Bulk material weighing method, system and equipment based on double sensors

Abstract

The invention provides a bulk material weighing method, system and device based on double sensors, and relates to the technical field of electronic scale verification, and the method comprises the steps: weighing a bulk material through double sensors, and obtaining an original weight signal corresponding to each sensor in the double sensors; performing preprocessing and scene filtering on each original weight signal to obtain effective weight data of each sensor on the bulk material; mutual correction analysis is carried out based on each effective weight data, and the data deviation state of the double sensors is judged; and according to the data deviation state, the final weight value of the bulk material is determined. According to the method, the problems of time difference, noise, sensor drifting and the like are effectively eliminated through the technical means of preprocessing, scene filtering, mutual correction analysis and the like, meanwhile, the consistency and the accuracy of data of the double sensors are ensured through mutual correction analysis, and finally, the precision and the stability of bulk material weighing are remarkably improved through comprehensive analysis and optimization processing.

Description

Technical Field

[0001] This invention relates to the field of electronic scale calibration technology, and more specifically, to a method, system, and device for weighing bulk materials based on dual sensors. Background Technology

[0002] In the field of industrial bulk material handling, weighing systems are core equipment for material measurement, cost control, and production efficiency management. Currently, the mainstream bulk material weighing solutions in the industry mainly rely on a dual-sensor architecture. The dual-sensor architecture has a certain degree of redundancy and data comparison capability. It typically collects analog signals of material weight through sensors, converts them from analog to digital signals, and then performs data processing and weight calculation.

[0003] In related technologies, existing dual-sensor systems, when weighing bulk materials, often use independent clocks to trigger data acquisition, which can easily lead to time differences during the acquisition process. This causes the weight of the material at the same moment in a dynamic scenario to be split into two sets of sampling data, directly causing measurement deviations. At the same time, the instantaneous spike noise generated by the impact of the material during weighing, as well as the sensor drift caused by dust and temperature changes in the industrial environment, further aggravate data fluctuations, resulting in a large difference between the weighing result and the actual weight of the bulk material, thus affecting the weighing effect. Summary of the Invention

[0004] The problem addressed by this invention is how to improve the weighing accuracy based on dual sensors.

[0005] To address the aforementioned problems, this invention provides a method, system, and device for weighing bulk materials based on dual sensors.

[0006] In a first aspect, the present invention provides a bulk material weighing method based on dual sensors, comprising: The bulk material is weighed using dual sensors to obtain the original weight signal corresponding to each of the dual sensors; Each of the original weight signals is preprocessed and scene-specific filtered to obtain the effective weight data of the bulk material for each sensor. Based on each of the effective weight data, a cross-calibration analysis is performed to determine the data deviation status of the dual sensors; Based on the data deviation, the final weight value of the bulk material is determined.

[0007] Optionally, the step of weighing the bulk material using dual sensors to obtain the original weight signal corresponding to each of the dual sensors includes: Each sensor is controlled by a hardware synchronous trigger signal to collect data according to a preset acquisition cycle, thereby obtaining the analog weight signal corresponding to the sensor. Wherein, the time difference between the acquisition of the two sensors is less than a preset duration; The analog weight signal output by each sensor is synchronously converted into a digital signal to obtain a first original weight signal and a second original weight signal.

[0008] Optionally, the dual sensors include a first sensor and a second sensor, wherein the weighing accuracy of the first sensor is higher than that of the second sensor, the measuring range of the first sensor is smaller than that of the second sensor, the first sensor obtains the first original weight signal, and the second sensor obtains the second original weight signal.

[0009] Optionally, the preprocessing and scene-specific filtering of each of the original weight signals to obtain the effective weight data of the bulk material for each sensor includes: Anomaly detection is performed on the first original weight signal and the second original weight signal respectively to obtain extreme data in the first original weight signal and the second original weight signal that exceed the preset signal interval; The extreme data is removed and replaced with the preceding valid data corresponding to the extreme data to obtain the preprocessed first original weight signal and the preprocessed second original weight signal. The fluctuation of the preprocessed first original weight signal and the preprocessed second original weight signal is calculated by continuously collecting data a preset number of times to obtain the fluctuation index corresponding to the preprocessed first original weight signal and the preprocessed second original weight signal respectively. The weighing status is determined based on the fluctuation index. If the weighing state is a static weighing state, a moving average filtering algorithm is used to filter the preprocessed first original weight signal and the preprocessed second original weight signal to obtain the first effective weight data and the second effective weight data corresponding to the preprocessed first original weight signal and the preprocessed second original weight signal, respectively. If the weighing state is a dynamic weighing state, an adaptive filtering algorithm is used to perform real-time filtering on the preprocessed first original weight signal and the preprocessed second original weight signal to obtain the first effective weight data and the second effective weight data corresponding to the preprocessed first original weight signal and the preprocessed second original weight signal, respectively.

[0010] Optionally, determining the weighing status based on the fluctuation index includes: Compare the volatility index with the preset volatility threshold; If the fluctuation index is less than or equal to the preset fluctuation threshold and continues for a preset duration, then the weighing state is determined to be a static weighing state. If the fluctuation index is greater than the preset fluctuation threshold, or if the fluctuation index is less than or equal to the preset fluctuation threshold but the duration does not reach the preset duration, then the weighing state is determined to be a dynamic weighing state.

[0011] Optionally, the step of performing cross-calibration analysis based on each of the effective weight data to determine the data deviation status of the dual sensors includes: The absolute deviation is determined based on the absolute value of the difference between the first effective weight data and the second effective weight data; The relative deviation rate is determined based on the ratio of the absolute deviation to the average of the first effective weight data and the second effective weight data; The absolute deviation and the relative deviation rate are compared with preset absolute deviation thresholds and preset relative deviation rate thresholds, respectively. If the absolute deviation is less than or equal to the preset absolute deviation threshold and the relative deviation rate is less than or equal to the preset relative deviation rate threshold, then the data deviation state of the dual sensors is determined to be normal. If the absolute deviation is greater than the preset absolute deviation threshold or the relative deviation rate is greater than the preset relative deviation rate threshold, then the number of deviations exceeding the limit is accumulated, and the data deviation state is determined to be either an accidental deviation state or a continuous fault state based on the number of deviations exceeding the limit.

[0012] Optionally, determining whether the data deviation state is an accidental deviation state or a continuous fault state based on the number of deviations exceeding the tolerance includes: If the number of deviations exceeding the limit is less than the preset threshold for the number of deviations exceeding the limit, then the data deviation state is determined to be the random deviation state. If the number of deviations exceeding the limit is greater than or equal to the preset threshold for the number of deviations exceeding the limit, then the data deviation state is determined to be the continuous fault state.

