A control index optimization method for cement production and burning system based on industrial big data

Abstract

The present invention discloses a control index optimization method for a cement production and burning system based on industrial big data, comprising the following steps: Step 1, collecting historical data of key process variables indicating kiln conditions. Step 2, normalizing the historical data, determining the weights of each key process variable by data fitting, and obtaining a formula for a comprehensive kiln condition index value. Step 3, collecting real-time data of the above key process variables during operation. Step 4, calculating the comprehensive kiln condition index value according to the formula obtained by fitting in step 2, and making a comprehensive evaluation of the kiln condition. Step 5, introducing the comprehensive kiln condition index value into a model algorithm to realize automatic optimization of a target value of a decomposition furnace temperature. The present invention automatically optimizes indicators including a target value of a decomposition furnace temperature and a target value of a Roots blower coal delivery pressure of a kiln head coal feeding scale, realizes adaptive adjustment of working conditions, reduces the influence of human factors, and achieves lower energy consumption while ensuring the quality of clinker.

Description

Technical Field

[0001] The present invention belongs to the field of cement clinker production, and more specifically, to a method for optimizing control indicators of a cement production and burning system based on industrial big data. Background Art

[0002] Cement clinker production is a typical process manufacturing with complex production technology. The main production links include raw material crushing, raw material grinding, coal grinding, burning (preheater, decomposition furnace, grate cooler), waste heat power generation and clinker storage. At present, most of the cement production process operations are performed by operators who monitor parameters through the DCS system to determine the operation status of the production line, and manually adjust the coal feeding amount, feeding amount, fan speed and other actuators based on process knowledge and experience to adjust the production conditions. In recent years, the continuous promotion of cement smart factories and the continuous improvement of construction levels have brought good development opportunities to the cement production expert automatic operation system, and its application scope has continued to expand.

[0003] At present, there is also an automatic operation system for cement production experts. Its main principle is to adjust the control experience based on traditional manual operation, with PID or MPC control algorithm as the calculation core, adding complex artificial constraint rules, and the operator sets the control target based on experience. The automatic operation system for cement production experts automatically adjusts the operating variables according to the set targets to make the control variables reach the set target values.

[0004] However, the above-mentioned prior art requires the operator to modify the target value of the corresponding parameter based on experience when making adjustments according to the cement production working conditions. Therefore, there are the following technical problems: 1. It is impossible to make adaptive adjustments according to the current working conditions. In the industrial control of cement production, the working conditions are extremely complex. Once the system is completed, it loses the ability to absorb new knowledge and can only passively solve the previous situations. As time goes by, any small change in equipment or process may make it lose its vitality and lead to safety risks. 2. Experience mining is relatively superficial. Most of the manual experience that the cement expert automatic operation system relies on is linear relationship, such as "increasing the grate speed to reduce the outlet pressure of the grate cooler fan", "increasing the high-temperature fan speed to increase the oxygen content at the preheater outlet", etc. These rules can only reflect the simple logical relationship between "1 to 1" or "2 to 1". If the relationship is more complicated, it will become more and more difficult to use the empirical rule constraint method, and it can even be said to be an impossible task. In this case, how to efficiently analyze and mine the nonlinear relationship between parameters through historical data is particularly important. 3. Energy saving and consumption reduction. The cement expert automatic operation system only considers safety automatic control, but cannot take into account energy saving and further reduce costs. This is also the main reason why traditional expert systems have not been widely used. How to combine big data analysis with production mechanisms to solve the complex coupling relationship between multiple variables, so as to achieve energy reduction and quality stability in the production process, is the development direction for upgrading and iterating the cement expert automatic operation system. Summary of the invention

[0005] The purpose of the present invention is to provide a method for optimizing control indicators of cement production and firing systems based on industrial big data, which is used to solve the technical problems that in the prior art, automatic operation and control of cement production lines must rely on manual adjustment to achieve optimization of parameter target values, and the corresponding control method cannot take into account energy saving considerations while considering safety control and ensuring clinker quality.

[0006] The method for optimizing control indicators of cement production and burning system based on industrial big data comprises the following steps.

[0007] Step 1: Collect historical data of key process variables that indicate kiln conditions.

[0008] Step 2: Normalize the historical data, determine the weights of each key process variable through data fitting, and obtain the formula for the comprehensive kiln condition index value.

[0009] Step 3: Collect real-time data of the above key process variables during operation.

[0010] Step 4: Calculate the comprehensive index value of the kiln condition according to the formula fitted in step 2, and make a comprehensive evaluation of the kiln condition.

