Intelligent dynamic regulation and control method for cleaning operation of rice and wheat combine harvester
Through intelligent dynamic control methods, the cleaning parameters of the rice and wheat combine harvester are monitored and automatically adjusted in real time, which solves the problems of unstable cleaning quality and low efficiency caused by traditional manual adjustment, and achieves efficient and stable cleaning effects.
Patent Information
- Application Number
- CN202511244968.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-02
- Publication Date
- 2025-10-17
AI Technical Summary
When the feed amount fluctuates or the material properties change, the cleaning system of traditional rice and wheat combine harvesters will lead to high grain loss, high impurity content and increased grain breakage rate. It is also highly dependent on the operator's experience and cannot adapt to the needs of precision agriculture and efficient operations.
An intelligent dynamic control method is adopted to form a closed-loop control by real-time monitoring of the cleaning loss rate and impurity rate, combined with automatic adjustment of operating parameters, including automatic adjustment of fan speed, air distributor angle and upper screen opening.
It improves the stability of cleaning quality and operating efficiency, reduces dependence on operator experience, adapts to complex environments under different crop conditions, and extends continuous operation time.
Smart Images

Figure CN120787623A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of agricultural machinery, more particularly, it relates to a method for intelligently dynamically regulating the cleaning operation of a rice-wheat combine harvester. BACKGROUND
[0002] The cleaning system of a rice-wheat combine harvester is one of the core components of the combine harvester, which is responsible for separating the mixture (grains, short stalks, husks, and residues, etc.) removed by the threshing device, sending clean grains into the grain tank, and discharging impurities outside the machine. The cleaning quality directly determines the harvesting quality (grain cleanliness, breakage rate) and harvesting efficiency (grain loss rate).
[0003] The traditional cleaning system mainly relies on the vibration screen (such as fish scale screen, woven screen) and the airflow generated by the fan for screening and air separation. The key adjustable parameters include screen opening, vibration frequency / amplitude, fan speed, etc. The traditional adjustment method relies on the operator to manually adjust the parameters according to crop types, variety characteristics, maturity, humidity, etc. in the parked state.
[0004] This open-loop control method has the following problems:
[0005] When the feeding amount fluctuates or the material characteristics change, it may lead to high grain loss (grains are discharged with residues outside the machine), high impurity content (impurities enter the grain tank), and increased grain breakage rate; frequent parking is required for adjustment, which seriously disrupts continuous operation; the operator often uses conservative parameter settings, which limits the potential of the machine to achieve maximum feeding capacity; the operator's experience is highly dependent; and it cannot adapt to the needs of precision agriculture and efficient operation.
[0006] The main reasons for these problems are: lack of real-time and accurate cleaning effect perception; the cleaning process is a multivariable, strongly coupled, nonlinear system, which is difficult to optimize manually; manual adjustment has a lag; open-loop control lacks a feedback mechanism; and the field working environment is harsh, requiring high-quality sensors.
[0007] In summary, the cleaning system of the rice-wheat combine harvester in the prior art has the problems of unstable cleaning quality, low operation efficiency, and strong dependence on the operator's experience, etc. Therefore, an intelligent dynamic regulation method that can real-time perceive the cleaning effect and automatically adjust the cleaning parameters is urgently needed to improve the cleaning quality and operation efficiency and reduce the dependence on the operator's experience. SUMMARY
[0008] The present application provides a method for intelligently dynamically regulating the cleaning operation of a rice-wheat combine harvester, which solves the technical problems of unstable cleaning quality and low operation efficiency caused by the traditional manual adjustment method in related technologies.
[0009] The present application provides a method for intelligently dynamically regulating the cleaning operation of a rice-wheat combine harvester, which includes:
[0010] setting an initial value of a working parameter of the cleaning device according to the working condition information;
[0011] monitoring the cleaning loss rate, the cleaning impurity rate and the working parameter in real time;
[0012] comparing the data obtained by real-time monitoring with preset threshold values to evaluate the cleaning working state;
[0013] judging the control priority based on the combination of the cleaning loss rate state and the cleaning impurity rate state obtained by evaluation;
[0014] generating a parameter control decision according to the priority judged and the cleaning working state matched with a corresponding rule set;
[0015] calculating and executing parameter adjustment based on the parameter control decision, and performing cyclic iteration control after evaluating the control effect.
[0016] Further, the real-time monitoring step comprises:
[0017] monitoring the grain content in the tail screen discharge by a cleaning loss monitoring unit;
[0018] monitoring the impurity content of the grain in the grain box by a cleaning impurity rate monitoring unit;
[0019] monitoring the fan rotating speed by a Hall rotating speed sensor;
[0020] monitoring the angle of the air distribution plate and the opening degree of the upper screen by an electric push rod position feedback sensor;
[0021] filtering the monitoring data to eliminate noise.
[0022] Further, the step of evaluating the cleaning working state comprises:
[0023] comparing the cleaning loss rate and the cleaning impurity rate with preset threshold values, wherein the threshold values are divided into a normal range threshold value 0.3% and an extremely high range threshold value 5%;
[0024] dividing the cleaning loss rate and the cleaning impurity rate into normal state, high state and very high state based on the comparison results.
