Abstract: The quantitative measurement of feeding parameters is an important prerequisite for automatic regulation of boiler. At present,most kinds of direct measuring equipment have limited applicability to industrial boilers due to the influence of concentration measuring range,cost and other factors.However,some of the sensor data of boiler have a significant relationship with the feeding system.In order to explore the feasibility of using these sensor data to calculate feed parameters,the sensor data of a 58 MW boiler in tianjin huayuan were collected at a frequency of 1 Hz.Under the condition of controlling other variables,the data related to the feed such as the weight of the intermediate bin,the total pressure of the primary air,the speed of the air lock valve,the pressure of the furnace and the oxygen content of the flue gas were screened out,and the relationship between the absolute value and standard deviation of the above variables and the feed parameters were quantitatively calculated by the least square method and the determination coefficient method.The results indicate that When the feeding valve of the intermediate bin is closed,the average feeding quantity within 300 s of the feeder can be calculated by weighing the middle bin with high accuracy.If the filling rate and filling density of the air lock valve are fixed,the feeding quantity can be accurately calculated by using the feeding speed of the air lock valve.The feeding curve at low speed conforms to this assumption.With the increase rate of the rotational speed,the marginal value of the feed decreases at the edge of the quadratic curve,and there are individual differences in the decline rule.The calculation of the total primary air pressure is a differential pressure method,and when the total primary air pressure is used to calculate the feed quantity,the adjusted determination coefficient R_(Adj)~2>0.98,and the calculation accuracy is very high.The disadvantage is that the coefficient k needs to be calculated based on the cumulative operation data of each device.The fluctuation of negative pressure in furnace can better reflect the fluctuation of feed,but the accuracy is lower than the total pressure of primary air.When the sampling time is 30 s,the standard deviation of supply-the standard deviation of furnace pressure R_(Adj)~2 reaches its maximum value of 0.890 1.The R2 Adjof the feedstock-oxygen consumption reached a maximum value of 0.296 1 at the sampling time of 200 s,and the R_(Adj)~2 of the standard deviation of feed-standard deviation of oxygen consumption reached its maximum value of 0.640 8 at the sampling time of 10 s.It indicates that oxygen consumption is of medium reliability in inferring feed fluctuation,while inferring feed quantity is of poor reliability,which may be related to the instability of the boiler pulverized coal mixing ratio is not stable,resulting in the unit weight of pulverized coal calorific value fluctuations.