RU2515382C2 - Method and device for determining average live weight of broilers in stock with their floor management - Google Patents

Abstract

FIELD: agriculture.

SUBSTANCE: method of determining average live weight of broilers in the stock with their floor management comprises automatic measurement of weight of the birds while drinking water at their state of minimum physical activity. Control is continuously carried out throughout the growing cycle. After weighing each broiler the current information on the number of weighted currently broiler chickens, their average live weight and the current margin of error of representation for the given sample size is provided. Reduction of influence of dynamic noises caused by physical activity of the bird on the measurement result is provided by carrying out a series of repeated measurements of weight of each bird weighed, at that the representation of the sample is achieved by weighing a certain number of broilers, which average weight characterises with sufficient accuracy and reliability the average live weight of the bird in the stock. The invention also relates to a device for implementing the said method.

EFFECT: increased accuracy of determining the average live weight of broilers in the stock at their floor management.

2 cl, 1 ex, 4 dwg

Description

The invention relates to agriculture, in particular to methods for controlling the productivity of poultry and devices for their implementation.

A device is known for weighing poultry during its rearing in cell batteries, consisting of a weighed cage with a sample group of birds based on force sensors, and a control unit, and a sample group of birds from a day old is contained in a control cage and continuously weighed throughout the cycle cultivation [a.s. USSR No. 1572470. A device for determining the average live weight of a bird / P.M. Slavin, G.A. Kharatyan, T.P. Volosovich // Discovery. Inventions 1990. No. 23]. To expand the range of mass measurement, a piecewise linear approximation of the characteristics of the primary mass transducer was applied.

A disadvantage of the known device is the inability to use it to control the live weight of the bird with its floor content, since it (the device) works on the principle of group stationary weighing of broilers, and the individual weighing mode is more suitable for the technology of floor bird maintenance.

Closest to the technical nature of the proposed method is a method and device for weighing a bird at its floor content, consisting of a load-bearing device in the form of a perch equipped with a load sensor, and a control unit [Grigorishvili T.G. Electrification and automation of the operation of weighing meat chickens in the process of fattening // Abstract of dissertation. Moscow (VIESH), 1972]. The bird, perching, with its weight acts on the load cell, the signal from which is fed to the input of the control unit, which registers the mass of the bird.

The disadvantage of this method of weighing broilers is that when a bird is perched, mechanical vibrations occur that adversely affect the measurement result. To reduce the influence of these noise on the accuracy of weighing, statistical processing of the information received from the primary transducer is not carried out, but only limited to waiting until these oscillations decay. In addition, this method can be applied mainly when weighing an adult bird. For chickens up to middle age, it is ineffective, as perch is not available for birds of this age. And finally, questions of representativeness of the sample from the general population, i.e. what should be the number of broilers to be weighed in order to judge with a certain accuracy and reliability the average live weight of the bird by the herd.

The objective of the invention is to increase the accuracy of determining the average live weight of broilers in a herd with their floor content, providing continuous monitoring of live weight of the poultry throughout the entire breeding cycle from the day old to its delivery to the slaughterhouse and providing personnel with information about the number of people currently weighed broiler chickens about their average live weight and current sampling error.

As a result of the use of the present invention, the accuracy of determining the average live weight of broilers in a herd is increased at their floor keeping, continuous monitoring of live weight of the bird is ensured from the day old until the end of its growing cycle with the issuance of information about the number of broilers chickens currently weighed, about their average live weight and the current error of sample representativeness.

The above technical result is achieved by the fact that in the proposed method for determining the average live weight of broilers when they are kept on the floor, including automatic individual weighing of individual birds using electronic scales installed in the house, measuring the weight of the bird with some delay after its attack on the weighing platform, expansion measuring range of electronic scales by piecewise linear approximation of the characteristics of the load cell, the weight of the bird is measured while drinking water , with its state of minimal locomotor activity on a weighing platform, moreover, the live weight of broiler chickens is monitored continuously throughout the entire growing cycle, starting from the day old until their delivery to the slaughter house, after weighing each next broiler, they provide current information on the number of the current moment of broiler chickens, on their average live weight and on the error of representativeness for a given sample size, and a decrease in the influence of dynamic noise caused by motor activity Tew of the bird on the measurement result is provided by a series of repeated measurements of the weighed mass of each bird, the representativeness of the sample by weighing provide a certain number of broilers, the average mass with sufficient accuracy and reliability characterizes the mean weight of live poultry flock.

The technical result is also achieved by the fact that in a device for determining the average live weight of broilers in a herd with their outdoor contents, containing a load-receiving device equipped with a load cell, and a control unit equipped with a delay element for delaying the process of measuring the mass of the bird until the initial transition processes occurring when the bird approaches the load-handling device, elements of the task and the choice of scale factors for piecewise linear approximation of the hara The characteristics of the load cell to expand the range of mass measurement additionally introduce a round weighing platform as a load receptacle, the diameter of which is designed for one broiler of any age and based on a load cell placed in a cylindrical housing to protect it from pollution, to which a step the path along which the chicken, rising to the weighing platform, drinks water from the water tank and is weighed at the same time, and the water tank is provided with attached to the vertical axis of the tripod with the possibility of adjusting its height relative to the weighing platform, which in turn can move along the horizontal base of the tripod, which will allow you to adjust the location of the weighing platform relative to water tank, depending on the age of the bird, as well as a counter for the number of broilers to be weighed, a subtraction element for calibrating the weights, and an element NT set the mass of the "tare", the elements of the task and the choice of scale factors, the element of summing and averaging the results of multiple measurements of one broiler, the element of summing and averaging the results of multiple weighings of many broilers and a display to display current information, while the output of the load cell through the delay element is connected to the input of the counter of the number of broilers to be weighed and to the first input of the subtraction element, the second input of which is connected to the output of the “tare” mass setting element, the output is connected to the information input of the scale factor selection circuit, the multi-channel inputs of which are connected to the outputs of the scale factor setting element, and the information output through the first element of summing and averaging the results of multiple mass measurements of one broiler is connected to the information input of the second element of summing and averaging the results of multiple weighings of multiple broilers whose control input is connected to the output of the counter of the number of weighed broilers, and inform Scintillation output coupled to an input display for displaying current information.

