate the time series attributes of the th time point combining the original data and the trend data ; Among them, represents the time series attribute of the th time point combining the original data and the trend data, and are weight coefficients, used to adjust the sensitivity to the difference between the two, represents the absolute difference between the original data and the trend data.

3. The electric furnace production management system based on big data according to claim 2, characterized in that, The specific obtaining step of the electric furnace temperature trend analysis result is as follows: Using the time series attributes of the th time point , a temperature control model is constructed, using the formula: Analyze the relationship between time series attributes and temperature, and output the temperature trend value ; Among them, represents the temperature trend value, that is, the output of the model at the time point ; is the model coefficient, which is used to convert the time series attribute into the temperature control trend, is the intercept, representing the baseline temperature level; Based on the temperature trend value , use the formula: Output the dynamic change of the temperature trend by integrating the model, and output the analysis result of the temperature trend of the electric furnace ; Among them, is a scaling factor used to adjust the model output.

4. The electric furnace production management system based on big data according to claim 1, characterized in that The specific obtaining step of the adjusted electric furnace temperature data is as follows: Based on the future temperature fluctuation data of the electric furnace The obtained temperature adjustment value , use the formula: Evaluate the demand for the current electric furnace heating power to obtain the adjusted heating power ; Among them, represents the adjusted heating power, represents the current heating power, is the sensitivity coefficient for adjusting the power according to the temperature difference, controls the curvature of the influence of the temperature difference on the power adjustment, is the current temperature; Based on the adjusted heating power and the temperature adjustment value , use the formula: Adjust the heat preservation time and integrate to obtain the adjusted electric furnace temperature data and ; Among them, represents the adjusted heat preservation time, is the current heat preservation time, is the coefficient for adjusting the time according to the temperature difference and power, is the target temperature.

Images