JPS6228319B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention comprises at least one pump and a water tank whose number is controlled, and the value of the amount of outflow from the water tank at each time is known in advance. This invention relates to a pump operation method for a pump water tank system.

[Prior Art] Conventionally, for example, in a pump storage tank system for agricultural water, one or more pumps P1, P are connected from a water suction tank 1 to a water storage tank 2, as shown in FIG.
2. It is now possible to pump water using Pn,
Discharge valves V1, V2, ...Vn are connected to these pumps P1, P2, ...Pn, respectively. The upper limit water level of this water tank 2 is HL, and the lower limit water level is LL, and from the water tank 2, a predicted outflow amount Q(t) of water is sent to a downstream demand end (not shown) through piping A, for example, for water for paddy fields. It is now being supplied as water for gardens and fields.

In such a pump water tank system, if the predicted value of the outflow amount Q(t) from the water tank at each time is obtained from the downstream demand end, it is possible to determine the desired requirements such as the number of times the pump is started or the number of pump operations required. The pump can be operated so that the amount of electric power is minimized, and the system can be operated very efficiently.

On the other hand, when discussing optimal control in a control system,
Since the optimal conditions are given by state changes, the optimal control of a control system is discussed as a relationship between state changes that are a function of time, and dynamic planning involves dividing time and controlling according to state changes at the previous or subsequent time. The law is known.

Therefore, when controlling the operation of the pump (on/off control in the case of one pump) as in the pump water tank system described above, since the water level information in the water tank is continuous, there is no difference between the upper and lower limits. Theoretically, information exists almost infinitely, so in practice there is a certain amount of time T.
When the amount of water is calculated, there is theoretically an almost infinite number of water level information before and after the ΔT time, so even if you try to use dynamic programming, there will be too much information and it will be difficult to determine the optimal pump operation method. It is difficult to do so.

In this way, when determining the water level that changes continuously over time, such as in a pump storage tank system, even though it is easy to decide to exclude those that exceed the upper or lower limit, it is difficult to use dynamic programming. It was not practical to perform scheduled operation of the pump. In other words, in the dynamic programming method, if the state takes a specific value, it can be controlled to take that specific value, but if the state changes continuously, like the water level in an aquarium, it does not take a specific value. Therefore, there are an infinite number of calculations, which is inappropriate.

[Object of the Invention] Therefore, an object of the present invention is to provide a pump schedule operation method that can perform optimal control by replacing continuously changing water level information with a finite number of pieces.

[Configuration of the Invention] According to the present invention, at least one
In a pump operation method for a pump water tank system, which is composed of two pumps and a water tank, and in which the value of the outflow amount from the water tank at each time is known in advance, the operating time width T of the system is divided into n time widths ΔT.
At the same time, set the upper limit water level HL and lower limit water level LL of the water tank water level, divide the upper and lower limit water levels into m water level widths ΔV, first find the actual water level of the tank and use this as the initial value. Then, calculate the predicted water level of the water tank after the time interval ΔT corresponding to all the pump operating conditions, and first exclude those that are outside the upper and lower limit ranges from among the obtained predicted water levels of the various water tanks. , and one water tank water level division width ΔV
If several predicted water levels are obtained, and if there are any with the same pump operating conditions, those with unsuitable operating factors are excluded, and in this way the predicted water levels of the various remaining water tanks are calculated. Water level width ΔV
Then, using the above procedure, calculate the water tank water level and operation factor after ΔT, repeat this calculation until the system operation stop time, calculate the pump operation schedule, and operate the pump according to the operation schedule. The predicted water level and the actual water level are compared at each of the n divided times ΔT, and if the comparison value differs by more than the set value, the actual water level is used as the initial value at that time and the I'm starting to redo the calculations.