[0025] Further, the step of judging the control priority comprises:
[0026] when the cleaning loss rate and the cleaning impurity rate are both in normal state, maintaining the current parameter unchanged;
[0027] when the cleaning loss rate is in normal state but the impurity rate is in high state, selecting to execute the impurity rate high rule set;
[0028] when the cleaning loss rate is in high state but the impurity rate is in normal state, selecting to execute the loss rate high rule set;
[0029] When the cleaning loss rate and the impurity content rate are both high, the loss rate high rule set is preferentially selected to be executed;
[0030] When either the cleaning loss rate or the impurity content rate is very high, the abnormal situation processing rule set is executed.
[0031] Further, the parameter regulation decision step includes:
[0032] Extracting corresponding rules from the knowledge base and matching them with the current state according to the priority order;
[0033] When multiple rules are matched successfully, the most suitable rule is selected through a conflict resolution strategy;
[0034] The conflict resolution strategy is based on the change trend of the cleaning loss rate and the impurity content rate and the regulation effect evaluation result of the last time;
[0035] Generating a specific regulation decision, including the parameters to be adjusted and the adjustment direction.
[0036] Further, the calculation and execution of parameter adjustment steps include:
[0037] Calculating the fan speed adjustment amount, the air distribution plate angle adjustment amount, and the upper screen opening adjustment amount;
[0038] Adjusting the air distribution plate angle and the upper screen opening through an electric push rod;
[0039] Adjusting the fan speed through a fan control unit;
[0040] Waiting for 10 to 15 seconds after parameter adjustment to allow the system to reach a new steady state;
[0041] Evaluating the regulation effect based on the change trend of the cleaning loss rate and the impurity content rate before and after adjustment;
[0042] Using the evaluation result as a reference for the next round of regulation decision.
[0043] Further, in the initial setting of the operation parameter step:
[0044] The operation condition information includes crop type, variety characteristics, maturity, humidity, and weed content;
[0045] Retrieving the most matched parameter combination from the historical case library as the initial parameter;
[0046] Normalizing the input parameters of different dimensions to map all parameters to the 0-1 interval;
[0047] Converting the crop type and variety characteristic categorical variables into numerical feature vectors through one-hot encoding.
[0048] Further, the rules include:
[0049] The high impurity rate processing rule includes reducing the upper sieve opening, increasing the fan speed, adjusting the air distribution plate angle and reducing the cleaning load.
[0050] The high loss rate processing rule includes increasing the upper sieve opening, adjusting the air distribution plate angle, reducing the fan speed and reducing the cleaning load.
[0051] The abnormal situation processing rule includes immediately reducing the cleaning load and executing the regulation and control decision in the high loss rate processing.
[0052] Further, the regulation and control effect evaluation includes:
[0053] The average value difference of the cleaning loss rate and the impurity rate in a time window before and after the parameter adjustment is calculated.
[0054] The difference value is compared with the expected improvement target and the stability of the change trend is considered.
[0055] According to the comprehensive score of the improvement degree, the stability and the adjustment amplitude, the effect is divided into three levels of effective, partially effective and ineffective.
[0056] The three different dimension indicators of the improvement degree, the stability and the adjustment amplitude are normalized by minimum and maximum value processing.
[0057] The present application provides a computer storage medium, including a memory and one or more processors, the memory stores executable code, and the one or more processors execute the executable code to implement the above-mentioned intelligent dynamic regulation and control method for rice and wheat combined harvester cleaning operation.
[0058] The beneficial effects of the present application are that by constructing a cleaning device operation quality and operation parameter dynamic monitoring and regulation system, real-time intelligent regulation and control of cleaning operation parameters is realized.
[0059] By real-time monitoring of cleaning loss rate and impurity rate and intelligent dynamic regulation and control, the stability of cleaning quality is improved, and the cleaning loss rate and impurity rate can be stably controlled within the ideal range under different crop conditions (including high humidity, high weeds, crop lodging and other complex environments), which is better than the traditional manual adjustment method.
[0060] Through the automatic parameter adjustment mechanism, the operation efficiency is improved, the unit area operation efficiency is increased, the continuous operation time is prolonged, the need for manual intervention is completely eliminated, and more precise and timely parameter regulation and control is realized.
[0061] Through the adaptive learning mechanism, the adaptability of the system to different crop types, variety characteristics, maturity, humidity, weed content and other conditions is enhanced, the cleaning parameters can be automatically adjusted, diversified harvesting operation scenes can be adapted, and excellent cleaning effect is shown in high humidity conditions (such as after the rain) and high weed environment.
