It is a natural wish of a customer to know as much as possible about the merchandise he is about to purchase. But if the merchandise is complex and expensive, the customer himself cannot form a correct opinion about it. For example, while considering the purchase of a car the customer does not perform crash tests to evaluate the safety of the car. While buying a house it might be difficult for the customer to estimate if the basement and draining have been done properly. How then could the customer obtain information about the merits and demerits of the merchandise in order to make a conscientious decision whether to buy the goods at the proposed price? In the field of developing of the simulators to train operators for fossil power plants this lack of objective information is practically universal. Convincing of the suppliers and developers cannot be considered as an objective source for the needed information. References from other customers who have already bought a simulator from a concrete supplier (developer) are not convincing either.
Making an analogy from car purchasing, one could say that the situation in the development of the simulators is as follows. Let us take an example of a car which is not equipped with power steering or safety cushions, a car which is impossible to repair because where are no available spare parts on the market. A seller of such a car says to the potential buyer: "Ask your neighbor who has been driving the same car for a few years if he likes the car or not." As for the neighbor he might still like the car. Firstly, he has not yet had any car accidents. Secondly, he is strong enough to steer the wheel without power steering and he does not even notice an inconvenience. Thirdly, the neighbor has not yet a high mileage on the car and does not need to repair it yet.
One might think that inviting an independent expert to test the simulators supplied by a given developer is a solution to this situation. However, an independent expert even if he has deep knowledge concerning the work of fossil power plants will find it difficult to evaluate the quality of a simulator for the equipment with which he is not familiar.
The "Power Plants Simulators" company considers such a situation in the field of simulator development unacceptable. Our firm sees the way forward in the following approach. The developer could himself perform strict tests on his simulators and make the results publicly available to the customer and the customer's independent experts. These results must contain objective information about the course of the test as well as explanations of the developer on why the test was performed in one way and not another. In our view an example of objective information could be plots of certain variables that the simulator draws versus time. There is a need for the developer’s explanations in order to facilitate the expert’s analysis of the test since the expert might not be familiar with the features of the equipment under test.
In order to realize our proposition, we developed and performed a number of tests on one of our simulators. From our point of view the tests will provide the specialists in the theory and practice of power units with the information that will permit them to evaluate the structure and fairness of our simulators.
All tests given in this material can be repeated at any time in our office in the presence of any interested parties.
2. Which simulator was tested?
For testing we used 200 MW gas and oil fired drum boiler at a drum pressure of 14 MPa with natural circulation and at a pressure before the turbine of 13 Mpa.
The oil fuel is pumped into the boiler furnace through 16 burners. From the furnace the flue gases pass into the horizontal gas duct and first encounter the platen superheater. Then the gases pass the 1st stage convective superheater (CSS-1). After the CSS-1 the flue gases pass into the 2nd stage convective superheater (CSS-2), which serves as an exit superheater of the live steam. After the CSS-2 along the gas duct the heating surfaces of the intermediate steam reheating, economizer and Regenerative Air Heater (RAH) are situated.
3. The goal of the testing
The main goal of the performed testing is to give to an expert an idea about the quality of the model and the degree of comprehensiveness of the description of the main physical phenomena of the technological process. In a number of cases we chose rather artificial tests. Nevertheless, they give an idea about the quality of the model.
4. The outcome of the test results
The results of each test are given as plots of the variables chosen for a particular test versus time. The test itself, the plotted variables and an explanation of the behavior of variables are given in detail.
5. The tests
5.1 Heat exchange from gas to metal while an exercise of the unit shutdown, safeguarding and subsequent start-up from a hot reserve conditions
The objective of this test was to show how the gas route of the boiler is modeled. The parameters of the flue gases and metal in the 1st stage convective steam superheater (CSS-1) were examined.
One can see the changes for the different variables on the plots:
BMAZSC – oil fuel flow to the boiler (kg/sec)
OXIG – oxygen content in the effluent gases (percent)
FDI14M – live steam flow, leg A (ton/hour)
RA11T2 – temperature of live steam in the live steam pipeline after the boiler (°Ñ)
TGMKH1 – gas temperature in the CSS-1 (°Ñ) (average)
KH1Tma – average temperature of the metal of heat exchangers in the CSS-1 area (°Ñ)
TZMKH1 – average temperature at the soiled wall of heat exchangers in the CSS-1 area (°Ñ)
WGKH1 – gas flow velocity in the CSS-1 area (m/sec)
ALKKH1 – convective heat transfer coefficient toward the clean metal (kkal/(m2×h×°Ñ))
KH1Nu – kinematic viscosity coefficient of the gases in the CSS-1 (n×106, m2/sec)
KH1Li – thermal conductivity coefficient for gases in the CSS-1 (l×102, kkal/(m×h×°Ñ)
KH1Pr – Prandtl criterion for gases in the CSS-1
BLKH1 – gas flow “blackness” degree in the CSS-1
ALLKH1 – coefficient of the heat irradiation toward the clean metal (kkal/(m2×h×°Ñ))
ALGKH1 – net heat transfer coefficient from the gases toward soiled metal (kkal/(m2×h×°Ñ))
The test was run as follows. Immediately after the beginning of the test a unit protection was activated. It stopped the boiler and the turbine. The operator ventilated the furnace after the shutdown, which took about 10 minutes. After this operation the unit was safeguarded in order to conserve heat for the subsequent start-up.