[0013] Optionally, determining the final weight value of the bulk material based on the data deviation status includes: If the data deviation status is normal, then determine whether the accuracy of the two sensors is consistent; If the accuracy of the two sensors is consistent, the average value of the effective weight data of the two sensors shall be taken as the final weight value of the bulk material. If the accuracy of the two sensors is different, the weight corresponding to the effective weight data of the sensors is determined according to the accuracy of the sensors, and the final weight value of the bulk material is obtained by weighted average calculation based on the weight. The value of the weight is positively correlated with the accuracy of the sensors. If the data deviation state is an accidental deviation state, then the variance of the effective weight data of each sensor within a preset time period is calculated respectively; The effective weight data with the smallest variance is taken as the final weight value of the bulk material; If the data deviation state is a continuous fault state, compare the degree of deviation between the effective weight data of each sensor and the historical normal weight data. The effective weight data that deviates from the historical normal weight data within a preset allowable range is taken as the final weight value of the bulk material.

[0014] Secondly, the present invention provides a bulk material weighing system based on dual sensors, comprising: The signal acquisition unit is used to weigh the bulk material through dual sensors and obtain the original weight signal corresponding to each of the dual sensors; The signal processing unit is used to preprocess and scene-specific filter each of the original weight signals to obtain the effective weight data of each sensor for the bulk material. A data verification unit is used to perform cross-calibration analysis based on each of the effective weight data to determine the data deviation status of the dual sensors. The weight determination unit is used to determine the final weight value of the bulk material based on the data deviation status.

[0015] Thirdly, the present invention provides a bulk material weighing device, including a memory and a processor; The memory is used to store computer programs; The processor is configured to implement the above-described dual-sensor-based bulk material weighing method when executing the computer program.

[0016] The present invention relates to a dual-sensor-based bulk material weighing method, system, and device. By introducing preprocessing and scenario-based filtering steps, the raw weight signals from each sensor are processed, enabling time alignment and noise suppression. This eliminates weight splitting issues caused by time differences, ensuring accurate acquisition and calculation of material weight at the same moment. Scenario-based filtering technology is used to design corresponding filtering algorithms for different working conditions (such as material impact and environmental interference), effectively suppressing instantaneous spike noise and interference caused by sensor drift. The preprocessing step further optimizes signal quality, ensuring data stability and reliability, thereby significantly improving weighing accuracy. Through a cross-calibration analysis step, the effective weight data from each sensor is compared and verified. By monitoring the data deviation between the two sensors in real time and dynamically adjusting the weights or selecting more reliable data based on the deviation, the final determined bulk material weight value has higher accuracy and reliability.

[0017] In summary, this invention effectively eliminates problems such as time difference, noise, and sensor drift through preprocessing, scene-based filtering, and cross-calibration analysis. At the same time, cross-calibration analysis ensures the consistency and accuracy of dual-sensor data. Finally, through comprehensive analysis and optimization, this invention solves the accuracy problem existing in the current dual-sensor weighing system, significantly improves the accuracy and stability of bulk material weighing, and meets the demand for high-precision weighing in industrial production. Attached Figure Description

[0018] Figure 1 This is a flowchart of a dual-sensor-based bulk material weighing method according to an embodiment of the present invention. Figure 2 This is a structural block diagram of a bulk material weighing system based on dual sensors according to an embodiment of the present invention; Figure 3 This is a structural block diagram of a bulk material weighing device according to an embodiment of the present invention. Detailed Implementation

[0019] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the accompanying drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0020] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0021] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to"; the term "based on" means "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; and the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first," "second," etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0022] It should be noted that the terms "a" and "a plurality of" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0023] Combination Figure 1 As shown in the figure, an embodiment of the present invention provides a bulk material weighing method based on dual sensors, comprising: The bulk material is weighed using dual sensors to obtain the original weight signal corresponding to each of the dual sensors.

[0024] Specifically, when starting the weighing operation, the dual load cells are first activated to ensure they are operational. The AD acquisition module is then initialized, configuring its sampling rate, resolution, and other parameters. A hardware synchronization signal (such as a synchronization clock signal) ensures the time difference between the two sensors' acquisitions is controlled within 1ms. This avoids weighing discrepancies caused by time differences during the dynamic flow of bulk materials, ensuring the comparability of the acquired weight data.

[0025] In industrial bulk material weighing scenarios, dual sensors are installed at key locations on the weighing equipment (such as the support points of a belt scale or the suspension points of a hopper) to sense real-time weight changes in the bulk material. These two sensors operate independently, each acquiring the weight signal of the bulk material. During each weighing, the sensors convert the sensed weight into electrical signals, which are the raw weight signals, denoted as W1_raw and W2_raw respectively. These signals contain the actual weight information of the bulk material. However, this embodiment considers that these signals may be affected by noise, interference, or the inherent characteristics of the sensors, therefore further processing is required.

[0026] Each of the original weight signals is preprocessed and scene-specific filtered to obtain the effective weight data of the bulk material for each sensor.

[0027] Specifically, firstly, each raw weight signal is amplified and filtered to remove high-frequency noise and interference. For example, a low-pass filter can effectively remove instantaneous spike noise caused by the impact of bulk materials. Next, an appropriate filtering algorithm is selected based on the weighing scenario (e.g., static or dynamic weighing). For static weighing, a moving average filtering algorithm is used to average the raw weight signals acquired multiple times consecutively, thus smoothing data fluctuations. For example, the average values ​​of W1_raw and W2_raw from five consecutive acquisitions are calculated to obtain W1_filtered and W2_filtered. These filtered data are the effective weight data, which can more accurately reflect the actual weight of the bulk material.

[0028] Based on the mutual calibration analysis of each of the effective weight data, the data deviation status of the dual sensors is determined.

[0029] Specifically, cross-calibration analysis determines the consistency and deviation of data by comparing the effective weight data of two sensors. First, the absolute deviation ΔW = |W1_filtered - W2_filtered| and the relative deviation rate R = (ΔW / W_avg) × 100% (where W_avg = (W1_filtered + W2_filtered) / 2) of the effective weight data from the two sensors are calculated. Then, based on preset deviation thresholds (such as the absolute deviation threshold ΔW_max and the relative deviation rate threshold R_max), it is determined whether the deviation is within the allowable range. If ΔW ≤ ΔW_max and R ≤ R_max, the data from the two sensors are considered consistent; otherwise, a data deviation is identified. Furthermore, this embodiment can further determine the persistence of the deviation by accumulating the number of deviation exceedances, thereby distinguishing between random and persistent deviations.

[0030] Based on the data deviation, the final weight value of the bulk material is determined.