[0062] Through the intelligent regulation mechanism, the dependence on the experience of the operator is reduced, so that the novice driver can complete high-quality operation without mastering professional debugging skills, the problem of skill shortage of agricultural labor is effectively alleviated, and the modernization of agricultural production is promoted. BRIEF DESCRIPTION OF DRAWINGS
[0063] Figure 1 is a flowchart of the intelligent dynamic regulation method for the cleaning operation of the rice-wheat combine harvester in the present application. DETAILED DESCRIPTION
[0064] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is merely meant to provide a better understanding of the subject matter described herein and can include changes to the function and arrangement of the elements discussed without departing from the scope of the content of this specification. Various examples can omit, substitute, or add various procedures or components as desired. In addition, features described in some examples can be combined in other examples.
[0065] At least one embodiment of the present application discloses an intelligent dynamic regulation method for the cleaning operation of a rice-wheat combine harvester, as shown in Figure 1 , including:
[0066] Step 1: Set the initial value of the cleaning device operation parameter according to the operation condition information;
[0067] Input: crop type (rice or wheat), variety characteristics, maturity, humidity, weed content and other current operation condition information, wherein the crop type and variety characteristics can be selected and input by the operator on the control panel; the maturity and humidity of the crop can be automatically detected by the humidity sensor installed on the harvester; the weed content can be input by the image recognition system or the operator after observation or automatically obtained from the field management system imported field data related information.
[0068] Processing process: the cleaning controller retrieves the parameter combination most matched with the current operation condition from the historical case library as the initial parameter of the cleaning device. The historical case library stores the optimal parameter combination of different regions, different crop varieties under various operation conditions, and the best matching case is selected through similarity calculation.
[0069] Before similarity calculation, the system normalizes the input parameters of different dimensions (such as maturity percentage, humidity percentage, weed content percentage) to the maximum and minimum value, maps all parameters to the 0-1 interval, and ensures that each parameter has the same weight in similarity calculation. At the same time, the classification variables such as crop species and variety characteristics are converted into numerical feature vectors through one-hot encoding, so that they can participate in similarity calculation.
[0070] Output: Initial setting values of the initial setting values of the operation parameters of the cleaning device such as fan speed, air distribution plate angle, upper sieve opening degree, vibration frequency, etc. are sent to the execution mechanism by the cleaning controller to complete the initialization setting.
[0071] Step 2, real-time monitoring of cleaning loss rate, cleaning impurity rate and operation parameters;
[0072] Input: Grain content sensor in cleaning loss monitoring unit, impurity content sensor in cleaning impurity rate monitoring unit, Hall speed sensor on fan shaft and electric push rod position feedback sensor. Real-time data collected.
[0073] The cleaning controller continuously receives data from the cleaning loss monitoring unit and the cleaning impurity rate monitoring unit through the CAN bus, and receives the Hall speed sensor and electric push rod position feedback signal at the same time, with a sampling frequency of 10Hz. After filtering to eliminate noise, the monitoring data is averaged to obtain the current cleaning operation quality and parameter state.
[0074] Through the above processing, the system obtains the real-time values of the key parameters such as cleaning loss rate (mass percentage of grains in tail sieve discharge), cleaning impurity rate (mass percentage of impurities in grain box), fan speed (r / min), air distribution plate angle (angle with horizontal plane, unit: degree) and upper sieve opening degree (size of sieve opening, unit: mm).
[0075] Step 3, compare the data obtained by real-time monitoring with the preset threshold value to evaluate the cleaning operation state;
[0076] Input: Cleaning loss rate and cleaning impurity rate values output in step 2.
[0077] Processing process: The cleaning controller compares the measured cleaning loss rate and impurity rate with the preset threshold value. The threshold value is divided into a normal range threshold (0.3%) and an extremely high range threshold (5%). By comparing the relationship between the measured value and the threshold value, the cleaning operation state category is determined.
[0078] The specific implementation of the state evaluation function is: the system first performs moving average filtering processing on the cleaning loss rate and impurity content rate data in each monitoring period to eliminate abnormal fluctuations; then the filtered data is compared with the preset threshold (0.3% and 5%) through a threshold comparator; the data trend of the last three time points is considered during the comparison process to avoid misjudgment caused by single-point fluctuations; when more than two consecutive data points exceed the corresponding threshold, state change is triggered to improve the stability of the evaluation.
[0079] The evaluation result generates the state determination results of the cleaning loss rate state (normal state, high state, very high state) and the cleaning impurity content rate state (normal state, high state, very high state), and the specific state determination criteria are:
[0080] Loss rate normal state: loss rate ≤ 0.3%;
[0081] Impurity content rate normal state: impurity content rate ≤ 0.3%;
[0082] Loss rate high state: 0.3% < loss rate ≤ 5%;
[0083] Impurity content rate high state: 0.3% < impurity content rate ≤ 5%;
[0084] Loss rate very high state: loss rate > 5%;
[0085] Impurity content rate very high state: impurity content rate > 5%.
[0086] Step 4, based on the evaluation obtained cleaning loss rate state and cleaning impurity content rate state combination to determine the priority of regulation and control;
[0087] Input: cleaning loss rate state and cleaning impurity content rate state output by step 3.
