A production scheduling optimization method for a cement raw material mill

Abstract

The application discloses a cement raw material mill equipment production scheduling optimization method, and steps include: using the historical mill time and power data of the raw material mill in the database, the working conditions of the raw material mill are divided into three kinds of working conditions through the clustering analysis method, and the maximum and minimum values of the corresponding mill time are found out from each kind of working condition; the first linear programming is used to plan the running time of the raw material mill; if the result shows that the ending time point of the current time and the starting time point of the next time interval are not more than the set value, the production scheduling task of the next time interval is migrated to the current time interval; the second linear programming is used to plan and arrange the mill time in the running time interval; the starting and ending time points of each time interval running time interval and the running mill time of each time interval are obtained; and the application can maximize the economic benefits of enterprises, limit the frequency of starting and stopping machines and reduce the equipment wear under the premise of meeting the total output.

Description

Technical Field

[0001] This invention relates to the field of production scheduling technology for industrial equipment operation, and in particular to a production scheduling optimization method for cement raw material grinding equipment. Background Technology

[0002] In the cement industry, due to the high utilization rate of high-power equipment such as raw material roller presses and cement mills, electricity and heat account for a large proportion of energy consumption in cement production, approximately 90% of which comes from coal and electricity. Therefore, the industry typically uses several indicators, including standard coal consumption for clinker burning and comprehensive electricity consumption for cement production, to reflect the actual energy consumption level of a cement production unit. The levels of these indicators can also assess a company's production and management technology level.

[0003] Currently, domestic cement plants typically schedule raw material mill production based on experience. There are generally two scheduling methods: one is to continuously operate the raw material mill before reaching the rated monthly output, taking into account the inventory in the homogenization silo, or to choose to run it continuously for a period and then shut it down for 2-3 days. This scheduling method, due to the prolonged continuous operation of the equipment, keeps the equipment running at high speed, which not only reduces the lifespan of the equipment but also places a significant load on the power grid during peak electricity consumption periods. Considering both the rational use of the national time-of-use electricity pricing mechanism for large industrial electricity consumption for cost control and the reduction of the burden on the national power grid through rational electricity consumption, this scheduling method does not achieve the optimal solution.

[0004] Experience-based production scheduling has the following problems:

[0005] 1. Because the constraints of complex production processes are difficult to quantify, it is impossible to fully describe them using precise modeling methods. Currently, when modeling production scheduling, the main features are usually extracted from multiple simplified features for analysis and modeling. This approach leads to the neglect of many details, resulting in various deviations in practical applications due to overly simplistic modeling parameter selection. These deviations require real-time manual adjustments, reducing the necessity for advance production scheduling.

[0006] 2. When using experience for production scheduling planning, historical production data is generally not used for targeted analysis of production scheduling. The role of data in production scheduling optimization is not emphasized, resulting in low scientific rigor and poor stability in actual operation. Summary of the Invention

[0007] To address the problems existing in the prior art, this invention provides a production scheduling optimization method for cement raw material mills that ensures current production can reasonably meet the needs of subsequent production without causing inventory pressure, maximizes enterprise economic benefits while meeting total output requirements, limits the frequency of start-up and shutdown, and reduces equipment wear.

[0008] The objective of this invention is achieved through the following technical solutions.

[0009] A method for optimizing the production scheduling of cement raw material grinding equipment, comprising the following steps:

[0010] Step 1: Using historical operating hours and electricity consumption data of the raw material mill in the database, the operating conditions of the raw material mill are divided into three types (a, b, and c) through cluster analysis. Simultaneously, the maximum and minimum operating hours for each type are identified, denoted as (ts...). max ,ts min ) (a，b，c) These three predicted operating conditions will be used to compare with the actual operating conditions of the equipment in the future.

[0011] Step 2: Using the time spent on the machine and the amount of electricity consumed, calculate the corresponding power. For these three features, use the k-means clustering method to obtain three clustering results. Each result includes the time spent on the machine (ts), the amount of electricity consumed (E), and the power (P).