[0031] Specifically, based on the results of the cross-calibration analysis, the final weight value of the bulk material is determined. If the data from the two sensors are consistent (i.e., the deviation is within the allowable range), the average of the effective weight data from the two sensors can be directly taken as the final weight value, i.e., W_final = W_avg. If the two sensors have different accuracies, weights can be assigned according to the sensor's accuracy, such as using a weighted average algorithm to calculate the final weight value. If data deviation exists, different processing strategies can be adopted according to the type of deviation. For random deviations, the sensor data with smaller variance can be selected as the final weight value; for persistent deviations, further investigation of sensor malfunctions is required, and the data from the normal sensor should be used as the final weight value, while issuing a fault alarm.

[0032] This embodiment of the bulk material weighing method based on dual sensors introduces preprocessing and scenario-based filtering steps to process the raw weight signals from each sensor. This enables time alignment and noise suppression, eliminating weight splitting issues caused by time differences and ensuring accurate acquisition and calculation of material weight at the same moment. Scenario-based filtering technology is used to design corresponding filtering algorithms for different working conditions (such as material impact and environmental interference), effectively suppressing instantaneous spike noise and interference caused by sensor drift. The preprocessing step further optimizes signal quality, ensuring data stability and reliability, thereby significantly improving weighing accuracy. Through a cross-calibration analysis step, the effective weight data from each sensor is compared and verified. By monitoring the data deviation between the two sensors in real time and dynamically adjusting the weights or selecting more reliable data based on the deviation, the final determined bulk material weight value has higher accuracy and reliability.

[0033] In summary, this embodiment effectively eliminates problems such as time difference, noise, and sensor drift by using techniques such as preprocessing, scene-based filtering, and cross-calibration analysis. At the same time, cross-calibration analysis ensures the consistency and accuracy of dual-sensor data. Finally, through comprehensive analysis and optimization, this embodiment solves the accuracy problem existing in the current dual-sensor weighing system, significantly improves the accuracy and stability of bulk material weighing, and meets the demand for high-precision weighing in industrial production.

[0034] Optionally, the step of weighing the bulk material using dual sensors to obtain the original weight signal corresponding to each of the dual sensors includes: Each sensor is controlled by a hardware synchronous trigger signal to collect data according to a preset acquisition cycle, thereby obtaining the analog weight signal corresponding to the sensor. Wherein, the time difference between the acquisition of the two sensors is less than a preset duration; The analog weight signal output by each sensor is synchronously converted into a digital signal to obtain a first original weight signal and a second original weight signal.

[0035] Specifically, to ensure that the two sensors can synchronously collect data during the weighing process, this embodiment uses a hardware synchronization trigger signal to control the data acquisition process of each sensor. This synchronization trigger signal can be a precise clock signal, which simultaneously triggers both sensors to start collecting data at the same time. The preset acquisition period refers to the time interval between sensor data acquisitions, for example, it can be set to collect data 10 times per second. In this way, each sensor will collect the weight signal of the bulk material at the same time and convert it into an analog signal output. This analog signal contains the weight information of the bulk material, but it is still a continuous electrical signal and needs to be further converted into a digital signal for subsequent processing.

[0036] Meanwhile, to ensure weighing accuracy, the time difference between sensor acquisitions must be controlled within a very small range. The preset duration is typically a very short time interval, such as 1 millisecond (ms). Precise control of the hardware synchronization trigger signal ensures that the acquisition times of the two sensors are almost perfectly synchronized. For example, if sensor 1 acquires data at a certain moment, the acquisition time of sensor 2 will not differ from it by more than 1 ms. This precision in time synchronization is particularly important in dynamic weighing scenarios because bulk materials may move rapidly during transport. A large time difference can cause the weight of the material at the same moment to be split into two sets of sampling data, leading to measurement deviations. This embodiment effectively avoids such deviations by controlling the acquisition time difference within the preset duration, ensuring the comparability and accuracy of the acquired weight data.

[0037] Furthermore, the analog weight signal output by the sensor is converted into a digital signal by an analog-to-digital converter (ADC) for subsequent digital signal processing. The ADC converts a continuous analog signal into a discrete digital signal. During the conversion process, it is ensured that the analog signals from both sensors are converted into digital signals synchronously, meaning that the ADC conversion processes of the two sensors are completed at the same time or very close to each other, thus guaranteeing the time synchronization between the digital signals. The converted digital signals are the first original weight signal W1_raw and the second original weight signal W2_raw. These digital signals contain the weight information of the bulk material, but also contain noise and other interference. Therefore, this embodiment further performs preprocessing and filtering operations to extract the effective weight data.

[0038] In this optional embodiment, a hardware synchronous trigger signal controls the dual sensors to collect data according to a preset acquisition cycle, ensuring that the acquisition time difference between the sensors is less than a preset duration (e.g., 1ms). This high-precision synchronous acquisition method effectively avoids measurement deviations caused by time differences, especially in dynamic scenarios, where the weight of material at the same moment is accurately recorded in two sets of sampling data, thus significantly improving weighing accuracy. In specific applications, such as in the weighing of industrial bulk materials, the dynamic flow of materials may lead to inaccurate weight data. This embodiment, through a hardware synchronous trigger signal, ensures that the two sensors acquire data at almost the same moment, thereby avoiding the problem of the weight of material at the same moment being split into two sets of sampling data due to time differences. This synchronous acquisition method can effectively reduce measurement errors in dynamic scenarios and improve the reliability of weighing results. The synchronously acquired digital signals have a high degree of time synchronization, which allows the data acquired by the two sensors to be directly compared and cross-calibrated, ensuring data consistency and accuracy, thereby improving the reliability of the entire weighing system. Controlling the acquisition process through a hardware synchronous trigger signal can reduce acquisition time deviations caused by external interference (such as electromagnetic interference, environmental noise, etc.). The synchronous acquisition mechanism ensures that both sensors collect data under identical conditions, thereby enhancing the system's anti-interference capability and reducing weight data fluctuations caused by interference. Simultaneously, controlling the acquisition process through hardware synchronous trigger signals reduces acquisition time deviations caused by sensor malfunctions or external interference. This synchronization mechanism enhances the system's robustness, ensuring high reliability and comparability of the acquired data even in cases of sensor failure or significant external interference, thus improving the overall stability and reliability of the weighing system.

[0039] This synchronous acquisition mechanism is applicable not only to static weighing scenarios but also to dynamic weighing scenarios. In dynamic weighing scenarios, the rapid flow and impact of materials may cause fluctuations in weight data. Synchronous acquisition can effectively reduce such fluctuations and ensure that the acquired data has high accuracy and stability.

[0040] Optionally, the dual sensors include a first sensor and a second sensor, wherein the weighing accuracy of the first sensor is higher than that of the second sensor, the measuring range of the first sensor is smaller than that of the second sensor, the first sensor obtains the first original weight signal, and the second sensor obtains the second original weight signal.