[0088] Processing process: the cleaning controller determines the priority of regulation and control and the category of rules to be executed according to the state combination of the cleaning loss rate and the impurity content rate. The specific determination method is as follows:
[0089] When the cleaning loss rate is in the normal state (≤0.3%) and the impurity content rate is in the normal state (≤0.3%), the system maintains the current parameters unchanged;
[0090] When the cleaning loss rate is in the normal state but the impurity content rate is in the high state (0.3%~5%), the system selects to execute the impurity content rate high rule set;
[0091] When the cleaning loss rate is in the high state (0.3%~5%) but the impurity content rate is in the normal state, the system selects to execute the loss rate high rule set;
[0092] When both the loss rate and the impurity rate are high, the system will preferentially execute the loss rate high rule set, because the loss rate has a greater impact on economic benefits.
[0093] When either the loss rate or the impurity rate is very high (> 5%), the system will immediately execute the abnormal situation processing rule set and issue a warning signal to the operator.
[0094] The judgment result determines the type of rule set to be executed (loss rate high rule set, impurity rate high rule set, abnormal situation processing rule set) and the execution priority, providing a basis for the next parameter regulation decision.
[0095] Step 5: According to the priority of the judgment and the cleaning operation state, match the corresponding rule set to generate parameter regulation decisions.
[0096] Input: Rule set type output in step 4, state judgment result output in step 3, and current operation parameters output in step 2.
[0097] Processing: The cleaning controller extracts the corresponding rules from the knowledge base according to the selected rule set type, and matches them with the current state in order of priority. When multiple rules match successfully, the most suitable rule is selected through conflict resolution strategies. Conflict resolution strategies are based on the change trend of cleaning loss rate and impurity rate at consecutive N times and the evaluation result of the last time regulation effect.
[0098] The specific implementation of the conflict resolution function is as follows: the system constructs a decision matrix, with rows representing selectable rules and columns representing evaluation indicators (including rule applicability, historical success rate, and expected improvement degree). Applicability is calculated by matching the current state with the rule conditions. Historical success rate is obtained from historical execution records under similar conditions. The expected improvement degree is estimated based on the effect of parameter adjustment predicted by the physical model. The three indicators are combined by weighted summation, and the weights are dynamically adjusted according to the crop type. Finally, the rule with the highest score is selected for execution.
[0099] Before constructing the decision matrix, the system standardizes the three indicators of rule applicability, historical success rate, and expected improvement degree with Z-score, making their mean value 0 and standard deviation 1, to eliminate the influence of dimension difference on weighted summation. For categorical variables such as crop type, the system uses one-hot encoding to convert them into numerical weight vectors, so that the characteristics of different crop types can be considered in the weight adjustment process.
[0100] Based on the principle of human-like intelligent control, the system realizes the adaptive selection mechanism of regulation strategy. By recording the effects of different regulation strategies under various operating conditions, the system can automatically learn which strategies are more effective under specific conditions and preferentially select these strategies in subsequent similar situations. At the same time, the system will adaptively adjust the regulation range of operating parameters based on the actual regulation effect. For example, when it is detected that certain parameter changes are sensitive to the cleaning effect, the adjustment step of that parameter will be automatically reduced to improve the regulation accuracy.
[0101] The decision result generates a specific regulation decision, including the parameters that need to be adjusted (fan speed, air distribution plate angle, upper sieve opening) and the adjustment direction (increase or decrease), as the basis for the execution of step 6. According to different situations, the regulation decision may include the following:
[0102] High impurity rate handling: when the cleaning loss rate is normal but the impurity rate is high, according to the current loss rate level and the current state of the fan and upper sieve opening, one of the following regulation decisions is output:
[0103] Decrease the upper sieve opening: when the impurity rate is high and the loss rate is below the normal upper limit;
[0104] Increase the fan speed: when the impurity rate is high, the loss rate is low, and the fan speed has not reached the upper limit;
[0105] Increase the fan speed and the upper sieve opening simultaneously: when the impurity rate is high and the loss rate is close to the normal upper limit;
[0106] Adjust the air distribution plate angle to the parallel position: when the current three kinds of regulation are ineffective;
[0107] Reduce the cleaning load (reduce the feed amount): when all regulation measures are ineffective;
[0108] High loss rate handling: when the cleaning loss rate is high but the impurity rate is normal, according to the current impurity rate level and the moisture content state, one of the following regulation decisions is output:
[0109] Increase the upper sieve opening: when the loss rate is high and the impurity rate is very low;
[0110] Adjust the air distribution plate angle to the parallel position: when the loss rate is high, the impurity rate is normal, and the moisture content is low;
[0111] Reduce the fan speed: when the loss rate is high, the impurity rate is normal, and the moisture content is normal or high;
[0112] Reduce the cleaning load: when the above regulation measures are ineffective;
[0113] Abnormal situation handling: when the cleaning loss rate and the impurity rate are both in abnormal state, one of the following regulation decisions is output:
[0114] Immediately reduce the cleaning load: when both loss rate and impurity rate are very high;
[0115] Firstly, execute the regulation decision in the case of high loss rate: when both loss rate and impurity rate are high, loss rate is processed first.
[0116] Step 6, based on the parameter regulation decision, calculate and execute parameter adjustment, and perform loop iteration control after evaluating the regulation effect.
