A dynamic nutrient requirement material tower, intelligent compound feeding system and intelligent compound feeding method

Abstract

The present application relates to a kind of material tower, specifically to a kind of dynamic nutrient requirement material tower, further to a kind of intelligent compound feeding system and intelligent compound feeding method, including support frame and hopper, the hopper is set on support frame;The hopper is conical, the bottom of hopper is connected with weighing bin, the bottom of weighing bin is provided with cantilever beam type weighing sensor for weighing the weight of material in weighing bin, weighing bin is connected with conveying pipeline, conveying pipeline is provided with auger for conveying material, the auger is connected with variable frequency speed regulation motor.The present application realizes that nutrition level is continuously adjustable with pig group weight, significantly improves feed conversion rate;Eliminate material replacement stress: smooth transition replaces discrete jump, pig group intake is stable, growth continuity is enhanced, and finishing time is shortened.

Description

Technical Field

[0001] This invention relates to a feed tower, specifically a feed tower with dynamic nutritional requirements, and also to an intelligent compound feeding system and an intelligent compound feeding method. Background Technology

[0002] Currently, large-scale pig farms generally adopt a "single feed tower + single feed line" feeding model: each pig house or breeding unit is equipped with one or more feed towers to store single-formula pelleted finished feed, such as early nursery feed, late nursery feed, and finishing feed, etc., and the feed is transported to the troughs of each pen through an auger or disc feed line. This model relies on human experience to judge the growth stage of the pig herd, manually switch feed towers or adjust feeding parameters, and the nutrient supply is stepped and jumpy, which cannot achieve continuous and precise matching.

[0003] Some large pig farms adopt a "multi-feed tower parallel" structure, but each feed tower operates independently and corresponds to pig houses at different growth stages. There is no synergistic mixing function between feed towers, and it is essentially still a physical superposition of a single formula supply.

[0004] The existing technology has the following key drawbacks: (1) Nutritional supply is seriously out of sync with physiological needs The energy, amino acid, and mineral requirements of pigs change continuously at different weight stages (e.g., 20kg, 50kg, 80kg), while the traditional "one feed tower, one formula" model can only provide discrete, stepped nutrition (e.g., feed A for 20-30kg, feed B for 30-60kg). When the average weight of the pig herd is at the boundary between the two stages, either there is excess nutrition leading to waste, or insufficient nutrition restricting growth, resulting in low feed conversion ratio and generally high feed-to-meat ratio.

[0005] (2) Stress and lag in formula switching Human judgment of the timing of switching feed to pigs is subject to subjective bias. Switching too early leads to insufficient nutrition, while switching too late leads to nutrient waste. In addition, when switching to pelleted feed, the old feed residue in the feed line mixes with the new feed, creating a 3-5 day "transition feed" confusion period, which causes fluctuations in the pigs' feed intake and hinders their growth.

[0006] (3) The structure of the material tower-material line has a single function. Existing feed towers mostly use fixed discharge ports or simple gates at the bottom, which can only control the flow and cannot achieve quantitative proportioning of various feeds; the feed line is designed as a single material conveying channel, lacking the function of mixing materials in the middle, and the physical structure limits the realization of precise nutrition technology.

[0007] (4) Data silos and decision lag The pig herd weight data relies on the frequency of manual sampling and weighing, or on feed changes based on age. This fails to reflect the real-time growth status of the herd; the feed tower control system operates independently without data interaction with the breeding management software, resulting in nutritional decisions that lag significantly behind the actual needs of the pigs. Summary of the Invention

[0008] The first objective of this invention is to provide a feed tower with dynamic nutritional requirements, which solves the problem that existing feed towers cannot accurately control the output.

[0009] To solve the above problems, the present invention adopts the following technical solution: A dynamic nutrient demand feed tower includes a support frame and a hopper, wherein the hopper is mounted on the support frame; the hopper is conical, and a weighing bin is connected to the bottom of the hopper. A cantilever beam load cell for weighing the material inside the weighing bin is installed at the bottom of the weighing bin. The weighing bin is connected to a conveying pipe, and an auger for conveying the material is installed inside the conveying pipe. The auger is connected to a variable frequency speed control motor.

[0010] Furthermore, the bottom of the hopper is provided with a baffle to control the opening and closing of the hopper outlet; a weighing platform is provided inside the weighing bin, the cantilever beam load cell is located at the bottom of the weighing platform, and the conveying pipe is used for conveying materials on the weighing platform.

[0011] The second objective of this invention is to provide an intelligent compound feeding system for dynamic nutritional requirements, thereby solving the problem of monotonous feed in feed towers.