[0011] Step 5: Introduce the comprehensive index value of kiln conditions into the model algorithm to realize automatic optimization of the target value of the decomposition furnace temperature.

[0012] Preferably, the key process variables include at least: the brightness of the kiln head flame, the secondary air temperature, the kiln main motor current, the kiln tail high temperature NO x Content and free calcium oxide content of clinker.

[0013] The formula for the comprehensive kiln condition index value used for fitting is: Comprehensive kiln condition index value = normalized value of flame brightness * coefficient 1 + normalized value of secondary air temperature * coefficient 2 + normalized value of kiln main motor current * coefficient 3 + normalized value of high temperature NOx at kiln tail * coefficient 4 + normalized value of free calcium oxide content in clinker * coefficient 5.

[0014] Preferably, the flame brightness is identified based on the video screen captured during each calculation, and is divided into five levels: bright, brighter, normal, darker, and dark; the corresponding flame brightness normalization values ​​are: bright (1.0), brighter (0.5), normal (0), darker (-0.5), and dark (-1.0).

[0015] Preferably, if the key process variables can be directly measured, the rolling average of the corresponding key process variables at time intervals is calculated, the interval time is t, and the rolling average of the corresponding key process variables in each interval time is used in the use interval time. The normalized formula is: key process variable normalized value = (key process variable rolling average （当前） -Rolling average of key process variables （t前） ) / maximum change range of rolling average of key process variables.

[0016] Rolling average of key process variables （当前） It is the rolling average of the key process variables calculated within a time interval from the current time. "t before" is the rolling average of the key process variables before an interval. （当前） The difference between the two is the change range of the rolling average of the key process variables before and after a time interval. The maximum value of the change range, that is, the maximum change range of the rolling average of the key process variables, can be directly calculated based on historical data.

[0017] The key process variables used in the above formula include secondary air temperature, kiln main motor current and kiln tail high temperature NO x content.

[0018] Preferably, for the free calcium oxide content of clinker, the system applies the free calcium oxide prediction algorithm of clinker, establishes a prediction model, uses the predicted value for control, and uses the test results corresponding to the predicted value in the historical data to correct the prediction model to maintain the accuracy of the predicted data. The corresponding formula is as follows:

[0019] Normalized value of free calcium oxide content in clinker = (predicted value of free calcium oxide content in clinker （当前） - target setting value of free calcium oxide content in clinker) / maximum difference between free calcium oxide content in clinker and target setting value.

[0020] Preferably, the kiln condition comprehensive index value calculated in step 4 is a value between -1 and 1, and the specific kiln condition corresponding to different kiln condition comprehensive index value intervals is set according to experiments and experience: kiln condition comprehensive index value > 0.5, indicating that the kiln condition is very good; 0.5> comprehensive index value of kiln condition > 0.2, indicating that the kiln condition is good; 0.2> kiln condition comprehensive index value>-0.2, indicating that the kiln condition is normal; -0.2 > The comprehensive index value of kiln condition > -0.5 indicates that the kiln condition is poor; -0.5 > The comprehensive index value of kiln condition indicates that the kiln condition is very bad; this comprehensive index value of kiln condition and the corresponding specific status of the kiln condition provide a basis for optimizing and adjusting the target values ​​of corresponding parameters.

[0021] Preferably, the optimization control method of the target value of the decomposition furnace temperature in step 5 includes: when the comprehensive index value of the kiln condition is > When the value is 0.5, the system judges the fluctuation of secondary air temperature by calculating the fluctuation range of secondary air temperature within a certain rolling time interval. The formula for the corresponding fluctuation range is: Secondary air temperature fluctuation range = current value of secondary air temperature - rolling average value of secondary air temperature in the previous t' minutes, where t' minutes is the rolling time interval.

[0022] If the fluctuation reaches the first state below: the absolute value of the secondary air temperature fluctuation amplitude > the set value of the secondary air temperature fluctuation amplitude （SP） , the system starts to accumulate time, and the time accumulation threshold is t". If the fluctuation lasts for t" in the first state, the target setting value of the decomposition furnace temperature will be increased by 0.1 degrees, and the accumulated time will be reset; if the fluctuation lasts for less than t" but remains in the first state, the target value of the decomposition furnace temperature will remain unchanged, and the time accumulation will continue; if the fluctuation changes and no longer remains in the first state, the target value of the decomposition furnace temperature will remain unchanged, and the accumulated time will be reset.