One hour after the shutting down the operator began the power unit start-up. At 1 hour 3 minutes the furnace ventilation before start-up was initiated. At 1 hour 20 minutes the boiler was ignited.
The test took about four hours, although the plots show only the first three hours.
The results of the test are shown in three pairs of plots. Each plot contains five variables. The plots with the index a) show the evolution of the chosen variables during the first 90 minutes of the test. The plots with the index b) show the evolution of the same variables during the following 90 minutes of the test.
What was the evolution of the chosen variables?
During the furnace ventilation after the shutdown (the first ten minutes) a relatively large amount of air necessary for the ventilation was passing through the boiler. The mass flow of air was even slightly more than the air flow in the nominal state. However, the gas flow velocity in the region of the CSS-1 (WGKH1) diminished considerably in comparison with the nominal value. During the furnace ventilation the gas flow velocity in the region of the CSS-1 (WGKH1) was about 2.6 m/sec (fig. 1-b), while the velocity was 8.8 m/sec for the nominal load of the boiler. This phenomenon is explained by the fact that gas temperature in the CSS-1 (TGMKH1) during ventilation was considerably lower than at the nominal loading.
At the same time the following was happening:
- the gas temperature in the CSS-1 (TGMKH1) dropped from the nominal value 1000°Ñ to approximately 220°Ñ and stabilized at this level (fig.1-a);
- the average temperature of the metal of the heat exchangers at the CSS-1 (KH1Tma) was dropping smoothly from the nominal value 490°Ñ to 370°Ñ (fig.2-a);
- the average temperature at the soiled wall of the heat exchangers (TZMKH1) was dropping as well and was all the time lower than the temperature of the clean metal (KH1Tma) – fig.2-a
- the kinematic viscosity coefficient of the gases at CSS-1 (KH1Nu) dropped from the nominal value 172.5 to 34 and stabilized at this value (fig.2-a);
- the thermal conductivity coefficient for gases at the CSS-1 (KH1Li) dropped from the nominal value 9.6 to 3.6 and stabilized at this value (fig.3-a);
- the Prandtl criterion for gases at the CSS-1 (KH1Pr) increased from the nominal value 0.58 to the value 0.66 (fig.3-a);
- the convective heat transfer coefficient (ALKKH1) dropped from the nominal value 57 to the value 32;
- the gas flow “blackness” degree (BLKH1) dropped from the nominal value 0.14 to the value 0.06, which corresponds to the absence of combustion products at temperatures 200-250°Ñ (fig.3-a);
- the coefficient of the heat irradiation (ALLKH1) practically dropped to zero from the nominal value 33
- the net heat transfer coefficient from the gases toward soiled metal (ALGKH1) dropped from the nominal value 57 to the value 20 (fig.3-a)
The convective heat transfer coefficient linearly depends on the thermal conductivity coefficient and on the quotient of the gas velocity over the kinematic viscosity coefficient to the 0.65 power. The convective heat transfer coefficient depends on the Prandtl criterion to the 0.33 power.
Given our plots let us verify that in the tested simulator the convective heat transfer coefficient fulfils the formula:
Alfa = k×l×(w/n)0.65×Pr0.33
l - thermal conductivity coefficient
w – gas velocity
n - kinematic viscosity coefficient for gases
Pr - Prandtl criterion for gases
In the nominal state we have:
The nominal value of the convective heat transfer coefficient (ALKKH1) is equal to 57. If our assumption is correct, then k=57/1.1593=49.2
So our assumption could be formulated like that: in the simulator under test the convective heat transfer coefficient was calculated according to the following formula:
Alfa = 49.2×l×(w/n)0.65×Pr0.33
Given the data from the figures 1-3 we have the following for the moment of time seven minutes:
According to the plot 2-a the convective heat transfer coefficient (ALKKH1) for the moment of time seven minutes is approximately equal to 32. This means our assumption is fulfilled.