[0012] To solve the above problems, the present invention adopts the following technical solution: A smart compound feeding system for dynamic nutritional requirements includes a confluence mixing chamber and multiple feed towers for dynamic nutritional requirements as described in the above embodiments. The conveying pipes are all connected to the confluence mixing chamber, and the confluence mixing chamber is connected to a feeding pipe, which is connected to a feed trough. A stirrer for stirring materials is provided inside the confluence mixing chamber.

[0013] The final objective of this invention is to provide an intelligent compound feeding method based on dynamic nutritional requirements, which solves the problem that pigs cannot rationally allocate feed when their average weight is at the boundary between two stages.

[0014] To solve the above problems, the present invention adopts the following technical solution: A method for intelligent compound feeding based on dynamic nutritional requirements, using the intelligent compound feeding system for dynamic nutritional requirements described in the above embodiments, includes: S1. Manually enter the name of the feed stored in each feed tower, including digestible energy. Lysine Crude protein ,calcium Feed information including phosphorus (P) nutritional indicators, applicable weight range, and inventory warning values; establishing relationships between the speed of variable frequency motors and the flow rate of conveying pipelines, and storing and controlling the system; configuring basic information on pig breeds, average weight at entry, target average weight at exit, and current inventory.

[0015] S2. The average weight of the pig herd is automatically collected daily using intelligent weighing equipment. Calculate daily weight gain ;Will and The nutritional requirement vector is obtained by inputting the nutritional model and outputting the nutritional model. ;by With the target as the objective and the nutritional indicators in each feed tower as variables, a system of equations is established, and the ratio of each feed tower is obtained by solving the equations. The system of equations has the following constraints: ; ; ,and ≤ ≤ ; In the above formula: Digest energy for the target; Target lysine; For the first Feed tower proportions; For the first Energy ratio for feed tower digestion; For the first Lysine ratio in the feed tower; No. Minimum proportions for the feed tower; For the first Maximum proportion of feed tower; The material ratio for the first feed tower; The material ratio for the second feed tower; The material ratio for the third feed tower; S3. The cloud management unit verifies the rationality of the formula based on the historical data of the pig herd and big data of similar pig herds, and can make fine adjustments if necessary; after verification, the final formula parameters, feeding time and total feeding amount are sent to the pig farm's control system via MQTT. The control system will adjust the ratio of each feed tower accordingly. and total amount of food fed per feeding Calculate the feed rate of each feed tower. ; ; Based on the relationship between the speed of the variable frequency drive motor and the flow rate of the conveying pipeline, the speed and running time of the variable frequency drive motor are determined; then, each feed tower is started sequentially, so that the feed in each feed tower flows into the mixing chamber; the cantilever beam weighing sensor measures the feed reaching the required level. Then, stop feeding materials; S4. After each feed tower collects feed into the mixing chamber, the agitator in the mixing chamber mixes the feed. After mixing, the feed is delivered to each feed trough. The data of this feeding is recorded and uploaded to the remote management unit. S5. New average weight collected the following day. Calculate the actual daily weight gain, and if the deviation exceeds 5%, make minor adjustments to the ratio; and update the nutrition model monthly to improve prediction accuracy; S6. The visual sensors in the cloud management unit monitor the remaining material in each material tower in real time, and push a purchase reminder when the material level is lower than the warning value.

[0016] Furthermore, the nutritional model expression is as follows: ; ; In the above formula: , , , , , , , All are variety-specific regression coefficients, determined by fitting historical data. The ambient temperature; This is to meet the daily energy requirements for digestion; To meet the daily lysine requirement; The output of the nutrition model is: ; ; In the above formula: This refers to the crude protein ratio; To predict daily feed intake.

[0017] Furthermore, after the feeds in each feed tower are mixed according to the formula, the digestible energy and lysine content of the mixture should be equal to the target value, and the total formula should be 100%, with the amount used in each feed tower within the applicable stage limit. ; ; ,and ≤ ≤ ; In the above formula: This refers to the number of material towers.

[0018] Furthermore, based on minimum feed cost The proportions of each feed tower were calculated. ; In the above formula: For the first Unit price of feed for each feed tower.

[0019] Furthermore, based on minimizing nutritional bias The proportions of each feed tower were calculated. ; In the above formula, For the first feed tower Nutritional indicators; For the first Target nutritional indicators.

[0020] Furthermore, the control system drives the variable frequency speed control motor at a speed of Start-up, cantilever beam load cell at 10 Frequency sampling, cumulative feeding amount is Calculation error The error is adjusted using PID control, as follows: Proportional term: ; Integral term: ; Differential term: ; Output adjustment amount: ; New speed command: ; In the above formula: The current rotational speed; For proportional gain; Integral coefficient; These are the differential coefficients; for Instantaneous error at a given moment; For a very short period of time; when Stop feeding materials when the time comes; like Record the error and compensate for it next time.