[0023] If the fluctuation reaches the following second state: the absolute value of the secondary air temperature fluctuation amplitude < the set value of the secondary air temperature fluctuation amplitude （SP） , the system starts to accumulate time. If the fluctuation continues in the second state for t", the target setting value of the decomposition furnace temperature will be lowered by 0.1 degree and the accumulated time will be reset; if the fluctuation continues for less than t" but remains in the second state, the target value of the decomposition furnace temperature will remain unchanged and the time accumulation will continue; if the fluctuation changes and no longer remains in the second state, the target value of the decomposition furnace temperature will remain unchanged and the accumulated time will be reset.

[0024] Preferably, the optimization control method of the decomposition furnace temperature target value in step 5 further includes: when the kiln condition comprehensive index value is <0.5, the formula for the decomposition furnace temperature target value is: decomposition furnace temperature target value = decomposition furnace temperature target value （前一次） - Comprehensive kiln condition index value*K, K is the coefficient fitted by historical data, the target value of the decomposition furnace temperature （前一次） It is the target value of the decomposition furnace temperature obtained by the previous calculation of the formula, and its initial value is the target value of the decomposition furnace temperature set by the operator in the initial state.

[0025] Preferably, the method for optimizing control indexes of cement production and firing system based on industrial big data also includes step six, optimizing the automatic setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale according to the comprehensive index value of the kiln condition.

[0026] The optimization of the target setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale is mainly based on the comprehensive indicators of the kiln operating conditions. If the comprehensive index value of the kiln condition is >0.8, the target setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale is reduced by 0.01; if the comprehensive index value of the kiln condition is <-0.8 and the current decomposition furnace temperature target value has reached the maximum, the head coal pressure is increased by 0.01; the maximum value of the decomposition furnace temperature target value is the maximum value determined by the operator based on historical data to avoid the decomposition furnace temperature target value being too high, resulting in energy consumption exceeding the upper limit.

[0027] The present invention has the following advantages: In view of the problems existing in the prior art, the present invention can accurately judge the working conditions of the cement production and burning system through an index optimization method, automatically optimize the indicators including the target value of the decomposition furnace temperature and the target value of the coal feeding pressure of the roots blower of the kiln head coal feeding scale, realize the adaptive adjustment of the working conditions, further reduce the labor intensity of the operator, reduce the influence of human factors, and through the optimization and adjustment of the main process control targets, can ensure the normal operation of the kiln condition, reduce energy consumption and reduce CO emissions under the premise of ensuring good clinker quality. 2 In order to accurately adapt to the real-time working conditions of the firing system, that is, the changes in kiln conditions, during optimization, this method selects key process variables that are closely related to the kiln conditions and an algorithm that can effectively judge the actual kiln conditions and take into account reasonable calculations to achieve the calculation of the corresponding kiln condition comprehensive index values, and then uses a reasonable and effective optimization control method to actually optimize the target value of the decomposition furnace temperature and the target value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale based on the kiln condition comprehensive index value. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 The network structure diagram of a system using a method for optimizing control indicators of a cement production and firing system based on industrial big data according to the present invention.

[0029] Figure 2The present invention is a flow chart of a method for optimizing control indicators of a cement production and firing system based on industrial big data. DETAILED DESCRIPTION

[0030] The specific implementation modes of the present invention are further explained in detail below by describing the embodiments with reference to the accompanying drawings, so as to help those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention.

[0031] The present invention provides a method for optimizing control indicators of cement production and burning system based on industrial big data. Figure 1 As shown in FIG. 1 , the system using this method includes a DCS control system, a cement expert automatic operation system, and a kiln optimization control system. The edge server of the kiln optimization control system is connected to the DCS system through OPC to achieve two-way data communication. The DCS system and the cement expert automatic operation have achieved two-way communication, so the kiln optimization system and the cement expert automatic operation system can also achieve two-way communication.

[0032] The kiln optimization system is equipped with a video acquisition device to collect video signals for kiln head flame monitoring. According to the strong real-time characteristics of the video signal, the kiln head flame monitoring video signal is diverted, one channel is connected to the central control video monitoring, and the other channel is connected to the edge server and uploaded to the cloud. By intercepting the video screen, the brightness of the flame is identified and divided into five levels: bright, brighter, normal, darker, and dark, to judge the condition of the firing zone. Bright means that the current kiln condition is good, and dark means that the current kiln condition is poor. This provides data support for system control, enhances the timeliness of control, and saves energy and reduces consumption.