Let us continue to study our test.
An operator completed the ventilation and safeguarded the boiler in order to conserve heat for the subsequent start-up. The process of natural cooling down started.
The average temperature of the metal of the heat exchangers in the CSS-1 area (KH1Tma) began to drop very slowly (fig.1-a). The reason for this slow decrease in temperature is the fact that the region under consideration is situated inside the hermetically closed at the moment boiler. At the same time the temperature of live steam in the steam pipeline at the boiler outlet (RA11T2) was dropping much faster (fig.1-a). The reason for this process is the heat loss through the isolation in the pipeline.
Let us point out that at the same time average temperature at the soiled wall of the heat exchanger in CSS-1 (TZMKH1) evened up with the temperature of the metal itself (KH1Tma) (fig.2-a).
The gas flow velocity at the CSS-1 (WGKH1) dropped practically to zero (fig.1-a). As a result the convective heat transfer coefficient toward the clean metal (ALKKH1) dropped considerably as well (fig.2-a). The temperature of the gases (TGMKH1) rose to 345°Ñ (fig.1-a) while the metal temperature (KH1Tma) reached 350°Ñ (fig.2-a). You know that if there is no gas flow, the temperature of the gases has to reach the same value as the metal temperature.
The kinematic viscosity coefficient of the gases (KH1Nu) stabilized at a new value slightly less than 50. This corresponds also to the new gas temperature of 350°Ñ.
Following the increase in the temperature of the gases:
- the thermal conductivity coefficient for the gases (KH1Li) increased as well to the new value equal to 4.5 (fig.3-a)
- the Prandtl criterion (KH1Pr) dropped to 0.65 (fig.3-a)
One hour after the beginning of the test the operator began the power unit start-up.
In the period between 1 hour 5 minutes and 1 hour 15 minutes the boiler was ventilated before the ignition. This ventilation was performed with less air flow in comparison with the ventilation, which took place between the first and tenth minutes after the unit shutdown. As a result of this new ventilation:
- the gas flow velocity in CSS-1 (WGKH1) rose to 1.6 m/sec, while during the first ventilation it was 2.6 m/sec (fig.2-a);
- the temperature of the gases at CSS-1 (TGMKH1) dropped to 250°Ñ. This value is higher than for the first ventilation because the air flow during the second ventilation was smaller (fig.1-a);
- the live steam temperature in the live steam pipeline (RA11T2) started a faster dropping. This was due to the fact that the process of natural cooling was coupled with the process of pressure dropping in the steam route as a result of the ventilation (fig.1-a);
- the CSS-1 metal temperature (KH1Tma) was dropping more actively due to the consumption of relatively cool air in the gas duct at the CSS-1 (fig.2-a);
- the average temperature at the soiled wall of the heat exchangers (TZMKH1) in the CSS-1 became lower than the temperature of the clean metal in the CSS-1 (KH1Tma) (fig.2-a);
- all other variables reacted realistically as well
At 1 hour 20 minutes the boiler was ignited. The oil fuel consumption to the furnace (BMAZSC) could be seen; the oxygen content (OXIG) reached 12 percent. A steam flow from the boiler appeared. Let us remember that, in the first part of the test, on one hand the ventilation of the furnace was performed twice, on the other hand, there was no steam flow from the boiler because it was sealed. As a result of these factors the metal inside the heated zones of the boiler cooled considerably more than the metal of the unheated steam pipelines since the steam pipelines cooled only due to a natural cooling. Therefore, before the ignition of the boiler the steam in the unheated steam pipelines was considerably hotter than the steam inside the boiler. Thus, after the boiler ignition, as soon as the relatively cool steam started to flow into the live steam pipeline, the temperature of the steam in the live steam pipeline dropped considerably. All these phenomena are shown in figure 2-a.
In addition, the following was happening:
- as a result of a fuel burning the temperature of the gases at CSS-1 (TGMKH1) rose to 600°Ñ (fig 2-b);
- as a consequence of the gas temperature rise as well as the increase in the quantity of gases the velocity of the gases (WGKH1) rose (fig 2-b);
- the convective heat transfer coefficient (ALKKH1) rose mainly due to the increase in the gas velocity (fig 2-b);
- the clean metal temperature (KH1Tma) was rising and the temperature at the soiled wall of the heat exchangers (TZMKH1) became bigger than the clean metal temperature (fig 2-b);
- the kinematic viscosity coefficient of the gases (KH1Nu) increased due to the gas temperature rise (fig 2-b);
- the gas flow “blackness” degree (BLKH1) changed due to the appearance of combustion products in the flue gases (fig.3-b);
- the coefficient of the heat irradiation (ALLKH1) increased mainly due to the rise of the gas temperature (TGMKH1) (fig.3-b);
- the thermal conductivity coefficient (KH1Li) and the Prandtl criterion (KH1Pr) for gases reacted adequately (fig.3-b).