[0021] Furthermore, the nutrition model is updated monthly, as detailed below: Input features: Derived features: , , ; In the above formula: For ambient humidity; Number the formula; For the season; Pig species; Number the pig farm; Random Forest Capture nonlinear relationships and predict Benchmark value; Handling high-dimensional sparse features and prediction Benchmark value; Network: Input the growth sequence of the most recent 30 days, and correct. and Static prediction; Weighted fusion: ; In the above formula: For nutrition prediction model scenarios; , , All are weighted fusion coefficients. The target average weight for the pig herd.

[0022] Compared with the prior art, the present invention has the following advantages and beneficial effects: 1. A weighing bin is installed at the bottom of the hopper, and a cantilever beam load cell is installed at the bottom of the weighing bin to weigh the material inside. The cantilever beam load cell can accurately measure the amount of material discharged from the hopper, solving the problem that existing material towers cannot accurately control the amount of material discharged.

[0023] 2. The feed from each feed tower is transported to the mixing chamber, which is equipped with a stirrer for mixing materials. The materials are stirred and mixed by the stirrer, and the mixed feed is transported to the feed trough through the feeding pipe.

[0024] 3. Achieve continuous and adjustable nutritional levels according to the weight of the pig herd, significantly improving feed conversion rate; eliminate feed change stress: smooth transition replaces discrete jumps, stabilizes pig herd feed intake, enhances growth continuity, and shortens slaughter time; data-driven decision-making: cloud-based aggregation of data from multiple pig farms, continuous evolution of nutritional models, and continuous optimization of herd feeding strategies; improved batch management: improved herd uniformity, high slaughter concentration, and facilitates all-in, all-out management. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of a feed tower that meets dynamic nutritional requirements.

[0026] Figure 2 This is a schematic diagram of an intelligent compound feeding system for dynamic nutritional needs.

[0027] Figure 3 This is a schematic diagram of the bus mixing chamber.

[0028] Figure 4 This is a schematic diagram of a screw conveyor.

[0029] Figure 5 This is a flowchart of an intelligent compound feeding method for dynamic nutritional needs. Detailed Implementation

[0030] Example 1 like Figure 1 As shown, a dynamic nutrient demand feed tower includes a support frame 1 and a hopper 2, the hopper 2 being mounted on the support frame 1; the hopper 2 is conical, and a weighing chamber 3 is connected to the bottom of the hopper 2. A cantilever beam load cell for weighing the material inside the weighing chamber 3 is installed at the bottom of the weighing chamber 3. The weighing chamber 3 is connected to a conveying pipe 4, such as... Figure 4 As shown, an auger for conveying materials is installed inside the conveying pipe 4, and the auger is connected to a variable frequency speed control motor 5. A baffle for controlling the opening and closing of the hopper 2's discharge port is installed at the bottom of the hopper 2; a weighing platform is installed inside the weighing bin 3, and a cantilever beam load cell is located at the bottom of the weighing platform. The conveying pipe 4 is used to convey materials on the weighing platform. After determining the feed ratio, the feed flow rate and feed quantity are controlled by controlling the speed of the variable frequency speed control motor 5. The feed quantity is accurately measured using the cantilever beam load cell, making the feed quantity more precise.

[0031] Example 2 like Figure 2 and Figure 3 As shown, an intelligent compound feeding system for dynamic nutritional requirements includes a confluence mixing chamber 6 and multiple feed towers for dynamic nutritional requirements as described in Embodiment 1. The conveying pipes 4 are all connected to the confluence mixing chamber 6, which is connected to a feeding pipe that is connected to a feed trough. An agitator for mixing materials is installed inside the confluence mixing chamber 6. When multiple feed towers feed materials according to the specified ratio, and the materials are weighed by a cantilever beam weighing sensor, they are conveyed to the confluence mixing chamber 6. The agitator mixes the materials to achieve a uniformity of over 85%, and then the mixture is conveyed to the feed trough.

[0032] Example 3 like Figure 5 As shown, an intelligent compound feeding method for dynamic nutritional requirements, using the intelligent compound feeding system for dynamic nutritional requirements described in Example 2, includes: S1. Manually enter the name of the feed stored in each feed tower, including digestible energy. Lysine Crude protein ,calcium Feed information including phosphorus (P) nutritional indicators, applicable weight range, and inventory warning values; establishing the relationship between the speed of the variable frequency drive motor and the flow rate of the conveying pipeline, and storing it in the control system; configuring basic information such as the breed of pigs, average weight at entry, target average weight at exit, and current number of pigs in stock.