[0012] [(ts1, E1, P1), (ts2, E2, P2), (ts3, E3, P3)], find the maximum and minimum timeout values ​​of the overall data ((ts1, E1, P1), (ts2, E2, P2), (ts3, E3, P3)). all ) max ,(ts all ) min );

[0013] Step 3: The first linear programming step aims to achieve the lowest electricity cost by planning the running time of the raw material mill:

[0014] The objective function of the linear programming is:

[0015] Where C represents the total electricity cost, t represents the time-of-use period, which includes (p1, p2, f1, f2, v1, v2), representing peak, flat, and valley electricity consumption times respectively, c represents the government-mandated time-of-use electricity price, and T... s T represents the starting time of a time-sharing period. e E indicates the end time of this time slot. t For time-of-use electricity consumption, via (-T) s +T e Calculate the running time, where T s and T e These are the parameters that need to be planned in this patent;

[0016] The constraints for linear programming are:

[0017] Among them, ts t For the operating time of the station during the time period t, Od For daily production, O d =O a / days n O a For the monthly production requirement entered by the user, days n The number of days in each month; the following formula ensures that the start and end times of each time segment are calculated at fixed positions and are not affected by other time periods.

[0018]

[0019] E t and ts t The clustering results obtained in step 2 will be substituted into the formula and 27 combined iterations will be performed to find the optimal start and end points of the running time and the combination of time and power consumption for each time period.

[0020] Step 4: Determine the start and end points of each time period. To reduce the additional wear and tear on the equipment caused by excessive start and stop and the power cost of instantaneous high power consumption, if the result shows that the interval between the end time of the current time period and the start time of the next time period does not exceed the set value, the production scheduling task of the next time period will be moved to the current time period to ensure the continuity of operation. If the interval is greater than or equal to the set value, production will be carried out according to the production plan calculated by the model. In this case, directly execute step 8.

[0021] Step 5: After the calculation in Step 3, the running time points of the time-sharing period are obtained. After the time merging in Step 4, if the daily output cannot meet the target requirements, a second linear programming is required.

[0022] Step 6: The second linear programming step still aims to achieve the lowest electricity cost, and plans and arranges the time of the station during the operating period:

[0023] The objective function formula is as follows:

[0024] Where C is the total electricity cost, t is the time-of-use period, which includes (p, f, v), representing the peak, off-peak, and valley time segments respectively, and T... s T represents the start time of the time segment calculated in step 4. e P represents the end time of the time-sharing period calculated in step 4. t For time-of-use power consumption data, ts t The operating time of the time-division segment to be solved;

[0025] The constraints of the objective function are:

[0026] -O c ≤-O dAmong them, O c For the calculated daily total output, O d ts is the daily output based on the rated daily demand. min and ts max Among operating conditions a, b, and c, the lowest ts is... min and the highest operating conditions TS max As the upper and lower limits of the timetable operation; ts min ≤ts t ,ts t ≤ts max ;

[0027] Step 7: Through steps 2 to 6, obtain the start and end times of each time-sharing operation period, as well as the running time of each time-sharing segment;

[0028] Step 8: After the actual operation according to the production schedule, the user can customize the automatic update time point. When the automatic update time point is reached, the output will be automatically reduced by the output already produced, and the production schedule duration will also be automatically reduced by the time already elapsed. The program will then be run again to schedule production, repeating steps 3 to 8 to ensure maximum profit.

[0029] The three operating conditions a, b, and c mentioned in step 1 are as follows: operating condition a is when the machine time is greater than 460, operating condition b is when the machine time is between 460 and 390, and operating condition c is when the machine time is less than 390.

[0030] The setting value mentioned in step 4 is 1.5 hours.

[0031] In step 7, the results, including the start and end times of each time-sharing operation period and the running time of each time-sharing period, will be plotted into a Gantt chart for display.

[0032] Compared to existing technologies, the advantages of this invention are as follows: Before using the production scheduling algorithm, the average monthly electricity expenditure for enterprises was around 980,000 yuan. After using the production scheduling algorithm, under ideal conditions, the average electricity expenditure can be reduced by 8%. At the same time, it reduces the error rate of manual calculation, making the production scheduling plan more scientific and feasible. This improves the economic efficiency of enterprises and reduces the peak load on the power grid. This invention ensures that current production can reasonably meet the needs of subsequent production without creating inventory pressure, maximizes the economic benefits of enterprises while meeting total output requirements, limits the frequency of start-up and shutdown, and reduces equipment wear and tear. Attached Figure Description

[0033] Figure 1 This is a flowchart of the present invention.

[0034] Figure 2 This is a Gantt chart showing the production scheduling results.