[0033] like Figure 2 As shown, the above index optimization method includes the following steps:

[0034] Step 1: Collect historical data of key process variables that indicate kiln conditions. Key process variables include at least: flame brightness at the kiln head, secondary air temperature, kiln main motor current, high temperature NO at the kiln tail, and so on. x Content and free calcium oxide content of clinker.

[0035] Step 2: Normalize the historical data, determine the weights of each key process variable through data fitting, and obtain the formula for the comprehensive index value of kiln conditions. The fitting formula is: comprehensive index value of kiln conditions = normalized value of flame brightness * coefficient 1 + normalized value of secondary air temperature * coefficient 2 + normalized value of kiln main motor current * coefficient 3 + normalized value of kiln tail high temperature NOx * coefficient 4 + normalized value of clinker free calcium oxide content * coefficient 5.

[0036] The flame brightness is identified based on the video footage captured during each calculation, and is divided into five levels: bright, brighter, normal, darker, and dark; the corresponding flame brightness normalization values ​​are: bright (1.0), brighter (0.5), normal (0), darker (-0.5), and dark (-1.0).

[0037] If the key process variables can be directly measured, calculate the rolling average of the corresponding key process variables at time intervals, the interval time is t, and the rolling average of the corresponding key process variables in each interval is used. The normalized formula is: Key process variable normalized value = (key process variable rolling average （当前） -Rolling average of key process variables （t前） ) / maximum change range of rolling average of key process variables.

[0038] Rolling average of key process variables （当前） It is the rolling average of the key process variables calculated within a time interval from the current time. "t before" is the rolling average of the key process variables before an interval. （当前） The difference between the two is the change range of the rolling average of the key process variables before and after a time interval. The maximum value of the change range, that is, the maximum change range of the rolling average of the key process variables, can be directly calculated based on historical data.

[0039] In this embodiment, t is 10 seconds. The key process variables using the above formula include secondary air temperature, kiln main motor current and kiln tail high temperature NO x The corresponding formula is as follows.

[0040] Secondary air temperature: Normalized value of secondary air temperature = (rolling average of secondary air temperature （当前） -Rolling average of secondary air temperature （10s前） ) / maximum variation of rolling average secondary air temperature.

[0041] Kiln main motor current: Kiln main motor current normalized value = (kiln main motor current rolling average （当前） -Rolling average of kiln main motor current （10s前） ) / Maximum variation of rolling average current of kiln main motor.

[0042] Kiln tail high temperature NO x Content: High temperature NO at the kiln tail x Content normalized value = (kiln tail high temperature NO x Content rolling average （当前） - High temperature NO at the kiln tail x Content rolling average （10s前） ) / kiln tail high temperature NO x The maximum change range of the rolling average content.

[0043] For the secondary air temperature, in this embodiment, the maximum change range of the rolling average of the secondary air temperature that can be calculated based on historical data is 0.28. The corresponding formula is first converted into: Secondary air temperature normalized value = (secondary air temperature rolling average （当前） -Rolling average of secondary air temperature （10s前） ) / 0.28. After that, the kiln main motor current and the kiln tail high temperature NO x The normalization formula for content can also be converted in a similar way.

[0044] For the free calcium oxide content of clinker, under conventional technology, this value is generally obtained based on manual sampling quality inspection with a cycle of 2 hours, which will have a large lag when used for control. To solve this problem, the system applies the free calcium oxide prediction algorithm for clinker, establishes a prediction model, uses the predicted value for control, and uses the test results corresponding to the predicted value in the historical data to correct the prediction model to maintain the accuracy of the predicted data. This can avoid the lag in the calculation of the normalized value and ensure that the current comprehensive index value of the kiln condition can be obtained together with the normalized values ​​of other variables currently calculated. The prediction algorithm is corrected by the content actually detected in the historical data to ensure its reliability in actual use. The corresponding formula is as follows.

[0045] Normalized value of free calcium oxide content in clinker = (predicted value of free calcium oxide content in clinker （当前） =(-clinker free calcium oxide content target setting value) / the maximum difference between the clinker free calcium oxide content and the target setting value. The maximum difference between the clinker free calcium oxide content and the target setting value can be adjusted according to actual conditions. In the above formula of this embodiment, according to historical data, it can be first calculated that the maximum difference between the clinker free calcium oxide content and the target setting value is 0.55, then the formula can be converted into: clinker free calcium oxide content normalized value = (clinker free calcium oxide content predicted value) （当前） - target setting value of free calcium oxide content in clinker) / 0.55 and then apply it to the actual calculation.