The essence of the invention is illustrated in figure 1, figure 2, figure 3 and figure 4.

Figure 1 shows a General diagram of a device for determining the average live weight of a bird in a herd.

Figure 2 shows a chart of the measured signal of the bird mass m (t) and its components M (t) and D (t).

Figure 3 shows a graph of the dependence of the sample size (n) on the relative error (Δ).

Figure 4 shows the structure of the formation of errors in determining the average live weight of broilers in a herd.

1. A device for determining the average live weight of a bird in a herd contains a weight platform 1, equipped with a load sensor 2, a cylindrical body 3, which protects the load sensor from pollution, a step track 5 for the convenience of lifting the bird on the weight platform 1, horizontal base of the tripod 4, drinking tank water 6, the vertical axis of the tripod 7, side rails 17 to provide access to drinking water exclusively from the side of the weighing platform 1 and the computing unit, in turn containing eleme nt delay 9, delaying the process of measuring the mass of the bird for a certain time to calm the weight platform from mechanical vibrations, a counter 10 to determine the number of broilers to be weighed, a subtraction element 11 to determine the mass of the bird without "packaging", the element for setting the mass of the "packaging" 12, element setting scale factors 13, an element for selecting scale factors 14, an element for summing and averaging the results of a series of N independent measurements of the mass of one chicken 15, an element for summing and averaging the results, and individual weighings of n broiler 16 chickens and a display 18 for displaying current information about the number of broiler chickens currently weighed, their average live weight, and representativeness error for a given sample size.

The device operates as follows.

.After the bird steps on the weighing platform, the signal from the force transducer 2 through the delay element 9 is fed to the first input of the subtraction element 11, the second input of which from the output of the mass element "tare" 12 receives the signal M ₀ proportional to the tare mass, and the value delay Δt ₁ , equal to the duration of the transients associated with the mechanical vibrations of the weighing platform, is determined experimentally based on the characteristics of dynamic noise (see figure 2). As a result, the output signal [m (t _i) -M _0] obtained by subtracting member 11, which is a result of the current i-th dimension without "tare" weight of birds which arrives at the information input of scaling factor selection element 14, on which the multichannel input the signal of the selected scale factor K _r is received from the element for setting the scale factors 13. As a result, at the output of element 14, a signal K _r [m (t _i ) -M ₀ ] is obtained, which is the result of the ith measurement of the mass of the bird already in mass units ( kg) which post falls on the input element of the summation and averaging of the results of multiple measurements of the mass of one broiler 15 for further processing. To improve the accuracy of weighing in the conditions of dynamic interference resulting from the movement of the bird on the weighing platform 1, a series of repeated measurements of the same chicken is performed N times with a sampling interval Δt. After the accumulation of the results of a series of measurements of N in element 15, they are summed and averaged. As a result, at the output of element 15, a signal is obtained proportional to the mass of one j-th broiler, i.e. $$m_j=\mathrm{one}/N{\displaystyle {\sum }_{i=\mathrm{one}}^N[m\left(t_i\right)-M_0]}$$

.

Next, in element 16, statistics are accumulated on the results of weighing n broiler chickens, their summation and averaging, after which its output gives the average live weight of the bird over the herd with a given accuracy and reliability, i.e.

$$m_{cp}=1/n{\displaystyle {\sum }_{j=1}^n{m}_j=1/n\{{\displaystyle {\sum }_{j=1}^{n}1/N{\displaystyle {\sum }_{i=1}^N[m\left(t_i\right)-M_0]\}}}}$$

, $$$$

$$m_{cp}=\mathrm{one}/n{\displaystyle {\sum }_{j=\mathrm{one}}^n{m}_j=\mathrm{one}/n\{{\displaystyle {\sum }_{j=\mathrm{one}}^n\mathrm{one}/N{\displaystyle {\sum }_{i=\mathrm{one}}^N[m\left(t_i\right)-M_0]\}}}}$$

,

where n is the number of weighed broilers formed at the output of the pulse counter 10, the input of which receives a spasmodic signal from the force transducer 2 at the time of the bird's approach to the weighing platform.

One of the advantages of the proposed device (as well as the method as a whole) is that personnel can directly monitor the process of weighing the bird. This means that after each j-th broiler to be weighed, according to the statistics available at the moment, the device determines the current value of the bird mass, which is just an estimate of the desired value of the average live weight of the bird by the herd. And with what accuracy and reliability this estimate characterizes the average live weight of the bird by the herd, it can be judged from the graph of the dependence Δ = f (n) stored in the memory of the block of computational operations, where Δ is the relative error of representativeness, and n is the sample size (cm Fig. 3). As a result, after each weighing, information on the number of broilers currently weighed, their average live weight, and with what accuracy it corresponds to the actual average live weight of the bird in the herd, is displayed on the display screen 18.