[0035] Figure 3 This is a comparison chart of electricity expenses. Detailed Implementation

[0036] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0037] like Figure 1 As shown, a method for optimizing the production scheduling of a cement raw material grinding equipment includes the following steps:

[0038] Step 1: Acquiring Basic Data

[0039] Step 1.1: After the cement plant releases its production plan for the raw material mill for the following month, this technology will automatically read the production plan data, including the total production output and number of production days for the next month. According to O d =O a / days n Calculate the lower limit of daily output.

[0040] Step 1.2: Retrieve historical production hours and electricity consumption for the past 3 months from the database. Calculate power consumption based on the hours and electricity consumption, perform data cleanup, and find the maximum and minimum values ​​of hourly and power consumption after data cleanup. all ) max ,(ts all ) min )

[0041] Step 1.3: Based on the data in Step 1.2, use clustering methods to obtain 3 clustering results, which are 3 different operating conditions. For example, the case where the machine time is greater than 460 is the case a, the case where the machine time is between 460 and 390 is the case b, and the case where the machine time is less than 390 is the case c.

[0042] Step 1.4: Using the k-means clustering method, split the time spent on the machine, the amount of electricity, and the power consumption from Step 1.2 into 3 groups. Each group's results include time spent on the machine (ts), the amount of electricity (E), and the power consumption.

[0043] [(ts1, E1, P1), (ts2, E2, P2), (ts3, E3, P3)].

[0044] Step 1.5: Based on the operating condition clustering results in Step 1.3, find the maximum and minimum values ​​of the downtime and power consumption corresponding to each operating condition, denoted as ((ts, P)). min (ts, P) max ) a，b，c .

[0045] Step 1.6: Read the electricity price for the corresponding time period from the database, as well as the start and end times of each time period.

[0046] Step 2: Initial Production Scheduling

[0047] Step 2.1: Substitute the time-of-use electricity price and time-of-use start and end times obtained from Step 1.6 into the objective function formula. middle.

[0048] Step 2.2: Substitute the clustering results from Step 1.4 into the constraints, and perform 27 iterations using a random combination method to find the optimal combination of time and power consumption.

[0049] Step 2.3: Combining the objective function and constraints, the interior point method is used to solve the problem. In 27 iterations, the objective function result of each iteration is compared with the previous one. Under the premise of satisfying the constraints, the combination result of production start and end time and unit hours with the lowest total electricity cost and electricity consumption is retained.

[0050] Step 2.4: The start and end times of the production time-sharing results from Step 2.3 are determined according to rules. If the end time of a time-sharing segment is less than 1.5 hours from the start time of the next time-sharing segment, the production scheduling task for the next time-sharing segment will be migrated to the current time-sharing segment to ensure operational continuity. If no migration operation is performed, meaning each interval exceeds 1.5 hours, proceed to Step 4; otherwise, continue with the next step.

[0051] Step 3: Due to the time shift, the total output does not meet the daily total output demand, therefore a second production scheduling plan is performed.

[0052] Step 3.1: Substitute the production start and end time results from Step 2.4 into the objective function formula. middle,

[0053] Step 3.2: In the constraints, the maximum and minimum time of the time are substituted from the calculation results of step 1.2.

[0054] Step 3.3: Solve the objective function and constraints using the interior point method. The results will include the start and end times of the operation for each time period and the station hours for each time period.

[0055] Step 4: Using the results from Step 3.3, draw a Gantt chart to display the results.

[0056] Step 5: After production is scheduled and put into operation, record the output already produced (O). x .

[0057] Step 6: Every hour, calculate the average operating hours of the past hour. Based on the calculation result, find the operating condition in Step 1.3. For example, if the average operating hours are 420, according to the result of Step 1.3, it is classified as operating condition b. The classification of other operating conditions is the same.

[0058] Step 7: Use total output O a Subtract O x The remaining unproduced output O r .

[0059] Step 8: Perform k-means clustering on the operating hours, power consumption, and power usage in operating condition b to obtain three results.

[0060] Step 9: Repeat steps 2 to 8. In this way, the actual operating conditions of the raw material mill can be tracked in real time, so that the production schedule can be aligned with reality.

[0061] like Figure 2 As shown, this is the daily operating schedule. Every hour, the schedule will be recalculated based on the current operating conditions and displayed again using a similar Gantt chart.