Let us by plots calculate the following expression for the moment of time 1 hour 30 minutes:
According to the plot in figure 2-b, for this moment of time the value of the convective heat transfer coefficient (ALKKH1) is equal about to 24. Thus, we can conclude that in the tested simulator the convective heat transfer coefficient from the gases to the metal under any conditions fulfils the formula:
It is worth mentioning that, while developing the simulator under test, the coefficient k was determined not from the observation of the real object but from the constructive parameters of the boiler. There is a special method for determining the coefficient k for a superheater of any type and for a boiler of any type. The Russian firm “Power Plants Simulators” masters this method.
A considerable part of the variables discussed in the test can’t be approximated or verified directly by the results of the real equipment testing since these variables are not measured on the real equipment. At the same time, for the calculation of the heat exchange between the flue gases and the metal following the fundamental physical laws and criteria equations of the heat exchange, many of these variables are necessary. If a simulator vendor isn’t able to demonstrate realistic plots for the majority of the discussed variables at a relatively complex test, it means that the vendor does not use the criteria heat transfer equations for the calculation of the heat transfer from the flue gases to the metal. If so all the declarations of such vendor that his simulators are very realistic mean really nothing even if his customers are pleased with his simulators.
Conclusions from this test.
The tested simulator models the following parameters correctly for all unit’s working regimes:
- the temperature of the gases in the gas ducts of the boiler;
- the heated surfaces of the boiler and the unheated surfaces without merging them into one unit;
- the temperature of the metal of the heat exchangers taking into account the soiling of the metal by burning products
- the velocity of flue gases in the boiler’s gas ducts with the regard to the temperature change and the flue gas flow;
- the heat flow from the flue gases to the metal based of the criteria heat exchange equations; no approximation are being used.
5.2 The teat exchange between the metal and the steam while an exercise of the unit shutdown, safeguarding and subsequent start-up from a hot reserve conditions
The main objective of the exercise is to demonstrate how the tested simulator implements the heat exchange from metal to steam. To do it we selected the same unit shutdown, safeguarding and subsequent start-up exercise like in the previous test.
During the test our attention was mainly focused at the metal and steam parameters in the 1st stage convective superheater (CSS-1), leg A. A few variables reflect temperatures in the next for CSS-1 convective steam superheater – CSS-2.
The following variables are presented at the plots:
FBI12M – the steam flow through the CSS-1, leg A (tons/hour)
KH14TS – the steam temperature at outlet of the heated (by the flue gases) zone of the CSS-1 (°Ñ)
KH14TM – the metal temperature at outlet of the heated zone of the CSS-1 (°Ñ)
KH14AC – the heat exchange coefficient from metal to steam at the CSS-1 heated zone outlet (kkal/(ì^(2)×h×°Ñ))
KH13TS – the steam temperature in the middle part of the heated zone of the CSS-1 – at the inlet of the zone where KH14TS and KH14TM are measured (°Ñ)
FAIRT – organized air flow to the boiler (n.meters^(3)/sec)
NK21TM – the metal temperature in the unheated (by the flue gases) header after CSS-1 and, at the same time, at the inlet of CSS-2 (°Ñ)
NK21TS - the steam temperature in the unheated (by the flue gases) header after CSS-1 and, at the same time, at the inlet of CSS-2 (°Ñ)
KH21TM – the metal temperature at the beginning of the heated zone of the CSS-2 (°Ñ)
KH21TS - the steam temperature at the beginning of the heated zone of the CSS-2 (°Ñ)
Because of the test was accomplished at the same exercise like in the previous test, all trends of the 1st test can be applied to the 2nd test too.
In addition 2 pairs of trends are being investigated (4-a, 4-b, 5-a, 5-b). Each of the pairs consists of 5 variables. In each pair the picture with a) index displays trends of the selected variables during first 90 minutes of the exercise while the picture with b) index displays trends of the selected variables during next 90 minutes of the exercise.
How did the selected variables were changing?
During the furnace ventilation after unit trip (time range 1-10 minutes) a relatively plenty of air were passed through the boiler. At the same time steam flow through the investigated CSS-1 was tiny.
During the time likewise it was in the nominal working mode, the steam in the CSS-1 was the superheated one. So the heat exchange coefficient from metal to steam decreased from 2800 nominal value till about 30 on the heals of the steam flow decrease – it is about 100 times decrease.