2. The essence of the proposed method is illustrated in figure 1, figure 2, figure 3 and figure 4.

The average live weight of broilers per herd is determined during the entire period of poultry rearing, starting from the day old until it is handed over to the slaughter house, and the weighing procedure is performed automatically while drinking water, when the bird is in the state of least physical activity, in which dynamic disturbances associated with with the movement of birds on the weighing platform, are insignificant and have little effect on the measurement result. For this purpose, a small round weighing platform 1 (with a diameter of not more than 10 cm) is used as a load-receiving device, the area calculated for finding no more than one broiler chicken of any age on it (see Fig. 1), and based on a tensometric load-measuring a sensor 2 located inside a cylindrical body 3 with a height of about 10-12 cm relative to the horizontal base 4 of the tripod, to the vertical axis 7 of which is attached a water tank 6 with side rails 17, providing bird access to de exclusively from the side of the weighing platform. This height of the weighing platform from the floor is due to the fact that it must be higher than the thickness of the bedding layer 8. To ensure access of the daily chicken to the weighing platform 1 for drinking water, a step track 5 with a width of not more than 7 cm is attached to the cylinder body, and the whole structure including a step track 5, a cylindrical body 3, a weight platform 1, a load cell 2, can move along a horizontal area with respect to the vertical axis 7 of the tripod, and the water tank 6 in turn can move along tikalnoy plane adjusting its height relative horizontal base 4 tripod. This allows you to adjust the relative position of the weighing platform 1 and the water reservoir 6 so that it is convenient for the bird to use water at any age. Thus, rising to the weighing platform 1 for drinking water, the bird is automatically weighed, and its mass is recorded in the block of computational operations.

Obviously, the process of measuring the mass of the bird and, accordingly, the design of the mechanical part of the device is made in such a way that the bird is weighed precisely during drinking water, when it is in the state of least physical activity. In this case, dynamic interference is practically absent and the weighing result is most consistent with the measured mass of the bird.

Figure 2 shows a chart of the measured signal of the bird mass m (t) and its components M (t) and D (t), and m (t) = M (t) + D (t), a M (t) = m ₀ + M ₀ (t), where m ₀ is the live weight of the bird; M ₀ (t) is the centered value of the function M (t), which characterizes the transient process after an abrupt change in the signal received from the primary mass transducer when the bird approaches the weighing platform; D (t) - dynamic interference caused by normal motor activity of the bird, i.e. the movement of the bird on the weighing platform.

As can be seen from the diagram, at time t ₁ , when the bird steps on the weighing platform, a jump in the measured signal m (t) is obtained, after which the transition process begins in the form of damped oscillations with a duration of Δt ₁ . After the attenuation of the initial mechanical vibrations M ₀ (t), small vibrations (dynamic disturbances - D (t)), due to the normal motor activity of the bird (time interval Δt ₂ ), continue along the movement of the bird to the water reservoir. From the moment of time t ₃ , when the bird stops and begins to drink water, its condition is characterized by the least motor activity (time interval Δt ₃ ), at which dynamic disturbances D (t) are practically absent. Therefore, it is desirable to carry out the weighing procedure of the bird in this time interval. Further, starting at time t _{4, the} bird stops drinking water and, passing back through the weighing platform (time interval Δt ₄ ), descends from it. At the same time, at time t ₅ , a reverse jump of the measured signal occurs, followed by a transition process, which completes the cycle of finding the bird on the weighing platform.

где $$k_{зат}=\frac{A_{i+1}}{A_i}100\%$$

- представляет собой степень уменьшения амплитуды колебаний за один период (T_c) следования синусоиды, а K_зат - заданный уровень затухания в процентах [Филипский Ю.К. Случайные сигналы в радиотехнике. - К.: Высшая школа. 1986. - 126 с.].It follows from the foregoing that, in order to simplify the procedure and filter the useful signal from interference, as well as increase the weighing accuracy, it is advisable to start the process of measuring the mass of the bird with a certain delay not less than the duration of the transient processes Δt ₁ that occurred when the bird stepped on the weighing platform, and the magnitude of this delay determined experimentally based on the numerical characteristics of the decaying component of the transient process itself (see figure 2). The duration of transients can be estimated using the expression $$\Delta t_{\mathrm{one}}=\frac{k_{s\mathrm{but}t}}{K_{s\mathrm{but}t}}{T}_c,$$

Where $$k_{s\mathrm{but}t}=\frac{A_{i+\mathrm{one}}}{A_i}\mathrm{one\; hundred}\%$$

- represents the degree of decrease in the amplitude of oscillations for one period (T _c ) of the sinusoid, and K _zat - the specified level of attenuation in percent [Filipsky Yu.K. Random signals in radio engineering. - K .: High school. 1986. - 126 p.].