[0041] Specifically, the dual-sensor system includes two types of sensors. The range of the first sensor is smaller than that of the second sensor, and the weighing accuracy of the first sensor is higher than that of the second sensor. In this embodiment, the first sensor is typically a high-precision sensor, which has high measurement accuracy and can accurately measure the weight of bulk materials within a small weight range. It is mainly used for weighing small weight segments, such as in chemical raw material packaging and food processing, where high-precision measurement of small quantities of materials is required. The output signal of the high-precision sensor is denoted as the first raw weight signal W1_raw and W2_raw. The second sensor is typically a large-range sensor, which has a large measurement range but relatively low accuracy. It is mainly used for weighing large weight segments, such as in bulk material transportation and grain silo metering, where the weight of large quantities of materials needs to be measured. The output signal of the large-range sensor is denoted as the second raw weight signal W2_raw.

[0042] The dual-sensor configuration in this embodiment can cover different weighing requirements. For example, in some industrial scenarios, it is necessary to handle both small and large weights of bulk materials simultaneously. For instance, in batch metering of grain silos, there are both small batches of additives (requiring high-precision measurement) and large batches of grains (requiring large-range measurement). By using both high-precision and large-range sensors, these two requirements can be met simultaneously. In static scenarios such as tank weighing, high-precision sensors can be used to handle small weights or batch dispensing, while large-range sensors can adapt to large weight measurements. This configuration ensures accurate weight data is obtained throughout the weighing process across different weight ranges. Furthermore, in fields such as chemical and food processing, the reliability of the weighing system is crucial. The dual-sensor design allows for rapid switching to the other sensor if one sensor fails, ensuring the continuity of the weighing process.

[0043] In this optional embodiment, the high-precision sensor accurately measures small-weight bulk materials, ensuring high accuracy in small-volume weighing, while the large-range sensor handles large-weight bulk materials, preventing damage to the high-precision sensor due to overload. By combining the high-precision sensor and the large-range sensor, the system can cover the full range of weighing needs from small to large weights. This configuration allows the system to adapt to various industrial scenarios, improving its versatility and applicability. In dynamic weighing scenarios, the combination of the high-precision sensor and the large-range sensor effectively copes with the rapid flow and impact of bulk materials, ensuring the stability and accuracy of weighing data. Furthermore, the redundant design of the dual sensors allows for rapid switching to the other sensor when one sensor fails, ensuring the continuity of the weighing process. This design reduces downtime caused by sensor failures and improves production efficiency. By comparing the data collected by the two sensors, the faulty sensor can be quickly located, enabling timely maintenance and repair. This fault diagnosis capability further improves the system's reliability and maintenance efficiency.

[0044] For example, in the chemical raw material packaging process, it is necessary to accurately control the weight of each batch of raw materials. High-precision sensors are used to measure small doses of raw materials (such as additives) to ensure the accuracy of packaging, while large-range sensors are used to measure the total weight of transport vehicles to ensure the safety and accuracy of the transportation process. In grain warehouse batch metering, there are both small batches of additives (requiring high-precision measurement) and large batches of grains (requiring large-range measurement). By using both high-precision and large-range sensors, these two needs can be met simultaneously, ensuring the accuracy and efficiency of metering. By configuring high-precision and large-range sensors, this embodiment can effectively improve the accuracy, adaptability, and reliability of the weighing system. This dual-sensor configuration can not only cover the full range of weighing needs from small to large weights, but also improve the system's stability and maintenance efficiency through redundant design and cross-calibration analysis.

[0045] Optionally, the preprocessing and scene-specific filtering of each of the original weight signals to obtain the effective weight data of the bulk material for each sensor includes: Anomaly detection is performed on the first original weight signal and the second original weight signal respectively to obtain extreme data in the first original weight signal and the second original weight signal that exceed the preset signal interval; The extreme data is removed and replaced with the preceding valid data corresponding to the extreme data to obtain the preprocessed first original weight signal and the preprocessed second original weight signal. The fluctuation of the preprocessed first original weight signal and the preprocessed second original weight signal is calculated by continuously collecting data a preset number of times to obtain the fluctuation index corresponding to the preprocessed first original weight signal and the preprocessed second original weight signal respectively. The weighing status is determined based on the fluctuation index. If the weighing state is a static weighing state, a moving average filtering algorithm is used to filter the preprocessed first original weight signal and the preprocessed second original weight signal to obtain the first effective weight data and the second effective weight data corresponding to the preprocessed first original weight signal and the preprocessed second original weight signal, respectively. If the weighing state is a dynamic weighing state, an adaptive filtering algorithm is used to perform real-time filtering on the preprocessed first original weight signal and the preprocessed second original weight signal to obtain the first effective weight data and the second effective weight data corresponding to the preprocessed first original weight signal and the preprocessed second original weight signal, respectively.

[0046] Specifically, when processing the first and second raw weight signals acquired by the dual sensors, invalid interference data is first eliminated, the weighing scenario is accurately determined, and then a corresponding filtering algorithm is matched to finally obtain effective data that truly reflects the weight of the bulk material. In detail: Step 1: Outlier detection. Identify extreme interference data in the signal. By setting a preset confidence interval (which is usually determined based on the sensor range, the fluctuation range of historical normal data, and the accuracy requirements of bulk material weighing; for example, in the case of granular material weighing, the confidence interval can be set to ±3 times the standard deviation of the normal weight fluctuation range), data in the first and second original weight signals that exceed this interval are marked as extreme data. Such data are mostly caused by instantaneous spike noise generated by material impact, sudden sensor drift, or external electromagnetic interference. This first step is to process the data to avoid affecting the measurement accuracy in the future.

[0047] Step 2: Extreme data processing. This embodiment uses a combination of "removal + replacement" to ensure data continuity. First, the extreme data identified in the first step is removed from the original signal sequence to avoid interfering with subsequent calculations. Then, using the preceding valid data corresponding to the extreme data (i.e., 3-5 sets of continuous and stable normal data before the occurrence of the extreme data; the specific number can be adjusted according to the sampling frequency, such as taking the first 5 sets of data when the sampling frequency is 10Hz) as a benchmark, the data at the removal position is replaced by mean filling or linear interpolation. Finally, the preprocessed first and second original weight signals are obtained, which not only eliminates abnormal interference but also ensures the integrity of the signal sequence, providing a reliable data foundation for subsequent fluctuation calculations.

[0048] Step 3: Quantify signal stability through fluctuation calculation to provide a basis for judging the weighing status. For the two sets of preprocessed signals, select a preset number of consecutive data acquisitions (the preset number of acquisitions is dynamically adjusted according to the weighing scenario; it can be set to 20-30 times in a static scenario and 10-15 times in a dynamic scenario due to the fast material flow). Calculate the fluctuation index of this set of data. The fluctuation index is usually expressed by variance, range, or coefficient of variation. For example, by calculating the variance of 15 consecutive data acquisitions, the smaller the variance, the more stable the signal; conversely, the larger the variance, the more violent the signal fluctuation.