[0033] S2. The average weight of the pig herd is automatically collected daily using intelligent weighing equipment. Calculate daily weight gain ;Will and The nutritional requirement vector is obtained by inputting the nutritional model and outputting the nutritional model. ;by With the target as the objective and the nutritional indicators in each feed tower as variables, a system of equations is established, and the ratio of each feed tower is obtained by solving the equations. The system of equations has the following constraints: ; ; ,and ≤ ≤ ; In the above formula: Digest energy for the target; Target lysine; For the first Feed tower proportions; For the first Energy ratio for feed tower digestion; For the first Lysine ratio in the feed tower; No. Minimum proportions for the feed tower; For the first Maximum proportion of feed tower; The material ratio for the first feed tower; The material ratio for the second feed tower; The material ratio for the third feed tower; S3. The cloud management unit verifies the rationality of the formula based on the historical data of the pig herd and big data of similar pig herds, and can make fine adjustments if necessary; after verification, the final formula parameters, feeding time and total feeding amount are sent to the pig farm's control system via MQTT. The control system will adjust the ratio of each feed tower accordingly. and total amount of food fed per feeding Calculate the feed rate of each feed tower. ; ; Based on the relationship between the speed of the variable frequency drive motor and the flow rate of the conveying pipeline, the speed and running time of the variable frequency drive motor are determined; then, each feed tower is started sequentially, so that the feed in each feed tower flows into the mixing chamber; the cantilever beam weighing sensor measures the feed reaching the required level. Then, stop feeding materials; S4. After each feed tower collects feed into the mixing chamber, the agitator in the mixing chamber mixes the feed. After mixing, the mixed feed is delivered to each feed trough. The data of this feeding is recorded and uploaded to the remote management unit. S5. New average weight collected the following day. Calculate the actual daily weight gain, and if the deviation exceeds 5%, make minor adjustments to the ratio; and update the nutrition model monthly to improve prediction accuracy; S6. The visual sensors in the cloud management unit monitor the remaining material in each material tower in real time, and push a purchase reminder when the material level is lower than the warning value.

[0034] Furthermore, the nutritional model expression is as follows: ; ; In the above formula: , , , , , , , All are variety-specific regression coefficients, determined by fitting historical data. The ambient temperature; This is to meet the daily energy requirements for digestion; To meet the daily lysine requirement; The output of the nutrition model is: ; ; In the above formula: This refers to the crude protein ratio; To predict daily feed intake.

[0035] Furthermore, after the feeds in each feed tower are mixed according to the formula, the digestible energy and lysine content of the mixture should be equal to the target value, and the total formula should be 100%, with the amount used in each feed tower within the applicable stage limit. ; ; ,and ≤ ≤ ; In the above formula: This refers to the number of material towers.

[0036] Furthermore, based on minimum feed cost The proportions of each feed tower were calculated. ; In the above formula: For the first Unit price of feed for each feed tower.

[0037] Furthermore, the control system drives the variable frequency speed control motor at a speed of Start-up, cantilever beam load cell at 10 Frequency sampling, cumulative feeding amount is Calculation error The error is adjusted using PID control, as follows: Proportional term: ; Integral term: ; Differential term: ; Output adjustment amount: ; New speed command: ; In the above formula: The current rotational speed; For proportional gain; Integral coefficient; These are the differential coefficients; for Instantaneous error at a given moment; For a very short period of time; when Stop feeding materials when the time comes; like Record the error and compensate for it next time.

[0038] PID parameter tuning: Initial tuning was performed using the Ziegler-Nichols method, followed by fine-tuning based on the material characteristics (granular material flowability) of the site. =2.0, =0.5, =0.1 is a typical value.

[0039] Furthermore, the nutrition model is updated monthly, as detailed below: Input features: Derived features: , , ; In the above formula: For ambient humidity; Number the formula; For the season; Pig species; Number the pig farm; Random Forest Capture nonlinear relationships and predict Benchmark value; Handling high-dimensional sparse features and prediction Benchmark value; Network: Input the growth sequence of the most recent 30 days, and correct. and Static prediction; Weighted fusion: ; In the above formula: For nutrition prediction model scenarios; , , All are weighted fusion coefficients. The target average weight for the pig herd.

[0040] The intelligent compound feeding system for dynamic nutritional requirements includes a mixing chamber and a feed tower for dynamic nutritional requirements as described in Example 1. All conveying pipes are connected to the mixing chamber, which is connected to a feeding pipe that is connected to a feed trough. An agitator for mixing materials is installed inside the mixing chamber. When multiple feed towers feed materials according to the specified ratio, and the materials are weighed by a cantilever beam weighing sensor, they are conveyed to the mixing chamber. The agitator mixes the materials to ensure a uniformity of over 85%, and then the mixture is conveyed to the feed trough.