During the furnace ventilation after the unit trip all metal temperatures in the boiler heated zones (KH14TM and KH21TM) were actively decreasing (pictures 4-a and 4-b). Because of the heat exchange coefficient from metal to steam in the CSS-1 didn’t became a zero despite it substantially decreased and at the same time there wasn’t a steam flow through the CSS-1, steam temperature in the heated zones of the CSS-1 (KH14TM and KH21TM) was decreasing promptly after the metal cooling down. At the same time steam and metal in the unheated zones of the CSS-1 (NK21TM and NK21TS) were very slowly cooling down – slower then steam and metal in the heated zones. It was stipulated by the fact that there wasn’t a steam flow at the time through the unheated zone and the fact that the unheated zone had no contiguity with the air going through the boiler for its ventilation. During all the time of boiler shut-down till the boiler ignition at 1 hour 5 minutes time moment when an operator provided a steam flow through the boiler, the unheated zone of the CSS-1 was working in the natural cooling down mode.
At 1 hour time moment an order to start the unit up was given. In 5 minutes at 1 hour 5 minutes time moment the operator began the boiler ventilation before ignition. An air flow through the boiler was organized. So the cooling down rate of all the metal pipes mounted inside the boiler (KH14TM, KH14TS, KH21TM, KH21TS) had increased. At 1 hour 15 minutes time moment the steam temperature at the outlet of the CSS-1 heated zone (KH14TS) reached down the 340°Ñ level. Having the pressure the steam in the CSS-1 had at that moment, the wet steam in the CSS-1 started to appear. As a consequence of the fact the heat exchange coefficient from metal to steam at outlet of CSS-1 heated zone (KH14AC) considerably increased despite the fact that at the moment there still wasn’t a steam flow through the zone – picture 4-a.
The heat exchange coefficient increase assisted the fact that during boiler ventilation at time range 1h5m - 1h15m the steam and metal temperatures at the outlet of the CSS-1 heated zone (KH14TS and KH14TM) were synchronously changing – picture 4-a.
In the tested boiler the CSS-2 follows the CSS-1 both by steam and by flue gases. Steam and metal at the CSS-2 initially were more fiery compare with the CSS-1. During the unit cooling down the CSS-2 was cooling down too but it had higher temperature level compare with the CSS-1. So the steam in the CSS-2 had no enough time to become a wet one. Therefore the heat exchange coefficient in the CSS-2 at time range 1h5m – 1h15m still was quite low. It explains the fact that during the furnace ventilation before boiler ignition the metal temperature at the beginning of the CSS-2 (KH21TM) was actively cooling down while the steam temperature in the area (KH21TS) was staying about the same.
Immediately after boiler ignition at 1 hour 20 minutes time moment the steam temperature (KH13TS) at the inlet of the investigated area had increased (picture 4-a) and a steam flow in the investigated area (FBI12M) had appeared – picture 4-a. As a consequence a dry steam come to the investigated area and the heat exchange coefficient from metal to steam (KH14AC) decreased. From the moment on till the very end of the exercise the heat exchange coefficient was more or less changing in accordance with steam flow through the investigated area (FBI12M) – picture 4-b.
It should be mentioned that as soon as a steam flow through the investigated area had appeared the steam temperature in the unheated zone after CSS-1 (NK21TS) followed by the metal temperature in the zone (NK21TM) started to going down.
The behavior of all the investigated variables during the unit start-up (time range 1 hour 20 minutes till the very end of the exercise) from our point of view was also realistic – pictures 4-b and 5-b.
Conclusions from the test:
The tested simulator in all the power unit working modes precisely calculates:
- heat exchange from metal to steam
- heated and unheated surfaces of the boiler without a joining of the surfaces up to a single whole
5.3 The furnace flame height influence to the boiler working modes
An important factor that influences at the working boiler properties is a position of the torch kernel in the furnace of the boil (the height of the torch). The torch kernel height position determines how the released heat is being distributed between the radiant heat transfer that mainly in charge for the steam generation and the convective heat transfer that mainly in charge for the steam reheating. The mentioned torch height depends in own turn from a number of factors: the distribution of the burners in the furnace by tiers, the current fuel flow distribution between tiers of burners, the burner slopes in the vertical direction (if hinged burners are being installed at the boiler), the properties of the currently combusted fuel (an oil fuel usually burns lower then a gas fuel), and so on. The operator ability to influence the mentioned factors is usually very limited. The main role here belongs to the boiler designer, that has selected the height positions where to mount the burners. If the height positions were abortively selected it could turns out that the boiler either generate too much steam but don’t superheat it till the required temperature (it could mean that the burners are mounted too low) or the boiler doesn’t generate enough steam but superheat it till the too high temperature (it could mean that the burners are mounted too high). Sometimes the problem could lead the redesign of the boiler, sometimes the desirable properties of the boiler could be archived by some changes of the burner height positions.