[0117] This step is based on the regulation decision output in step 5 and the current operation parameters output in step 2 to perform specific execution and evaluation.
[0118] Step 6.1, parameter adjustment calculation and execution;
[0119] The cleaning controller calculates the specific adjustment amount according to the regulation decision. The adjustment amount calculation is based on the current parameter value and the target change direction, while considering the step limit and upper and lower limit constraints of parameter adjustment. For example, the fan speed adjustment amplitude should not exceed 50 r / min, and the upper sieve opening should not exceed 0.5 mm each time, to avoid system instability caused by parameter mutation.
[0120] The calculation result determines the specific parameter adjustment instructions, including fan speed adjustment amount (± X r / min), air distribution plate angle adjustment amount (± Y°), and upper sieve opening adjustment amount (± Z mm). These instructions are sent to the corresponding electric push rod and fan control unit through the cable line to execute specific parameter adjustment actions.
[0121] Step 6.2, regulation effect evaluation and loop iteration;
[0122] The system receives the cleaning operation parameters after step 6.1 execution and the cleaning loss rate and impurity rate data continuously monitored in step 2.
[0123] Process: The cleaning controller waits for 10 to 15 seconds (the time of material passing through the cleaning system) after parameter adjustment, allowing the system to reach a new steady state, and then evaluating the regulation effect. The effect evaluation is based on the change trend of cleaning loss rate and impurity rate before and after adjustment to determine whether the regulation is effective. The evaluation result will affect the conflict resolution strategy in the next regulation decision, forming a closed-loop feedback control.
[0124] The specific implementation of the effect evaluation function is as follows: the system first calculates the average difference in cleaning loss rate and impurity content rate within a certain time window (usually 10-15 seconds) before and after parameter adjustment; then compares the difference value with the expected improvement target; meanwhile, considering the stability of the trend, the fluctuation degree is evaluated by calculating the standard deviation; finally, according to the comprehensive score of the improvement degree, stability and adjustment amplitude of the three factors, the effect is divided into three levels: "effective" (improvement and stability), "partially effective" (improvement but not stable or improvement amplitude small) and "ineffective" (no improvement or deterioration), providing basis for the next round of decision-making.
[0125] Before calculating the comprehensive score, the system performs min-max normalization on the three indicators of improvement degree (percentage value), stability (standard deviation value) and adjustment amplitude (physical unit) of different dimensions, mapping them to the 0-1 interval. For the improvement degree of loss rate and impurity content rate, the system considers the relative change rate rather than the absolute difference, in order to more accurately reflect the control effect. This normalization ensures the comparability of the three factors in the comprehensive score, avoiding the dominance of certain parameters due to large dimensions.
[0126] After evaluation, the control effect evaluation result (effective, partially effective, ineffective) is generated as a reference for the next round of control decision-making. At the same time, the system returns to step 2 to continue monitoring the cleaning operation quality and parameters, forming a complete intelligent dynamic control closed loop.
[0127] To further optimize the dynamic control model of cleaning operation parameters, this embodiment also constructs a multi-parameter, intelligent cleaning simulation mechanical platform. The platform includes the following components:
[0128] Physical simulation module: based on computational fluid dynamics and discrete element method, simulates the motion trajectory and separation effect of materials in the cleaning system under different operation parameters;
[0129] Parameter sensitivity analysis module: analyzes the influence degree and interaction of different cleaning parameters on cleaning effect through orthogonal experimental design method;
[0130] Rule verification and optimization module: verifies the rules extracted in actual operation in the simulation environment, and optimizes the rules based on a large number of simulation results;
[0131] Through this simulation platform, the system can verify the effectiveness of the control strategy in advance before actual application, avoiding the high cost and high risk of field tests, and providing a theoretical basis for the adaptability adjustment of new models and new varieties. In addition, the simulation platform can also provide physical parameters that are difficult to measure directly inside the cleaning system, such as material distribution density and air flow velocity distribution, providing data support for further improvement of the control model.
[0132] Through the above six consecutive and interrelated steps, the embodiment realizes intelligent dynamic regulation and control of the cleaning operation of the rice-wheat combine harvester, effectively solving the problems of unstable cleaning quality and low operation efficiency existing in the traditional manual adjustment mode. The entire regulation and control process does not require manual intervention and can automatically adjust the cleaning parameters according to changes in field conditions, ensuring that the cleaning loss rate and impurity content are stable within the target range.
[0133] A computer storage medium includes a memory and one or more processors, the memory having stored therein executable code, the one or more processors executing the executable code to implement the above-mentioned intelligent dynamic regulation and control method for the cleaning operation of the rice-wheat combine harvester.
[0134] Herein, the present application provides an implementation example:
[0135] To verify the effectiveness and practicality of the intelligent dynamic regulation and control method for the cleaning operation of the rice-wheat combine harvester provided by the embodiment, a detailed application example is given in this section. The example is based on the actual application data of the Liewe Guoshen 4LZ-10 rice-wheat combine harvester in rice harvesting operations in Nantong, Jiangsu.