[0049] Step 4: Determine the weighing status based on the fluctuation index and set a fluctuation index threshold (e.g., set the threshold to 0.5 for a static scenario). The dynamic scene threshold is set to 2. (Specific values ​​can be calibrated according to sensor accuracy and material characteristics). When the fluctuation index is below the threshold, it is determined to be a static weighing state (such as bulk material temporary storage and static metering scenarios, where the material does not flow significantly and the weight signal is stable). When the fluctuation index is above the threshold, it is determined to be a dynamic weighing state (such as bulk material conveying and continuous feeding scenarios, where the material flows continuously and the weight signal fluctuates significantly with the movement of the material).

[0050] Step 5: Match differentiated filtering algorithms to different weighing states to achieve scenario-based accurate filtering. For static weighing states, a moving average filtering algorithm is used. This algorithm selects a sliding window of fixed length (e.g., the window size is set to 5-10 sets of data), and takes the average of the preprocessed signal data within the window as the output. This can smooth out small random noise in static scenarios (such as slight sensor drift caused by dust in industrial environments) without affecting the real-time performance of static measurement due to filtering delay, ultimately obtaining the first and second effective weight data. For dynamic weighing states, an adaptive filtering algorithm is used. This algorithm can adjust the filtering parameters in real time according to signal fluctuations (e.g., dynamically adjust the filtering coefficient; decrease the coefficient when the signal fluctuation is severe to quickly track weight changes, and increase the coefficient when the signal fluctuation is gentle to enhance the filtering effect). This avoids the problems of excessive smoothing leading to weight lag or insufficient filtering and residual noise that occur in traditional fixed parameter filtering in dynamic scenarios, thus outputting the first and second effective weight data that accurately reflect the weight of dynamic materials in real time.

[0051] In this optional embodiment, a closed-loop operation of outlier handling, fluctuation analysis, and scenario-based filtering is employed. First, a pre-set signal interval is used to accurately locate and eliminate extreme data caused by material impact spikes and sensor drift. Then, preceding valid data is used to ensure the integrity of the signal sequence, cutting off the impact of invalid interference on measurement from the data source. Next, by calculating the fluctuation index of continuously collected data for a preset number of times, accurate distinction is achieved between static (e.g., bulk material temporary storage weighing) and dynamic (e.g., bulk material conveying weighing) scenarios, avoiding the problem of traditional single filtering algorithms failing to adapt to different scenarios. Finally, for static scenarios, a moving average filter is used to smooth out small random noise, and for dynamic scenarios, an adaptive filter is used to adjust parameters in real time to track weight changes. This embodiment solves the accuracy deviation caused by sensor drift in static weighing and avoids the weight lag or noise residue problems caused by fixed filtering parameters in dynamic weighing. Ultimately, it significantly improves the accuracy and stability of the effective weight data output by dual sensors, effectively improving the measurement deviation problem caused by data instability in existing dual-sensor weighing technology.

[0052] Optionally, determining the weighing status based on the fluctuation index includes: Compare the volatility index with the preset volatility threshold; If the fluctuation index is less than or equal to the preset fluctuation threshold and continues for a preset duration, then the weighing state is determined to be a static weighing state. If the fluctuation index is greater than the preset fluctuation threshold, or if the fluctuation index is less than or equal to the preset fluctuation threshold but the duration does not reach the preset duration, then the weighing state is determined to be a dynamic weighing state.

[0053] Specifically, the preset fluctuation threshold in this embodiment is determined comprehensively based on the sensor accuracy level, the physical characteristics of the bulk material, and the weighing accuracy requirements. For example, for weighing scenarios involving granular materials, the threshold can be set to... For lumpy materials, the increase can be appropriately increased to The essence of the preset fluctuation threshold is to define the quantitative boundary of signal stability; the preset duration is set according to the dynamic characteristics of the weighing scenario. For example, it can be set to 1-2 seconds in a static scenario and shortened to 0.5 seconds in a dynamic scenario, which is used to distinguish between the real stable state and the instantaneous fluctuation.

[0054] In practical judgment, this embodiment adopts a dual judgment standard of fluctuation index value + duration. Specifically, when the calculated fluctuation index is less than or equal to the preset fluctuation threshold and the duration of the state reaches the preset duration, it is judged as a static weighing state. This logic can effectively exclude occasional short-term stable signals that occur during dynamic processes, such as the instantaneous pause in the material conveying gap, ensuring the reliability of static judgment. It is suitable for scenarios where the weight signal is stable for a long time, such as bulk material static storage and batch metering. When the fluctuation index is greater than the preset fluctuation threshold, it is directly judged as a dynamic weighing state, which corresponds to scenarios where the material is continuously flowing and the weight signal changes in real time with the conveying process, such as belt conveyor feeding and screw feeder feeding. For cases where the fluctuation index does not exceed the threshold but the duration does not reach the preset duration, such as short-term signal stability caused by a sudden decrease in material flow during dynamic weighing, it is still judged as a dynamic weighing state. This can effectively avoid misjudging a short-term stable state as static and using an incompatible filtering algorithm.

[0055] In this optional embodiment, the dual verification mechanism of numerical value and time can accurately capture the real static weighing state and reliably identify various dynamic scenarios, such as continuous fluctuations, short-term stabilization followed by fluctuations, and other complex situations. This provides an accurate basis for switching subsequent scenario-based filtering algorithms, avoids the decline in filtering effect due to misjudgment of the state, and further ensures the reliability of effective weight data under different weighing scenarios.

[0056] Optionally, the step of performing cross-calibration analysis based on each of the effective weight data to determine the data deviation status of the dual sensors includes: The absolute deviation is determined based on the absolute value of the difference between the first effective weight data and the second effective weight data; The relative deviation rate is determined based on the ratio of the absolute deviation to the average of the first effective weight data and the second effective weight data; The absolute deviation and the relative deviation rate are compared with preset absolute deviation thresholds and preset relative deviation rate thresholds, respectively. If the absolute deviation is less than or equal to the preset absolute deviation threshold and the relative deviation rate is less than or equal to the preset relative deviation rate threshold, then the data deviation state of the dual sensors is determined to be normal. If the absolute deviation is greater than the preset absolute deviation threshold or the relative deviation rate is greater than the preset relative deviation rate threshold, then the number of deviations exceeding the limit is accumulated, and the data deviation state is determined to be either an accidental deviation state or a continuous fault state based on the number of deviations exceeding the limit.