While developing a model for a new boiler the “Power plant simulators” company behaves about the same way a real boiler designer could and sometimes do. First of all a model of the boiler with some average torch height has been developed. After that the first version of the boiler with operator actions is brought to the desirable nominal parameters. It could turn out that some parameters of this nominal conditions of the simulated boiler are not conform to the same parameters of the real one. In that case one of the actions the simulator developer could undertake to archive the desirable values of the parameters is to change the torch kernel height position that as though corresponds to a change of the burner mounting height position. Naturally in order to a simulator developer could develop the simulator in such a manner it is required the boiler model would adequately take into consideration the torch height position.
The main objective of the test is to demonstrate that all the boiler model component parts adequately react on the torch height position changing.
The test is organized in a simple way. Initially the power unit was working in a steady working mode. At the 8 minutes time moment the torch height position in the furnace was moved up.
The test results are demonstrated at the picture 6. At the picture there are the following variables:
TG2 – the flue gases temperature at the furnace outlet (°Ñ)
TGMKH1 – the flue gases temperature in the 1st stage convective steam superheater (CSS-1) (°Ñ)
WGKH1 – flue gases velocity in the area of the CSS-1 (m/s)
KH1Tma – the average metal temperature of the heat-exchangers in the area of the CSS-1 (°Ñ)
TGOEKA – the flue gases temperature at the boiler outlet (°Ñ)
PDRA – steam pressure in the boiler’s drum (ata)
J14TSc – the live steam temperature at the boiler outlet (°Ñ)
After the torch height position was moved up at the 8 minutes time the smaller amount of heat became left in the furnace compare with the moment before that. In the furnace of the tested boiler are mounted the heat-exchangers of the steam-generating circuit only. So because as a result of the torch height increasing a smaller heat amount became left in the furnace, the steam-generating circuit started to accept the smaller heat and hence the amount of the generated steam in the steam-generating circuit diminished. As a result of the diminishing the smaller amount of steam started to come to the boiler drum. So the drum pressure (PDRA) went steeply down – picture 6.
Because of the smaller amount of heat became left in the furnace the flue gases temperature at the furnace outlet (TG2) went up. As a result of the fact all the flue gases temperatures along the flue gases duct (at the picture 6 this are the TGMKH1 and TGOEKA variables) went up. Because of the flue gases temperature increase the flue gases velocity throughout the flue gases duct, in particularly in the area of CSS-1 (WGKH1), went up – picture 6. The flue gases temperature increase resulted in the fact that the metal temperatures of the steam duct, in particularly in the area of CSS-1 (KH1Tma) went up too. The metal temperature increase resulted in an increase of the live steam temperature at the boiler outlet (J14TSc) – picture 6.
Our conclusions from the test:
- The equation system of the tested simulator takes into account the torch kernel height position in the boiler furnace.
- The reaction of the tested simulator upon a change of the torch kernel height position in the furnace is quite adequate
5.4 Continuous integration of temperatures and pressures
The test is devoted to a flay of a simulation approach used by some our competitors in Russia. The essence of the approach is that a dynamic problem is separated from the static one. The approach idea is that after the next operator action the simulator calculates a static state that the simulated power unit has to come if nothing more shell be done. Having the calculated static state the simulator starts a “transfer” of the current simulation state to the just calculated “final” one. The mentioned “transfer” is usually performed by some exponential functions with somehow defined time constants. If during the “transfer” an operator performs a new action, the simulator calculates the new “final” state and immediately starts a new “transfer” from the current state. For a wide spectrum of the non-steady processes (start-ups, cooling down) the approach is certainly inadmissible. But to rapidly understand that a simulator uses such the approach is not always easy.
The next hypothetical test is used to demonstrate that the tested simulator continuously performs a direct integration in time of temperatures and pressures but not to perform a “transfer” from one state to another.
The test was organized as follows. The power unit was working in the nominal conditions. At the 45 minutes time moment a piece of covering of the live steam pipeline was disturbed. As a result a heat lost of the pipeline section to the environment increased. The disturbance was continuing during the 100 seconds only. At the 46 minutes and 40 seconds time moment the covering of the live steam pipeline was fully restored.
In order to demonstrate the test results the following variables were selected:
NT12TM – the metal temperature of the selected part of the live steam pipeline (°Ñ)
NT12QM – the amount of the heat transferred in the selected pipeline from the metal to steam (kkal/sec); during all the test the heat was negative – it means that really it is exactly the steam transferred a heat to the metal but not vice verse.