[Operations and Effects of the Invention] Therefore, the point that the operating time width T is divided into n time widths ΔT is the same as the conventionally known dynamic programming method, but the water level information of the water tank is divided between the upper and lower limit water levels. is divided into a finite number of pumps (m), so at that point, among the operating states of several pumps belonging to the divided water level division width ΔV, the operating factors such as the number of starts and the amount of power consumption (internal combustion engine drive In this case, select the one with the lowest fuel consumption). In this way, it is only necessary to judge the operating state of the pump within a certain water level division width ΔV, and therefore a finite number of judgments can be made.

Therefore, if the time T is divided into n parts and the water level is divided into l parts, the value to be calculated at the final stage is m of l.
The calculation becomes easy because the numbers exceeding the upper and lower limits are removed.

Therefore, when the water demand is predicted in advance, such as in the case of agricultural water, the optimum operating state of the pump can be easily determined, and operating factors such as the number of activations and power consumption can be reduced.

If the actual water demand is significantly different from the predicted outflow, naturally there will be a difference between the predicted water level and the actual water level of the aquarium. If the difference is greater than a certain value, the calculation is recalculated using the actual water level at that time as the initial value. For example, if it is predicted that there will be rain from a certain time, a correction value may be input in advance or at that time. In any case, even if the calculation is re-calculated, the calculation will not be infinite, but will be a finite number, so calculation is easy.

When implementing the present invention, in the case of one pump,
Since the number of units is controlled from 0 to 1, that is, on and off, the operation factor means the number of activations, and 2
In addition to the number of startups, the amount of electricity and fuel consumption are also operational factors when the number of units is more than one.

Optimal operating factors therefore mean optimal conditions for operating one or more pumps, ie minimizing operating costs and the number of starts.

[Example] Hereinafter, an example of the present invention will be described with reference to FIGS. 2 and 3.

FIG. 2 shows an overall block diagram for carrying out the pump schedule operation method according to the present invention, and all parts with respect to FIG. 1 are designated by the same reference numerals.
The water level in the water tank 2 is detected by a water level detector 3, and the water level information is sent to the central processing unit 5 via a line 4. In the central processing unit 5, this water level information is divided into a finite number m between the upper limit water level HL and the lower limit water level LL. In addition, the operating status of the pumps P1, P2,...Pn, that is, the number of activations and operating times are sent to this central processing unit 5 via a line 6.
In addition, from line 8, predicted information on the amount of outflow Q(t) flowing through pipe A leading to the downstream demand end is inputted according to the demand for agricultural pumped water, for example, how many ^m3 per minute from what time to what time? . This central processing unit converts the information from lines 4, 6, and 8 into the predicted outflow amount Q(t).
Based on this, a command to start and stop each pump is issued via line 7 in a manner that will be described later.

Next, referring to FIG. 3, a case where the present invention is implemented in the system shown in FIG. 2 will be described. In order to make the explanation easier to understand, a specific example will be described in which the desired requirement is to minimize the total number of activations of a plurality of pumps per unit time.

Now all pumps P1, P2,...Pl (assuming all pumps have the same efficiency, discharge volume, etc.)
The number of on/off combinations for these l pumps is 2 to the l power, and in the figure, 1 is the number 1 operation, 2 is the number 1 and 2 operation, and 3 is the number 1, 2, and 3 pumps. In the same manner as below, the numbers indicate the number of vehicles in operation.

Therefore, the objective function to minimize the total number of activations of each pump is: It is determined that Here, Ki is the number of times the pump Pi is activated. This value G will be referred to as the operating factor hereinafter.

Now, the number of divisions of the water level of the water tank is m as mentioned above.
(This m is determined appropriately based on the capacity of the water tank),
The initial conditions for operation (0 o'clock or 24 o'clock) are determined as follows, for example.

(i) The initial water amount shall be the upper limit water amount.

(ii) All pumps shall be stopped.

(iii) Therefore, the initial operation factor (number of starts) is set to 0.

As mentioned above, when implementing the present invention, the water level H
Since the information of . Optimization (optimization of operating factors) can be easily determined for each ΔT).