[0136] The initial condition data is shown in Table 1:
[0137] Table 1: Initial condition data table
[0138] Parameter item Parameter value Acquisition method Crop type Rice Operator control panel input Variety characteristics Nanjing 46 (medium-grain type) Operator control panel input Maturity 92% Humidity sensor automatic detection Humidity 19.5% Humidity sensor automatic detection Weed content 7.2% Image recognition system automatic detection Estimated yield 650 kg / acre Farmland management system data import
[0139] The initial parameter settings are shown in Table 2:
[0140] Table 2: Initial parameter setting table
[0141] Parameter item Initial set value Parameter range Fan speed 780 r / min 600 to 900 r / min Distributor angle 45° 30° to 60° Upper screen opening 8.5 mm 6 to 12 mm Vibration frequency 4.2 Hz 3.5 to 5.0 Hz Feeding amount setting 9.8 kg / s 8.0 to 12.0 kg / s
[0142] The key monitoring parameters are shown in Table 3:
[0143] Table 3: Key monitoring parameter table (initial 10 monitoring points)
[0144]
[0145] The operator selects "rice" as the crop type on the combine harvester control panel and selects the "Nanjing 46" variety from the drop-down menu. The system automatically detects the crop maturity to be 92% and the humidity to be 19.5% through the humidity sensor, and detects the weed content to be about 7.2% through the onboard image recognition system. At the same time, the system imports the estimated yield data of the current field from the farmland management system, and the data is 650 kg / acre.
[0146] The cleaning controller matches these input parameters with the parameter combinations stored in the historical case library. In the historical case library, the system retrieves 3 cases similar to the current conditions:
[0147] Case 1: Nangjing 46, maturity 95%, humidity 18%, weed content 5%;
[0148] Case 2: Nangjing 46, maturity 90%, humidity 20%, weed content 8%;
[0149] Case 3: Nangjing 9311, maturity 93%, humidity 19%, weed content 6%.
[0150] By calculating the comprehensive similarity (variety characteristic weight 0.3, maturity weight 0.2, humidity weight 0.3, weed content weight 0.2), the similarity of Case 2 is the highest, reaching 0.92. The system adopts the parameter combination of Case 2 as the initial parameters and fine-tunes them in combination with the current specific conditions to finally determine the initial parameter settings in Table 2.
[0151] After the harvester starts working, various sensors installed on the cleaning device start working. The grain content sensor in the cleaning loss monitoring unit collects the grain proportion in the tail sieve discharge, the image recognition system in the cleaning impurity content monitoring unit monitors the impurity proportion in the grain box, and the Hall rotation speed sensor and the electric push rod position feedback sensor monitor the actual values of the fan rotation speed and the air distribution plate angle, respectively.
[0152] The system collects data at a sampling frequency of 10 Hz, collecting once every 0.1 second. Every 5 seconds, 50 sample point data are filtered, and the moving average filtering algorithm is used to remove outliers, and then the average value is calculated as the monitoring result at the current time. As shown in Table 3, the monitoring data from 0 seconds to 25 seconds shows that the cleaning loss rate gradually rises from 0.28% to 0.32%, still within the normal range; while the impurity content rises from 0.42% to 1.05%, which has exceeded the normal range threshold of 0.3% and entered the high state.
[0153] Taking the monitoring data at 25 seconds as an example, the cleaning controller receives the data of cleaning loss rate 0.32% and impurity content 1.05%. The system compares these two values with the preset threshold: the normal range threshold is 0.3%, and the extremely high range threshold is 5%.
[0154] The comparison result is: cleaning loss rate 0.32% > 0.3%, judged as "loss rate high state"; impurity content 1.05% > 0.3%, judged as "impurity content high state"; both of them do not exceed the extremely high range threshold of 5%.
[0155] Therefore, the evaluation result of the cleaning device operation state at the 25th second is that the loss rate is in a high state, the impurity rate is in a high state, and the system needs to be regulated.
[0156] Based on the state evaluation result, the cleaning controller needs to determine the priority of regulation. At this time, the cleaning loss rate is 0.32% (high state), and the impurity rate is 1.05% (high state), both of which are in a high state.
[0157] According to the preset priority determination rule, when the cleaning loss rate and the impurity rate are in a high state at the same time, the loss rate problem is given priority. This is because the loss rate reflects the direct waste of grain and has a greater impact on economic benefits. Therefore, the system selects the "loss rate high rule set" as the basis for the next regulation.
[0158] At the same time, the system records the state of the high impurity rate, and then handles the impurity rate problem after the loss rate problem is solved.
[0159] Based on the priority determination, the system extracts the loss rate high rule set from the knowledge base. In the rule set, there are three possible rules:
[0160] Rule 1: When the loss rate is high and the impurity rate is very low, increase the upper sieve opening;
[0161] Rule 2: When the loss rate is high, the impurity rate is normal, and the moisture content is low, adjust the air distribution plate angle to the parallel position;
[0162] Rule 3: When the loss rate is high, the impurity rate is normal, and the moisture content is normal or high, reduce the fan speed.