[0041] Example 4 A method for intelligent compound feeding based on dynamic nutritional requirements, using the intelligent compound feeding system for dynamic nutritional requirements described in Example 2 above, includes: S1. Manually enter the name of the feed stored in each feed tower, including digestible energy. Lysine Crude protein ,calcium Feed information including phosphorus (P) nutritional indicators, applicable weight range, and inventory warning values; establishing relationships between the speed of variable frequency motors and the flow rate of conveying pipelines, and storing and controlling the system; configuring basic information on pig breeds, average weight at entry, target average weight at exit, and current inventory.

[0042] S2. The average weight of the pig herd is automatically collected daily using intelligent weighing equipment. Calculate daily weight gain ;Will and The nutritional requirement vector is obtained by inputting the nutritional model and outputting the nutritional model. ;by With the target as the objective and the nutritional indicators in each feed tower as variables, a system of equations is established, and the ratio of each feed tower is obtained by solving the equations. The system of equations has the following constraints: ; ; ,and ≤ ≤ ; In the above formula: Digest energy for the target; Target lysine; For the first Feed tower proportions; For the first Energy ratio for feed tower digestion; For the first Lysine ratio in the feed tower; No. Minimum proportions for the feed tower; For the first Maximum proportion of feed tower; The material ratio for the first feed tower; The material ratio for the second feed tower; The material ratio for the third feed tower; S3. The cloud management unit verifies the rationality of the formula based on the historical data of the pig herd and big data of similar pig herds, and can make fine adjustments if necessary; after verification, the final formula parameters, feeding time and total feeding amount are sent to the pig farm's control system via MQTT. The control system will adjust the ratio of each feed tower accordingly. and total amount of food fed per feeding Calculate the feed rate of each feed tower. ; ; Based on the relationship between the speed of the variable frequency drive motor and the flow rate of the conveying pipeline, the speed and running time of the variable frequency drive motor are determined; then, each feed tower is started sequentially, so that the feed in each feed tower flows into the mixing chamber; the cantilever beam weighing sensor measures the feed reaching the required level. Then, stop feeding materials; S4. After each feed tower collects feed into the mixing chamber, the agitator in the mixing chamber mixes the feed. After mixing, the feed is delivered to each feed trough. The data of this feeding is recorded and uploaded to the remote management unit. S5. New average weight collected the following day. Calculate the actual daily weight gain, and if the deviation exceeds 5%, make minor adjustments to the ratio; and update the nutrition model monthly to improve prediction accuracy; S6. The visual sensors in the cloud management unit monitor the remaining material in each material tower in real time, and push a purchase reminder when the material level is lower than the warning value.

[0043] Furthermore, the nutritional model expression is as follows: ; ; In the above formula: , , , , , , , All are variety-specific regression coefficients, determined by fitting historical data. The ambient temperature; This is to meet the daily energy requirements for digestion; To meet the daily lysine requirement; The output of the nutrition model is: ; ; In the above formula: This refers to the crude protein ratio; To predict daily feed intake.

[0044] Furthermore, after the feeds in each feed tower are mixed according to the formula, the digestible energy and lysine content of the mixture should be equal to the target value, and the total formula should be 100%, with the amount used in each feed tower within the applicable stage limit. ; ; ,and ≤ ≤ ; In the above formula: This refers to the number of material towers.

[0045] Furthermore, based on minimizing nutritional bias The proportions of each feed tower were calculated. ; In the above formula, For the first feed tower Nutritional indicators; For the first Target nutritional indicators.

[0046] Furthermore, the control system drives the variable frequency speed control motor at a speed of Start-up, cantilever beam load cell at 10 Frequency sampling, cumulative feeding amount is Calculation error The error is adjusted using PID control, as follows: Proportional term: ; Integral term: ; Differential term: ; Output adjustment amount: ; New speed command: ; In the above formula: The current rotational speed; For proportional gain; Integral coefficient; These are the differential coefficients; for Instantaneous error at a given moment; For a very short period of time; when Stop feeding materials when the time comes; like Record the error and compensate for it next time.

[0047] PID parameter tuning: Initial tuning was performed using the Ziegler-Nichols method, followed by fine-tuning based on the material characteristics (granular material flowability) of the site. =2.0, =0.5, =0.1 is a typical value.