In Fig. 3, the horizontal axis represents time and the vertical axis represents the amount of water in the water tank, i.e., the water level.
It is assumed that the column is divided into n stages (2...n), and the vertical axis is divided into m columns (1, 2...m).

First, in the 0th stage (T=0), if l combinations of pumps are adopted from the above-mentioned initial water level, the water level at the 1st stage, that is, after time ΔT, can be determined in l ways. The general formula for determining this water level is expressed as X ^(j) = X ^(j-1) + (Qin ^(j) - Qout ^(j) ) x ΔT.

Here, X ^(j) is the water level at the j-th stage, Qin ^(j) is the inflow amount to the water tank, that is, the pump discharge amount,
Qout ^(j) is the predicted outflow from the reservoir.

The operating factors (number of activations and pump operating state) of each pump are given to each of the l points in the first stage.

Next, from the water level obtained in the first stage (T = ΔT), the water level of the second stage (T = 2 × ΔT) (there are l × l ways) is determined from the previous equation, and each Give the operating factor for each pump at the point.

If the water level of each stage is determined sequentially according to the above formula, the water level of the Kth stage, that is, the water level after K×ΔT time will be found.
The conditions (water levels) of several points determined above are included in one of the divided widths. Among those states, the operating factors are compared for those in which the pump operating states are the same, and one of the states with the optimal operating factor is selected and made the only optimal state for that pump combination in that division width. be.

For example, in the Kth stage, if the predicted value is in the (m-1)th column (assuming that this unexpected value is optimal), points a, b, c, d,
Assume that e and f have been obtained, and points a and d, points c and e
Assuming that the pump operating conditions of and are equal, compare the operating factors of a and d, c and e,
If the operating factors are a<d, c<e (smaller is more appropriate), exclude d and e, and set the four points a, b, c, and f to the optimal state in this stage and this division width. It is something.

In addition, when implementing the present invention, cases in which the amount of water exceeds the upper and lower limits may be excluded.

Next, a specific example of the present invention will be briefly described with reference to FIG. The illustrated example shows on/off control of two pumps.

In the figure, the horizontal axis is time and the vertical axis is predicted water volume Q.
(t) (lower side) and the aquarium water level (upper side) are shown. In reality, the water level is detected every ΔT for a short period of time, such as 30 seconds or 1 minute, but to make the explanation easier to understand, ΔT is assumed to be 1 hour, and two units with a discharge rate of 1 m ³ /min are used. Using a pump, the effective capacity of the water tank is 72 m ³ , the cross-sectional area is 18 m ² , and the effective water level (H) is 4 m. Assume that the water level is at the upper limit water level HL at 0:00, and the predicted runoff Q(t) is 1.167 m ³ /min from 0:00 to 2:00, zero from 2:00 to 3:00, and 0.833 from 3:00 to 5:00. m ³ /min (this is entered in advance). Regarding the water level, the solid line indicates that all pumps are stopped, the one-dot chain line indicates that one pump is in operation, and the two-dot chain line indicates that two pumps are in operation.