[0163] The system first checks the conditions of rule 1: although the loss rate is high (0.32%), the impurity rate is not very low but high (1.05%), so rule 1 does not match.
[0164] Then check the conditions of rule 2: the loss rate is high, but the impurity rate is not normal but high, and the humidity is 19.5% which is not low, so rule 2 does not match.
[0165] Finally, check the conditions of rule 3: the loss rate is high, but the impurity rate is not normal but high, so rule 3 does not completely match.
[0166] At this time, the system detects a special situation: the loss rate and the impurity rate are both high. According to the conflict resolution rules in the knowledge base, a more complex regulation strategy needs to be considered at this time. The system analyzes the recent parameter adjustment history and effect evaluation record, and finds that in similar situations, the combination strategy of appropriately reducing the fan speed and slightly increasing the upper sieve opening has good effect.
[0167] Therefore, the system decides to adopt a combined control strategy: increase the fan speed by 20 r / min (from 780 to 800 r / min), while keeping the angle of the air distribution plate unchanged, and reduce the upper sieve opening by 0.3 mm (from 8.5 mm to 8.2 mm), in order to improve the loss rate and the impurity rate at the same time.
[0168] According to the control decision, the system calculates the specific parameter adjustment amount:
[0169] Fan speed calculation process: current value 780 r / min, target increase 20 r / min. The system checks that the fan speed adjustment step limit is 50 r / min, and the current adjustment value is within the limit range. At the same time, the target value 800 r / min is also within the fan speed range (600 to 900 r / min).
[0170] Upper sieve opening calculation process: current value 8.5 mm, target decrease 0.3 mm. The system checks that the upper sieve opening adjustment step limit is 0.5 mm, and the current adjustment value is within the limit range. At the same time, the target value 8.2 mm is also within the upper sieve opening range (6 to 12 mm).
[0171] The system generates specific parameter adjustment instructions: fan speed adjustment amount +20 r / min, upper sieve opening adjustment amount -0.3 mm, air distribution plate angle adjustment amount 0°. These instructions are sent to the corresponding electric push rod and fan control unit through the cable line for execution.
[0172] After adjusting the parameters, the system waits for 15 seconds for the material to pass through the cleaning system, at which time new monitoring data is collected (see Table 3, data from 35 to 45 seconds). It can be seen that after increasing the fan speed to 800 r / min and reducing the upper sieve opening to 8.2 mm, the cleaning loss rate is basically maintained between 0.32% and 0.33%, but the impurity rate is reduced from 1.45% to 1.10%.
[0173] The system evaluation considers that this control has no obvious improvement on the loss rate, but has certain improvement effect on the impurity rate, and the evaluation result is "partially effective". Therefore, the system decides to further reduce the upper sieve opening to 8.0 mm in the next round of control, so as to further reduce the impurity rate.
[0174] Subsequently, the system returns to the real-time monitoring of the working quality and parameters, continues to monitor the cleaning working quality and parameters, and enters the next round of control cycle.
[0175] The control effect comparison is shown in Table 4:
[0176] Table 4: Comparison of control effect
[0177] Performance index Traditional manual adjustment Method of the present application Improvement effect Average cleaning loss rate (%) 1.85 0.42 Reduced by 77.3% Average impurity content (%) 2.16 0.65 Reduced by 69.9% Parameter adjustment frequency (times / hour) 4.5 12.2 Increased by 171.1% Manual intervention frequency (times / hour) 4.5 0 Reduced by 100% Continuous operation time (hours) 3.8 8.5 Increased by 123.7% Unit area operation efficiency (acre / hour) 5.2 6.8 Improved by 30.8%
[0178] The adaptability of different crops is shown in Table 5:
[0179] Table 5: Different crop condition adaptability table
[0180]
[0181] As can be seen from the data in Tables 4 and 5, the rice-wheat combined harvester cleaning operation intelligent dynamic control method provided by the present application has significant advantages compared to the traditional manual adjustment method. Under the same operating conditions, the average cleaning loss rate and impurity content of the present method are reduced by 77.3% and 69.9%, respectively, the continuous operation time is extended by 123.7%, and the unit area operation efficiency is improved by 30.8%.
[0182] Especially in complex and variable operating environments, such as high humidity conditions or crop lodging conditions, the present method can still maintain a low cleaning loss rate and impurity content, showing good adaptability. The cleaning loss rate and impurity content under all test conditions are controlled within 1%, which is better than the 3% control target proposed in the technical disclosure, fully proving the technical advantages and practical value of the present application.
[0183] The above describes the embodiments of the present application, but the embodiments are not limited to the specific implementation described above, which is only illustrative and not restrictive. Those skilled in the art can make more forms of equivalent embodiments under the inspiration of the embodiments, which are all within the protection of the embodiments.