[0048] Furthermore, the nutrition model is updated monthly, as detailed below: Input features: Derived features: , , ; In the above formula: For ambient humidity; Number the formula; For the season; Pig species; Number the pig farm; Random Forest Capture nonlinear relationships and predict Benchmark value; Handling high-dimensional sparse features and prediction Benchmark value; Network: Input the growth sequence of the most recent 30 days, and correct. and Static prediction; Weighted fusion: ; In the above formula: For nutrition prediction model scenarios; , , All are weighted fusion coefficients. The target average weight for the pig herd.

[0049] The intelligent compound feeding system for dynamic nutritional requirements includes a mixing chamber and a feed tower for dynamic nutritional requirements as described in Example 1. All conveying pipes are connected to the mixing chamber, which is connected to a feeding pipe that is connected to a feed trough. An agitator for mixing materials is installed inside the mixing chamber. When multiple feed towers feed materials according to the specified ratio, and the materials are weighed by a cantilever beam weighing sensor, they are conveyed to the mixing chamber. The agitator mixes the materials to ensure a uniformity of over 85%, and then the mixture is conveyed to the feed trough.

[0050] Example 5 A method for intelligent compound feeding based on dynamic nutritional requirements, using the intelligent compound feeding system for dynamic nutritional requirements described in Example 2 above, includes: S1. Manually enter the name of the feed stored in each feed tower, including digestible energy. Lysine Crude protein ,calcium Feed information including phosphorus (P) nutritional indicators, applicable weight range, and inventory warning values; establishing relationships between the speed of variable frequency motors and the flow rate of conveying pipelines, and storing and controlling the system; configuring basic information on pig breeds, average weight at entry, target average weight at exit, and current inventory.

[0051] S2. Using intelligent weighing equipment, such as weighbridges, cage scales, and AI-powered weight estimation cameras, the average weight of the pig herd is obtained in real time. Gaining weight every day The value, =35 kg, 75 Ambient temperature It is 25℃; Call the variety model to calculate =14.2 , =18.5 ; Based on predicted feed intake =1.8 Calculate the target nutrient concentration; 14.2 / 1.8=7.89 ; The crude protein content is 0.16. =18.5 / (1.8×0.16)=0.88%; Given: Target: =7.89 MJ / kg, =0.88% Tower #1 (High-nutrient feed): =14.5, =1.15, =3.5, =0.1, =0.5; Tower #2 (Medium Nutrient Feed): =14.0, =0.95, =3.2, =0.2, =0.7; Tower #3 (low nutrient feed): =13.5, =0.75, =2.9, =0, =0.6; Solution: =3.5 +3.2 +2.9 ; Constraint: 14.5 +14.0 +13.5 =7.89 1.15 +0.95 +0.75 =0.88 + + =10.1≤ ≤0.5, 0.2≤ ≤0.7,0≤ ≤0.6.

[0052] result: =0.25, =0.55, =0.20, which means 25% high-nutrient feed + 55% medium-nutrient feed + 20% low-nutrient feed.

[0053] Furthermore, the control system drives the variable frequency speed control motor at a speed of Start-up, cantilever beam load cell at 10 Frequency sampling, cumulative feeding amount is Calculation error The error is adjusted using PID control, as follows: Proportional term: ; Integral term: ; Differential term: ; Output adjustment amount: ; New speed command: ; In the above formula: The current rotational speed; For proportional gain; Integral coefficient; These are the differential coefficients; for Instantaneous error at a given moment; For a very short period of time; when Stop feeding materials when the time comes; like Record the error and compensate for it next time.

[0054] PID parameter tuning: Initial tuning was performed using the Ziegler-Nichols method, followed by fine-tuning based on the material characteristics (granular material flowability) of the site. =2.0, =0.5, =0.1 is a typical value.

[0055] Furthermore, the nutrition model is updated monthly, as detailed below: Input features: Derived features: , , ; In the above formula: For ambient humidity; Number the formula; For the season; Pig species; Number the pig farm; Random Forest Capture nonlinear relationships and predict Benchmark value; Handling high-dimensional sparse features and prediction Benchmark value; Network: Input the growth sequence of the most recent 30 days, and correct. and Static prediction; Weighted fusion: ; In the above formula: For nutrition prediction model scenarios; , , All are weighted fusion coefficients. The target average weight for the pig herd.

[0056] The intelligent compound feeding system for dynamic nutritional requirements includes a mixing chamber and a feed tower for dynamic nutritional requirements as described in Example 1. All conveying pipes are connected to the mixing chamber, which is connected to a feeding pipe that is connected to a feed trough. An agitator for mixing materials is installed inside the mixing chamber. When multiple feed towers feed materials according to the specified ratio, and the materials are weighed by a cantilever beam weighing sensor, they are conveyed to the mixing chamber. The agitator mixes the materials to ensure a uniformity of over 85%, and then the mixture is conveyed to the feed trough.