[0046] The normalized values ​​calculated above are introduced into the fitting formula, and the values ​​of each coefficient are obtained after fitting a large amount of data. In this embodiment, coefficient 1=0.2, coefficient 2=0.2, coefficient 3=0.1, coefficient 4=0.3, and coefficient 5=0.2 are obtained after fitting using historical data. So there is a formula for the kiln condition comprehensive index value after fitting: kiln condition comprehensive index value = flame brightness normalized value * 0.2 + secondary air temperature normalized value * = 0.2 + kiln main motor current normalized value * 0.1 + kiln tail high temperature NO x Normalized value of content*0.3+normalized value of free calcium oxide content in clinker*0.2.

[0047] Step 3: Collect the real-time data of the above key process variables during operation. The data is transmitted to the control center of the system in real time, including the brightness of the kiln head flame, secondary air temperature, kiln main motor current, high temperature NO at the kiln tail, etc. x The content and free calcium oxide content of clinker, where the brightness of the kiln head flame is obtained by identifying the flame brightness in the kiln through the recognition module after intercepting the real-time video screen, and the free calcium oxide content of clinker is obtained according to the free calcium oxide prediction algorithm of clinker. The input values ​​of the free calcium oxide prediction algorithm of clinker include the decomposition furnace outlet temperature, feed rate, secondary air temperature, kiln current average value, cement raw material silicate rate and other input variables related to the free calcium oxide index of cement burning process, thereby obtaining the predicted value of the free calcium oxide content of clinker. In the previous step 2, the prediction model for the prediction of free calcium oxide in clinker has been corrected by using historical data as training samples, so that the prediction value can be guaranteed to have sufficient reliability.

[0048] Step 4: Calculate the comprehensive index value of kiln condition according to the formula fitted in step 2, and make a comprehensive evaluation of kiln condition. Process the real-time data collected in the previous step according to the normalization method of step 2, and introduce the normalized value into the formula of the comprehensive index value of kiln condition fitted in step 2 to obtain the comprehensive index value of kiln condition as the evaluation result of kiln condition.

[0049] The above formula for the comprehensive kiln condition index value can be calculated as a value between -1 and 1. The specific kiln condition corresponding to different kiln condition comprehensive index value intervals is set according to experiments and experience, for example: > 0.5, indicating that the kiln condition is very good; 0.5> comprehensive index value of kiln condition > 0.2, indicating that the kiln condition is good; 0.2> kiln condition comprehensive index value>-0.2, indicating that the kiln condition is normal; -0.2 > The comprehensive index value of kiln condition > -0.5 indicates that the kiln condition is poor; -0.5 > The comprehensive kiln condition index value indicates that the kiln condition is very bad. This comprehensive kiln condition index value and the corresponding specific kiln condition provide a basis for the next step of optimizing and adjusting the system control index (i.e. the target value of the corresponding parameter).

[0050] Step 5: Introduce the comprehensive index value of kiln conditions into the model algorithm to realize automatic optimization of the target value of the decomposition furnace temperature.

[0051] The decomposition furnace temperature is a key control variable in the cement clinker production process. Improving the stability to the optimal value can effectively improve the product quality of the clinker and achieve the goal of reducing energy consumption. In the normal production process, the decomposition furnace temperature is mainly adjusted by adding or subtracting the amount of coal in the decomposition furnace coal feeding scale. The existing cement expert automatic operation system realizes automatic control according to the decomposition furnace temperature target value manually set by the operator, and the temperature target value requires the operator to manually adjust it according to the kiln working conditions. In this scheme, since the kiln condition can be comprehensively judged by calculating the comprehensive kiln condition index, introducing it into the following model algorithm can realize the automatic recommendation of the decomposition furnace temperature target value, and achieve full automatic control of the decomposition furnace temperature, without the operator manually setting or adjusting the decomposition furnace temperature target value according to the changes in the kiln condition, and intervening in the automatic control process.

[0052] The specific optimization control methods for the target value of the decomposition furnace temperature include: when the kiln condition is very good (that is, the comprehensive index value of the kiln condition is > 0.5), the system judges the secondary air temperature fluctuation by calculating the fluctuation amplitude of the secondary air temperature within a certain rolling time interval, and the corresponding fluctuation amplitude is calculated as follows: Secondary air temperature fluctuation amplitude = current value of secondary air temperature - rolling average value of secondary air temperature in the previous t' minutes, where t' minutes is the rolling time interval, which is set to 5 minutes in this embodiment.