GCN14 – heat lost to the environment in the selected pipeline section (kkal/sec); the negative sign of the heat means the heat is taken from the metal.
Before to describe the selected variables behavior the following has to be said. During the simulator design the complete heat capacity of the investigated pipeline section at the 500°Ñ temperature was calculated. The calculation was done on the base of the section mass and the used steel type. The heat capacity is equal to 2045 kkal/°Ñ . It is well-known that the higher the metal temperature is the higher the complete heat capacity of the same piece of steel is. It means that it is easier to heat a piece of steel from 100 till 101°Ñ than from 500 till 501°Ñ. At the 540°Ñ that the metal of the pipeline section had at the 45 minutes time the complete heat capacity of the investigated pipeline section was 2105 kkal/°Ñ. It means that in order to heat the piece of steel up to 1°Ñ at that time the 2105 kkal of energy were required.
How did the test was developed? At the 45 minutes time the values of GCN14 and NT12QM were equal – 15.2 kkal/sec. As a consequence of the fact that for the metal the heat supply was equal to the heat withdraw, the metal temperature didn’t change – picture 7.
At the moment of the covering disturbance the heat lost had immediately increased till the 30.6 kkal/s. As a result the metal temperature started to moving down. As a result the heat that steam was supplying to the metal (NT12QM) started to increase – picture 7. In accordance with the picture 7 during the 100 seconds from 45 minutes time moment till the 46 minutes and 40 seconds time moment the NT12QM variable had been changed from 15.2 till 22.4.
During the 100 seconds the metal had an excess of the withdrawn heat upon the supplied heat. Let us try with the help of some geometry knowledge to calculate by picture 7 how much heat the metal lost during the time period. The heat numerically equal to the square of the ABCD geometrical figure at the picture 7 that is limited with green and red lines.
It was above mentioned that at the test moment the complete heat capacity of the investigated pipeline section was equal to 2105 kkal/°Ñ. Than having returned the 1180 kkal, the metal had to get cold down by 1180/2105=0.56°Ñ. By the picture 7 we could see that during the time the metal temperature (NT12TM) dropped down from the 540.091°Ñ till the 539.531°Ñ – it means that it dropped exactly by 0.56°Ñ.
Let us investigate what was happening during the 100 seconds right after the 46 minutes 40 seconds time when the covering was fully restored.
At the moment of the covering restoring the heat lost from the metal to the environment (GCN14) come to the original value 15.2. As a result of the heat lost reduction the metal temperature (NT12TM) started to increase. As a consequence the heat (NT12QM) transferred from metal to steam started to decrease. However during the next 100 seconds after the covering restoring from the 46 minutes 40 seconds time till the 48 minutes 20 seconds time the heat was bigger then the heat lost (GCN14).
Let us try with the help of some geometry knowledge to calculate by picture 7 how much heat the metal obtained during the time period. The heat numerically equal to the square of the DEFG geometrical figure at the picture 7.
It was above mentioned that at the test moment the complete heat capacity of the investigated pipeline section was equal to 2105 kkal/°Ñ. Than having obtained the 545 kkal, the metal had to get warm up by 545/2105=0.259°Ñ. By the picture 7 we could see that during the time the metal temperature (NT12TM) increased from the 539.531°Ñ till the 539.79°Ñ – it means that it increased exactly by 0.259°Ñ.
Our conclusions from the test.
The tested simulator in all the working modes of the power unit:
- with high precision calculates temperatures and pressures on the base of the direct numeric integration method
- for metals takes into account the heat capacity dependency from the temperatures
5.5 The steam pressure calculation takes into account not the substance balance only but the energy balance too
From the everyday life we know that if to heat a sealed vessel where amount of the substance stays the same, the pressure in the vessel goes up. It means that a steam pressure in an area is not a result of steam supply and withdrawal to the area only but it is a function of a heat balance in the area too. A correct implementation of the simple feature of the nature in a simulator could be a real problem. And we know that a lot of simulators don’t implement the feature of the nature.
The next test demonstrates that in the tested simulator the steam pressure calculation takes into consideration not the substance balance only but the energy balance too.
The test was organized as follows. The power unit was working in a steady mode. At the 10 minutes time the high pressure heater number 7 (HPH-7) insulation was broken. As a result the heat lost from the HPH-7 to the environment went up.
At the picture 8 the only variable for the test is presented.