If two pumps are operated at 0:00, the upper limit water level HL will be exceeded, so this is excluded, and changes in water level are shown by line a0 when one pump is in operation, and line b0 when all pumps are stopped.
It is shown in At 1 o'clock, the water level for line a0 is 1a, and for line b0, the water level is 1b.
At point 1a, if two pumps are operated, the water level will exceed the upper limit water level at 2 o'clock, and if all pumps are stopped, the water level will fall below the lower limit water level at 2 o'clock, so these are excluded and only one pump is operated (as shown by line a1). The water level at 2 o'clock stops at point 2a. Also, if two machines are operated at point 1b, the point will be point 2a at 2 o'clock. Line a1 at 2 o'clock
In this case, one pump is in operation, and in line b1, pump 2 is in operation.
Since the pumps are operated, the water level is the same 2a, but the pump operation status (number of pumps) is different, so line a0-a1 and line b0
−b1 comparison is not performed. Now, if one or two pumps are operated at 2 o'clock, all the pumps will be stopped at 3 o'clock because the water level will exceed the upper limit. The water level when all vehicles are stopped is indicated by line a2, and the water level at 3 o'clock is ha. So, looking at the progress up to ha, there are two routes, line a0-a1-a2 and line b0-b1-b2, and both routes are at the water level ha (of course within the same water volume division) at 3 o'clock, and the pump is All machines are stopped. Therefore, if we compare the number of startups between lines a0, a1, a2 and lines b0, b1, and a2 at 3 o'clock, the former is activated only once at the beginning, but the latter is 2 units in operation at point 1b. , 2 times, so lines a0, a1, a2 are adopted, and since the operating factor is optimal with fewer starts, lines b0, b1, a2
Exclude. In other words, both the two-vehicle operation and the one-vehicle operation between 2 o'clock and 3 o'clock exceed the upper limit water level, so the zero-vehicle operation must be performed at 2 o'clock, so lines b0 and b1 are not desirable. Considering the flow up to 5 o'clock in this way, the one with the least number of startups is a0, a1, a2, a3, a4 or
a0, a1, a2, a3, b4 or a0, a1, a2, b3,
The d4 line is the most preferable (optimal state).
If we give the condition that the water level close to the upper limit water level at 5 o'clock is selected, only the former will be selected.

Therefore, for example, in Figure 3 (m-1)
The reason for selecting the item in column (m-1) is that it was determined that the item in column (m-1) is the most preferable by calculating backwards from the final value (work end time).

If the control is turned off at the normal upper limit water level and turned on at the lower limit water level, this flow pattern will be as shown in Figure 4.
They become b0 and b1. However, in the present invention, since lines a0 and a1 are selected, the number of activations can be saved by one, and the water level will be the same at 2 o'clock.

Next, a specific example of scheduled operation to minimize pump power consumption will be described with reference to FIG.

In FIG. 5, as in FIG. 3, the horizontal axis shows time and the vertical axis shows the predicted water amount Q(t) (lower side) and the aquarium water level (upper side). In practice, the water level is detected every short time ΔT, such as 30 seconds or 1 minute, but to make the explanation easier to understand, ΔT is assumed to be 1 hour. And discharge amount 1m ³ /
min pump (No. 1 pump) and discharge rate 2m ³ /min
pump (No. 2 pump), the effective capacity of the water tank is 72 m ³ , the cross-sectional area is 18 m ² , and the effective water level (H).
The length shall be 4m. Also, it is assumed that two pumps, No. 1 and No. 2, will not be operated. Assume that the water level is at the upper limit water level HL at 0 o'clock, and the predicted runoff amount Q
(t) is 1.167m ³ /min from 0:00 to 2:00, zero from 2:00 to 3:00, and 0.833m ^{3 /} min from 3:00 to 5:00.
Set to min (this is pre-filled).

Here, P(t) indicates the cumulative amount of power consumed by the pump. Regarding the water level, the solid line indicates that all pumps are stopped, the one-dot chain line indicates that the No. 1 pump is operating, and the two-dot chain line indicates that the No. 2 pump is operating. Then, the power consumption of pump No. 1 is 10KW, and the power consumption of pump No. 2 is
The power is 22KW.

If the No. 2 pump is operated at 0:00, it will exceed the upper limit water level HL, so this is excluded, and the change in water level is shown by the line a0 when the No. 1 pump is operating, and the line b0 when all units are stopped. . At 1 o'clock, the water level for line a0 is 1a, and for line b0 the water level is 1b.
becomes. At point 1a, when No. 2 pump is operated, the upper limit water level is exceeded at 2 o'clock, and when all pumps are stopped, the water level reaches 2 o'clock.
Since the water level is below the lower limit at the time, these are excluded and the No. 1 pump is operated (indicated by line a1).
The water level at time stops at point 2a. Also at point 1b 2
When the No. 1 pump is operated, it will be at point 2a at 2 o'clock.