[0057] Specifically, in this embodiment, when performing cross-calibration analysis based on the effective weight data of the two sensors to determine the data deviation status, the consistency of the sensor data is accurately determined by using dual quantitative indicators of absolute deviation and relative deviation rate, combined with the cumulative statistics of the number of deviations exceeding the tolerance. Specifically: First, the first step is to calculate the absolute deviation, which is obtained by taking the absolute value of the difference between two valid weight data. This indicator can intuitively reflect the actual numerical difference between the weighing results of the two sensors. For example, when the first valid weight data is 100.2kg and the second is 99.7kg, the absolute deviation is 0.5kg, which can directly reflect the "magnitude" of the data deviation. It is suitable for scenarios with high requirements for the accuracy of the absolute weight value, such as weighing bulk materials in batches with a fixed range.

[0058] The second step is to calculate the relative deviation rate. This is done by dividing the absolute deviation by the average of the two valid weight data points. In the example above, the average of the two data points is 99.95 kg, and the relative deviation rate is approximately 0.5%. This indicator can eliminate the influence of the absolute weight value on the deviation judgment. For example, the same absolute deviation of 0.5 kg has a relative deviation rate of only 0.5% in a 100 kg weighing scenario, but reaches 5% in a 10 kg weighing scenario. The relative deviation rate can more objectively reflect the severity of the deviation, and is especially suitable for dynamic weighing scenarios with large fluctuations in the weight of bulk materials, such as continuous weighing from 50 kg to 200 kg in belt conveyors.

[0059] In a preferred embodiment of the present invention, when setting the judgment criteria, the preset absolute deviation threshold and the preset relative deviation rate threshold are determined by combining the sensor accuracy level, the industry standard for bulk material weighing (such as grain weighing accuracy ≤0.1%, ore weighing accuracy ≤0.5%), and actual application requirements. For example, for high-precision chemical raw material weighing, the preset absolute deviation threshold can be set to 0.2kg and the preset relative deviation rate threshold can be set to 0.1%; for ordinary building sand and gravel weighing, the absolute deviation threshold can be relaxed to 0.5kg and the relative deviation rate threshold to 0.3%, ensuring that the threshold setting matches the actual accuracy requirements.

[0060] When entering the deviation status determination stage, this embodiment handles two cases: The first is when both thresholds are met, i.e., absolute deviation ≤ preset absolute deviation threshold and relative deviation rate ≤ preset relative deviation rate threshold. In this case, the deviation status of the dual sensor data is determined to be normal, indicating good consistency between the data collected by the two sensors. Subsequent weight measurement can be directly based on the valid data, such as taking the average of the two data as the final weighing result. The second is when either threshold is not met, i.e., absolute deviation > preset absolute deviation threshold or relative deviation rate > preset relative deviation rate threshold. In this case, the cumulative statistics of deviation exceeding the tolerance are initiated, for example, setting "3 consecutive times". "Out of tolerance" is the judgment threshold. If there is only one instance of out of tolerance, such as a single data deviation caused by instantaneous airflow interference, it is judged as an accidental deviation state. It can be verified by cross-calibrating with subsequent data, such as taking the next set of valid data for re-analysis, to avoid misjudging the sensor failure due to accidental interference. If the number of out-of-tolerance instances reaches or exceeds the set limit, such as three consecutive out-of-tolerance instances, it is judged as a continuous failure state, indicating that at least one sensor has problems such as drift or decreased sensitivity, such as dust adhesion causing sensor reading deviation. At this time, the alarm mechanism is triggered, such as on-site audible and visual alarms or remote system prompts, to remind staff to calibrate or repair the sensor.

[0061] In this optional embodiment, the judgment logic of dual deviation index + out-of-tolerance count can avoid the limitations of single index judgment, such as only looking at the absolute deviation and ignoring the severity of deviation in small weight scenarios, or only looking at the relative deviation rate and ignoring the actual numerical deviation in large weight scenarios. It can also distinguish between accidental interference and continuous faults, which can not only ensure the accuracy of data mutual calibration in normal weighing scenarios, but also promptly identify sensor abnormalities, providing dual protection for the accuracy, reliability and timeliness of operation and maintenance of the dual sensor weighing system.

[0062] Optionally, determining whether the data deviation state is an accidental deviation state or a continuous fault state based on the number of deviations exceeding the tolerance includes: If the number of deviations exceeding the limit is less than the preset threshold for the number of deviations exceeding the limit, then the data deviation state is determined to be the random deviation state. If the number of deviations exceeding the limit is greater than or equal to the preset threshold for the number of deviations exceeding the limit, then the data deviation state is determined to be the continuous fault state.

[0063] Specifically, the basis for setting the preset threshold for the number of deviations exceeding the limit is first clarified. In this embodiment, the threshold is determined by comprehensively considering historical sensor fault data, the interference frequency of the weighing scenario, and the system's fault tolerance requirements. For example, in refined weighing scenarios with less dust and stable material flow, such as weighing pharmaceutical raw materials, the probability of accidental interference such as instantaneous airflow or slight vibration is low, so the threshold can be set to 2 times to avoid triggering unnecessary maintenance due to minor interference. In industrial scenarios with more dust and frequent material impact, such as weighing ore conveying, accidental interference is more common. In this case, the threshold can be appropriately increased to 3-4 times to prevent misjudgment of sensor faults. At the same time, the threshold setting is also matched with the data acquisition frequency. If the sampling frequency is 10Hz, that is, 10 sets of data are collected per second, then 3 deviations can correspond to continuous anomalies within 0.3 seconds, which can both capture continuous faults in a timely manner and avoid misjudgment due to a single instantaneous interference.

[0064] In the actual judgment process, when the number of deviations exceeding the limit is less than the preset deviation exceeding the limit, it is judged as an accidental deviation state. This type of situation is mostly caused by short-term, non-continuous interference, such as a single data deviation caused by a sudden jamming of material particles during dynamic weighing, or two consecutive deviations caused by temporary electromagnetic interference in the industrial environment. At this time, the sensor itself is not faulty, and there is no need to stop the machine for maintenance. The system can continue to weigh normally, and subsequent data cross-calibration can be performed, such as taking the next set of valid weight data for re-analysis, to verify whether the deviation has disappeared, so as to avoid the impact of accidental interference on weighing efficiency. When the number of deviations exceeding the limit is greater than or equal to the preset deviation exceeding the limit, it is judged as a continuous fault state, that is, the deviation is not an accidental interference, but a continuous abnormality of the sensor. It may be that the sensor's sensitivity drift is caused by dust adhesion, data deviation is caused by aging of internal components, or long-term inaccurate readings are caused by loose installation position. At this time, the system's early warning mechanism is triggered, such as a fault prompt popping up on the on-site display screen or a maintenance notification being sent by the remote monitoring platform, reminding the staff to calibrate, clean or replace the sensor in time to prevent the fault from continuing to cause a significant decrease in weighing accuracy, or even causing errors in the measurement of production materials.