For example, at the required K _zat = 5% - signal attenuation level, attenuation coefficient k _zat = 75% and the repetition period of the natural oscillations of the system T _c = 0.2 sec for the duration of the transient process, we obtain Δt ₁ = 3 sec. This means that when weighing a bird, the process of measuring its mass should not begin earlier than 3 seconds after its attack on the weighing platform. During this time, she can easily overcome the 10 cm space of the weighing platform and be at the water reservoir. So the time interval Δt ₂ may not be, i.e. Δt ₂ = 0, or it may take several seconds if the chicken walks on the weight platform. But judging by visual observations, this rarely happens, and the chicken, as a rule, necessarily reaches the reservoir of water in 3-5 seconds. Visual observations also showed that the process of drinking water for different bird ages can last from 5 to 10 seconds. This gives grounds for the assumption that after Δt ₁ = 3 sec of delay in measuring the mass of the bird, the probability that during this time the chicken will be near the water tank and begin to drink water is very large, and during the weighing process at this time it will make it possible to obtain high accuracy measurements. And if the time-based weighing process covers the intervals Δt ₂ , Δt ₃ , Δt ₄ , the system provides for multiple measurements of bird mass with certain sampling parameters and further averaging of the results of these measurements. This allows you to increase the accuracy of weighing by reducing the influence of dynamic noise associated with the movement of birds on the weighing platform. To achieve this, when weighing each broiler, a series of multiple measurements is made, followed by averaging of their results, and the number (N) and the interval (Δt) of these measurements (sampling parameters of the measured analog signal - m (t)) depend on the numerical characteristics of the flowing oscillations.

We define the discretization parameters of the measured analog signal m (t), considering it as a random process. The duration of a series of measurements, which is T = N Δt, is determined using the expression [Romanenko A.F., Sergeev G.A. Issues of applied analysis of random processes. - M.: Soviet Radio, 1968. - 256 p.]

$$T\ge 2D{\tau }_{\mathrm{to}\mathrm{about}R}/{\epsilon }^2,(\mathrm{one})$$

where D is the variance of the measured signal, τ _cor is the correlation interval, ε is the permissible error in absolute units, and the variance D and the correlation interval τ _{cor are} determined based on the processing of statistical data obtained experimentally. For this purpose, considering the change in the signal m (t) in time as a random process, according to the obtained statistical data and according to a well-known technique, using the table of typical functions given in [Gmurman V.E. Theory of Probability and Mathematical Statistics. - M.: Vysshaya Shkola, 1977. 479 pp.], An analytical expression is obtained for the normalized correlation function - ρ _{cor (τ)} , and then the value of the correlation interval - τ _cor , using the formula [Romanenko AF, Sergeev G. BUT. Issues of applied analysis of random processes. - M.: Soviet Radio, 1968. - 256 p.]

. $${\tau }_{\mathrm{to}\mathrm{about}R}={\displaystyle {\int }_0^{\infty }{\rho }_{\mathrm{to}\mathrm{about}R\left(\tau \right)d\left(\tau \right)}}$$

.

Further, from the known values of the variance D and the correction interval τ _cor for a given value of the required accuracy ε, using formula (1), the duration of a series of measurements is determined T.

The value of the sampling interval is determined using Kotelnikov’s 2nd theorem, according to which Δt≤1 / 2 f _cp , where f _cp = ω _cf / 2π is the cutoff frequency (Nyquist frequency), with respect to which high harmonics are “discarded”, and ω _cp is the angular cutoff frequency, which in turn is determined using the expression

, $$2{\displaystyle {\int }_{{\omega }_{\mathrm{from}R}}^{\infty }S}\left(\omega \right)d\omega \le \delta$$

,

where δ is the required accuracy,

or based on the spectral density graph constructed from typical correlation functions given in [Romanenko AF, Sergeev GA Issues of applied analysis of random processes. - M.: Soviet Radio, 1968. - 256 p.].

From the known values of the duration of the measurement series T and the sampling interval Δt, the number of measurements is determined from the expression N = T / Δt.

Thus, to ensure the required accuracy of weighing the bird ε in the presence of dynamic interference caused by its normal motor activity on the weighing platform, it is necessary to make N independent readings with a time interval Δt and, accordingly, a duration of T = N Δt.

The proposed method for weighing birds and, accordingly, the design of the mechanical part of the device that implements this method, are designed to measure the mass of broiler chickens from the age of one day until the end of the cycle of their rearing. A stepped path leading to the weighing platform, a small-sized weighing platform, capable of accepting only one broiler chicken and having the ability to move on a horizontal plane relative to the vertical axis of the tripod, where the water tank is attached with side rails providing access to water exclusively from the side of the weighing platform, and the water tank, in turn, can move along a vertical plane relative to the weight platform, - all these circumstances allow weighing at the same time There is only one broiler chicken of any age, starting from the daily allowance to an adult bird.

Since the mass of the daily chicken and the mass of the adult bird differ from each other by about 50-60 times, then weighing them with the same weights with the same accuracy is almost impossible due to the non-linearity of the characteristics of the primary mass transducer, especially in its lower and upper sections. Therefore, in this case, it is necessary to expand the weighing range by piecewise linear approximation of the characteristics of the primary mass transducer, the essence of which is that its characteristic is divided into numerous elementary sections, each of which is considered linear, and depending on the size of the input signal, the corresponding an elementary region with an appropriate scale factor for converting mass into an electrical signal.

The representativeness (representativeness) of the sample is ensured by weighing a certain number of broilers in such a way that their average weight with sufficient accuracy and reliability characterizes the average live weight of the bird in the herd. From the theory of mathematical statistics it is known that the sample size from the general population depends on the degree of dispersion (dispersion or standard deviation) of the measured parameter relative to its average value. For the normal distribution of a random variable, which is the distribution of bird mass over the herd, since all participants in this “community” are kept in the same conditions and their mass depends on many small factors, each of which does not have a clear advantage over the others (normal distribution condition random variable), the sample size is determined by the formula

$$n={\left(\frac{\sigma t}{\delta }\right)}^2,(2)$$

where σ is the mean square deviation of the bird mass from the average value; δ is the required accuracy in absolute units; t is the argument of the Laplace function, which is determined from the equality 2Ф (t) = γ given in the form of a table, where γ is the confidence probability [Bendat J., Pirsol A. Applied analysis of random data: Per. from English - M .: Mir, 1989. - 540 p.].