P7 – steam pressure in the steam space of the HPH-7
From the picture 8 it can be seen that till the 10 minutes time moment the steam pressure in the HPH-7 was quite stable. It means that steam supply flow to the steam space of the HPH-7 was about equal to the steam condensing flow inside the HPH-7. At the first moment after the heat lost to the environment increased the steam balance in the steam space of the HPH-7 was not broken, but the steam pressure started to go down. It tells that the tested simulator to calculate a steam pressure really takes into consideration the heat balance too.
Our conclusions from the test:
- The tested simulator the steam pressure calculation algorithm take into consideration not the substance balance only but the energy balance too
We image that some vendors of simulators for fossil power plants after reading the material could immediately say: “It is not necessary at all. The tests are artificial. We have been developing excellent simulators without all this rubbish for many years and our customers satisfy with our simulators”.
Let us cite of a satirist: “If someone never seen another shoes – our shoes (Russian made) are the best in the world”. People, try another shoes!
Any developer of a simulator for a fossil power plant would like his simulator is considered to be a fully-variable one. It is obvious – a fully-variable simulator only is really able to effectively train intellectual skills of the boiler and turbine operators. We are sure that in order to a simulator be a fully-variable one it has to have a lot of features that were mentioned in the material. If a simulator does have such features the simulator developers ought to demonstrate such the features to the potential customers.
The simulators developed by the Russian “Power plant simulators” company are the truly fully-variable simulators.
Dear customers! A simulator is a very expensive commodity. It could be very effective but it could be the useless. Prior to sign a contract for a simulator development with a vendor you have the right and you have to know as much as possible about the vendor. The proposed tests are focused to enable you to have a better notion about how a vendor develops its simulators, how his simulators are arranged inside. You have the right to know as much as possible about what you are purchasing.
The 1st test has demonstrated that the heat exchange from flue gases to metal in the tested simulator has been correctly simulated in all the working modes of a fossil power unit on the base of the project and constructive data of the energy equipment being simulated. The 2nd test has demonstrated that the heat exchange from the metal to the steam has been correctly simulated in all the working modes of a fossil power unit on the base of the project and constructive data of the energy equipment being simulated. The 4th test has demonstrated that if for a steam/water pipeline section the supplied and withdrawal energy are correctly simulated in all the working modes of a fossil power unit then the tested simulator based upon its constructive date only is able to correctly simulate the metal temperature of the pipeline section.
Jointly tests 1, 2 and 4 have proved that the simulation technology used in the tested simulator allows a development of a highly adequate model of the air-gas and the steam-water ducts of the boilers based upon the constructive and project data of the equipment simulated only. It is exactly the technology the “Power plant simulators” company from Russia uses to develop its unique simulators.
The main objective of the “Power plant simulators” company while it is developing a new simulator for a new power unit consists in the idea that in all the working modes of the simulated power plant all the essential internal simulator variables have to be not only realistic but they have to have exact numeric values.
If we are able to archive the objective it turns out that we became partners with the customer at the time of simulator testing at a new higher technical level. The customer has an experience (or he thinks that he knows hot it works) of his power unit behavior but the simulator knows all the basic physical lows and at any situation the simulator has exact values of a lot of essential internal variables that could help to comprehend any intricate situation at the simulator in case the customer and the developer are not agree with each other.
So there is a question. What would happen if in a situation the customer doesn’t satisfied with the values of some power unit parameters he sees at the simulator? First of all the developer can’t just twist some coefficients around to satisfy the customer because the customer could be satisfied but some internal simulator variables could not be realistic. In the case the developer has to either:
- find and correct a mistake in the model (for example, it could turn out that, flue gases heat exchange for leg A is being calculated with the metal temperature of the leg B, so till the time both legs worked synchronously it was not visible but when the legs started to work in different conditions the operator had detected a problem that leaded to the reason),
- find a constructive reason why such the simulator is not working adequately. We had such a case in our experience. For a few weeks the customer had been testing one of our simulators, we were changing some our coefficients that could be changed (torch height, air suctions flows in different elements of the gas-air duct, degree of dirtying for some heat-exchangers, coefficients how effective some heat-exchange surfaces are and so on) and no matter what we did it finally turned out that gas-air duct of the boiler didn’t behave realistically in some situations. We told to the customer that there should be a something different with the equipment compare with the information he supplied us. The customer was thinking for a week. And finally he had thought that about 10 years ago during an equipment renewal they weld on blades of the FD-fans, but they just forgot about it and they didn’t know where were the documents about it. The constructive reason why the simulator didn’t correspond the real power unit was found. The simulator developer quite quickly corrected the simulator so that it became adequate to the real power unit in all its working modes
- convince the customer that despite of his huge experience with the power unit the simulator is working correctly while the customer has a wrong estimation of the situation – we have a lot of such examples in our practical work