[0065] Furthermore, in the process of counting the number of deviations exceeding the tolerance, this embodiment also clarifies the accumulation rules. This embodiment adopts a continuous accumulation method, that is, starting from the first occurrence of deviation, if subsequent data still exceeds the tolerance, it continues to accumulate. If a non-excessive data occurs, the accumulation count is reset. For example, the preset threshold is 3 times. If there is 1 normal data after 2 consecutive deviations, the accumulation count is cleared to zero, so as to avoid miscounting non-continuous occasional deviations in the fault statistics. If there are 3 consecutive deviations, it is directly determined as a continuous fault, ensuring rapid response to equipment problems.

[0066] In this optional embodiment, the threshold quantization + continuous accumulation judgment logic can accurately distinguish between random deviations and continuous faults, and balance the stability and timeliness of operation and maintenance of the weighing system, thus ensuring the long-term reliability of the dual-sensor weighing accuracy.

[0067] Optionally, determining the final weight value of the bulk material based on the data deviation status includes: If the data deviation status is normal, then determine whether the accuracy of the two sensors is consistent; If the accuracy of the two sensors is consistent, the average value of the effective weight data of the two sensors shall be taken as the final weight value of the bulk material. If the accuracy of the two sensors is different, the weight corresponding to the effective weight data of the sensors is determined according to the accuracy of the sensors, and the final weight value of the bulk material is obtained by weighted average calculation based on the weight. The value of the weight is positively correlated with the accuracy of the sensors. If the data deviation state is an accidental deviation state, then the variance of the effective weight data of each sensor within a preset time period is calculated respectively; The effective weight data with the smallest variance is taken as the final weight value of the bulk material; If the data deviation state is a continuous fault state, compare the degree of deviation between the effective weight data of each sensor and the historical normal weight data. The effective weight data that deviates from the historical normal weight data within a preset allowable range is taken as the final weight value of the bulk material.

[0068] Specifically, in the first scenario: when the data deviation is within the normal range, the weighted average calculation based on sensor accuracy is used. When the data deviation of both sensors is within the normal range, i.e., both the absolute deviation and the relative deviation rate meet the standards, the appropriate calculation method is selected according to the difference in sensor accuracy to avoid measurement deviation caused by inconsistency in accuracy. The determination is made by comprehensively considering the sensor's factory calibration report, periodic verification data (such as accuracy error records for the past 3 months), and the stability of real-time acquired data (such as the fluctuation index during static weighing). If the accuracy class (such as both being 0.1 grade), error range (such as both being ≤±0.2kg), and long-term stability deviation (such as monthly drift amount being ≤0.1kg) of the two sensors are consistent, then they are considered to have consistent accuracy. If there is a significant difference (such as the first sensor being 0.1 grade and the second being 0.3 grade), then they are considered to have different accuracy. At this point, the reliability of the data from the two sensors is comparable. The arithmetic mean of the first and second effective weight data is taken as the final weight value. For example, if the first data is 100.1 kg and the second is 99.9 kg, the final weight value is (100.1 + 99.9) / 2 = 100.0 kg. This method can reduce the influence of random noise through data complementarity and further improve measurement accuracy. For weighted average calculations with different accuracies, the weight is determined according to the accuracy level. Higher-precision sensors have higher weights, and the weight is inversely proportional to the accuracy error. For example, the smaller the accuracy error, the larger the weight. For instance, if the accuracy error of the first sensor is 0.1% (weight set to 0.6) and the second is 0.3% (weight set to 0.4), and the first data is 100.2 kg and the second is 99.7 kg, the final weight value is 100.2 × 0.6 + 99.7 × 0.4 = 100.0 kg. This method can highlight the dominant role of high-precision sensor data and avoid low-precision data lowering the overall measurement accuracy.

[0069] The second scenario: When the data deviation is considered random, variance optimization based on data stability is used. When it's determined to be random deviation (i.e., the number of deviations does not reach the threshold, and the deviation is caused by short-term interference), the core is to select the data with higher stability and less susceptibility to interference from two sets of valid data as the final result. A preset time period covers the stable data interval before and after the random deviation occurs. For example, it might be set to 10 sets of data: 5 sets before the deviation occurs + 5 sets after the deviation occurs. This ensures a complete reflection of the data fluctuation trend. For example, a sampling frequency of 10Hz and a preset time period of 1 second can effectively capture the impact of short-term interference on data stability. The smaller the variance, the smaller the data fluctuation and the higher the stability. For example, the variance of the 10 sets of data from the first sensor is 0.02. The second is 0.15. This indicates that the first data is less susceptible to accidental interference. Therefore, the effective weight data with the smallest variance is selected as the final weight value. For example, 100.1 kg from the first sensor is taken as the result to avoid the measurement accuracy being affected by data with greater interference.

[0070] The third scenario: When the data deviation is in a continuous fault state, reliability screening is based on historical data. When a continuous fault state is determined, i.e., the number of deviations exceeds the threshold and at least one sensor is faulty, the data from the non-faulty sensors is identified by comparing with historical normal data to ensure the reliability of the final weight value. Historical data with fault-free sensors and consistent weighing scenarios, such as the same material and the same conveying speed, are selected. For example, normal data from static weighing of the same batch of bulk materials within the past week (mean 100.0 kg, fluctuation range ±0.3 kg) ensures that historical data is comparable to the current weighing task. The deviation between the two sets of valid weight data and the mean of historical normal data is calculated (e.g., the first data is 100.2 kg, deviation 0.2 kg; the second is 101.5 kg, deviation 1.5 kg), and compared with the preset allowable range (based on the fluctuation range of historical data, such as ±0.5 kg). Data with deviations within the allowable range (e.g., 100.2 kg from the first sensor) is used as the final weight value, excluding abnormal data from faulty sensors (e.g., 101.5 kg from the second sensor) to avoid measurement errors caused by faulty data.

[0071] In this optional embodiment, this state-adaptation strategy can fully utilize dual-sensor data to improve accuracy when the data is normal, filter stable data when there are random deviations, and eliminate abnormal data when there are continuous faults, ensuring the accuracy and reliability of the final weight value of bulk materials under different scenarios, and effectively solving the measurement problem of dual-sensor weighing systems under complex working conditions.

[0072] Combination Figure 2 As shown, a bulk material weighing system based on dual sensors according to the present invention includes: The signal acquisition unit is used to weigh the bulk material through dual sensors and obtain the original weight signal corresponding to each of the dual sensors; The signal processing unit is used to preprocess and scene-specific filter each of the original weight signals to obtain the effective weight data of each sensor for the bulk material. A data verification unit is used to perform cross-calibration analysis based on each of the effective weight data to determine the data deviation status of the dual sensors. The weight determination unit is used to determine the final weight value of the bulk material based on the data deviation status.

[0073] The dual-sensor-based bulk material weighing system of the present invention has the same advantages over the prior art as the aforementioned dual-sensor-based bulk material weighing method, and will not be repeated here.