In practice, to determine the sample size it is convenient to use the formula n = (vt / Δ) ² , where v = (σ / m ₀ ) 100% is the coefficient of variation, Δ = (δ / Δm ₀ ) 100% is the relative error, m ₀ - average value of the measured parameter. Figure 3 shows a graph of the dependence of the sample size (n) on the relative error (Δ) obtained as a result of processing the statistical data obtained by us during the weekly control measurements of live weight of about 400 heads of broiler chickens throughout the entire growing period at the Bronnitskaya poultry farm Moscow region. As can be seen from the graph in figure 3, with a sample size of 100 goals, as recommended by the technological standards for monitoring live weight of birds according to GOST 18292-85, the methodological error of representativeness is 2.5-3.0%, depending on the value of the coefficient of variation, which at different ages of the bird takes on different meanings. For example, in daily age, v = 12.6%, and at the end of the growing period, increasing, reaches a level of v = 15.6%.

It should be noted that the results obtained in Fig. 3 are characteristic of a bird of a given species (cross-country), when kept under these specific conditions (type of equipment, microclimate parameters, etc.). For a bird of a different species in other growing conditions, other data will be obtained. Therefore, for each specific case, its own statistical data must be processed and the corresponding numerical characteristics determined, allowing to construct a similar graph of the dependence n = f (Δ) to determine the sample size in the new conditions of detention.

Moreover, statistical data for the study and determination of the numerical characteristics of dynamic noise, as well as similar characteristics for determining the sample size can be obtained using the proposed device, when it is installed on the object and configured for all of the above parameters, providing the necessary accuracy and reliability of the information on average live weight broilers in a herd.

From the foregoing, it follows that in order to determine these parameters, the following sequence of actions must be observed:

1) Registration of the analog signal received from the force transducer m (t) using a recorder;

2) The study of the measured signal m (t) and its components M (t) and D (t) by the methods of theories of mechanical and electrical vibrations, as well as the theory of random processes (see figure 2):

где $$k_{зат}=\frac{A_{i+1}}{A_i}100\%$$

- представляет собой степень уменьшения амплитуды колебаний за один период (T_c) следования синусоиды, а K_зат - заданный уровень затухания в процентах;a) Determining the value of the attenuation interval Δt _{1 of the} mechanical vibrations M (t) that occur when the bird approaches the weighing platform, using the expression $$\Delta t_{\mathrm{one}}=\frac{k_{s\mathrm{but}t}}{K_{s\mathrm{but}t}}{T}_c,$$

Where $$k_{s\mathrm{but}t}=\frac{A_{i+\mathrm{one}}}{A_i}\mathrm{one\; hundred}\%$$

- represents the degree of decrease in the amplitude of oscillations for one period (T _c ) of the sinusoid, and K _zat - the specified level of attenuation in percent;

b) Determination of the sampling parameters of the measured analog signal m (t): the duration of the series of measurements - T, the sampling interval - Δt and the number of samples - N, for digital filtering of the useful signal from dynamic noise - D (t):

1. Definition of estimates of mathematical expectation - m ₀ and variance - D of the measured signal m (t), considering it as a random process;

2. The definition of the correlation (or normalized correlation) function of the random process m (t), using the expression

, $$K\left(q\right)=\frac{\mathrm{one}}{n-q}{\displaystyle {\sum }_{i=\mathrm{one}}^{n-q}m\left(t_i\right)-m_0][m\left(t_{i+q}\right)-m_0]}$$

,

- оценка математического ожидания измеряемого сигнала m(t) при достаточно большом количестве отсчетов n; q=0, 1, 2, 3,…n;Where $$m_0=\frac{\mathrm{one}}{n}{\displaystyle {\sum }_{i=\mathrm{one}}^{n}m\left(t_i\right)}$$

- assessment of the expected value of the measured signal m (t) with a sufficiently large number of samples n; q = 0, 1, 2, 3, ... n;

3. Obtaining the analytical expression of the correlation (or normalized correlation) function K (τ) using the least squares method, the essence of which is that the best agreement between the expected theoretical curve and experimental points can be achieved by minimizing the sum of the squared deviations of the experimental points from the smoothing crooked;

4. Determination of the correlation interval τ _cor using tables given in the literature, based on the analytical expression of the correlation (or normalized correlation) function

. $${\tau }_{\mathrm{to}\mathrm{about}R}={\displaystyle {\int }_0^{\infty }{\rho }_{\mathrm{to}\mathrm{about}R\left(\tau \right)d\left(\tau \right)}}$$

.