[0053] If the fluctuation reaches the first state below: the absolute value of the secondary air temperature fluctuation amplitude > the set value of the secondary air temperature fluctuation amplitude （SP） (SP stands for Set Point), the system starts to accumulate time, and the time threshold for time accumulation is t", which is 30 minutes in this embodiment. If the fluctuation continues for t" in the first state, the target setting value of the decomposition furnace temperature is increased by 0.1 degrees, and the accumulated time is reset; if the fluctuation continues for less than t" but remains in the first state, the target value of the decomposition furnace temperature remains unchanged, and the time accumulation continues. If the fluctuation changes and no longer maintains the first state, the target value of the decomposition furnace temperature remains unchanged, and the accumulated time is reset.

[0054] If the fluctuation reaches the following second state: the absolute value of the secondary air temperature fluctuation amplitude < the set value of the secondary air temperature fluctuation amplitude （SP） , the system starts to accumulate time. If the fluctuation lasts for t" in the second state, the target setting value of the decomposition furnace temperature is lowered by 0.1 degrees, and the accumulated time is reset; if the fluctuation lasts for less than t" but remains in the second state, the target value of the decomposition furnace temperature remains unchanged, and the time accumulation continues; if the fluctuation changes and no longer remains in the second state, the target value of the decomposition furnace temperature remains unchanged, and the accumulated time is reset.

[0055] There are many factors that affect the secondary air temperature. The above optimization method takes into account that when the kiln condition is very good, all indicators of the key process variables are positive. The correlation between the secondary air temperature and the kiln condition can characterize the change of the kiln condition, and the fluctuation range of the secondary air temperature reflects the relationship between the clinker quality and energy consumption in this case. When the fluctuation range of the secondary air temperature is at the set value (that is, the optimal point determined by the experiment) or fluctuates around the set value, the firing system can take into account both the clinker quality and energy consumption at the same time to achieve the optimization purpose.

[0056] In other cases (i.e., the comprehensive kiln condition index value < 0.5), the formula for the target temperature value of the decomposition furnace is: target temperature value of the decomposition furnace = target temperature value of the decomposition furnace （前一次） - Comprehensive kiln condition index value*K, K is the coefficient fitted by historical data, the target value of the decomposition furnace temperature （前一次） The target value of the decomposition furnace temperature obtained by the previous calculation of the formula, and its initial value is the target value of the decomposition furnace temperature set by the operator in the initial state. In this embodiment, the formula of the comprehensive index value of the kiln condition is obtained after fitting in step 2, and the comprehensive index value of the kiln condition corresponding to the historical data in time is immediately calculated based on it. The historical data and the corresponding comprehensive index value of the kiln condition are introduced into the formula of the target value of the decomposition furnace temperature, and the coefficient K=3 can be fitted. In this way, in step 4, when the kiln condition is not very good, the formula of the target value of the decomposition furnace temperature can be used: target value of the decomposition furnace temperature = target value of the decomposition furnace temperature （操作员设定） - Comprehensive kiln condition index value*3, automatically calculate the target value of the decomposition furnace temperature, and optimize the target setting value of the decomposition furnace temperature.

[0057] The above situation is that the kiln condition is not in a very good state. Since it is impossible to judge the kiln condition and the corresponding energy consumption changes through a single key process variable, the optimization of the decomposition furnace temperature target value is to control the decomposition furnace temperature target value to make it close to or reach the state of the kiln condition comprehensive index value = 0. At this time, the kiln condition is in a normal state, and the clinker quality and energy consumption are relatively balanced.

[0058] Step 6: Optimize the automatic setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale according to the comprehensive index value of the kiln condition.

[0059] Affected by the start and stop of the coal mill and the switching of coal feeding into the pulverized coal bin, the actual amount of coal discharged from the coal feeding scale will vary greatly when the set coal feeding amount remains unchanged. The amount of coal fed to the decomposition furnace can be quickly reflected by the change of the decomposition furnace temperature, and the coal feeding amount can be increased or decreased in time to reduce the impact. However, there is no indicator that can quickly measure the amount of coal fed to the kiln head. Here, the process control variable of the Roots blower coal feeding pressure of the kiln head coal feeding scale is introduced to reflect the actual amount of coal fed to the kiln head. The amount of coal fed to the kiln head is increased or decreased by adjusting the target setting value of the Roots blower coal feeding pressure of the kiln head coal feeding scale. If the pressure is stable, the coal amount is stable.