[0057] Example 6 Assuming a large-scale pig farm with 5000 fattening pigs, originally configured with 4 independent feed towers (20t each), storing: -1# Tower: Post-conservation feed ( It is 14.5 MJ / kg. 1.15%, suitable for 15-30kg) -2# Tower: Early growth material ( 14.0 MJ / kg 0.95%, suitable for 30-50kg) -3# Tower: Late-stage growth material ( 13.5 MJ / kg 0.75%, suitable for 50-80kg) -4# Tower: Fertilizer ( 13.0 MJ / kg 0.65%, applicable to 80kg slaughter weight. The original model was a step-by-step feed change: switch to No. 2 feed at 30kg, No. 3 feed at 50kg, and No. 4 feed at 80kg. During the feed change period, the pigs experienced significant stress, and the feed conversion ratio was 2.95.

[0058] Modification Plan: Retain 4 feed towers, install a weighing bin at the bottom of each tower, and connect the outlet to a newly added mixing chamber, followed by a feed trough; a new control system will be implemented, connecting the weighbridge data to a cloud platform. Combined with the Watcher series inspection robots, automatic weight estimation can be achieved, and the cloud platform will automatically analyze and make decisions regarding the application of feed data.

[0059] (I) Implementation Steps Phase 1: System deployment (1-2 days); Install variable frequency material shoe: Remove the original fixed material inlet and install the weighing bin; The mixing chamber is made of stainless steel and has built-in static mixing blades; it is installed between the feed tower group and the feed line. Control system: Connects to the weighbridge RS485 signal; Cloud-based account opening: Configure pig farm information, feed tower records, and basic pig herd data.

[0060] Phase Two: Operation and Debugging (Days 3-7); The current average weight of the pigs is 35kg, and the target is to reach an average weight of 110kg at slaughter. The control system calculates nutritional requirements: =14.2MJ / kg, =0.88%; Proportioning solution: Material #2 ( 14.95)×60%+3#material ( 13.5, 0.75) × 40% = 13.8 (slightly low), adjusted to 70% #2 material + 30% #3 material = 13.85, close to the target; Cloud-based verification and distribution: After confirming that the ratio is appropriate, the execution command is issued. Initial feeding: Total quantity per batch is 500kg, 350kg under tower #2 (rotation speed 80rpm, time 26min), 150kg under tower #3 (rotation speed 60rpm, time 15min), mixed and then conveyed, with a weighing error of ±1.5%.

[0061] Phase 3: Continuous operation (days 8-120); Daily automatic weight collection and dynamic adjustment of mix proportions: At 50kg: 40% #2 feed + 60% #3 feed; At 65kg: 80% #3 feed + 20% #4 feed; At 85kg: 30% of #3 feed + 70% of #4 feed; At 100kg: 100% of #4 feed; The nutritional supply transitioned smoothly, with no stress from feed changes; The model is regularly optimized in the cloud, and the energy demand curve is corrected based on the field data.

[0062] Of course, the above description is not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A feed tower for dynamic nutrient demand, characterized in that, The device includes a support frame and a hopper, with the hopper mounted on the support frame. The hopper is conical, and a weighing bin is connected to the bottom of the hopper. A cantilever beam load cell for weighing the material inside the weighing bin is installed at the bottom of the weighing bin. The weighing bin is connected to a conveying pipe, and an auger for conveying the material is installed inside the conveying pipe. The auger is connected to a variable frequency speed control motor.

2. The feed tower for dynamic nutrient demand according to claim 1, characterized in that, The bottom of the hopper is equipped with a baffle to control the opening and closing of the hopper outlet; the weighing bin is equipped with a weighing platform, the cantilever beam load cell is located at the bottom of the weighing platform, and the conveying pipe is used to convey the material on the weighing platform.

3. An intelligent compound feeding system for dynamic nutritional requirements, characterized in that, It includes a mixing chamber and multiple feed towers for dynamic nutrient demand as described in claim 1 or 2, wherein the conveying pipes are all connected to the mixing chamber, the mixing chamber is connected to a feeding pipe, and the feeding pipe is connected to a feed trough; the mixing chamber is equipped with a stirrer for stirring materials.