5. Determination of duration measurement bird weight by expression _armature T≥2D τ / ε ² for known values of dispersion D, the correlation interval τ _armature and the required accuracy ε;

6. Determination of the analytical expression of the spectral density S (ω) of the measured signal m (t) based on its normalized correlation function ρ (τ) using the Fourier transform

$$S\left(\omega \right)=\frac{2}{\pi }{\displaystyle {\int }_0^{\infty }\rho \left(\tau \right)\cos \omega \tau d\tau }$$

or using tables for typical correlation (or normalized correlation) functions given in the above literature sources;

7. Determination of the cutoff frequency ω _cf (Nyquist frequency) using the expression

, $$2{\displaystyle {\int }_{{\omega }_{\mathrm{from}R}}^{\infty }S\left(\omega \right)d\omega \le \delta }$$

,

where δ is the required accuracy;

8. Determination of the sampling interval by Kotelnikov's theorem using the expression

, где $$f_{ср}=\raisebox{1ex}{${\omega }_{с}$}\!\left/ \!\raisebox{-1ex}{$2\pi $}\right.$$

. $$\Delta t\le \raisebox{1ex}{$\mathrm{one}$}\!\left/ \!\raisebox{-1ex}{$2f_{\mathrm{from}R}$}\right.$$

where $$f_{\mathrm{from}R}=\raisebox{1ex}{${\omega }_{\mathrm{from}}$}\!\left/ \!\raisebox{-1ex}{$2\pi $}\right.$$

.

;9. Determination of the number of samples or the number of measurements of the weight of the weighed bird using the expression $$N=\raisebox{1ex}{$T$}\!\left/ \!\raisebox{-1ex}{$\Delta t$}\right.$$

;

3) Determination of the number of birds n weighed, ensuring sample representativeness:

a) Collection and processing of statistical data on the mass of birds of at least 3 age categories;

b) Determination of the numerical characteristics of the mass of the bird as a random variable: m ₀ - the average value of the mass of the bird in a given age; σ is the mean square deviation of the bird mass from the average value; coefficient of variation of bird mass $$v=\frac{\sigma }{m_0}\text{one hundred\%;}$$

, где t - -аргумент функции Лапласа, который определяется из равенства 2Ф(t)=γ, заданного в виде таблицы.c) Determining the number of birds to be weighed - n, the average weight of which with sufficient accuracy - Δ and reliability (confidence probability) - γ characterizes the average live weight of the bird in the herd, using the expression $$n={\left(\raisebox{1ex}{$vt$}\!\left/ \!\raisebox{-1ex}{$\Delta $}\right.\right)}^2$$

, where t is the argument of the Laplace function, which is determined from the equality 2Ф (t) = γ given in the form of a table.

Thus, the parameter values determined by this technique are the duration of a series of measurements of bird mass - T, the sampling interval - Δt and the number of samples - N, as well as the number of broilers to be weighed - n, which is necessary to provide the required sample size, in the form of n = f ( Δ), - are entered into the memory of the block of computational operations as initial data for use during the operation of the proposed device in these specific conditions of the house for this particular type (cross) of the bird.

SPECIFIC EXAMPLE OF PERFORMANCE OF THE PROPOSED METHOD

Let us analyze all the errors forming the total error of "determining the average live weight of the bird by the herd."

Figure 4 shows a block diagram of the structure of the formation of the error "determine the average live weight of broilers for the herd."

As can be seen from the flowchart in figure 4, the total error of "determining the average live weight of broilers by the herd" is formed from the "instrumental error of the weighed device" and "methodical error of determining the average live weight of the broilers by the herd", which consists of the "error of representativeness of the sample ”And“ methodological error of the weighing process ”, which in turn is formed from the“ error of limiting the measurement duration ”, due to the choice of the value of the parameter for the duration of the measurement series th T, and “quantization errors of the continuous signal”, due to the choice of the value of the parameter of the sampling interval Δt.

In the present invention, the technical result of increasing the accuracy of determining the average live weight of broilers in a herd is achieved solely by reducing the "methodological error in determining the average live weight of broilers in a herd," rather than "the instrumental error of a weighed device."

Consider the individual components of the "methodological error in determining the average live weight of broilers for the herd" and conduct a comparative analysis between the proposed invention, "analog" and "prototype" for these components of the error.

a) Sampling representative error

Based on the above formula (2), the error of representativeness of the sample is determined from the expression

$$\Delta =\frac{vt}{\sqrt{n}},(P-\mathrm{one})$$

- коэффициент вариации; σ - среднее квадратическое отклонение случайной величины от его среднего значения m₀; t - аргумент функции Лапласа, который определяется из равенства 2Ф(t)=γ, где γ - доверительная вероятность; или с помощью построенного на базе выражения (2) графика зависимости n=f(Δ) на фиг.3.where n is the sample size, i.e. number of weighed broilers; $$v=\frac{\sigma }{m_0}\text{one hundred\%}$$

- the coefficient of variation; σ is the mean square deviation of a random variable from its mean value m ₀ ; t is the argument of the Laplace function, which is determined from the equality 2Ф (t) = γ, where γ is the confidence probability; or using a plot of n = f (Δ) in FIG. 3 constructed on the basis of expression (2).

To compare the errors of representativeness for two different sample sizes n ₁ and n _2, all other things being equal, we obtain

$$\frac{{\Delta }_2}{{\Delta }_{\mathrm{one}}}=\sqrt{\frac{n_{\mathrm{one}}}{n_2}}.(P-2)$$

For example, if in the “analogue” [a.s. USSR No. 1572470. A device for determining the average live weight of a bird / P.M. Slavin, G.A. Kharatyan, T.P. Volosovich // Discoveries, Inventions. 1990. No. 23] the number of broilers to be weighed is limited by the size of the experimental cage and is equal, for example, to equipment for cage type 2B-3 120 goals, then in the present invention it is possible to weigh an unlimited number of birds. For example, with a sample size of 150 goals, it follows from the expression (P-2) that, when applying the present invention, the “sampling representative error” decreases 1.12 times as compared to the “analog”.