In this way, the operating conditions of the No. 1 pump and the No. 2 pump are substantially the same as the operating conditions of one and two pumps in FIG. Adopt.

Comparing the power consumption of lines a0, a1, a2 and lines b0, b1, a2, the power consumption of lines a0, a1, a2 is 10KW x 2H = 20KWH, and lines b0, b1 , the power consumption of a2 is 22KW x 1H = 22KWH.

Therefore, lines a0, a1, a2 are adopted, and lines b0,
Exclude b1 and a2. Considering the operating conditions until 5 o'clock in this way, the line with the least amount of power consumption is line a0, a1, a2, b3, d4 or a0, a1, a2.
It turns out that they are a3 and b4.

All of the above are examples where the prediction and reality almost matched, but if the water level at that time rises due to rain etc. and exceeds a certain set value (for example ΔH),
At that time, water level information is input from the water level detector 3 to the central processing unit 5 via line 4, and from that time, another calculation is performed as described above, and the optimum pump operating state is selected. Furthermore, if it is known that the runoff amount Q(t) changes at a certain time due to, for example, rainfall, the predicted information on the runoff amount Q(t) is inputted via line 8 at the time when the rain can be predicted, and the The processing device also performs calculations similar to those described above based on that information.

The above explanation has been given using the number of starts and the amount of power consumption as operating factors, but it is obvious that, for example, when a pump is operated by an internal combustion engine, the pump operating state can be determined using fuel consumption as an operating factor.

[Summary] As described above, according to the present invention, continuously changing water level information is divided into a finite number of pieces, and the operating factors of the pump operating status belonging to the division width are compared to determine the optimal state. Even in the final judgment, it is sufficient to make a judgment within 1 to the m power, which is easy to calculate, and is suitable for scheduled operation of the pump. Furthermore, even if the prediction differs from the actual situation, it is easy to correct it.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing an example of a pump water tank system, Fig. 2 is a diagram of a system similar to Fig. 1 in which the present invention is implemented, and Fig. 3 is a diagram for explaining an embodiment of the present invention. , FIG. 4 is a diagram showing an example of minimizing the number of activations, and FIG. 5 is a diagram showing an example of minimizing power consumption. 1...Water absorption tank, 2...Water tank, 3...Water level detector, 5...Central processing unit, P1, P2...Pn...
Pump, HL... Upper limit water level, LL... Lower limit water level.

Claims (1)

[Claims] 1. In a pump operation method for a pump water tank system that is composed of at least one pump and a water tank that are controlled in quantity and in which the value of the outflow amount from the water tank at each time is known in advance, the system operating time width T is divided into n time widths ΔT, and the upper limit water level HL and lower limit water level LL of the water tank water level are divided into n time widths ΔT.
is set, and the area between the upper and lower limit water levels is divided into m water level widths ΔV. First, the actual water level of the water tank is determined and used as the initial value, and then the time width ΔT corresponding to all pump operating conditions is determined. The predicted water level of the subsequent water tank is calculated, and among the obtained predicted water levels of the various water tanks, those that are outside the upper and lower limits are first excluded, and then several predictions are calculated within one water tank water level division width ΔV. Among the water levels obtained, if there are pumps with the same operating conditions, exclude those with unsuitable operating factors, and calculate the water level width ΔV of each of the predicted water levels of the remaining various water tanks in this way. , Using the above procedure, further determine the water tank water level and operation factor after ΔT, repeat this calculation until the system operation stop time, determine the pump operation schedule, and operate the pump according to the operation schedule. , the predicted water level and the actual water level are compared at each of the n divided times ΔT, and if the comparison value differs by more than the set value, the above calculation is performed using the actual water level as the initial value at that time. A method of operating a pump on a scheduled basis, which is characterized by re-scheduling.