4. A method for intelligent compound feeding based on dynamic nutritional requirements, using the intelligent compound feeding system for dynamic nutritional requirements as described in claim 3, characterized in that, include: S1. Manually enter the name of the feed stored in each feed tower, including digestible energy. Lysine Crude protein ,calcium Feed information including phosphorus (P) nutritional indicators, applicable weight range, and inventory warning values; establishing the relationship between the speed of the variable frequency drive motor and the flow rate of the conveying pipeline, and storing it in the control system; configuring basic information on the breed of pigs, average weight at entry, target average weight at exit, and current number of pigs in stock. S2. The average weight of the pig herd is automatically collected daily using intelligent weighing equipment. Calculate daily weight gain ;Will and The nutritional requirement vector is obtained by inputting the nutritional model and outputting the nutritional model. ;by With the target as the objective and the nutritional indicators in each feed tower as variables, a system of equations is established, and the ratio of each feed tower is obtained by solving the equations. The system of equations has the following constraints: ； ； ,and ≤ ≤ ; In the above formula: Digest energy for the target; Target lysine; For the first Feed tower proportions; For the first Energy ratio for feed tower digestion; For the first Lysine ratio in the feed tower; No. Minimum proportions for the feed tower; For the first Maximum proportion of feed tower; The material ratio for the first feed tower; The material ratio for the second feed tower; The material ratio for the third feed tower; S3. The cloud management unit verifies the rationality of the formula based on the historical data of the pig herd and big data of similar pig herds, and can make fine adjustments if necessary; after verification, the final formula parameters, feeding time and total feeding amount are sent to the pig farm's control system via MQTT. The control system will adjust the ratio of each feed tower accordingly. and total amount of food fed per feeding Calculate the feed rate of each feed tower. ; ； Based on the relationship between the speed of the variable frequency speed control motor and the flow rate of the conveying pipeline, the speed and running time of the variable frequency speed control motor are determined; then, each material tower is started in sequence so that the feed in each material tower flows into the mixing chamber. The cantilever beam load cell measured the feed to reach Then, stop feeding materials; S4. After each feed tower collects feed into the mixing chamber, the agitator in the mixing chamber mixes the feed. After mixing, the feed is delivered to each feed trough. The data of this feeding is recorded and uploaded to the remote management unit. S5. New average weight collected the following day. Calculate the actual daily weight gain, and if the deviation exceeds 5%, make minor adjustments to the ratio; and update the nutrition model monthly to improve prediction accuracy; S6. The visual sensors in the cloud management unit monitor the remaining material in each material tower in real time, and push a purchase reminder when the material level is lower than the warning value.

5. The intelligent compound feeding method for dynamic nutritional requirements according to claim 4, characterized in that, The nutritional model expression is: ； ； In the above formula: , , , , , , , All are variety-specific regression coefficients, determined by fitting historical data. Ambient temperature; This is to meet the daily energy requirements for digestion; To meet the daily lysine requirement; The output of the nutrition model is: ； ； In the above formula: This refers to the crude protein ratio; To predict daily feed intake.

6. The intelligent compound feeding method for dynamic nutritional requirements according to claim 4, characterized in that, After the feeds in each feed tower are mixed according to the formula, the digestible energy and lysine content of the mixture should be equal to the target value, and the total formula should be 100%. The amount of each feed tower should be within the applicable stage limit. ； ； ,and ≤ ≤ ; In the above formula: This refers to the number of material towers.

7. The intelligent compound feeding method for dynamic nutritional requirements according to claim 4, characterized in that, Based on minimum feed cost The proportions of each feed tower were calculated. ; In the above formula: For the first Unit price of feed for each feed tower.

8. The intelligent compound feeding method for dynamic nutritional requirements according to claim 4, characterized in that, Based on minimizing nutritional deviation The proportions of each feed tower were calculated. ; In the above formula, For the first feed tower Nutritional indicators; For the first Target nutritional indicators.

9. The intelligent compound feeding method for dynamic nutritional requirements according to claim 4, characterized in that, The control system drives the variable frequency speed control motor at a speed of Start-up, cantilever beam load cell at 10 Frequency sampling, cumulative feeding amount is Calculation error The error is adjusted using PID control, as follows: Proportional term: ; Integral term: ; Differential term: ; Output adjustment amount: ; New speed command: ; In the above formula: The current rotational speed; For proportional gain; Integral coefficient; These are the differential coefficients; for Instantaneous error at a given moment; For a very short period of time; ; when Stop feeding materials when the time comes; like Record the error and compensate for it next time.

10. The intelligent compound feeding method for dynamic nutritional requirements according to claim 4, characterized in that, The nutrition model is updated monthly, as detailed below: Input features: Derived features: , , ; In the above formula: For ambient humidity; Number the formula; For the season; Pig species; Number the pig farm; Random Forest Capture nonlinear relationships and predict Benchmark value; Handling high-dimensional sparse features and prediction Benchmark value; Network: Input the growth sequence of the most recent 30 days, and correct. and Static prediction; Weighted fusion: ; In the above formula: For nutrition prediction model scenarios; , , The weighted fusion coefficient, The target average weight for the pig herd.

Images