And in the "prototype" [Grigorishvili T.G. Electrification and automation of the operation of weighing meat chickens in the process of fattening // Abstract of dissertation. Moscow (VIESH), 1972] weighing results are recorded in the form of an analog signal without filtering it from dynamic noise associated with perched bird movement and without taking into account the question of sample representativeness. When processing the recorded material, a certain sample size is still recruited, but this number in no way reaches the sample size required by GOST 18292-85 - 100 goals. If we take the sample size of the "prototype" n ₁ = 100 goals, and the proposed invention - 150 goals., Then from the expression (P-2) it turns out that the "sampling representative error" when applying the proposed invention will decrease in comparison with the "prototype" by 1.225 times.

b) Methodological error of the weighing process

"Methodological error of the weighing process" occurs when processing the measured signal in order to filter the useful signal from dynamic noise caused by normal motor activity of the bird on the weighing platform. In analog processing of the measured signal, its average value, which is the average live weight of the bird, is determined by integrating and averaging this signal over a sufficiently large time interval T.

$$m_{0(\mathrm{but}n)}=\frac{\mathrm{one}}{T}{\displaystyle {\int }_0^{T}m\left(t\right)dt}(P-3)$$

In practice, T (duration of measurements) cannot be infinitely large, therefore, instead of the mathematical expectation of the measured signal, we obtain its estimate, and the more this estimate differs from the mathematical expectation, the greater the value of the so-called “measurement duration limitation error” (see Fig. four).

In digital information processing, the integral (P-3) is replaced by the sum

$$m_{0\left(c\mathrm{and}fR\right)}=\frac{\mathrm{one}}{N}{\displaystyle {\sum }_{i=\mathrm{one}}^{N}m\left(t_i\right),(P-\mathrm{four})}$$

where m (t _i ) is the value of the measured signal at time t _i ; N is the number of samples in the time interval T = Δt N, where Δt is the sampling step (or interval). The smaller the sampling interval Δt, the summation result m _{0 (digits)} determined by the formula (P-4), closer to the result of integration m _{0 (en)} by the formula (P-3), and how much they will differ from each other, in this way, the magnitude of the so-called “quantization error of the continuous signal” will be estimated (see FIG. 4).

Thus, the “methodological error of the weighing process” is formed of two components: “errors of limiting the duration of measurements” and “quantization errors of a continuous signal”.

"The error of the limitation of the duration of measurements" - ε is determined based on the above expression (1)

$$\epsilon \ge \sqrt{2D\frac{{\tau }_{\mathrm{to}\mathrm{about}R}}{T}},(P-5)$$

where D is the dispersion, and τ _cor is the correlation interval of the measured signal; T is the duration of the measurements.

This error is present in both analog and digital data processing, and its value does not depend on the method of processing information, but depends on the spectral and correlation properties of dynamic noise:

variance - D, correlation interval - τ _cor , and also from the measurement duration - T.

As for the “error of quantization of a continuous signal”, in contrast to the “error of limiting the duration of measurements” it takes place exclusively in digital processing of information and depends on the frequency characteristics of dynamic noise caused by the movement of the bird on the weighing platform. The magnitude of this error depends on the choice of the sampling interval Δt, determined by the 2nd Kotelnikov theorem Δt = 1 / 2f _c , where f _c is the cutoff frequency with respect to which the high harmonics are “discarded”. It is evaluated using the expression

$$2{\displaystyle {\int }_{{\omega }_{\mathrm{from}R}}^{\infty }S}\left(\omega \right)d\omega \le \delta ,gde\delta -tReb\mathrm{at}em\mathrm{but}\mathrm{I\; am}t\mathrm{about}hn\mathrm{about}\mathrm{from}tb.(P-6)$$

As a result, the total “methodological error of the weighing process” is ε _met , consisting of the “error of limitation of the measurement duration” - ε and the “quantization error of the continuous signal” - δ, is determined by the well-known formula of “summing” of independent errors

$${\epsilon }_{met}=\sqrt{{\epsilon }^2+{\delta }^2}.(P-7)$$

We will conduct a comparative analysis of the “methodological error of the weighing process” and its components when applying the “analog”, “prototype” and the present invention.

1. Comparative analysis of the invention with "analog"

In both cases, the value of the “methodological error of the weighing process” depends on the characteristics of dynamic interference and on the “quality” of filtering the useful signal from these interference, i.e. from the correct choice of sampling parameters of the measured signal: sampling interval - Δt; the number of samples - N and the measurement duration - T = N Δt, - providing the required accuracy - ε _met .

In the present invention, the weighing of the bird occurs during drinking water, with its least physical activity, when dynamic interference is practically absent and the measured signal with a fairly high accuracy represents the static mass of the weighed bird.

Filtering the useful signal from dynamic noise in this case is additional in nature to maintain high accuracy of weighing, in the case when the chicken before or after drinking water continues to be on the weighing platform and walk on it. As for the “analogue”, due to the fact that the control group of broilers is continuously weighed during the entire growing cycle, there are constantly dynamic disturbances that negatively affect the measurement result. Moreover, in this case, the spectral and correlation properties of the measured signal as a random process are not thoroughly studied, as a result of which the selected discretization parameters (T, Δt and N) do not provide sufficient accuracy for the statistical processing of information.