[0060] The optimization of the target setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale is mainly based on the comprehensive indicators of the kiln working conditions. If the comprehensive index value of the kiln condition is >0.8, the target setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale is reduced by 0.01; if the comprehensive index value of the kiln condition is <-0.8 and the current target value of the decomposition furnace temperature has reached the maximum, the head coal pressure is increased by 0.01. The maximum value of the target value of the decomposition furnace temperature is the maximum value determined by the operator based on historical data to avoid the target value of the decomposition furnace temperature being too high, resulting in energy consumption exceeding the upper limit.

[0061] The above technical solution to the technical problem is based on the following principles:

[0062] 1. Do "subtraction" on the data. In a complex production process, not all production parameters will have a direct and critical impact on clinker quality or energy consumption. With the help of the big data platform and the super "computing power" of cloud computing, through data modeling, quantitative analysis of process parameters is carried out, and the logical relationship between different variables is analyzed to find the parameter combination that can bring the most value to production. This method thus determines that the key process variables related to the comprehensive index value of the kiln condition include: the brightness of the flame at the kiln head, the secondary air temperature, the kiln main motor current, and the high temperature NO at the kiln tail. x Content and free calcium oxide content of clinker.

[0063] 2. Introduce a data-driven automatic control technology solution for cement production lines. This solution adopts a research approach that is significantly different from any previous industrial control. It combines deep learning technology with traditional control for the first time, and is adaptive and self-learning. It continuously uses historical data to fully explore the correlation between parameters and find the most suitable control parameter combination for the current working conditions, overcoming the rigid and complex rule constraints of the cement expert automatic operating system. On the other hand, the model algorithm can update itself at any time, and optimize and adjust the algorithm parameters in real time according to the current working conditions, fundamentally overcoming the pain point of the existing technology that the system cannot adapt to changes in working conditions for a long time.

[0064] The present invention is described above by way of example in conjunction with the accompanying drawings. It is obvious that the specific implementation of the present invention is not limited to the above-mentioned method. As long as various non-substantial improvements are made using the inventive concept and technical solution of the present invention, or the inventive concept and technical solution are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. A method for optimizing control indicators of cement production and burning system based on industrial big data. Features: The following steps are involved: Step 1: Collect historical data of key process variables that indicate kiln conditions; Step 2: Normalize the historical data, determine the weights of each key process variable through data fitting, and obtain the formula for the comprehensive index value of the kiln condition; Step 3: Collect real-time data of the above key process variables during operation; Step 4: Calculate the comprehensive index value of kiln condition according to the formula fitted in step 2, and make a comprehensive evaluation of the kiln condition; Step 5: Introduce the comprehensive index value of kiln condition into the model algorithm to realize automatic optimization of the target value of decomposition furnace temperature; The key process variables include at least: the brightness of the kiln head flame, the secondary air temperature, the kiln main motor current, the high temperature NO x content and free calcium oxide content of clinker; The formula for the kiln condition comprehensive index value used for fitting is: kiln condition comprehensive index value = flame brightness normalized value * coefficient 1 + secondary air temperature normalized value * coefficient 2 + kiln main motor current normalized value * coefficient 3 + kiln tail high temperature NOx normalized value * coefficient 4 + clinker free calcium oxide content normalized value * coefficient 5; The optimization control method of the target value of the decomposition furnace temperature in step 5 includes: > When the value is 0.5, the system determines the fluctuation of secondary air temperature by calculating the fluctuation range of secondary air temperature within a certain rolling time interval. The formula for the corresponding fluctuation range is: secondary air temperature fluctuation range = current value of secondary air temperature - rolling average value of secondary air temperature in the previous t' minutes, where t' minutes is the rolling time interval; If the fluctuation reaches the first state below: the absolute value of the secondary air temperature fluctuation amplitude > the set value of the secondary air temperature fluctuation amplitude （SP） , the system starts to accumulate time, and the time accumulation threshold is t". If the fluctuation lasts for t" in the first state, the target setting value of the decomposition furnace temperature is increased by 0.1 degrees, and the accumulated time is reset; if the fluctuation lasts for less than t" but remains in the first state, the target value of the decomposition furnace temperature remains unchanged, and the time accumulation continues; if the fluctuation changes and no longer maintains the first state, the target value of the decomposition furnace temperature remains unchanged, and the accumulated time is reset; If the fluctuation reaches the following second state: the absolute value of the secondary air temperature fluctuation amplitude < the set value of the secondary air temperature fluctuation amplitude （SP） , the system starts to accumulate time. If the fluctuation continues in the second state for t", the target setting value of the decomposition furnace temperature will be lowered by 0.1 degree and the accumulated time will be reset; if the fluctuation continues for less than t" but remains in the second state, the target value of the decomposition furnace temperature will remain unchanged and the time accumulation will continue; if the fluctuation changes and no longer remains in the second state, the target value of the decomposition furnace temperature will remain unchanged and the accumulated time will be reset.