Claims
1. An intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester, characterized in that: include: Set the initial values of the cleaning device operating parameters according to the operating condition information; Real-time monitoring of cleaning loss rate, cleaning impurity rate and operation parameters; Compare the data obtained from real-time monitoring with the preset threshold to evaluate the status of the cleaning operation; Determine the control priority based on the combination of the assessed cleaning loss rate and cleaning impurity rate; Based on the judged priority and the cleaning job status, the corresponding rule set is matched to generate parameter control decisions; Based on the parameter control decision, the parameter adjustment is calculated and executed, and the control effect is evaluated and then cyclic iterative control is performed.
2. The intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester according to claim 1, characterized in that: The real-time monitoring step includes: Monitoring of the kernel content in the tailings screen discharge by means of a cleaning loss monitoring unit; Monitor the impurity content of grain in the grain tank through the cleaning impurity rate monitoring unit; Monitor the fan speed through the Hall speed sensor; Monitor the angle of the air distributor and the opening of the upper screen through the electric push rod position feedback sensor; The monitoring data is filtered to eliminate noise.
3. The intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester according to claim 1, characterized in that: The step of evaluating the cleaning operation status includes: Comparing the cleaning loss rate and cleaning impurity rate with preset thresholds, wherein the thresholds are divided into a normal range threshold of 0.3% and an extremely high range threshold of 5%; Based on the comparison results, the cleaning loss rate and cleaning impurity rate are divided into normal state, high state and very high state respectively.
4. The intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester according to claim 1, characterized in that: The step of determining the control priority includes: When the cleaning loss rate and cleaning impurity rate are both in normal state, the current parameters are maintained unchanged; When the cleaning loss rate is normal but the impurity rate is high, the high impurity rate rule set is executed; When the cleaning loss rate is high but the impurity rate is normal, the high loss rate rule set is executed; When both the cleaning loss rate and impurity rate are high, the high loss rate rule set is executed first; When any of the cleaning loss rate and impurity rate is at a very high level, the abnormal situation handling rule set is executed.
5. The intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester according to claim 1, characterized in that: The parameter control decision-making step includes: Extract corresponding rules from the knowledge base and match them with the current status in order of priority; When multiple rules are matched successfully, the most appropriate rule is selected through conflict resolution strategy; The conflict resolution strategy is based on the changing trends of the cleaning loss rate and impurity rate and the evaluation results of the control effect executed at the last moment; Generate specific control decisions, including the parameters that need to be adjusted and the adjustment direction.
6. The intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester according to claim 1, characterized in that: The step of calculating and executing parameter adjustment comprises: Calculate the fan speed adjustment, air distributor angle adjustment and upper screen opening adjustment; Adjust the angle of the air distributor and the opening of the upper screen through the electric push rod; Adjust the fan speed through the fan control unit; Wait 10 to 15 seconds after adjusting the parameters to allow the system to reach the new steady state; Evaluate the control effect based on the changing trends of the cleaning loss rate and impurity rate before and after the adjustment; The evaluation results will be used as a reference for the next round of regulatory decisions.
7. The intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester according to claim 1, characterized in that: In the initial setting step of the operation parameters: Operational condition information includes crop type, variety characteristics, maturity, moisture content, and weed content; By retrieving the parameter combination that best matches the current operating conditions from the historical case library as the initial parameters; Normalize the input parameters of different dimensions and map all parameters to the range of 0-1; The categorical variables of crop species and variety characteristics are converted into numerical feature vectors through one-hot encoding.
8. The intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester according to claim 5, characterized in that: The rules include: The treatment rules for high impurity content include reducing the upper screen opening, increasing the fan speed, adjusting the angle of the air distributor and reducing the cleaning load; The treatment rules for high loss rate include increasing the upper screen opening, adjusting the angle of the air distributor, reducing the fan speed and reducing the cleaning load; Abnormal situation handling rules include immediately reducing the cleaning load and first implementing the control decisions in handling situations where the loss rate is too high.
9. The intelligent dynamic control method for cleaning operation of a rice and wheat combine harvester according to claim 5, characterized in that: The regulatory effect evaluation includes: Calculate the difference in the average values of the cleaning loss rate and trash content within a time window before and after the parameter adjustment; Compare the difference value with the expected improvement target and consider the stability of the change trend; Based on the comprehensive scores of improvement degree, stability and adjustment range, the effects are divided into three levels: effective, partially effective and ineffective; The minimum and maximum values of the three indicators of different dimensions, namely, degree of improvement, stability and adjustment range, are normalized.
10. A computer storage medium, characterized in that It comprises a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, it is used to implement the intelligent dynamic control method for the cleaning operation of the rice and wheat combine harvester according to any one of claims 1 to 9.
Citation Information
Patent Citations
Cleaning working condition control method based on angle of air distribution plate, opening degree of louver sieve and rotating speed of fan
CN110347045A
Cleaning fuzzy reasoning device of rice and wheat combine harvester and automatic control method
CN113796198A
Self-adaptive adjustment method and system for working parameters of combine harvester
CN115399140A
Control method and device for harvester, storage medium and processor
CN116491293A
Harvesting control method and device of harvester
CN119014206A
Cited By
Combine harvester intelligent regulation and control method and system based on multi-dimensional weed information perception
CN121621122A