As a result, from the point of view of the degree of influence of dynamic noise on the result of measuring the mass of the bird, the present invention looks more preferable in comparison with the “analogue”, since in the present invention, the process of weighing the bird occurs directly during drinking water, when the bird is in the state of least physical activity at which dynamic interference is practically absent.

From the foregoing, it follows that the "methodological error of the weighing process" of the present invention is less than that of the "analog".

2. Comparative analysis of the invention with a "prototype"

The "prototype" does not provide for filtering the useful signal from dynamic noise caused by the movement of the bird on a perch.

In this case, when the bird is weighed, only the instantaneous value of the measured signal is recorded without taking into account the influence of random mechanical vibrations on the measurement result, the amplitude of which, judging by the experimental data obtained by testing the automated broiler group weighing system, when they are caged at the Bronnitskaya poultry farm in Moscow Region , can reach 5-10% of the average weight of the bird, depending on its motor activity.

With regard to the present invention, due to the deep analysis of the correlation and spectral properties of dynamic noise, the parameters of the statistical processing of the measured signal are determined, i.e. the measurement duration is T, the sampling interval is Δt, and the number of samples is N, which ensures the specified weighing accuracy.

This allows you to process the measured analog signal by digital methods, allowing you to filter the useful signal from dynamic noise caused by the movement of the bird on the weighing platform, and to ensure high accuracy of weighing.

For example, according to the results of group weighing of broilers with their cage during testing at the Bronnitskaya poultry farm, the “methodological error of the weighing process”, due to the discretization of the measured continuous signal in order to get rid of dynamic noise, was approximately 0.5%, at a time when without filtering these interferences and without statistical data processing, as occurs in the “prototype”, the error in measuring the mass of the bird is several times higher, since the maximum deviation m The estimated value of the measured value from its average value according to experimental data is 5-10%. Moreover, according to the well-known “three sigma rule” (3σ) from the course “probability theory and mathematical statistics”, the average quadratic error of measuring the mass of a bird using the “prototype” will be approximately 1.6-3.3% depending on the motor activity of the bird .

Thus, we conducted a detailed comparative analysis of the invention, “analogue” and “prototype” for all components of the “methodological error in determining the average live weight of broilers in a herd”.

The results of the analysis show that in terms of the "error of representativeness of the sample" "analogue" and "prototype" exceed the proposed invention by 1.12 and 1.225 times, respectively, and in terms of the "methodological error of the weighing process" "analogue" and "prototype" also exceed the proposed invention, moreover, in the "prototype" this error is approximately 1.6-3.3%, and in the present invention - approximately 0.5%.

Claims (2)

1. A method for determining the average live weight of broilers in a herd when they are kept on the floor, including automatic individual weighing of individual birds using electronic weights installed in the house, measuring the weight of the bird with some delay after it steps onto the weighing platform, expanding the range of measurement of electronic weights by piecewise -linear approximation of the characteristics of the load cell, characterized in that the mass of the bird is measured while drinking water, with its state of minimum motor ac activity on a weighing platform, moreover, the live weight of broiler chickens is monitored continuously throughout the entire growing cycle, starting from the day old until they are delivered to the slaughter house, after weighing each next broiler, they provide current information about the number of broilers chickens weighed at the moment, about their average live weight and the error of representativeness for a given sample size, and reducing the effect of dynamic noise caused by the motor activity of the bird on the measurement result provides They are collected by carrying out a series of repeated measurements of the weight of each weighed bird, and the representativeness of the sample is provided by weighing a certain number of broilers, the average weight of which characterizes the average live weight of the bird with sufficient accuracy and reliability over the herd. 2. A device for determining the average live weight of broilers in a herd with their floor content, containing a load receiving device equipped with a force measuring sensor, and a control unit equipped with a delay element for delaying the process of measuring the mass of the bird until the initial transient processes that occur when the bird approaches the load receiving device , elements of the task and the choice of scale factors in the piecewise linear approximation of the characteristics of the load cell for the range of the mass measurement, characterized in that it further comprises a round weighing platform as a load receptacle, the diameter calculated for one broiler chicken and supported by a load cell placed in a cylindrical housing to protect it from pollution, to which a step track is attached, according to which the chicken, rising to the weighing platform, drinks water from the water tank and is weighed at the same time, and the water tank equipped with side rails provides allowing birds access to water exclusively from the side of the weighing platform, in turn, it is attached to the vertical axis of the tripod with the possibility of adjusting its height relative to the weighing platform, which in turn can move along the horizontal base of the tripod, which allows you to adjust the location of the weighing platform relative to the water tank depending from the age of the bird, as well as into the device, a counter for the number of weighed broilers, a subtraction element for calibrating the weights, an element for specifying the “tare” mass, e the element of summing and averaging the results of multiple measurements of one broiler and the element of summing and averaging the results of multiple weighings of many broilers and a display for displaying current information, while the output of the load cell through the delay element is connected to the input of the counter of the number of weighed broilers and to the first input of the subtraction element, the second input which is connected to the output of the element for setting the mass of "tare", and the output is connected to the information input of the element for selecting scale factors NTV, the multi-channel inputs of which are connected to the outputs of the element for setting scale factors, and the information output through the first element of summing and averaging the results of multiple measurements of the mass of one broiler is connected to the information input of the second element of summing and averaging the results of multiple weighings of many broilers, the control input of which is connected to the output of the counter number of weighed broilers, and the information output is connected to the display input to display the current information ui.

Images