[0062] like Figure 3 As shown, the line graph represents monthly output, the light-colored histogram represents monthly electricity costs before using production scheduling, and the dark-colored histogram represents electricity costs after using it. It can be seen that, under the same output conditions, using production scheduling can theoretically reduce monthly electricity costs by at least 5%, and up to 15%, thus improving the company's economic efficiency.

Claims

1. A method of production scheduling optimization for a cement raw mill plant, characterized by the steps of Comprising: Step 1: Using the historical data of the raw mill in the database, the working conditions of the raw mill are divided into three working conditions a, b, and c by clustering analysis method, and the maximum and minimum values corresponding to the time are found from each working condition, represented as (ts max , ts min ) (a，b，c) The three predicted working conditions will be used for subsequent comparison with the actual working conditions of the equipment; Step 2: using the time and the power, calculate the corresponding power, for the three characteristic data, using kmeans clustering method, get 3 groups of clustering results, each group of results contains time ts, power E, power P, that is [(ts1, E1, P1), (ts2, E2, P2), (ts3, E3, P3)], find the maximum and minimum value of the overall data time ((ts all ) max , (ts all ) min ); Step 3: First linear programming to plan the running time of the raw material mill for the purpose of achieving the lowest power cost: The objective function of linear programming is: Where C represents the total power cost, t represents the time-of-use period, (p1, p2, f1, f2, v1, v2) represents the peak, flat, valley period respectively, c represents the time-of-use electricity price regulated by the government, T s represents the starting time point of the time-of-use period, T e represents the ending time point of the time-of-use period, E t represents the electricity consumption of the time-of-use period, which is calculated by (-T s +T e ), where T s and T e are the parameters that need to be planned in this patent; The constraints of linear programming are: Where ts t is the operating hours at t-minute interval, O d is the daily production, O d = O a / days n , O a is the monthly production requirement input by the user, days n is the number of days in each month; the following formula is to ensure the calculation of fixed positions of the start and end time of each interval, which is not affected by other intervals; E t and ts t The clustering results obtained in step 2 are substituted into the formula, and 27 combination iteration calculations are performed to find the optimal start and end points of the running time and the combination of station time and power consumption of each time period. Step 4: Determine the start and end points of each time period. In order to reduce the additional loss of equipment and the instantaneous high power consumption of excessive start and stop, if the result shows that the end time point of the current time and the start time point of the next time period interval is not more than the set value, the next time period will be migrated to the current time period to ensure the continuity of operation. If the interval is greater than or equal to the set value, the production will be carried out according to the production planning calculated by the model. In this case, step 8 is directly executed; Step 5: After the calculation of step 3, the running time point of the time period is obtained. After the time merging of step 4, if the daily output cannot meet the target requirement, a second linear programming is needed; Step 6: The second linear programming still aims to achieve the lowest power cost to plan the running time period: The objective function is as follows Where C is the total electricity cost, t is the time-of-use period, (p, f, v) are the peak, flat, and valley time partitions, respectively, and T s is the start time point of the time-of-use period calculated in step 4, T e is the end time point of the time-of-use period calculated in step 4, P t is the power consumption data of the time-of-use period, ts t is the operating hours of the time-of-use period to be solved; The constraint condition of the objective function is: -O c ≤-O d , where O c is the total daily production of the result, O d is the daily rated demand day production, ts min and ts max are a, b, c conditions, the lowest ts min of the condition, and the highest ts max of the condition, as the upper and lower limits of the station hour operation; ts min ≤ ts t , ts t ≤ ts max ; Step 7: Through steps 2 to 6, the start and end time points of each time period and the running time of each time period are obtained; Step 8: After the actual operation according to the production planning, the hourly time is averaged, and the corresponding working condition is found according to the average value. Under this working condition, the total daily output is reduced by the already produced output, and steps 3 to 8 are repeatedly executed.

2. The method of optimizing production scheduling of a cement raw mill plant according to claim 1, characterized in that The a, b, and c working conditions in step 1 include: in the case of more than 460 hours, it is a working condition, in the case of between 460 and 390 hours, it is a working condition, and in the case of less than 390 hours, it is a working condition.

3. The method of optimizing production scheduling of a cement raw mill plant according to claim 1, wherein The set value in step 4 is 1.5 hours.

4. The method of claim 1, wherein In step 7, the start and end time points of each time period and the running time of each time period are drawn into a Gantt chart for display.

Images