[0074] Combination Figure 3 As shown, a bulk material weighing device of the present invention includes a memory and a processor; The memory is used to store computer programs; The processor is configured to implement the above-described dual-sensor-based bulk material weighing method when executing the computer program.

[0075] The bulk material weighing device of the present invention has the same advantages over the prior art as the aforementioned bulk material weighing method based on dual sensors, and will not be repeated here.

[0076] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A two-sensor based method of weighing bulk material, characterized by, The method comprises the following steps: The raw weight signals corresponding to each sensor of the double sensors are obtained by weighing the bulk material through the double sensors; The effective weight data of the bulk material corresponding to each sensor is obtained by preprocessing and scenario filtering of each raw weight signal; The data deviation state of the double sensors is judged based on each effective weight data; The final weight value of the bulk material is determined according to the data deviation state.

2. The dual sensor based bulk material weighing method of claim 1, wherein, The raw weight signals corresponding to each sensor of the double sensors are obtained by weighing the bulk material through the double sensors, which comprises the following steps: The analog weight signals corresponding to each sensor are obtained by controlling each sensor to collect data according to a preset collection period through a hardware synchronization trigger signal; The collection time difference between the two sensors is less than a preset time length; The analog weight signals output by each sensor are converted into digital signals synchronously to obtain a first raw weight signal and a second raw weight signal.

3. The dual sensor based bulk material weighing method of claim 2, wherein, The double sensors comprise a first sensor and a second sensor, the weighing precision of the first sensor is higher than that of the second sensor, the range of the first sensor is smaller than that of the second sensor, the first raw weight signal is obtained by the first sensor, and the second raw weight signal is obtained by the second sensor.

4. The dual sensor based bulk material weighing method as claimed in claim 2, wherein, The effective weight data of the bulk material corresponding to each sensor is obtained by preprocessing and scenario filtering of each raw weight signal, which comprises the following steps: The extreme data exceeding the preset signal interval in the first raw weight signal and the second raw weight signal is obtained by respectively performing abnormal value detection on the first raw weight signal and the second raw weight signal; The extreme data is eliminated, and the preprocessed first raw weight signal and the preprocessed second raw weight signal are obtained by replacing the preceding effective data corresponding to the extreme data; The fluctuation index corresponding to the preprocessed first raw weight signal and the preprocessed second raw weight signal is obtained by respectively performing fluctuation calculation on the preprocessed first raw weight signal and the preprocessed second raw weight signal for a continuous preset number of collection data; The weighing state is judged according to the fluctuation index; If the weighing state is a static weighing state, the preprocessed first raw weight signal and the preprocessed second raw weight signal are filtered by using a sliding average filtering algorithm to obtain the first effective weight data and the second effective weight data corresponding to the preprocessed first raw weight signal and the preprocessed second raw weight signal respectively; If the weighing state is a dynamic weighing state, the preprocessed first raw weight signal and the preprocessed second raw weight signal are filtered in real time by using an adaptive filtering algorithm to obtain the first effective weight data and the second effective weight data corresponding to the preprocessed first raw weight signal and the preprocessed second raw weight signal respectively.

5. The dual sensor based bulk material weighing method of claim 4, wherein, The weighing state is judged according to the fluctuation index, which comprises the following steps: comparing the fluctuation index with the preset fluctuation threshold value; if the fluctuation index is less than or equal to the preset fluctuation threshold value and lasts for a preset time length, determining that the weighing state is a static weighing state; if the fluctuation index is greater than the preset fluctuation threshold value, or the fluctuation index is less than or equal to the preset fluctuation threshold value but lasts for a time length less than the preset time length, determining that the weighing state is a dynamic weighing state.

6. The dual sensor based bulk material weighing method of claim 4, wherein, the mutual calibration analysis based on each of the effective weight data to determine the data deviation state of the double sensors, comprising: determining an absolute deviation according to the absolute value of the difference between the first effective weight data and the second effective weight data; determining a relative deviation rate according to the ratio of the absolute deviation to the average of the first effective weight data and the second effective weight data; comparing the absolute deviation and the relative deviation rate with preset absolute deviation threshold value and preset relative deviation rate threshold value respectively; if the absolute deviation is less than or equal to the preset absolute deviation threshold value and the relative deviation rate is less than or equal to the preset relative deviation rate threshold value, determining that the data deviation state of the double sensors is a normal state; if the absolute deviation is greater than the preset absolute deviation threshold value or the relative deviation rate is greater than the preset relative deviation rate threshold value, accumulating the number of deviation overruns, and determining the data deviation state according to the number of deviation overruns as an occasional deviation state or a continuous fault state.

7. The dual sensor based bulk material weighing method of claim 6, wherein, the determination of the data deviation state as an occasional deviation state or a continuous fault state according to the number of deviation overruns, comprising: if the number of deviation overruns is less than a preset number of deviation overrun threshold value, determining that the data deviation state is the occasional deviation state; if the number of deviation overruns is greater than or equal to the preset number of deviation overrun threshold value, determining that the data deviation state is the continuous fault state.

8. The dual sensor based bulk material weighing method of claim 1, wherein, the determination of the final weight value of the bulk material according to the data deviation state, comprising: if the data deviation state is a normal state, determining whether the accuracy of the double sensors is consistent; if the accuracy of the double sensors is consistent, taking the average of the effective weight data of the double sensors as the final weight value of the bulk material; if the accuracy of the double sensors is different, determining the weight corresponding to the effective weight data of the sensor according to the accuracy of the sensor, and obtaining the final weight value of the bulk material according to the weight through weighted average calculation, and the value of the weight is positively related to the accuracy of the sensor; if the data deviation state is an occasional deviation state, calculating the variance of the effective weight data of each sensor within a preset time period respectively; taking the effective weight data with the smallest variance as the final weight value of the bulk material; if the data deviation state is a continuous fault state, comparing the deviation degree of the effective weight data of each sensor with the historical normal weight data; taking the effective weight data with a deviation within a preset allowable range from the historical normal weight data as the final weight value of the bulk material.

9. A two-sensor based bulk material weighing system, characterized in that ​ A signal acquisition unit is configured to weigh the bulk material by using the double sensors to obtain an original weight signal corresponding to each sensor of the double sensors; A signal processing unit is configured to pre-process and filter each original weight signal to obtain effective weight data of the bulk material corresponding to each sensor; A data verification unit is configured to analyze mutual verification based on each effective weight data to determine a data deviation state of the double sensors; A weight determination unit is configured to determine a final weight value of the bulk material according to the data deviation state.

10. A bulk material weighing apparatus characterised in that, comprising a memory and a processor; the memory is configured to store a computer program; the processor is configured to implement the bulk material weighing method based on the double sensors according to any one of claims 1 to 8 when the computer program is executed.

Images