2. According to claim 1, a method for optimizing control indicators of cement production and burning system based on industrial big data, Features: The flame brightness is identified based on the video footage captured during each calculation, and is divided into five levels: bright, brighter, normal, darker, and dark; the corresponding flame brightness normalization values ​​are: 1.0, 0.5, 0, -0.5, and -1.

0.

3. According to claim 1, a method for optimizing control indicators of cement production and burning system based on industrial big data, Features: If the key process variables can be directly measured, calculate the rolling average of the corresponding key process variables at time intervals, the interval time is t, and the rolling average of the corresponding key process variables in each interval is used. The normalized formula is: Key process variable normalized value = (key process variable rolling average （当前） -Rolling average of key process variables （t前） ) / Maximum variation of rolling average of key process variables; Rolling average of key process variables （当前） It is the rolling average of the key process variables calculated within a time interval from the current time. "t before" refers to the rolling average of the key process variables before an interval. （当前） The difference between the two is the change range between the rolling average values ​​of the key process variables before and after a time interval. The maximum value of the change range, i.e. the maximum change range of the rolling average of the key process variables, can be directly calculated based on historical data. The key process variables used in the above formula include secondary air temperature, kiln main motor current and kiln tail high temperature NO x content.

4. According to claim 1, a method for optimizing control indicators of cement production and burning system based on industrial big data, Features: For the free calcium oxide content in clinker, the system applies the free calcium oxide prediction algorithm in clinker to establish a prediction model, uses the predicted value for control, and uses the test results corresponding to the predicted value in the historical data to correct the prediction model to maintain the accuracy of the predicted data. The corresponding formula is as follows: Normalized value of free calcium oxide content in clinker = (predicted value of free calcium oxide content in clinker （当前） - target setting value of free calcium oxide content in clinker) / maximum difference between free calcium oxide content in clinker and target setting value.

5. A method for optimizing control indicators of cement production and burning system based on industrial big data according to any one of claims 1 to 4, Features: The kiln condition comprehensive index value calculated in step 4 is a value between -1 and 1. The specific kiln condition corresponding to different kiln condition comprehensive index value intervals is set according to experiments and experience: > 0.5, indicating that the kiln condition is very good; 0.5> comprehensive index value of kiln condition > 0.2, indicating that the kiln condition is good; 0.2> kiln condition comprehensive index value>-0.2, indicating that the kiln condition is normal; -0.2 > The comprehensive index value of kiln condition > -0.5 indicates that the kiln condition is poor; -0.5 > The comprehensive index value of kiln condition indicates that the kiln condition is very bad; this comprehensive index value of kiln condition and the corresponding specific status of the kiln condition provide a basis for optimizing and adjusting the target values ​​of corresponding parameters.

6. According to claim 1, a method for optimizing control indicators of cement production and burning system based on industrial big data, Features: The optimization control method of the decomposition furnace temperature target value in step 5 also includes: when the kiln condition comprehensive index value is less than 0.5, the formula for the decomposition furnace temperature target value is: decomposition furnace temperature target value = decomposition furnace temperature target value （前一次） - Comprehensive kiln condition index value*K, K is the coefficient fitted by historical data, the target value of the decomposition furnace temperature （前一次） It is the target value of the decomposition furnace temperature obtained by the previous calculation of the formula, and its initial value is the target value of the decomposition furnace temperature set by the operator in the initial state.

7. A method for optimizing control indicators of cement production and burning system based on industrial big data according to any one of claims 1 to 4, Features: The step also includes step 6, optimizing the automatic setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale according to the comprehensive index value of the kiln condition; The optimization of the target setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale is mainly based on the comprehensive index of the kiln working condition. If the comprehensive index value of the kiln condition is >0.8, the target setting value of the coal delivery pressure of the Roots blower of the kiln head coal feeding scale is reduced by 0.01; if the comprehensive index value of the kiln condition is <-0.8 and the current target value of the decomposition furnace temperature has reached the maximum, the head coal pressure is increased by 0.01; The maximum value of the decomposition furnace temperature target value is the maximum value determined by the operator based on historical data to avoid the decomposition furnace temperature target value being too high, resulting in energy consumption exceeding the upper limit.

Images